Release 0.6.5

- Basic support for OpenStack: Cinder driver, patches for Nova and libvirt
- Add missing "image" and "config_path" QEMU options
- Calculate aggregate per-pool statistics in monitor
- Implement writes with Check-And-Set semantics
- Add a C wrapper library with public header

.dockerignore

@@ -0,0 +1,19 @@
+.git
+build
+packages
+mon/node_modules
+*.o
+*.so
+osd
+stub_osd
+stub_uring_osd
+stub_bench
+osd_test
+dump_journal
+nbd_proxy
+rm_inode
+fio
+qemu
+rpm/*.Dockerfile
+debian/*.Dockerfile
+Dockerfile

.gitignore

@@ -0,0 +1,18 @@
+*.o
+*.so
+package-lock.json
+fio
+qemu
+osd
+stub_osd
+stub_uring_osd
+stub_bench
+osd_test
+osd_peering_pg_test
+dump_journal
+nbd_proxy
+rm_inode
+test_allocator
+test_blockstore
+test_shit
+osd_rmw_test

.gitmodules

@@ -0,0 +1,6 @@
+[submodule "cpp-btree"]
+	path = cpp-btree
+	url = ../cpp-btree.git
+[submodule "json11"]
+	path = json11
+	url = ../json11.git

CMakeLists.txt

@@ -0,0 +1,7 @@
+cmake_minimum_required(VERSION 2.8)
+
+project(vitastor)
+
+set(VERSION "0.6.5")
+
+add_subdirectory(src)

GPL-2.0.txt

@@ -0,0 +1,339 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Lesser General Public License instead.)  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
+
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                    GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+                            NO WARRANTY
+
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+    Gnomovision version 69, Copyright (C) year name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  <signature of Ty Coon>, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.

LICENSE

@@ -0,0 +1,27 @@
+Copyright (c) Vitaliy Filippov (vitalif [at] yourcmc.ru), 2019+
+
+All server-side code (OSD, Monitor and so on) is licensed under the terms of
+Vitastor Network Public License 1.1 (VNPL 1.1), a copyleft license based on
+GNU GPLv3.0 with the additional "Network Interaction" clause which requires
+opensourcing all programs directly or indirectly interacting with Vitastor
+through a computer network and expressly designed to be used in conjunction
+with it ("Proxy Programs"). Proxy Programs may be made public not only under
+the terms of the same license, but also under the terms of any GPL-Compatible
+Free Software License, as listed by the Free Software Foundation.
+This is a stricter copyleft license than the Affero GPL.
+
+Please note that VNPL doesn't require you to open the code of proprietary
+software running inside a VM if it's not specially designed to be used with
+Vitastor.
+
+Basically, you can't use the software in a proprietary environment to provide
+its functionality to users without opensourcing all intermediary components
+standing between the user and Vitastor or purchasing a commercial license
+from the author 😀.
+
+Client libraries (cluster_client and so on) are dual-licensed under the same
+VNPL 1.1 and also GNU GPL 2.0 or later to allow for compatibility with GPLed
+software like QEMU and fio.
+
+You can find the full text of VNPL-1.1 in the file [VNPL-1.1.txt](VNPL-1.1.txt).
+GPL 2.0 is also included in this repository as [GPL-2.0.txt](GPL-2.0.txt).

Make-gen.pl

@@ -1,46 +0,0 @@
-#!/usr/bin/perl
-
-use strict;
-
-my $deps = {};
-for my $line (split /\n/, `grep '^#include "' *.cpp *.h`)
-{
-    if ($line =~ /^([^:]+):\#include "([^"]+)"/s)
-    {
-        $deps->{$1}->{$2} = 1;
-    }
-}
-
-my $added;
-do
-{
-    $added = 0;
-    for my $file (keys %$deps)
-    {
-        for my $dep (keys %{$deps->{$file}})
-        {
-            if ($deps->{$dep})
-            {
-                for my $subdep (keys %{$deps->{$dep}})
-                {
-                    if (!$deps->{$file}->{$subdep})
-                    {
-                        $added = 1;
-                        $deps->{$file}->{$subdep} = 1;
-                    }
-                }
-            }
-        }
-    }
-} while ($added);
-
-for my $file (sort keys %$deps)
-{
-    if ($file =~ /\.cpp$/)
-    {
-        my $obj = $file;
-        $obj =~ s/\.cpp$/.o/s;
-        print "$obj: $file ".join(" ", sort keys %{$deps->{$file}})."\n";
-        print "\tg++ \$(CXXFLAGS) -c -o \$\@ \$\<\n";
-    }
-}

Makefile

@@ -1,153 +0,0 @@
-BLOCKSTORE_OBJS := allocator.o blockstore.o blockstore_impl.o blockstore_init.o blockstore_open.o blockstore_journal.o blockstore_read.o \
-	blockstore_write.o blockstore_sync.o blockstore_stable.o blockstore_rollback.o blockstore_flush.o crc32c.o ringloop.o
-# -fsanitize=address
-CXXFLAGS := -g -O3 -Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fPIC -fdiagnostics-color=always
-all: libfio_blockstore.so osd libfio_sec_osd.so libfio_cluster.so stub_osd stub_uring_osd stub_bench osd_test dump_journal
-clean:
-	rm -f *.o
-
-dump_journal: dump_journal.cpp crc32c.o blockstore_journal.h
-	g++ $(CXXFLAGS) -o $@ $< crc32c.o
-
-libblockstore.so: $(BLOCKSTORE_OBJS)
-	g++ $(CXXFLAGS) -o $@ -shared $(BLOCKSTORE_OBJS) -ltcmalloc_minimal -luring
-libfio_blockstore.so: ./libblockstore.so fio_engine.o json11.o
-	g++ $(CXXFLAGS) -shared -o $@ fio_engine.o json11.o ./libblockstore.so -ltcmalloc_minimal -luring
-
-OSD_OBJS := osd.o osd_secondary.o msgr_receive.o msgr_send.o osd_peering.o osd_flush.o osd_peering_pg.o \
-	osd_primary.o osd_primary_subops.o etcd_state_client.o messenger.o osd_cluster.o http_client.o pg_states.o \
-	osd_rmw.o json11.o base64.o timerfd_manager.o
-osd: ./libblockstore.so osd_main.cpp osd.h osd_ops.h $(OSD_OBJS)
-	g++ $(CXXFLAGS) -o $@ osd_main.cpp $(OSD_OBJS) ./libblockstore.so -ltcmalloc_minimal -luring
-
-stub_osd: stub_osd.o rw_blocking.o
-	g++ $(CXXFLAGS) -o $@ stub_osd.o rw_blocking.o -ltcmalloc_minimal
-
-STUB_URING_OSD_OBJS := stub_uring_osd.o epoll_manager.o messenger.o msgr_send.o msgr_receive.o ringloop.o timerfd_manager.o json11.o
-stub_uring_osd: $(STUB_URING_OSD_OBJS)
-	g++ $(CXXFLAGS) -o $@ -ltcmalloc_minimal $(STUB_URING_OSD_OBJS) -luring
-stub_bench: stub_bench.cpp osd_ops.h rw_blocking.o
-	g++ $(CXXFLAGS) -o $@ stub_bench.cpp rw_blocking.o -ltcmalloc_minimal
-osd_test: osd_test.cpp osd_ops.h rw_blocking.o
-	g++ $(CXXFLAGS) -o $@ osd_test.cpp rw_blocking.o -ltcmalloc_minimal
-osd_peering_pg_test: osd_peering_pg_test.cpp osd_peering_pg.o
-	g++ $(CXXFLAGS) -o $@ $< osd_peering_pg.o -ltcmalloc_minimal
-
-libfio_sec_osd.so: fio_sec_osd.o rw_blocking.o
-	g++ $(CXXFLAGS) -ltcmalloc_minimal -shared -o $@ fio_sec_osd.o rw_blocking.o
-
-FIO_CLUSTER_OBJS := fio_cluster.o cluster_client.o epoll_manager.o etcd_state_client.o \
-	messenger.o msgr_send.o msgr_receive.o ringloop.o json11.o http_client.o pg_states.o timerfd_manager.o base64.o
-libfio_cluster.so: $(FIO_CLUSTER_OBJS)
-	g++ $(CXXFLAGS) -ltcmalloc_minimal -shared -o $@ $(FIO_CLUSTER_OBJS) -luring
-
-test_blockstore: ./libblockstore.so test_blockstore.cpp timerfd_interval.o
-	g++ $(CXXFLAGS) -o test_blockstore test_blockstore.cpp timerfd_interval.o ./libblockstore.so -ltcmalloc_minimal -luring
-test: test.cpp osd_peering_pg.o
-	g++ $(CXXFLAGS) -o test test.cpp osd_peering_pg.o -luring -lm
-test_allocator: test_allocator.cpp allocator.o
-	g++ $(CXXFLAGS) -o test_allocator test_allocator.cpp allocator.o
-
-crc32c.o: crc32c.c crc32c.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-json11.o: json11/json11.cpp
-	g++ $(CXXFLAGS) -c -o json11.o json11/json11.cpp
-
-# Autogenerated
-
-allocator.o: allocator.cpp allocator.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-base64.o: base64.cpp base64.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore.o: blockstore.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_flush.o: blockstore_flush.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_impl.o: blockstore_impl.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_init.o: blockstore_init.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_journal.o: blockstore_journal.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_open.o: blockstore_open.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_read.o: blockstore_read.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_rollback.o: blockstore_rollback.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_stable.o: blockstore_stable.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_sync.o: blockstore_sync.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-blockstore_write.o: blockstore_write.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-cluster_client.o: cluster_client.cpp cluster_client.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd_id.h osd_ops.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-dump_journal.o: dump_journal.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-epoll_manager.o: epoll_manager.cpp epoll_manager.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-etcd_state_client.o: etcd_state_client.cpp base64.h etcd_state_client.h http_client.h json11/json11.hpp object_id.h osd_id.h osd_ops.h pg_states.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-fio_cluster.o: fio_cluster.cpp cluster_client.h epoll_manager.h etcd_state_client.h fio/fio.h fio/optgroup.h http_client.h json11/json11.hpp messenger.h object_id.h osd_id.h osd_ops.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-fio_engine.o: fio_engine.cpp blockstore.h fio/fio.h fio/optgroup.h json11/json11.hpp object_id.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-fio_sec_osd.o: fio_sec_osd.cpp fio/fio.h fio/optgroup.h object_id.h osd_id.h osd_ops.h rw_blocking.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-http_client.o: http_client.cpp http_client.h json11/json11.hpp timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-messenger.o: messenger.cpp json11/json11.hpp messenger.h object_id.h osd_id.h osd_ops.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-msgr_receive.o: msgr_receive.cpp json11/json11.hpp messenger.h object_id.h osd_id.h osd_ops.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-msgr_send.o: msgr_send.cpp json11/json11.hpp messenger.h object_id.h osd_id.h osd_ops.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd.o: osd.cpp blockstore.h cpp-btree/btree_map.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_cluster.o: osd_cluster.cpp base64.h blockstore.h cpp-btree/btree_map.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_flush.o: osd_flush.cpp blockstore.h cpp-btree/btree_map.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_main.o: osd_main.cpp blockstore.h cpp-btree/btree_map.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_peering.o: osd_peering.cpp base64.h blockstore.h cpp-btree/btree_map.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_peering_pg.o: osd_peering_pg.cpp cpp-btree/btree_map.h object_id.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_peering_pg_test.o: osd_peering_pg_test.cpp cpp-btree/btree_map.h object_id.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_primary.o: osd_primary.cpp blockstore.h cpp-btree/btree_map.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd.h osd_id.h osd_ops.h osd_peering_pg.h osd_primary.h osd_rmw.h pg_states.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_primary_subops.o: osd_primary_subops.cpp blockstore.h cpp-btree/btree_map.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd.h osd_id.h osd_ops.h osd_peering_pg.h osd_primary.h osd_rmw.h pg_states.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_rmw.o: osd_rmw.cpp object_id.h osd_id.h osd_rmw.h xor.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_rmw_test.o: osd_rmw_test.cpp object_id.h osd_id.h osd_rmw.cpp osd_rmw.h test_pattern.h xor.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_secondary.o: osd_secondary.cpp blockstore.h cpp-btree/btree_map.h etcd_state_client.h http_client.h json11/json11.hpp messenger.h object_id.h osd.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-osd_test.o: osd_test.cpp object_id.h osd_id.h osd_ops.h rw_blocking.h test_pattern.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-pg_states.o: pg_states.cpp pg_states.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-ringloop.o: ringloop.cpp ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-rw_blocking.o: rw_blocking.cpp rw_blocking.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-stub_bench.o: stub_bench.cpp object_id.h osd_id.h osd_ops.h rw_blocking.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-stub_osd.o: stub_osd.cpp object_id.h osd_id.h osd_ops.h rw_blocking.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-stub_uring_osd.o: stub_uring_osd.cpp epoll_manager.h json11/json11.hpp messenger.h object_id.h osd_id.h osd_ops.h ringloop.h timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-test.o: test.cpp allocator.h blockstore.h blockstore_flush.h blockstore_impl.h blockstore_init.h blockstore_journal.h cpp-btree/btree_map.h crc32c.h object_id.h osd_id.h osd_ops.h osd_peering_pg.h pg_states.h ringloop.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-test_allocator.o: test_allocator.cpp allocator.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-test_blockstore.o: test_blockstore.cpp blockstore.h object_id.h ringloop.h timerfd_interval.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-timerfd_interval.o: timerfd_interval.cpp ringloop.h timerfd_interval.h
-	g++ $(CXXFLAGS) -c -o $@ $<
-timerfd_manager.o: timerfd_manager.cpp timerfd_manager.h
-	g++ $(CXXFLAGS) -c -o $@ $<

README-ru.md

@@ -0,0 +1,579 @@
+## Vitastor
+
+[Read English version](README.md)
+
+## Идея
+
+Я всего лишь хочу сделать качественную блочную SDS!
+
+Vitastor - распределённая блочная SDS, прямой аналог Ceph RBD и внутренних СХД популярных
+облачных провайдеров. Однако, в отличие от них, Vitastor быстрый и при этом простой.
+Только пока маленький :-).
+
+Архитектурная схожесть с Ceph означает заложенную на уровне алгоритмов записи строгую консистентность,
+репликацию через первичный OSD, симметричную кластеризацию без единой точки отказа
+и автоматическое распределение данных по любому числу дисков любого размера с настраиваемыми схемами
+избыточности - репликацией или с произвольными кодами коррекции ошибок.
+
+## Возможности
+
+Vitastor на данный момент находится в статусе предварительного выпуска, расширенные
+возможности пока отсутствуют, а в будущих версиях вероятны "ломающие" изменения.
+
+Однако следующее уже реализовано:
+
+- Базовая часть - надёжное кластерное блочное хранилище без единой точки отказа
+- Производительность ;-D
+- Несколько схем отказоустойчивости: репликация, XOR n+1 (1 диск чётности), коды коррекции ошибок
+  Рида-Соломона на основе библиотеки jerasure с любым числом дисков данных и чётности в группе
+- Конфигурация через простые человекочитаемые JSON-структуры в etcd
+- Автоматическое распределение данных по OSD, с поддержкой:
+  - Математической оптимизации для лучшей равномерности распределения и минимизации перемещений данных
+  - Нескольких пулов с разными схемами избыточности
+  - Дерева распределения, выбора OSD по тегам / классам устройств (только SSD, только HDD) и по поддереву
+  - Настраиваемых доменов отказа (диск/сервер/стойка и т.п.)
+- Восстановление деградированных блоков
+- Ребаланс, то есть перемещение данных между OSD (дисками)
+- Поддержка "ленивого" fsync (fsync не на каждую операцию)
+- Сбор статистики ввода/вывода в etcd
+- Клиентская библиотека режима пользователя для ввода/вывода
+- Драйвер диска для QEMU (собирается вне дерева исходников QEMU)
+- Драйвер диска для утилиты тестирования производительности fio (также собирается вне дерева исходников fio)
+- NBD-прокси для монтирования образов ядром ("блочное устройство в режиме пользователя")
+- Утилита удаления образов/инодов (vitastor-rm)
+- Пакеты для Debian и CentOS
+- Статистика операций ввода/вывода и занятого места в разрезе инодов
+- Именование инодов через хранение их метаданных в etcd
+- Снапшоты и copy-on-write клоны
+- Сглаживание производительности случайной записи в SSD+HDD конфигурациях
+- Поддержка RDMA/RoCEv2 через libibverbs
+- CSI-плагин для Kubernetes
+- Базовая поддержка OpenStack: драйвер Cinder, патчи для Nova и libvirt
+
+## Планы развития
+
+- Поддержка удаления снапшотов (слияния слоёв)
+- Более корректные скрипты разметки дисков и автоматического запуска OSD
+- Другие инструменты администрирования
+- Плагины для OpenNebula, Proxmox и других облачных систем
+- iSCSI-прокси
+- Более быстрое переключение при отказах
+- Фоновая проверка целостности без контрольных сумм (сверка реплик)
+- Контрольные суммы
+- Поддержка SSD-кэширования (tiered storage)
+- Поддержка NVDIMM
+- Web-интерфейс
+- Возможно, сжатие
+- Возможно, поддержка кэширования данных через системный page cache
+
+## Архитектура
+
+Так же, как и в Ceph, в Vitastor:
+
+- Есть пулы (pools), PG, OSD, мониторы, домены отказа, дерево распределения (аналог crush-дерева).
+- Образы делятся на блоки фиксированного размера (объекты), и эти объекты распределяются по OSD.
+- У OSD есть журнал и метаданные и они тоже могут размещаться на отдельных быстрых дисках.
+- Все операции записи тоже транзакционны. В Vitastor, правда, есть режим отложенного/ленивого fsync
+  (коммита), в котором fsync не вызывается на каждую операцию записи, что делает его более
+  пригодным для использования на "плохих" (десктопных) SSD. Однако все операции записи
+  в любом случае атомарны.
+- Клиентская библиотека тоже старается ждать восстановления после любого отказа кластера, то есть,
+  вы тоже можете перезагрузить хоть весь кластер разом, и клиенты только на время зависнут,
+  но не отключатся.
+
+Некоторые базовые термины для тех, кто не знаком с Ceph:
+
+- OSD (Object Storage Daemon) - процесс, который хранит данные на одном диске и обрабатывает
+  запросы чтения/записи от клиентов.
+- Пул (Pool) - контейнер для данных, имеющих одну и ту же схему избыточности и правила распределения по OSD.
+- PG (Placement Group) - группа объектов, хранимых на одном и том же наборе реплик (OSD).
+  Несколько PG могут храниться на одном и том же наборе реплик, но объекты одной PG
+  в норме не хранятся на разных наборах OSD.
+- Монитор - демон, хранящий состояние кластера.
+- Домен отказа (Failure Domain) - группа OSD, которым вы разрешаете "упасть" всем вместе.
+  Иными словами, это группа OSD, в которые СХД не помещает разные копии одного и того же
+  блока данных. Например, если домен отказа - сервер, то на двух дисках одного сервера
+  никогда не окажется 2 и более копий одного и того же блока данных, а значит, даже
+  если в этом сервере откажут все диски, это будет равносильно потере только 1 копии
+  любого блока данных.
+- Дерево распределения (Placement Tree / CRUSH Tree) - иерархическая группировка OSD
+  в узлы, которые далее можно использовать как домены отказа. То есть, диск (OSD) входит в
+  сервер, сервер входит в стойку, стойка входит в ряд, ряд в датацентр и т.п.
+
+Чем Vitastor отличается от Ceph:
+
+- Vitastor в первую очередь сфокусирован на SSD. Также Vitastor, вероятно, должен неплохо работать
+  с комбинацией SSD и HDD через bcache, а в будущем, возможно, будут добавлены и нативные способы
+  оптимизации под SSD+HDD. Однако хранилище на основе одних лишь жёстких дисков, вообще без SSD,
+  не в приоритете, поэтому оптимизации под этот кейс могут вообще не состояться.
+- OSD Vitastor однопоточный и всегда таким останется, так как это самый оптимальный способ работы.
+  Если вам не хватает 1 ядра на 1 диск, просто делите диск на разделы и запускайте на нём несколько OSD.
+  Но, скорее всего, вам хватит и 1 ядра - Vitastor не так прожорлив к ресурсам CPU, как Ceph.
+- Журнал и метаданные всегда размещаются в памяти, благодаря чему никогда не тратится лишнее время
+  на чтение метаданных с диска. Размер метаданных линейно зависит от размера диска и блока данных,
+  который задаётся в конфигурации кластера и по умолчанию составляет 128 КБ. С блоком 128 КБ метаданные
+  занимают примерно 512 МБ памяти на 1 ТБ дискового пространства (и это всё равно меньше, чем нужно Ceph-у).
+  Журнал вообще не должен быть большим, например, тесты производительности в данном документе проводились
+  с журналом размером всего 16 МБ. Большой журнал, вероятно, даже вреден, т.к. "грязные" записи (записи,
+  не сброшенные из журнала) тоже занимают память и могут немного замедлять работу.
+- В Vitastor нет внутреннего copy-on-write. Я считаю, что реализация CoW-хранилища гораздо сложнее,
+  поэтому сложнее добиться устойчиво хороших результатов. Возможно, в один прекрасный день
+  я придумаю красивый алгоритм для CoW-хранилища, но пока нет - внутреннего CoW в Vitastor не будет.
+  Всё это не относится к "внешнему" CoW (снапшотам и клонам).
+- Базовый слой Vitastor - простое блочное хранилище с блоками фиксированного размера, а не сложное
+  объектное хранилище с расширенными возможностями, как в Ceph (RADOS).
+- В Vitastor есть режим "ленивых fsync", в котором OSD группирует запросы записи перед сбросом их
+  на диск, что позволяет получить лучшую производительность с дешёвыми настольными SSD без конденсаторов
+  ("Advanced Power Loss Protection" / "Capacitor-Based Power Loss Protection").
+  Тем не менее, такой режим всё равно медленнее использования нормальных серверных SSD и мгновенного
+  fsync, так как приводит к дополнительным операциям передачи данных по сети, поэтому рекомендуется
+  всё-таки использовать хорошие серверные диски, тем более, стоят они почти так же, как десктопные.
+- PG эфемерны. Это означает, что они не хранятся на дисках и существуют только в памяти работающих OSD.
+- Процессы восстановления оперируют отдельными объектами, а не целыми PG.
+- PGLOG-ов нет.
+- "Мониторы" не хранят данные. Конфигурация и состояние кластера хранятся в etcd в простых человекочитаемых
+  JSON-структурах. Мониторы Vitastor только следят за состоянием кластера и управляют перемещением данных.
+  В этом смысле монитор Vitastor не является критичным компонентом системы и больше похож на Ceph-овский
+  менеджер (MGR). Монитор Vitastor написан на node.js.
+- Распределение PG не основано на консистентных хешах. Вместо этого все маппинги PG хранятся прямо в etcd
+  (ибо нет никакой проблемы сохранить несколько сотен-тысяч записей в памяти, а не считать каждый раз хеши).
+  Перераспределение PG по OSD выполняется через математическую оптимизацию,
+  а конкретно, сведение задачи к ЛП (задаче линейного программирования) и решение оной с помощью утилиты
+  lp_solve. Такой подход позволяет обычно выравнивать распределение места почти идеально - равномерность
+  обычно составляет 96-99%, в отличие от Ceph, где на голом CRUSH-е без балансировщика обычно выходит 80-90%.
+  Также это позволяет минимизировать объём перемещения данных и случайность связей между OSD, а также менять
+  распределение вручную, не боясь сломать логику перебалансировки. В таком подходе есть и потенциальный
+  недостаток - есть предположение, что в очень большом кластере он может сломаться - однако вплоть до
+  нескольких сотен OSD подход точно работает нормально. Ну и, собственно, при необходимости легко
+  реализовать и консистентные хеши.
+- Отдельный слой, подобный слою "CRUSH-правил", отсутствует. Вы настраиваете схемы отказоустойчивости,
+  домены отказа и правила выбора OSD напрямую в конфигурации пулов.
+
+## Понимание сути производительности систем хранения
+
+Вкратце: для быстрой хранилки задержки важнее, чем пиковые iops-ы.
+
+Лучшая возможная задержка достигается при тестировании в 1 поток с глубиной очереди 1,
+что приблизительно означает минимально нагруженное состояние кластера. В данном случае
+IOPS = 1/задержка. Ни числом серверов, ни дисков, ни серверных процессов/потоков
+задержка не масштабируется... Она зависит только от того, насколько быстро один
+серверный процесс (и клиент) обрабатывают одну операцию.
+
+Почему задержки важны? Потому, что некоторые приложения *не могут* использовать глубину
+очереди больше 1, ибо их задача не параллелизуется. Важный пример - это все СУБД
+с поддержкой консистентности (ACID), потому что все они обеспечивают её через
+журналирование, а журналы пишутся последовательно и с fsync() после каждой операции.
+
+fsync, кстати - это ещё одна очень важная вещь, про которую почти всегда забывают в тестах.
+Смысл в том, что все современные диски имеют кэши/буферы записи и не гарантируют, что
+данные реально физически записываются на носитель до того, как вы делаете fsync(),
+который транслируется в команду сброса кэша операционной системой.
+
+Дешёвые SSD для настольных ПК и ноутбуков очень быстрые без fsync - NVMe диски, например,
+могут обработать порядка 80000 операций записи в секунду с глубиной очереди 1 без fsync.
+Однако с fsync, когда они реально вынуждены писать каждый блок данных во флеш-память,
+они выжимают лишь 1000-2000 операций записи в секунду (число практически постоянное
+для всех моделей SSD).
+
+Серверные SSD часто имеют суперконденсаторы, работающие как встроенный источник
+бесперебойного питания и дающие дискам успеть сбросить их DRAM-кэш в постоянную
+флеш-память при отключении питания. Благодаря этому диски с чистой совестью
+*игнорируют fsync*, так как точно знают, что данные из кэша доедут до постоянной
+памяти.
+
+Все наиболее известные программные СХД, например, Ceph и внутренние СХД, используемые
+такими облачными провайдерами, как Amazon, Google, Яндекс, медленные в смысле задержки.
+В лучшем случае они дают задержки от 0.3мс на чтение и 0.6мс на запись 4 КБ блоками
+даже при условии использования наилучшего возможного железа.
+
+И это в эпоху SSD, когда вы можете пойти на рынок и купить там SSD, задержка которого
+на чтение будет 0.1мс, а на запись - 0.04мс, за 100$ или даже дешевле.
+
+Когда мне нужно быстро протестировать производительность дисковой подсистемы, я
+использую следующие 6 команд, с небольшими вариациями:
+
+- Линейная запись:
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4M -iodepth=32 -rw=write -runtime=60 -filename=/dev/sdX`
+- Линейное чтение:
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4M -iodepth=32 -rw=read -runtime=60 -filename=/dev/sdX`
+- Запись в 1 поток (T1Q1):
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4k -iodepth=1 -fsync=1 -rw=randwrite -runtime=60 -filename=/dev/sdX`
+- Чтение в 1 поток (T1Q1):
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4k -iodepth=1 -rw=randread -runtime=60 -filename=/dev/sdX`
+- Параллельная запись (numjobs используется, когда 1 ядро CPU не может насытить диск):
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4k -iodepth=128 [-numjobs=4 -group_reporting] -rw=randwrite -runtime=60 -filename=/dev/sdX`
+- Параллельное чтение (numjobs - аналогично):
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4k -iodepth=128 [-numjobs=4 -group_reporting] -rw=randread -runtime=60 -filename=/dev/sdX`
+
+## Теоретическая максимальная производительность Vitastor
+
+При использовании репликации:
+- Задержка чтения в 1 поток (T1Q1): 1 сетевой RTT + 1 чтение с диска.
+- Запись+fsync в 1 поток:
+  - С мгновенным сбросом: 2 RTT + 1 запись.
+  - С отложенным ("ленивым") сбросом: 4 RTT + 1 запись + 1 fsync.
+- Параллельное чтение: сумма IOPS всех дисков либо производительность сети, если в сеть упрётся раньше.
+- Параллельная запись: сумма IOPS всех дисков / число реплик / WA либо производительность сети, если в сеть упрётся раньше.
+
+При использовании кодов коррекции ошибок (EC):
+- Задержка чтения в 1 поток (T1Q1): 1.5 RTT + 1 чтение.
+- Запись+fsync в 1 поток:
+  - С мгновенным сбросом: 3.5 RTT + 1 чтение + 2 записи.
+  - С отложенным ("ленивым") сбросом: 5.5 RTT + 1 чтение + 2 записи + 2 fsync.
+- Под 0.5 на самом деле подразумевается (k-1)/k, где k - число дисков данных,
+  что означает, что дополнительное обращение по сети не нужно, когда операция
+  чтения обслуживается локально.
+- Параллельное чтение: сумма IOPS всех дисков либо производительность сети, если в сеть упрётся раньше.
+- Параллельная запись: сумма IOPS всех дисков / общее число дисков данных и чётности / WA либо производительность сети, если в сеть упрётся раньше.
+  Примечание: IOPS дисков в данном случае надо брать в смешанном режиме чтения/записи в пропорции, аналогичной формулам выше.
+
+WA (мультипликатор записи) для 4 КБ блоков в Vitastor обычно составляет 3-5:
+1. Запись метаданных в журнал
+2. Запись блока данных в журнал
+3. Запись метаданных в БД
+4. Ещё одна запись метаданных в журнал при использовании EC
+5. Запись блока данных на диск данных
+
+Если вы найдёте SSD, хорошо работающий с 512-байтными блоками данных (Optane?),
+то 1, 3 и 4 можно снизить до 512 байт (1/8 от размера данных) и получить WA всего 2.375.
+
+Кроме того, WA снижается при использовании отложенного/ленивого сброса при параллельной
+нагрузке, т.к. блоки журнала записываются на диск только когда они заполняются или явным
+образом запрашивается fsync.
+
+## Пример сравнения с Ceph
+
+Железо - 4 сервера, в каждом:
+- 6x SATA SSD Intel D3-4510 3.84 TB
+- 2x Xeon Gold 6242 (16 cores @ 2.8 GHz)
+- 384 GB RAM
+- 1x 25 GbE сетевая карта (Mellanox ConnectX-4 LX), подключённая к свитчу Juniper QFX5200
+
+Экономия энергии CPU отключена. В тестах и Vitastor, и Ceph развёрнуто по 2 OSD на 1 SSD.
+
+Все результаты ниже относятся к случайной нагрузке 4 КБ блоками (если явно не указано обратное).
+
+Производительность голых дисков:
+- T1Q1 запись ~27000 iops (задержка ~0.037ms)
+- T1Q1 чтение ~9800 iops (задержка ~0.101ms)
+- T1Q32 запись ~60000 iops
+- T1Q32 чтение ~81700 iops
+
+Ceph 15.2.4 (Bluestore):
+- T1Q1 запись ~1000 iops (задержка ~1ms)
+- T1Q1 чтение ~1750 iops (задержка ~0.57ms)
+- T8Q64 запись ~100000 iops, потребление CPU процессами OSD около 40 ядер на каждом сервере
+- T8Q64 чтение ~480000 iops, потребление CPU процессами OSD около 40 ядер на каждом сервере
+
+Тесты в 8 потоков проводились на 8 400GB RBD образах со всех хостов (с каждого хоста запускалось 2 процесса fio).
+Это нужно потому, что в Ceph несколько RBD-клиентов, пишущих в 1 образ, очень сильно замедляются.
+
+Настройки RocksDB и Bluestore в Ceph не менялись, единственным изменением было отключение cephx_sign_messages.
+
+На самом деле, результаты теста не такие уж и плохие для Ceph (могло быть хуже).
+Собственно говоря, эти серверы как раз хорошо сбалансированы для Ceph - 6 SATA SSD как раз
+утилизируют 25-гигабитную сеть, а без 2 мощных процессоров Ceph-у бы не хватило ядер,
+чтобы выдать пристойный результат. Собственно, что и показывает жор 40 ядер в процессе
+параллельного теста.
+
+Vitastor:
+- T1Q1 запись: 7087 iops (задержка 0.14ms)
+- T1Q1 чтение: 6838 iops (задержка 0.145ms)
+- T2Q64 запись: 162000 iops, потребление CPU - 3 ядра на каждом сервере
+- T8Q64 чтение: 895000 iops, потребление CPU - 4 ядра на каждом сервере
+- Линейная запись (4M T1Q32): 2800 МБ/с
+- Линейное чтение (4M T1Q32): 1500 МБ/с
+
+Тест на чтение в 8 потоков проводился на 1 большом образе (3.2 ТБ) со всех хостов (опять же, по 2 fio с каждого).
+В Vitastor никакой разницы между 1 образом и 8-ю нет. Естественно, примерно 1/4 запросов чтения
+в такой конфигурации, как и в тестах Ceph выше, обслуживалась с локальной машины. Если проводить
+тест так, чтобы все операции всегда обращались к первичным OSD по сети - тест сильнее упирался
+в сеть и результат составлял примерно 689000 iops.
+
+Настройки Vitastor: `--disable_data_fsync true --immediate_commit all --flusher_count 8
+  --disk_alignment 4096 --journal_block_size 4096 --meta_block_size 4096
+  --journal_no_same_sector_overwrites true --journal_sector_buffer_count 1024
+  --journal_size 16777216`.
+
+### EC/XOR 2+1
+
+Vitastor:
+- T1Q1 запись: 2808 iops (задержка ~0.355ms)
+- T1Q1 чтение: 6190 iops (задержка ~0.16ms)
+- T2Q64 запись: 85500 iops, потребление CPU - 3.4 ядра на каждом сервере
+- T8Q64 чтение: 812000 iops, потребление CPU - 4.7 ядра на каждом сервере
+- Линейная запись (4M T1Q32): 3200 МБ/с
+- Линейное чтение (4M T1Q32): 1800 МБ/с
+
+Ceph:
+- T1Q1 запись: 730 iops (задержка ~1.37ms latency)
+- T1Q1 чтение: 1500 iops с холодным кэшем метаданных (задержка ~0.66ms), 2300 iops через 2 минуты прогрева (задержка ~0.435ms)
+- T4Q128 запись (4 RBD images): 45300 iops, потребление CPU - 30 ядер на каждом сервере
+- T8Q64 чтение (4 RBD images): 278600 iops, потребление CPU - 40 ядер на каждом сервере
+- Линейная запись (4M T1Q32): 1950 МБ/с в пустой образ, 2500 МБ/с в заполненный образ
+- Линейное чтение (4M T1Q32): 2400 МБ/с
+
+### NBD
+
+NBD расшифровывается как "сетевое блочное устройство", но на самом деле оно также
+работает просто как аналог FUSE для блочных устройств, то есть, представляет собой
+"блочное устройство в пространстве пользователя".
+
+NBD - на данный момент единственный способ монтировать Vitastor ядром Linux.
+NBD немного снижает производительность, так как приводит к дополнительным копированиям
+данных между ядром и пространством пользователя. Тем не менее, способ достаточно оптимален,
+а производительность случайного доступа вообще затрагивается слабо.
+
+Vitastor с однопоточной NBD прокси на том же стенде:
+- T1Q1 запись: 6000 iops (задержка 0.166ms)
+- T1Q1 чтение: 5518 iops (задержка 0.18ms)
+- T1Q128 запись: 94400 iops
+- T1Q128 чтение: 103000 iops
+- Линейная запись (4M T1Q128): 1266 МБ/с (в сравнении с 2800 МБ/с через fio)
+- Линейное чтение (4M T1Q128): 975 МБ/с (в сравнении с 1500 МБ/с через fio)
+
+## Установка
+
+### Debian
+
+- Добавьте ключ репозитория Vitastor:
+  `wget -q -O - https://vitastor.io/debian/pubkey | sudo apt-key add -`
+- Добавьте репозиторий Vitastor в /etc/apt/sources.list:
+  - Debian 11 (Bullseye/Sid): `deb https://vitastor.io/debian bullseye main`
+  - Debian 10 (Buster): `deb https://vitastor.io/debian buster main`
+- Для Debian 10 (Buster) также включите репозиторий backports:
+  `deb http://deb.debian.org/debian buster-backports main`
+- Установите пакеты: `apt update; apt install vitastor lp-solve etcd linux-image-amd64 qemu`
+
+### CentOS
+
+- Добавьте в систему репозиторий Vitastor:
+  - CentOS 7: `yum install https://vitastor.io/rpms/centos/7/vitastor-release-1.0-1.el7.noarch.rpm`
+  - CentOS 8: `dnf install https://vitastor.io/rpms/centos/8/vitastor-release-1.0-1.el8.noarch.rpm`
+- Включите EPEL: `yum/dnf install epel-release`
+- Включите дополнительные репозитории CentOS:
+  - CentOS 7: `yum install centos-release-scl`
+  - CentOS 8: `dnf install centos-release-advanced-virtualization`
+- Включите elrepo-kernel:
+  - CentOS 7: `yum install https://www.elrepo.org/elrepo-release-7.el7.elrepo.noarch.rpm`
+  - CentOS 8: `dnf install https://www.elrepo.org/elrepo-release-8.el8.elrepo.noarch.rpm`
+- Установите пакеты: `yum/dnf install vitastor lpsolve etcd kernel-ml qemu-kvm`
+
+### Установка из исходников
+
+- Установите ядро 5.4 или более новое, для поддержки io_uring. Желательно 5.8 или даже новее,
+  так как в 5.4 есть как минимум 1 известный баг, ведущий к зависанию с io_uring и контроллером HP SmartArray.
+- Установите liburing 0.4 или более новый и его заголовки.
+- Установите lp_solve.
+- Установите etcd, версии не ниже 3.4.15. Более ранние версии работать не будут из-за различных багов,
+  например [#12402](https://github.com/etcd-io/etcd/pull/12402). Также вы можете взять версию 3.4.13 с
+  этим конкретным исправлением из ветки release-3.4 репозитория https://github.com/vitalif/etcd/.
+- Установите node.js 10 или новее.
+- Установите gcc и g++ 8.x или новее.
+- Склонируйте данный репозиторий с подмодулями: `git clone https://yourcmc.ru/git/vitalif/vitastor/`.
+- Желательно пересобрать QEMU с патчем, который делает необязательным запуск через LD_PRELOAD.
+  См `patches/qemu-*.*-vitastor.patch` - выберите версию, наиболее близкую вашей версии QEMU.
+- Установите QEMU 3.0 или новее, возьмите исходные коды установленного пакета, начните его пересборку,
+  через некоторое время остановите её и скопируйте следующие заголовки:
+   - `<qemu>/include` &rarr; `<vitastor>/qemu/include`
+   - Debian:
+      * Берите qemu из основного репозитория
+      * `<qemu>/b/qemu/config-host.h` &rarr; `<vitastor>/qemu/b/qemu/config-host.h`
+      * `<qemu>/b/qemu/qapi` &rarr; `<vitastor>/qemu/b/qemu/qapi`
+   - CentOS 8:
+      * Берите qemu из репозитория Advanced-Virtualization. Чтобы включить его, запустите
+        `yum install centos-release-advanced-virtualization.noarch` и далее `yum install qemu`
+      * `<qemu>/config-host.h` &rarr; `<vitastor>/qemu/b/qemu/config-host.h`
+      * Для QEMU 3.0+: `<qemu>/qapi` &rarr; `<vitastor>/qemu/b/qemu/qapi`
+      * Для QEMU 2.0+: `<qemu>/qapi-types.h` &rarr; `<vitastor>/qemu/b/qemu/qapi-types.h`
+   - `config-host.h` и `qapi` нужны, т.к. в них содержатся автогенерируемые заголовки
+- Установите fio 3.7 или новее, возьмите исходники пакета и сделайте на них симлинк с `<vitastor>/fio`.
+- Соберите и установите Vitastor командой `mkdir build && cd build && cmake .. && make -j8 && make install`.
+  Обратите внимание на переменную cmake `QEMU_PLUGINDIR` - под RHEL её нужно установить равной `qemu-kvm`.
+
+## Запуск
+
+Внимание: процедура пока что достаточно нетривиальная, задавать конфигурацию и смещения
+на диске нужно почти вручную. Это будет исправлено в ближайшем будущем.
+
+- Желательны SATA SSD или NVMe диски с конденсаторами (серверные SSD). Можно использовать и
+  десктопные SSD, включив режим отложенного fsync, но производительность однопоточной записи
+  в этом случае пострадает.
+- Быстрая сеть, минимум 10 гбит/с
+- Для наилучшей производительности нужно отключить энергосбережение CPU: `cpupower idle-set -D 0 && cpupower frequency-set -g performance`.
+- Пропишите нужные вам значения вверху файлов `/usr/lib/vitastor/mon/make-units.sh` и `/usr/lib/vitastor/mon/make-osd.sh`.
+- Создайте юниты systemd для etcd и мониторов: `/usr/lib/vitastor/mon/make-units.sh`
+- Создайте юниты для OSD: `/usr/lib/vitastor/mon/make-osd.sh /dev/disk/by-partuuid/XXX [/dev/disk/by-partuuid/YYY ...]`
+- Вы можете поменять параметры OSD в юнитах systemd. Смысл некоторых параметров:
+  - `disable_data_fsync 1` - отключает fsync, используется с SSD с конденсаторами.
+  - `immediate_commit all` - используется с SSD с конденсаторами.
+  - `disable_device_lock 1` - отключает блокировку файла устройства, нужно, только если вы запускаете
+    несколько OSD на одном блочном устройстве.
+  - `flusher_count 256` - "flusher" - микропоток, удаляющий старые данные из журнала.
+    Не волнуйтесь об этой настройке, 256 теперь достаточно практически всегда.
+  - `disk_alignment`, `journal_block_size`, `meta_block_size` следует установить равными размеру
+    внутреннего блока SSD. Это почти всегда 4096.
+  - `journal_no_same_sector_overwrites true` запрещает перезапись одного и того же сектора журнала подряд
+    много раз в процессе записи. Большинство (99%) SSD не нуждаются в данной опции. Однако выяснилось, что
+    диски, используемые на одном из тестовых стендов - Intel D3-S4510 - очень сильно не любят такую
+    перезапись, и для них была добавлена эта опция. Когда данный режим включён, также нужно поднимать
+    значение `journal_sector_buffer_count`, так как иначе Vitastor не хватит буферов для записи в журнал.
+- Запустите все etcd: `systemctl start etcd`
+- Создайте глобальную конфигурацию в etcd: `etcdctl --endpoints=... put /vitastor/config/global '{"immediate_commit":"all"}'`
+  (если все ваши диски - серверные с конденсаторами).
+- Создайте пулы: `etcdctl --endpoints=... put /vitastor/config/pools '{"1":{"name":"testpool","scheme":"replicated","pg_size":2,"pg_minsize":1,"pg_count":256,"failure_domain":"host"}}'`.
+  Для jerasure EC-пулов конфигурация должна выглядеть так: `2:{"name":"ecpool","scheme":"jerasure","pg_size":4,"parity_chunks":2,"pg_minsize":2,"pg_count":256,"failure_domain":"host"}`.
+- Запустите все OSD: `systemctl start vitastor.target`
+- Ваш кластер должен быть готов - один из мониторов должен уже сконфигурировать PG, а OSD должны запустить их.
+- Вы можете проверить состояние PG прямо в etcd: `etcdctl --endpoints=... get --prefix /vitastor/pg/state`. Все PG должны быть 'active'.
+
+### Задать имя образу
+
+```
+etcdctl --endpoints=<etcd> put /vitastor/config/inode/<pool>/<inode> '{"name":"<name>","size":<size>[,"parent_id":<parent_inode_number>][,"readonly":true]}'
+```
+
+Например:
+
+```
+etcdctl --endpoints=http://10.115.0.10:2379/v3 put /vitastor/config/inode/1/1 '{"name":"testimg","size":2147483648}'
+```
+
+Если вы зададите parent_id, то образ станет CoW-клоном, т.е. все новые запросы записи пойдут в новый инод, а запросы
+чтения будут проверять сначала его, а потом родительские слои по цепочке вверх. Чтобы случайно не перезаписать данные
+в родительском слое, вы можете переключить его в режим "только чтение", добавив флаг `"readonly":true` в его запись
+метаданных. В таком случае родительский образ становится просто снапшотом.
+
+Таким образом, для создания снапшота вам нужно просто переименовать предыдущий inode (например, из testimg в testimg@0),
+сделать его readonly и создать новый слой с исходным именем образа (testimg), ссылающийся на только что переименованный
+в качестве родительского.
+
+### Запуск тестов с fio
+
+Пример команды для запуска тестов:
+
+```
+fio -thread -ioengine=libfio_vitastor.so -name=test -bs=4M -direct=1 -iodepth=16 -rw=write -etcd=10.115.0.10:2379/v3 -image=testimg
+```
+
+Если вы не хотите обращаться к образу по имени, вместо `-image=testimg` можно указать номер пула, номер инода и размер:
+`-pool=1 -inode=1 -size=400G`.
+
+### Загрузить образ диска ВМ в/из Vitastor
+
+Используйте qemu-img и строку `vitastor:etcd_host=<HOST>:image=<IMAGE>` в качестве имени файла диска. Например:
+
+```
+qemu-img convert -f qcow2 debian10.qcow2 -p -O raw 'vitastor:etcd_host=10.115.0.10\:2379/v3:image=testimg'
+```
+
+Обратите внимание, что если вы используете немодифицированный QEMU, потребуется установить переменную окружения
+`LD_PRELOAD=/usr/lib/x86_64-linux-gnu/qemu/block-vitastor.so`.
+
+Если вы не хотите обращаться к образу по имени, вместо `:image=<IMAGE>` можно указать номер пула, номер инода и размер:
+`:pool=<POOL>:inode=<INODE>:size=<SIZE>`.
+
+### Запустить ВМ
+
+Для запуска QEMU используйте опцию `-drive file=vitastor:etcd_host=<HOST>:image=<IMAGE>` (аналогично qemu-img)
+и физический размер блока 4 KB.
+
+Например:
+
+```
+qemu-system-x86_64 -enable-kvm -m 1024
+  -drive 'file=vitastor:etcd_host=10.115.0.10\:2379/v3:image=testimg',format=raw,if=none,id=drive-virtio-disk0,cache=none
+  -device virtio-blk-pci,scsi=off,bus=pci.0,addr=0x5,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1,write-cache=off,physical_block_size=4096,logical_block_size=512
+  -vnc 0.0.0.0:0
+```
+
+Обращение по номерам (`:pool=<POOL>:inode=<INODE>:size=<SIZE>` вместо `:image=<IMAGE>`) работает аналогично qemu-img.
+
+### Удалить образ
+
+Используйте утилиту vitastor-rm. Например:
+
+```
+vitastor-rm --etcd_address 10.115.0.10:2379/v3 --pool 1 --inode 1 --parallel_osds 16 --iodepth 32
+```
+
+### NBD
+
+Чтобы создать локальное блочное устройство, используйте NBD. Например:
+
+```
+vitastor-nbd map --etcd_address 10.115.0.10:2379/v3 --image testimg
+```
+
+Команда напечатает название устройства вида /dev/nbd0, которое потом можно будет форматировать
+и использовать как обычное блочное устройство.
+
+Для обращения по номеру инода, аналогично другим командам, можно использовать опции
+`--pool <POOL> --inode <INODE> --size <SIZE>` вместо `--image testimg`.
+
+### Kubernetes
+
+У Vitastor есть CSI-плагин для Kubernetes, поддерживающий RWO-тома.
+
+Для установки возьмите манифесты из директории [csi/deploy/](csi/deploy/), поместите
+вашу конфигурацию подключения к Vitastor в [csi/deploy/001-csi-config-map.yaml](001-csi-config-map.yaml),
+настройте StorageClass в [csi/deploy/009-storage-class.yaml](009-storage-class.yaml)
+и примените все `NNN-*.yaml` к вашей инсталляции Kubernetes.
+
+```
+for i in ./???-*.yaml; do kubectl apply -f $i; done
+```
+
+После этого вы сможете создавать PersistentVolume. Пример смотрите в файле [csi/deploy/example-pvc.yaml](csi/deploy/example-pvc.yaml).
+
+## Известные проблемы
+
+- Запросы удаления объектов могут в данный момент приводить к "неполным" объектам в EC-пулах,
+  если в процессе удаления произойдут отказы OSD или серверов, потому что правильная обработка
+  запросов удаления в кластере должна быть "трёхфазной", а это пока не реализовано. Если вы
+  столкнётесь с такой ситуацией, просто повторите запрос удаления.
+
+## Принципы реализации
+
+- Я люблю архитектурно простые решения. Vitastor проектируется именно так и я намерен
+  и далее следовать данному принципу.
+- Если вы пришли сюда за идеальным кодом на C++, вы, вероятно, не по адресу. "Общепринятые"
+  практики написания C++ кода меня не очень волнуют, так как зачастую, опять-таки, ведут к
+  излишним усложнениям и код получается красивый... но медленный.
+- По той же причине в коде иногда можно встретить велосипеды типа собственного упрощённого
+  HTTP-клиента для работы с etcd. Зато эти велосипеды маленькие и компактные и не требуют
+  использования десятка внешних библиотек.
+- node.js для монитора - не случайный выбор. Он очень быстрый, имеет встроенную событийную
+  машину, приятный нейтральный C-подобный язык программирования и развитую инфраструктуру.
+
+## Автор и лицензия
+
+Автор: Виталий Филиппов (vitalif [at] yourcmc.ru), 2019+
+
+Заходите в Telegram-чат Vitastor: https://t.me/vitastor
+
+Лицензия: VNPL 1.1 на серверный код и двойная VNPL 1.1 + GPL 2.0+ на клиентский.
+
+VNPL - "сетевой копилефт", собственная свободная копилефт-лицензия
+Vitastor Network Public License 1.1, основанная на GNU GPL 3.0 с дополнительным
+условием "Сетевого взаимодействия", требующим распространять все программы,
+специально разработанные для использования вместе с Vitastor и взаимодействующие
+с ним по сети, под лицензией VNPL или под любой другой свободной лицензией.
+
+Идея VNPL - расширение действия копилефта не только на модули, явным образом
+связываемые с кодом Vitastor, но также на модули, оформленные в виде микросервисов
+и взаимодействующие с ним по сети.
+
+Таким образом, если вы хотите построить на основе Vitastor сервис, содержаший
+компоненты с закрытым кодом, взаимодействующие с Vitastor, вам нужна коммерческая
+лицензия от автора 😀.
+
+На Windows и любое другое ПО, не разработанное *специально* для использования
+вместе с Vitastor, никакие ограничения не накладываются.
+
+Клиентские библиотеки распространяются на условиях двойной лицензии VNPL 1.0
+и также на условиях GNU GPL 2.0 или более поздней версии. Так сделано в целях
+совместимости с таким ПО, как QEMU и fio.
+
+Вы можете найти полный текст VNPL 1.1 в файле [VNPL-1.1.txt](VNPL-1.1.txt),
+а GPL 2.0 в файле [GPL-2.0.txt](GPL-2.0.txt).

README.md

@@ -0,0 +1,530 @@
+## Vitastor
+
+[Читать на русском](README-ru.md)
+
+## The Idea
+
+Make Software-Defined Block Storage Great Again.
+
+Vitastor is a small, simple and fast clustered block storage (storage for VM drives),
+architecturally similar to Ceph which means strong consistency, primary-replication, symmetric
+clustering and automatic data distribution over any number of drives of any size
+with configurable redundancy (replication or erasure codes/XOR).
+
+## Features
+
+Vitastor is currently a pre-release, a lot of features are missing and you can still expect
+breaking changes in the future. However, the following is implemented:
+
+- Basic part: highly-available block storage with symmetric clustering and no SPOF
+- Performance ;-D
+- Multiple redundancy schemes: Replication, XOR n+1, Reed-Solomon erasure codes
+  based on jerasure library with any number of data and parity drives in a group
+- Configuration via simple JSON data structures in etcd
+- Automatic data distribution over OSDs, with support for:
+  - Mathematical optimization for better uniformity and less data movement
+  - Multiple pools
+  - Placement tree, OSD selection by tags (device classes) and placement root
+  - Configurable failure domains
+- Recovery of degraded blocks
+- Rebalancing (data movement between OSDs)
+- Lazy fsync support
+- I/O statistics reporting to etcd
+- Generic user-space client library
+- QEMU driver (built out-of-tree)
+- Loadable fio engine for benchmarks (also built out-of-tree)
+- NBD proxy for kernel mounts
+- Inode removal tool (vitastor-rm)
+- Packaging for Debian and CentOS
+- Per-inode I/O and space usage statistics
+- Inode metadata storage in etcd
+- Snapshots and copy-on-write image clones
+- Write throttling to smooth random write workloads in SSD+HDD configurations
+- RDMA/RoCEv2 support via libibverbs
+- CSI plugin for Kubernetes
+- Basic OpenStack support: Cinder driver, Nova and libvirt patches
+
+## Roadmap
+
+- Snapshot deletion (layer merge) support
+- Better OSD creation and auto-start tools
+- Other administrative tools
+- Plugins for OpenNebula, Proxmox and other cloud systems
+- iSCSI proxy
+- Faster failover
+- Scrubbing without checksums (verification of replicas)
+- Checksums
+- Tiered storage
+- NVDIMM support
+- Web GUI
+- Compression (possibly)
+- Read caching using system page cache (possibly)
+
+## Architecture
+
+Similarities:
+
+- Just like Ceph, Vitastor has Pools, PGs, OSDs, Monitors, Failure Domains, Placement Tree.
+- Just like Ceph, Vitastor is transactional (even though there's a "lazy fsync mode" which
+  doesn't implicitly flush every operation to disks).
+- OSDs also have journal and metadata and they can also be put on separate drives.
+- Just like in Ceph, client library attempts to recover from any cluster failure so
+  you can basically reboot the whole cluster and only pause, but not crash, your clients
+  (I consider this a bug if the client crashes in that case).
+
+Some basic terms for people not familiar with Ceph:
+
+- OSD (Object Storage Daemon) is a process that stores data and serves read/write requests.
+- PG (Placement Group) is a container for data that (normally) shares the same replicas.
+- Pool is a container for data that has the same redundancy scheme and placement rules.
+- Monitor is a separate daemon that watches cluster state and handles failures.
+- Failure Domain is a group of OSDs that you allow to fail. It's "host" by default.
+- Placement Tree groups OSDs in a hierarchy to later split them into Failure Domains.
+
+Architectural differences from Ceph:
+
+- Vitastor's primary focus is on SSDs. Proper SSD+HDD optimizations may be added in the future, though.
+- Vitastor OSD is (and will always be) single-threaded. If you want to dedicate more than 1 core
+  per drive you should run multiple OSDs each on a different partition of the drive.
+  Vitastor isn't CPU-hungry though (as opposed to Ceph), so 1 core is sufficient in a lot of cases.
+- Metadata and journal are always kept in memory. Metadata size depends linearly on drive capacity
+  and data store block size which is 128 KB by default. With 128 KB blocks metadata should occupy
+  around 512 MB per 1 TB (which is still less than Ceph wants). Journal doesn't have to be big,
+  the example test below was conducted with only 16 MB journal. A big journal is probably even
+  harmful as dirty write metadata also take some memory.
+- Vitastor storage layer doesn't have internal copy-on-write or redirect-write. I know that maybe
+  it's possible to create a good copy-on-write storage, but it's much harder and makes performance
+  less deterministic, so CoW isn't used in Vitastor.
+- The basic layer of Vitastor is block storage with fixed-size blocks, not object storage with
+  rich semantics like in Ceph (RADOS).
+- There's a "lazy fsync" mode which allows to batch writes before flushing them to the disk.
+  This allows to use Vitastor with desktop SSDs, but still lowers performance due to additional
+  network roundtrips, so use server SSDs with capacitor-based power loss protection
+  ("Advanced Power Loss Protection") for best performance.
+- PGs are ephemeral. This means that they aren't stored on data disks and only exist in memory
+  while OSDs are running.
+- Recovery process is per-object (per-block), not per-PG. Also there are no PGLOGs.
+- Monitors don't store data. Cluster configuration and state is stored in etcd in simple human-readable
+  JSON structures. Monitors only watch cluster state and handle data movement.
+  Thus Vitastor's Monitor isn't a critical component of the system and is more similar to Ceph's Manager.
+  Vitastor's Monitor is implemented in node.js.
+- PG distribution isn't based on consistent hashes. All PG mappings are stored in etcd.
+  Rebalancing PGs between OSDs is done by mathematical optimization - data distribution problem
+  is reduced to a linear programming problem and solved by lp_solve. This allows for almost
+  perfect (96-99% uniformity compared to Ceph's 80-90%) data distribution in most cases, ability
+  to map PGs by hand without breaking rebalancing logic, reduced OSD peer-to-peer communication
+  (on average, OSDs have fewer peers) and less data movement. It also probably has a drawback -
+  this method may fail in very large clusters, but up to several hundreds of OSDs it's perfectly fine.
+  It's also easy to add consistent hashes in the future if something proves their necessity.
+- There's no separate CRUSH layer. You select pool redundancy scheme, placement root, failure domain
+  and so on directly in pool configuration.
+
+## Understanding Storage Performance
+
+The most important thing for fast storage is latency, not parallel iops.
+
+The best possible latency is achieved with one thread and queue depth of 1 which basically means
+"client load as low as possible". In this case IOPS = 1/latency, and this number doesn't
+scale with number of servers, drives, server processes or threads and so on.
+Single-threaded IOPS and latency numbers only depend on *how fast a single daemon is*.
+
+Why is it important? It's important because some of the applications *can't* use
+queue depth greater than 1 because their task isn't parallelizable. A notable example
+is any ACID DBMS because all of them write their WALs sequentially with fsync()s.
+
+fsync, by the way, is another important thing often missing in benchmarks. The point is
+that drives have cache buffers and don't guarantee that your data is actually persisted
+until you call fsync() which is translated to a FLUSH CACHE command by the OS.
+
+Desktop SSDs are very fast without fsync - NVMes, for example, can process ~80000 write
+operations per second with queue depth of 1 without fsync - but they're really slow with
+fsync because they have to actually write data to flash chips when you call fsync. Typical
+number is around 1000-2000 iops with fsync.
+
+Server SSDs often have supercapacitors that act as a built-in UPS and allow the drive
+to flush its DRAM cache to the persistent flash storage when a power loss occurs.
+This makes them perform equally well with and without fsync. This feature is called
+"Advanced Power Loss Protection" by Intel; other vendors either call it similarly
+or directly as "Full Capacitor-Based Power Loss Protection".
+
+All software-defined storages that I currently know are slow in terms of latency.
+Notable examples are Ceph and internal SDSes used by cloud providers like Amazon, Google,
+Yandex and so on. They're all slow and can only reach ~0.3ms read and ~0.6ms 4 KB write latency
+with best-in-slot hardware.
+
+And that's in the SSD era when you can buy an SSD that has ~0.04ms latency for 100 $.
+
+I use the following 6 commands with small variations to benchmark any storage:
+
+- Linear write:
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4M -iodepth=32 -rw=write -runtime=60 -filename=/dev/sdX`
+- Linear read:
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4M -iodepth=32 -rw=read -runtime=60 -filename=/dev/sdX`
+- Random write latency (T1Q1, this hurts storages the most):
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4k -iodepth=1 -fsync=1 -rw=randwrite -runtime=60 -filename=/dev/sdX`
+- Random read latency (T1Q1):
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4k -iodepth=1 -rw=randread -runtime=60 -filename=/dev/sdX`
+- Parallel write iops (use numjobs if a single CPU core is insufficient to saturate the load):
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4k -iodepth=128 [-numjobs=4 -group_reporting] -rw=randwrite -runtime=60 -filename=/dev/sdX`
+- Parallel read iops (use numjobs if a single CPU core is insufficient to saturate the load):
+  `fio -ioengine=libaio -direct=1 -invalidate=1 -name=test -bs=4k -iodepth=128 [-numjobs=4 -group_reporting] -rw=randread -runtime=60 -filename=/dev/sdX`
+
+## Vitastor's Theoretical Maximum Random Access Performance
+
+Replicated setups:
+- Single-threaded (T1Q1) read latency: 1 network roundtrip + 1 disk read.
+- Single-threaded write+fsync latency:
+  - With immediate commit: 2 network roundtrips + 1 disk write.
+  - With lazy commit: 4 network roundtrips + 1 disk write + 1 disk flush.
+- Saturated parallel read iops: min(network bandwidth, sum(disk read iops)).
+- Saturated parallel write iops: min(network bandwidth, sum(disk write iops / number of replicas / write amplification)).
+
+EC/XOR setups:
+- Single-threaded (T1Q1) read latency: 1.5 network roundtrips + 1 disk read.
+- Single-threaded write+fsync latency:
+  - With immediate commit: 3.5 network roundtrips + 1 disk read + 2 disk writes.
+  - With lazy commit: 5.5 network roundtrips + 1 disk read + 2 disk writes + 2 disk fsyncs.
+  - 0.5 in actually (k-1)/k which means that an additional roundtrip doesn't happen when
+    the read sub-operation can be served locally.
+- Saturated parallel read iops: min(network bandwidth, sum(disk read iops)).
+- Saturated parallel write iops: min(network bandwidth, sum(disk write iops * number of data drives / (number of data + parity drives) / write amplification)).
+  In fact, you should put disk write iops under the condition of ~10% reads / ~90% writes in this formula.
+
+Write amplification for 4 KB blocks is usually 3-5 in Vitastor:
+1. Journal block write
+2. Journal data write
+3. Metadata block write
+4. Another journal block write for EC/XOR setups
+5. Data block write
+
+If you manage to get an SSD which handles 512 byte blocks well (Optane?) you may
+lower 1, 3 and 4 to 512 bytes (1/8 of data size) and get WA as low as 2.375.
+
+Lazy fsync also reduces WA for parallel workloads because journal blocks are only
+written when they fill up or fsync is requested.
+
+## Example Comparison with Ceph
+
+Hardware configuration: 4 nodes, each with:
+- 6x SATA SSD Intel D3-4510 3.84 TB
+- 2x Xeon Gold 6242 (16 cores @ 2.8 GHz)
+- 384 GB RAM
+- 1x 25 GbE network interface (Mellanox ConnectX-4 LX), connected to a Juniper QFX5200 switch
+
+CPU powersaving was disabled. Both Vitastor and Ceph were configured with 2 OSDs per 1 SSD.
+
+All of the results below apply to 4 KB blocks and random access (unless indicated otherwise).
+
+Raw drive performance:
+- T1Q1 write ~27000 iops (~0.037ms latency)
+- T1Q1 read ~9800 iops (~0.101ms latency)
+- T1Q32 write ~60000 iops
+- T1Q32 read ~81700 iops
+
+Ceph 15.2.4 (Bluestore):
+- T1Q1 write ~1000 iops (~1ms latency)
+- T1Q1 read ~1750 iops (~0.57ms latency)
+- T8Q64 write ~100000 iops, total CPU usage by OSDs about 40 virtual cores on each node
+- T8Q64 read ~480000 iops, total CPU usage by OSDs about 40 virtual cores on each node
+
+T8Q64 tests were conducted over 8 400GB RBD images from all hosts (every host was running 2 instances of fio).
+This is because Ceph has performance penalties related to running multiple clients over a single RBD image.
+
+cephx_sign_messages was set to false during tests, RocksDB and Bluestore settings were left at defaults.
+
+In fact, not that bad for Ceph. These servers are an example of well-balanced Ceph nodes.
+However, CPU usage and I/O latency were through the roof, as usual.
+
+Vitastor:
+- T1Q1 write: 7087 iops (0.14ms latency)
+- T1Q1 read: 6838 iops (0.145ms latency)
+- T2Q64 write: 162000 iops, total CPU usage by OSDs about 3 virtual cores on each node
+- T8Q64 read: 895000 iops, total CPU usage by OSDs about 4 virtual cores on each node
+- Linear write (4M T1Q32): 2800 MB/s
+- Linear read (4M T1Q32): 1500 MB/s
+
+T8Q64 read test was conducted over 1 larger inode (3.2T) from all hosts (every host was running 2 instances of fio).
+Vitastor has no performance penalties related to running multiple clients over a single inode.
+If conducted from one node with all primary OSDs moved to other nodes the result was slightly lower (689000 iops),
+this is because all operations resulted in network roundtrips between the client and the primary OSD.
+When fio was colocated with OSDs (like in Ceph benchmarks above), 1/4 of the read workload actually
+used the loopback network.
+
+Vitastor was configured with: `--disable_data_fsync true --immediate_commit all --flusher_count 8
+  --disk_alignment 4096 --journal_block_size 4096 --meta_block_size 4096
+  --journal_no_same_sector_overwrites true --journal_sector_buffer_count 1024
+  --journal_size 16777216`.
+
+### EC/XOR 2+1
+
+Vitastor:
+- T1Q1 write: 2808 iops (~0.355ms latency)
+- T1Q1 read: 6190 iops (~0.16ms latency)
+- T2Q64 write: 85500 iops, total CPU usage by OSDs about 3.4 virtual cores on each node
+- T8Q64 read: 812000 iops, total CPU usage by OSDs about 4.7 virtual cores on each node
+- Linear write (4M T1Q32): 3200 MB/s
+- Linear read (4M T1Q32): 1800 MB/s
+
+Ceph:
+- T1Q1 write: 730 iops (~1.37ms latency)
+- T1Q1 read: 1500 iops with cold cache (~0.66ms latency), 2300 iops after 2 minute metadata cache warmup (~0.435ms latency)
+- T4Q128 write (4 RBD images): 45300 iops, total CPU usage by OSDs about 30 virtual cores on each node
+- T8Q64 read (4 RBD images): 278600 iops, total CPU usage by OSDs about 40 virtual cores on each node
+- Linear write (4M T1Q32): 1950 MB/s before preallocation, 2500 MB/s after preallocation
+- Linear read (4M T1Q32): 2400 MB/s
+
+### NBD
+
+NBD is currently required to mount Vitastor via kernel, but it imposes additional overhead
+due to additional copying between the kernel and userspace. This mostly hurts linear
+bandwidth, not iops.
+
+Vitastor with single-thread NBD on the same hardware:
+- T1Q1 write: 6000 iops (0.166ms latency)
+- T1Q1 read: 5518 iops (0.18ms latency)
+- T1Q128 write: 94400 iops
+- T1Q128 read: 103000 iops
+- Linear write (4M T1Q128): 1266 MB/s (compared to 2800 MB/s via fio)
+- Linear read (4M T1Q128): 975 MB/s (compared to 1500 MB/s via fio)
+
+## Installation
+
+### Debian
+
+- Trust Vitastor package signing key:
+  `wget -q -O - https://vitastor.io/debian/pubkey | sudo apt-key add -`
+- Add Vitastor package repository to your /etc/apt/sources.list:
+  - Debian 11 (Bullseye/Sid): `deb https://vitastor.io/debian bullseye main`
+  - Debian 10 (Buster): `deb https://vitastor.io/debian buster main`
+- For Debian 10 (Buster) also enable backports repository:
+  `deb http://deb.debian.org/debian buster-backports main`
+- Install packages: `apt update; apt install vitastor lp-solve etcd linux-image-amd64 qemu`
+
+### CentOS
+
+- Add Vitastor package repository:
+  - CentOS 7: `yum install https://vitastor.io/rpms/centos/7/vitastor-release-1.0-1.el7.noarch.rpm`
+  - CentOS 8: `dnf install https://vitastor.io/rpms/centos/8/vitastor-release-1.0-1.el8.noarch.rpm`
+- Enable EPEL: `yum/dnf install epel-release`
+- Enable additional CentOS repositories:
+  - CentOS 7: `yum install centos-release-scl`
+  - CentOS 8: `dnf install centos-release-advanced-virtualization`
+- Enable elrepo-kernel:
+  - CentOS 7: `yum install https://www.elrepo.org/elrepo-release-7.el7.elrepo.noarch.rpm`
+  - CentOS 8: `dnf install https://www.elrepo.org/elrepo-release-8.el8.elrepo.noarch.rpm`
+- Install packages: `yum/dnf install vitastor lpsolve etcd kernel-ml qemu-kvm`
+
+### Building from Source
+
+- Install Linux kernel 5.4 or newer, for io_uring support. 5.8 or later is highly recommended because
+  there is at least one known io_uring hang with 5.4 and an HP SmartArray controller.
+- Install liburing 0.4 or newer and its headers.
+- Install lp_solve.
+- Install etcd, at least version 3.4.15. Earlier versions won't work because of various bugs,
+  for example [#12402](https://github.com/etcd-io/etcd/pull/12402). You can also take 3.4.13
+  with this specific fix from here: https://github.com/vitalif/etcd/, branch release-3.4.
+- Install node.js 10 or newer.
+- Install gcc and g++ 8.x or newer.
+- Clone https://yourcmc.ru/git/vitalif/vitastor/ with submodules.
+- Install QEMU 3.0+, get its source, begin to build it, stop the build and copy headers:
+   - `<qemu>/include` &rarr; `<vitastor>/qemu/include`
+   - Debian:
+      * Use qemu packages from the main repository
+      * `<qemu>/b/qemu/config-host.h` &rarr; `<vitastor>/qemu/b/qemu/config-host.h`
+      * `<qemu>/b/qemu/qapi` &rarr; `<vitastor>/qemu/b/qemu/qapi`
+   - CentOS 8:
+      * Use qemu packages from the Advanced-Virtualization repository. To enable it, run
+        `yum install centos-release-advanced-virtualization.noarch` and then `yum install qemu`
+      * `<qemu>/config-host.h` &rarr; `<vitastor>/qemu/b/qemu/config-host.h`
+      * For QEMU 3.0+: `<qemu>/qapi` &rarr; `<vitastor>/qemu/b/qemu/qapi`
+      * For QEMU 2.0+: `<qemu>/qapi-types.h` &rarr; `<vitastor>/qemu/b/qemu/qapi-types.h`
+   - `config-host.h` and `qapi` are required because they contain generated headers
+- You can also rebuild QEMU with a patch that makes LD_PRELOAD unnecessary to load vitastor driver.
+  See `patches/qemu-*.*-vitastor.patch`.
+- Install fio 3.7 or later, get its source and symlink it into `<vitastor>/fio`.
+- Build & install Vitastor with `mkdir build && cd build && cmake .. && make -j8 && make install`.
+  Pay attention to the `QEMU_PLUGINDIR` cmake option - it must be set to `qemu-kvm` on RHEL.
+
+## Running
+
+Please note that startup procedure isn't currently simple - you specify configuration
+and calculate disk offsets almost by hand. This will be fixed in near future.
+
+- Get some SATA or NVMe SSDs with capacitors (server-grade drives). You can use desktop SSDs
+  with lazy fsync, but prepare for inferior single-thread latency.
+- Get a fast network (at least 10 Gbit/s).
+- Disable CPU powersaving: `cpupower idle-set -D 0 && cpupower frequency-set -g performance`.
+- Check `/usr/lib/vitastor/mon/make-units.sh` and `/usr/lib/vitastor/mon/make-osd.sh` and
+  put desired values into the variables at the top of these files.
+- Create systemd units for the monitor and etcd: `/usr/lib/vitastor/mon/make-units.sh`
+- Create systemd units for your OSDs: `/usr/lib/vitastor/mon/make-osd.sh /dev/disk/by-partuuid/XXX [/dev/disk/by-partuuid/YYY ...]`
+- You can edit the units and change OSD configuration. Notable configuration variables:
+  - `disable_data_fsync 1` - only safe with server-grade drives with capacitors.
+  - `immediate_commit all` - use this if all your drives are server-grade.
+  - `disable_device_lock 1` - only required if you run multiple OSDs on one block device.
+  - `flusher_count 256` - flusher is a micro-thread that removes old data from the journal.
+    You don't have to worry about this parameter anymore, 256 is enough.
+  - `disk_alignment`, `journal_block_size`, `meta_block_size` should be set to the internal
+    block size of your SSDs which is 4096 on most drives.
+  - `journal_no_same_sector_overwrites true` prevents multiple overwrites of the same journal sector.
+    Most (99%) SSDs don't need this option. But Intel D3-4510 does because it doesn't like when you
+    overwrite the same sector twice in a short period of time. The setting forces Vitastor to never
+    overwrite the same journal sector twice in a row which makes D3-4510 almost happy. Not totally
+    happy, because overwrites of the same block can still happen in the metadata area... When this
+    setting is set, it is also required to raise `journal_sector_buffer_count` setting, which is the
+    number of dirty journal sectors that may be written to at the same time.
+- `systemctl start vitastor.target` everywhere.
+- Create global configuration in etcd: `etcdctl --endpoints=... put /vitastor/config/global '{"immediate_commit":"all"}'`
+  (if all your drives have capacitors).
+- Create pool configuration in etcd: `etcdctl --endpoints=... put /vitastor/config/pools '{"1":{"name":"testpool","scheme":"replicated","pg_size":2,"pg_minsize":1,"pg_count":256,"failure_domain":"host"}}'`.
+  For jerasure pools the configuration should look like the following: `2:{"name":"ecpool","scheme":"jerasure","pg_size":4,"parity_chunks":2,"pg_minsize":2,"pg_count":256,"failure_domain":"host"}`.
+- At this point, one of the monitors will configure PGs and OSDs will start them.
+- You can check PG states with `etcdctl --endpoints=... get --prefix /vitastor/pg/state`. All PGs should become 'active'.
+
+### Name an image
+
+```
+etcdctl --endpoints=<etcd> put /vitastor/config/inode/<pool>/<inode> '{"name":"<name>","size":<size>[,"parent_id":<parent_inode_number>][,"readonly":true]}'
+```
+
+For example:
+
+```
+etcdctl --endpoints=http://10.115.0.10:2379/v3 put /vitastor/config/inode/1/1 '{"name":"testimg","size":2147483648}'
+```
+
+If you specify parent_id the image becomes a CoW clone. I.e. all writes go to the new inode and reads first check it
+and then upper layers. You can then make parent readonly by updating its entry with `"readonly":true` for safety and
+basically treat it as a snapshot.
+
+So to create a snapshot you basically rename the previous upper layer (for example from testimg to testimg@0), make it readonly
+and create a new top layer with the original name (testimg) and the previous one as a parent.
+
+### Run fio benchmarks
+
+fio command example:
+
+```
+fio -thread -ioengine=libfio_vitastor.so -name=test -bs=4M -direct=1 -iodepth=16 -rw=write -etcd=10.115.0.10:2379/v3 -image=testimg
+```
+
+If you don't want to access your image by name, you can specify pool number, inode number and size
+(`-pool=1 -inode=1 -size=400G`) instead of the image name (`-image=testimg`).
+
+### Upload VM image
+
+Use qemu-img and `vitastor:etcd_host=<HOST>:image=<IMAGE>` disk filename. For example:
+
+```
+qemu-img convert -f qcow2 debian10.qcow2 -p -O raw 'vitastor:etcd_host=10.115.0.10\:2379/v3:image=testimg'
+```
+
+Note that the command requires to be run with `LD_PRELOAD=/usr/lib/x86_64-linux-gnu/qemu/block-vitastor.so qemu-img ...`
+if you use unmodified QEMU.
+
+You can also specify `:pool=<POOL>:inode=<INODE>:size=<SIZE>` instead of `:image=<IMAGE>`
+if you don't want to use inode metadata.
+
+### Start a VM
+
+Run QEMU with `-drive file=vitastor:etcd_host=<HOST>:image=<IMAGE>` and use 4 KB physical block size.
+
+For example:
+
+```
+qemu-system-x86_64 -enable-kvm -m 1024
+  -drive 'file=vitastor:etcd_host=10.115.0.10\:2379/v3:image=testimg',format=raw,if=none,id=drive-virtio-disk0,cache=none
+  -device virtio-blk-pci,scsi=off,bus=pci.0,addr=0x5,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1,write-cache=off,physical_block_size=4096,logical_block_size=512
+  -vnc 0.0.0.0:0
+```
+
+You can also specify `:pool=<POOL>:inode=<INODE>:size=<SIZE>` instead of `:image=<IMAGE>`,
+just like in qemu-img.
+
+### Remove inode
+
+Use vitastor-rm. For example:
+
+```
+vitastor-rm --etcd_address 10.115.0.10:2379/v3 --pool 1 --inode 1 --parallel_osds 16 --iodepth 32
+```
+
+### NBD
+
+To create a local block device for a Vitastor image, use NBD. For example:
+
+```
+vitastor-nbd map --etcd_address 10.115.0.10:2379/v3 --image testimg
+```
+
+It will output the device name, like /dev/nbd0 which you can then format and mount as a normal block device.
+
+Again, you can use `--pool <POOL> --inode <INODE> --size <SIZE>` insteaf of `--image <IMAGE>` if you want.
+
+### Kubernetes
+
+Vitastor has a CSI plugin for Kubernetes which supports RWO volumes.
+
+To deploy it, take manifests from [csi/deploy/](csi/deploy/) directory, put your
+Vitastor configuration in [csi/deploy/001-csi-config-map.yaml](001-csi-config-map.yaml),
+configure storage class in [csi/deploy/009-storage-class.yaml](009-storage-class.yaml)
+and apply all `NNN-*.yaml` manifests to your Kubernetes installation:
+
+```
+for i in ./???-*.yaml; do kubectl apply -f $i; done
+```
+
+After that you'll be able to create PersistentVolumes. See example in [csi/deploy/example-pvc.yaml](csi/deploy/example-pvc.yaml).
+
+## Known Problems
+
+- Object deletion requests may currently lead to 'incomplete' objects in EC pools
+  if your OSDs crash during deletion because proper handling of object cleanup
+  in a cluster should be "three-phase" and it's currently not implemented.
+  Just repeat the removal request again in this case.
+
+## Implementation Principles
+
+- I like architecturally simple solutions. Vitastor is and will always be designed
+  exactly like that.
+- I also like reinventing the wheel to some extent, like writing my own HTTP client
+  for etcd interaction instead of using prebuilt libraries, because in this case
+  I'm confident about what my code does and what it doesn't do.
+- I don't care about C++ "best practices" like RAII or proper inheritance or usage of
+  smart pointers or whatever and I don't intend to change my mind, so if you're here
+  looking for ideal reference C++ code, this probably isn't the right place.
+- I like node.js better than any other dynamically-typed language interpreter
+  because it's faster than any other interpreter in the world, has neutral C-like
+  syntax and built-in event loop. That's why Monitor is implemented in node.js.
+
+## Author and License
+
+Copyright (c) Vitaliy Filippov (vitalif [at] yourcmc.ru), 2019+
+
+Join Vitastor Telegram Chat: https://t.me/vitastor
+
+All server-side code (OSD, Monitor and so on) is licensed under the terms of
+Vitastor Network Public License 1.1 (VNPL 1.1), a copyleft license based on
+GNU GPLv3.0 with the additional "Network Interaction" clause which requires
+opensourcing all programs directly or indirectly interacting with Vitastor
+through a computer network and expressly designed to be used in conjunction
+with it ("Proxy Programs"). Proxy Programs may be made public not only under
+the terms of the same license, but also under the terms of any GPL-Compatible
+Free Software License, as listed by the Free Software Foundation.
+This is a stricter copyleft license than the Affero GPL.
+
+Please note that VNPL doesn't require you to open the code of proprietary
+software running inside a VM if it's not specially designed to be used with
+Vitastor.
+
+Basically, you can't use the software in a proprietary environment to provide
+its functionality to users without opensourcing all intermediary components
+standing between the user and Vitastor or purchasing a commercial license
+from the author 😀.
+
+Client libraries (cluster_client and so on) are dual-licensed under the same
+VNPL 1.1 and also GNU GPL 2.0 or later to allow for compatibility with GPLed
+software like QEMU and fio.
+
+You can find the full text of VNPL-1.1 in the file [VNPL-1.1.txt](VNPL-1.1.txt).
+GPL 2.0 is also included in this repository as [GPL-2.0.txt](GPL-2.0.txt).

VNPL-1.1.txt

@@ -0,0 +1,648 @@
+                     VITASTOR NETWORK PUBLIC LICENSE
+                     Version 1.1,  6 February 2021
+
+ Copyright (C) 2021 Vitaliy Filippov <vitalif@yourcmc.ru>
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The Vitastor Network Public License is a free, copyleft license for
+software and other kinds of works, specifically designed to ensure
+cooperation with the community in the case of network server software.
+
+  The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works.  By contrast,
+GNU General Public Licenses and Vitastor Network Public License are
+intended to guarantee your freedom to share and change all versions
+of a program--to make sure it remains free software for all its users.
+
+  When we speak of free software, we are referring to freedom, not
+price.  GNU General Public Licenses and Vitastor Network Public License
+are designed to make sure that you have the freedom to distribute copies
+of free software (and charge for them if you wish), that you receive
+source code or can get it if you want it, that you can change the software
+or use pieces of it in new free programs, and that you know you can do these
+things.
+
+  Developers that use GNU General Public Licenses and Vitastor
+Network Public License protect your rights with two steps:
+(1) assert copyright on the software, and (2) offer
+you this License which gives you legal permission to copy, distribute
+and/or modify the software.
+
+  A secondary benefit of defending all users' freedom is that
+improvements made in alternate versions of the program, if they
+receive widespread use, become available for other developers to
+incorporate.  Many developers of free software are heartened and
+encouraged by the resulting cooperation.  However, in the case of
+software used on network servers, this result may fail to come about.
+The GNU General Public License permits making a modified version and
+letting the public access it on a server without ever releasing its
+source code to the public. Even the GNU Affero General Public License
+permits running a modified version in a closed environment where
+public users only interact with it through a closed-source proxy, again,
+without making the program and the proxy available to the public
+for free.
+
+  The Vitastor Network Public License is designed specifically to
+ensure that, in such cases, the modified program and the proxy stays
+available to the community. It requires the operator of a network server to
+provide the source code of the original program and all other programs
+communicating with it running there to the users of that server.
+Therefore, public use of a modified version, on a server accessible
+directly or indirectly to the public, gives the public access to the source
+code of the modified version.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                       TERMS AND CONDITIONS
+
+  0. Definitions.
+
+  "This License" refers to version 1 of the Vitastor Network Public License.
+
+  "Copyright" also means copyright-like laws that apply to other kinds of
+works, such as semiconductor masks.
+
+  "The Program" refers to any copyrightable work licensed under this
+License.  Each licensee is addressed as "you".  "Licensees" and
+"recipients" may be individuals or organizations.
+
+  To "modify" a work means to copy from or adapt all or part of the work
+in a fashion requiring copyright permission, other than the making of an
+exact copy.  The resulting work is called a "modified version" of the
+earlier work or a work "based on" the earlier work.
+
+  A "covered work" means either the unmodified Program or a work based
+on the Program.
+
+  To "propagate" a work means to do anything with it that, without
+permission, would make you directly or secondarily liable for
+infringement under applicable copyright law, except executing it on a
+computer or modifying a private copy.  Propagation includes copying,
+distribution (with or without modification), making available to the
+public, and in some countries other activities as well.
+
+  To "convey" a work means any kind of propagation that enables other
+parties to make or receive copies.  Mere interaction with a user through
+a computer network, with no transfer of a copy, is not conveying.
+
+  An interactive user interface displays "Appropriate Legal Notices"
+to the extent that it includes a convenient and prominently visible
+feature that (1) displays an appropriate copyright notice, and (2)
+tells the user that there is no warranty for the work (except to the
+extent that warranties are provided), that licensees may convey the
+work under this License, and how to view a copy of this License.  If
+the interface presents a list of user commands or options, such as a
+menu, a prominent item in the list meets this criterion.
+
+  1. Source Code.
+
+  The "source code" for a work means the preferred form of the work
+for making modifications to it.  "Object code" means any non-source
+form of a work.
+
+  A "Standard Interface" means an interface that either is an official
+standard defined by a recognized standards body, or, in the case of
+interfaces specified for a particular programming language, one that
+is widely used among developers working in that language.
+
+  The "System Libraries" of an executable work include anything, other
+than the work as a whole, that (a) is included in the normal form of
+packaging a Major Component, but which is not part of that Major
+Component, and (b) serves only to enable use of the work with that
+Major Component, or to implement a Standard Interface for which an
+implementation is available to the public in source code form.  A
+"Major Component", in this context, means a major essential component
+(kernel, window system, and so on) of the specific operating system
+(if any) on which the executable work runs, or a compiler used to
+produce the work, or an object code interpreter used to run it.
+
+  The "Corresponding Source" for a work in object code form means all
+the source code needed to generate, install, and (for an executable
+work) run the object code and to modify the work, including scripts to
+control those activities.  However, it does not include the work's
+System Libraries, or general-purpose tools or generally available free
+programs which are used unmodified in performing those activities but
+which are not part of the work.  For example, Corresponding Source
+includes interface definition files associated with source files for
+the work, and the source code for shared libraries and dynamically
+linked subprograms that the work is specifically designed to require,
+such as by intimate data communication or control flow between those
+subprograms and other parts of the work.
+
+  The Corresponding Source need not include anything that users
+can regenerate automatically from other parts of the Corresponding
+Source.
+
+  The Corresponding Source for a work in source code form is that
+same work.
+
+  2. Basic Permissions.
+
+  All rights granted under this License are granted for the term of
+copyright on the Program, and are irrevocable provided the stated
+conditions are met.  This License explicitly affirms your unlimited
+permission to run the unmodified Program.  The output from running a
+covered work is covered by this License only if the output, given its
+content, constitutes a covered work.  This License acknowledges your
+rights of fair use or other equivalent, as provided by copyright law.
+
+  You may make, run and propagate covered works that you do not
+convey, without conditions so long as your license otherwise remains
+in force.  You may convey covered works to others for the sole purpose
+of having them make modifications exclusively for you, or provide you
+with facilities for running those works, provided that you comply with
+the terms of this License in conveying all material for which you do
+not control copyright.  Those thus making or running the covered works
+for you must do so exclusively on your behalf, under your direction
+and control, on terms that prohibit them from making any copies of
+your copyrighted material outside their relationship with you.
+
+  Conveying under any other circumstances is permitted solely under
+the conditions stated below.  Sublicensing is not allowed; section 10
+makes it unnecessary.
+
+  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+
+  No covered work shall be deemed part of an effective technological
+measure under any applicable law fulfilling obligations under article
+11 of the WIPO copyright treaty adopted on 20 December 1996, or
+similar laws prohibiting or restricting circumvention of such
+measures.
+
+  When you convey a covered work, you waive any legal power to forbid
+circumvention of technological measures to the extent such circumvention
+is effected by exercising rights under this License with respect to
+the covered work, and you disclaim any intention to limit operation or
+modification of the work as a means of enforcing, against the work's
+users, your or third parties' legal rights to forbid circumvention of
+technological measures.
+
+  4. Conveying Verbatim Copies.
+
+  You may convey verbatim copies of the Program's source code as you
+receive it, in any medium, provided that you conspicuously and
+appropriately publish on each copy an appropriate copyright notice;
+keep intact all notices stating that this License and any
+non-permissive terms added in accord with section 7 apply to the code;
+keep intact all notices of the absence of any warranty; and give all
+recipients a copy of this License along with the Program.
+
+  You may charge any price or no price for each copy that you convey,
+and you may offer support or warranty protection for a fee.
+
+  5. Conveying Modified Source Versions.
+
+  You may convey a work based on the Program, or the modifications to
+produce it from the Program, in the form of source code under the
+terms of section 4, provided that you also meet all of these conditions:
+
+    a) The work must carry prominent notices stating that you modified
+    it, and giving a relevant date.
+
+    b) The work must carry prominent notices stating that it is
+    released under this License and any conditions added under section
+    7.  This requirement modifies the requirement in section 4 to
+    "keep intact all notices".
+
+    c) You must license the entire work, as a whole, under this
+    License to anyone who comes into possession of a copy.  This
+    License will therefore apply, along with any applicable section 7
+    additional terms, to the whole of the work, and all its parts,
+    regardless of how they are packaged.  This License gives no
+    permission to license the work in any other way, but it does not
+    invalidate such permission if you have separately received it.
+
+    d) If the work has interactive user interfaces, each must display
+    Appropriate Legal Notices; however, if the Program has interactive
+    interfaces that do not display Appropriate Legal Notices, your
+    work need not make them do so.
+
+  A compilation of a covered work with other separate and independent
+works, which are not by their nature extensions of the covered work,
+and which are not combined with it such as to form a larger program,
+in or on a volume of a storage or distribution medium, is called an
+"aggregate" if the compilation and its resulting copyright are not
+used to limit the access or legal rights of the compilation's users
+beyond what the individual works permit.  Inclusion of a covered work
+in an aggregate does not cause this License to apply to the other
+parts of the aggregate.
+
+  6. Conveying Non-Source Forms.
+
+  You may convey a covered work in object code form under the terms
+of sections 4 and 5, provided that you also convey the
+machine-readable Corresponding Source under the terms of this License,
+in one of these ways:
+
+    a) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by the
+    Corresponding Source fixed on a durable physical medium
+    customarily used for software interchange.
+
+    b) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by a
+    written offer, valid for at least three years and valid for as
+    long as you offer spare parts or customer support for that product
+    model, to give anyone who possesses the object code either (1) a
+    copy of the Corresponding Source for all the software in the
+    product that is covered by this License, on a durable physical
+    medium customarily used for software interchange, for a price no
+    more than your reasonable cost of physically performing this
+    conveying of source, or (2) access to copy the
+    Corresponding Source from a network server at no charge.
+
+    c) Convey individual copies of the object code with a copy of the
+    written offer to provide the Corresponding Source.  This
+    alternative is allowed only occasionally and noncommercially, and
+    only if you received the object code with such an offer, in accord
+    with subsection 6b.
+
+    d) Convey the object code by offering access from a designated
+    place (gratis or for a charge), and offer equivalent access to the
+    Corresponding Source in the same way through the same place at no
+    further charge.  You need not require recipients to copy the
+    Corresponding Source along with the object code.  If the place to
+    copy the object code is a network server, the Corresponding Source
+    may be on a different server (operated by you or a third party)
+    that supports equivalent copying facilities, provided you maintain
+    clear directions next to the object code saying where to find the
+    Corresponding Source.  Regardless of what server hosts the
+    Corresponding Source, you remain obligated to ensure that it is
+    available for as long as needed to satisfy these requirements.
+
+    e) Convey the object code using peer-to-peer transmission, provided
+    you inform other peers where the object code and Corresponding
+    Source of the work are being offered to the general public at no
+    charge under subsection 6d.
+
+  A separable portion of the object code, whose source code is excluded
+from the Corresponding Source as a System Library, need not be
+included in conveying the object code work.
+
+  A "User Product" is either (1) a "consumer product", which means any
+tangible personal property which is normally used for personal, family,
+or household purposes, or (2) anything designed or sold for incorporation
+into a dwelling.  In determining whether a product is a consumer product,
+doubtful cases shall be resolved in favor of coverage.  For a particular
+product received by a particular user, "normally used" refers to a
+typical or common use of that class of product, regardless of the status
+of the particular user or of the way in which the particular user
+actually uses, or expects or is expected to use, the product.  A product
+is a consumer product regardless of whether the product has substantial
+commercial, industrial or non-consumer uses, unless such uses represent
+the only significant mode of use of the product.
+
+  "Installation Information" for a User Product means any methods,
+procedures, authorization keys, or other information required to install
+and execute modified versions of a covered work in that User Product from
+a modified version of its Corresponding Source.  The information must
+suffice to ensure that the continued functioning of the modified object
+code is in no case prevented or interfered with solely because
+modification has been made.
+
+  If you convey an object code work under this section in, or with, or
+specifically for use in, a User Product, and the conveying occurs as
+part of a transaction in which the right of possession and use of the
+User Product is transferred to the recipient in perpetuity or for a
+fixed term (regardless of how the transaction is characterized), the
+Corresponding Source conveyed under this section must be accompanied
+by the Installation Information.  But this requirement does not apply
+if neither you nor any third party retains the ability to install
+modified object code on the User Product (for example, the work has
+been installed in ROM).
+
+  The requirement to provide Installation Information does not include a
+requirement to continue to provide support service, warranty, or updates
+for a work that has been modified or installed by the recipient, or for
+the User Product in which it has been modified or installed.  Access to a
+network may be denied when the modification itself materially and
+adversely affects the operation of the network or violates the rules and
+protocols for communication across the network.
+
+  Corresponding Source conveyed, and Installation Information provided,
+in accord with this section must be in a format that is publicly
+documented (and with an implementation available to the public in
+source code form), and must require no special password or key for
+unpacking, reading or copying.
+
+  7. Additional Terms.
+
+  "Additional permissions" are terms that supplement the terms of this
+License by making exceptions from one or more of its conditions.
+Additional permissions that are applicable to the entire Program shall
+be treated as though they were included in this License, to the extent
+that they are valid under applicable law.  If additional permissions
+apply only to part of the Program, that part may be used separately
+under those permissions, but the entire Program remains governed by
+this License without regard to the additional permissions.
+
+  When you convey a copy of a covered work, you may at your option
+remove any additional permissions from that copy, or from any part of
+it.  (Additional permissions may be written to require their own
+removal in certain cases when you modify the work.)  You may place
+additional permissions on material, added by you to a covered work,
+for which you have or can give appropriate copyright permission.
+
+  Notwithstanding any other provision of this License, for material you
+add to a covered work, you may (if authorized by the copyright holders of
+that material) supplement the terms of this License with terms:
+
+    a) Disclaiming warranty or limiting liability differently from the
+    terms of sections 15 and 16 of this License; or
+
+    b) Requiring preservation of specified reasonable legal notices or
+    author attributions in that material or in the Appropriate Legal
+    Notices displayed by works containing it; or
+
+    c) Prohibiting misrepresentation of the origin of that material, or
+    requiring that modified versions of such material be marked in
+    reasonable ways as different from the original version; or
+
+    d) Limiting the use for publicity purposes of names of licensors or
+    authors of the material; or
+
+    e) Declining to grant rights under trademark law for use of some
+    trade names, trademarks, or service marks; or
+
+    f) Requiring indemnification of licensors and authors of that
+    material by anyone who conveys the material (or modified versions of
+    it) with contractual assumptions of liability to the recipient, for
+    any liability that these contractual assumptions directly impose on
+    those licensors and authors.
+
+  All other non-permissive additional terms are considered "further
+restrictions" within the meaning of section 10.  If the Program as you
+received it, or any part of it, contains a notice stating that it is
+governed by this License along with a term that is a further
+restriction, you may remove that term.  If a license document contains
+a further restriction but permits relicensing or conveying under this
+License, you may add to a covered work material governed by the terms
+of that license document, provided that the further restriction does
+not survive such relicensing or conveying.
+
+  If you add terms to a covered work in accord with this section, you
+must place, in the relevant source files, a statement of the
+additional terms that apply to those files, or a notice indicating
+where to find the applicable terms.
+
+  Additional terms, permissive or non-permissive, may be stated in the
+form of a separately written license, or stated as exceptions;
+the above requirements apply either way.
+
+  8. Termination.
+
+  You may not propagate or modify a covered work except as expressly
+provided under this License.  Any attempt otherwise to propagate or
+modify it is void, and will automatically terminate your rights under
+this License (including any patent licenses granted under the third
+paragraph of section 11).
+
+  However, if you cease all violation of this License, then your
+license from a particular copyright holder is reinstated (a)
+provisionally, unless and until the copyright holder explicitly and
+finally terminates your license, and (b) permanently, if the copyright
+holder fails to notify you of the violation by some reasonable means
+prior to 60 days after the cessation.
+
+  Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+  Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License.  If your rights have been terminated and not permanently
+reinstated, you do not qualify to receive new licenses for the same
+material under section 10.
+
+  9. Acceptance Not Required for Having Copies.
+
+  You are not required to accept this License in order to receive or
+run a copy of the Program.  Ancillary propagation of a covered work
+occurring solely as a consequence of using peer-to-peer transmission
+to receive a copy likewise does not require acceptance.  However,
+nothing other than this License grants you permission to propagate or
+modify any covered work.  These actions infringe copyright if you do
+not accept this License.  Therefore, by modifying or propagating a
+covered work, you indicate your acceptance of this License to do so.
+
+  10. Automatic Licensing of Downstream Recipients.
+
+  Each time you convey a covered work, the recipient automatically
+receives a license from the original licensors, to run, modify and
+propagate that work, subject to this License.  You are not responsible
+for enforcing compliance by third parties with this License.
+
+  An "entity transaction" is a transaction transferring control of an
+organization, or substantially all assets of one, or subdividing an
+organization, or merging organizations.  If propagation of a covered
+work results from an entity transaction, each party to that
+transaction who receives a copy of the work also receives whatever
+licenses to the work the party's predecessor in interest had or could
+give under the previous paragraph, plus a right to possession of the
+Corresponding Source of the work from the predecessor in interest, if
+the predecessor has it or can get it with reasonable efforts.
+
+  You may not impose any further restrictions on the exercise of the
+rights granted or affirmed under this License.  For example, you may
+not impose a license fee, royalty, or other charge for exercise of
+rights granted under this License, and you may not initiate litigation
+(including a cross-claim or counterclaim in a lawsuit) alleging that
+any patent claim is infringed by making, using, selling, offering for
+sale, or importing the Program or any portion of it.
+
+  11. Patents.
+
+  A "contributor" is a copyright holder who authorizes use under this
+License of the Program or a work on which the Program is based.  The
+work thus licensed is called the contributor's "contributor version".
+
+  A contributor's "essential patent claims" are all patent claims
+owned or controlled by the contributor, whether already acquired or
+hereafter acquired, that would be infringed by some manner, permitted
+by this License, of making, using, or selling its contributor version,
+but do not include claims that would be infringed only as a
+consequence of further modification of the contributor version.  For
+purposes of this definition, "control" includes the right to grant
+patent sublicenses in a manner consistent with the requirements of
+this License.
+
+  Each contributor grants you a non-exclusive, worldwide, royalty-free
+patent license under the contributor's essential patent claims, to
+make, use, sell, offer for sale, import and otherwise run, modify and
+propagate the contents of its contributor version.
+
+  In the following three paragraphs, a "patent license" is any express
+agreement or commitment, however denominated, not to enforce a patent
+(such as an express permission to practice a patent or covenant not to
+sue for patent infringement).  To "grant" such a patent license to a
+party means to make such an agreement or commitment not to enforce a
+patent against the party.
+
+  If you convey a covered work, knowingly relying on a patent license,
+and the Corresponding Source of the work is not available for anyone
+to copy, free of charge and under the terms of this License, through a
+publicly available network server or other readily accessible means,
+then you must either (1) cause the Corresponding Source to be so
+available, or (2) arrange to deprive yourself of the benefit of the
+patent license for this particular work, or (3) arrange, in a manner
+consistent with the requirements of this License, to extend the patent
+license to downstream recipients.  "Knowingly relying" means you have
+actual knowledge that, but for the patent license, your conveying the
+covered work in a country, or your recipient's use of the covered work
+in a country, would infringe one or more identifiable patents in that
+country that you have reason to believe are valid.
+
+  If, pursuant to or in connection with a single transaction or
+arrangement, you convey, or propagate by procuring conveyance of, a
+covered work, and grant a patent license to some of the parties
+receiving the covered work authorizing them to use, propagate, modify
+or convey a specific copy of the covered work, then the patent license
+you grant is automatically extended to all recipients of the covered
+work and works based on it.
+
+  A patent license is "discriminatory" if it does not include within
+the scope of its coverage, prohibits the exercise of, or is
+conditioned on the non-exercise of one or more of the rights that are
+specifically granted under this License.  You may not convey a covered
+work if you are a party to an arrangement with a third party that is
+in the business of distributing software, under which you make payment
+to the third party based on the extent of your activity of conveying
+the work, and under which the third party grants, to any of the
+parties who would receive the covered work from you, a discriminatory
+patent license (a) in connection with copies of the covered work
+conveyed by you (or copies made from those copies), or (b) primarily
+for and in connection with specific products or compilations that
+contain the covered work, unless you entered into that arrangement,
+or that patent license was granted, prior to 28 March 2007.
+
+  Nothing in this License shall be construed as excluding or limiting
+any implied license or other defenses to infringement that may
+otherwise be available to you under applicable patent law.
+
+  12. No Surrender of Others' Freedom.
+
+  If conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot convey a
+covered work so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you may
+not convey it at all.  For example, if you agree to terms that obligate you
+to collect a royalty for further conveying from those to whom you convey
+the Program, the only way you could satisfy both those terms and this
+License would be to refrain entirely from conveying the Program.
+
+  13. Remote Network Interaction.
+
+  A "Proxy Program" means a separate program which is specially designed to
+be used in conjunction with the covered work and interacts with it directly
+or indirectly through any kind of API (application programming interfaces),
+a computer network, an imitation of such network, or another Proxy Program
+itself.
+
+  Notwithstanding any other provision of this License, if you provide any user
+with an opportunity to interact with the covered work through a computer
+network, an imitation of such network, or any number of "Proxy Programs",
+you must prominently offer that user an opportunity to receive the
+Corresponding Source of the covered work and all Proxy Programs from a
+network server at no charge, through some standard or customary means of
+facilitating copying of software. The Corresponding Source for the covered
+work must be made available under the conditions of this License, and
+the Corresponding Source for all Proxy Programs must be made available
+under the conditions of either this License or any GPL-Compatible
+Free Software License, as described by the Free Software Foundation
+in their "GPL-Compatible License List".
+
+  14. Revised Versions of this License.
+
+  Vitastor Author may publish revised and/or new versions of
+the Vitastor Network Public License from time to time.  Such new versions
+will be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+  Each version is given a distinguishing version number.  If the
+Program specifies that a certain numbered version of the Vitastor Network
+Public License "or any later version" applies to it, you have the
+option of following the terms and conditions either of that numbered
+version or of any later version. If the Program does not specify a version
+number of the Vitastor Network Public License, you may choose any version
+ever published.
+
+  Later license versions may give you additional or different
+permissions.  However, no additional obligations are imposed on any
+author or copyright holder as a result of your choosing to follow a
+later version.
+
+  15. Disclaimer of Warranty.
+
+  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
+APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
+HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
+OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
+THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
+IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
+ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+  16. Limitation of Liability.
+
+  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
+THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
+GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
+USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
+DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
+PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
+EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
+SUCH DAMAGES.
+
+  17. Interpretation of Sections 15 and 16.
+
+  If the disclaimer of warranty and limitation of liability provided
+above cannot be given local legal effect according to their terms,
+reviewing courts shall apply local law that most closely approximates
+an absolute waiver of all civil liability in connection with the
+Program, unless a warranty or assumption of liability accompanies a
+copy of the Program in return for a fee.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+state the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the Vitastor Network Public License as published by
+    the Vitastor Author, either version 1 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    Vitastor Network Public License for more details.
+
+Also add information on how to contact you by electronic and paper mail.
+
+  If your software can interact with users remotely through a computer
+network, you should also make sure that it provides a way for users to
+get its source.  For example, if your program is a web application, its
+interface could display a "Source" link that leads users to an archive
+of the code.  There are many ways you could offer source, and different
+solutions will be better for different programs; see section 13 for the
+specific requirements.

blockstore_write.cpp

@@ -1,512 +0,0 @@
-#include "blockstore_impl.h"
-
-bool blockstore_impl_t::enqueue_write(blockstore_op_t *op)
-{
-    // Check or assign version number
-    bool found = false, deleted = false, is_del = (op->opcode == BS_OP_DELETE);
-    bool is_inflight_big = false;
-    uint64_t version = 1;
-    if (dirty_db.size() > 0)
-    {
-        auto dirty_it = dirty_db.upper_bound((obj_ver_id){
-            .oid = op->oid,
-            .version = UINT64_MAX,
-        });
-        dirty_it--; // segfaults when dirty_db is empty
-        if (dirty_it != dirty_db.end() && dirty_it->first.oid == op->oid)
-        {
-            found = true;
-            version = dirty_it->first.version + 1;
-            deleted = IS_DELETE(dirty_it->second.state);
-            is_inflight_big = dirty_it->second.state >= ST_D_IN_FLIGHT &&
-                dirty_it->second.state < ST_D_SYNCED ||
-                dirty_it->second.state == ST_J_WAIT_BIG;
-        }
-    }
-    if (!found)
-    {
-        auto clean_it = clean_db.find(op->oid);
-        if (clean_it != clean_db.end())
-        {
-            version = clean_it->second.version + 1;
-        }
-        else
-        {
-            deleted = true;
-        }
-    }
-    if (op->version == 0)
-    {
-        op->version = version;
-    }
-    else if (op->version < version)
-    {
-        // Invalid version requested
-        op->retval = -EEXIST;
-        return false;
-    }
-    if (deleted && is_del)
-    {
-        // Already deleted
-        op->retval = 0;
-        return false;
-    }
-    if (is_inflight_big && !is_del && !deleted && op->len < block_size &&
-        immediate_commit != IMMEDIATE_ALL)
-    {
-        // Issue an additional sync so that the previous big write can reach the journal
-        blockstore_op_t *sync_op = new blockstore_op_t;
-        sync_op->opcode = BS_OP_SYNC;
-        sync_op->callback = [this, op](blockstore_op_t *sync_op)
-        {
-            delete sync_op;
-        };
-        enqueue_op(sync_op);
-    }
-#ifdef BLOCKSTORE_DEBUG
-    if (is_del)
-        printf("Delete %lu:%lu v%lu\n", op->oid.inode, op->oid.stripe, op->version);
-    else
-        printf("Write %lu:%lu v%lu offset=%u len=%u\n", op->oid.inode, op->oid.stripe, op->version, op->offset, op->len);
-#endif
-    // No strict need to add it into dirty_db here, it's just left
-    // from the previous implementation where reads waited for writes
-    dirty_db.emplace((obj_ver_id){
-        .oid = op->oid,
-        .version = op->version,
-    }, (dirty_entry){
-        .state = (uint32_t)(
-            is_del
-                ? ST_DEL_IN_FLIGHT
-                : (op->len == block_size || deleted ? ST_D_IN_FLIGHT : (is_inflight_big ? ST_J_WAIT_BIG : ST_J_IN_FLIGHT))
-        ),
-        .flags = 0,
-        .location = 0,
-        .offset = is_del ? 0 : op->offset,
-        .len = is_del ? 0 : op->len,
-        .journal_sector = 0,
-    });
-    return true;
-}
-
-// First step of the write algorithm: dequeue operation and submit initial write(s)
-int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
-{
-    if (PRIV(op)->op_state)
-    {
-        return continue_write(op);
-    }
-    auto dirty_it = dirty_db.find((obj_ver_id){
-        .oid = op->oid,
-        .version = op->version,
-    });
-    assert(dirty_it != dirty_db.end());
-    if (dirty_it->second.state == ST_J_WAIT_BIG)
-    {
-        return 0;
-    }
-    else if (dirty_it->second.state == ST_D_IN_FLIGHT)
-    {
-        blockstore_journal_check_t space_check(this);
-        if (!space_check.check_available(op, unsynced_big_writes.size() + 1, sizeof(journal_entry_big_write), JOURNAL_STABILIZE_RESERVATION))
-        {
-            return 0;
-        }
-        // Big (redirect) write
-        uint64_t loc = data_alloc->find_free();
-        if (loc == UINT64_MAX)
-        {
-            // no space
-            if (flusher->is_active())
-            {
-                // hope that some space will be available after flush
-                PRIV(op)->wait_for = WAIT_FREE;
-                return 0;
-            }
-            op->retval = -ENOSPC;
-            FINISH_OP(op);
-            return 1;
-        }
-        BS_SUBMIT_GET_SQE(sqe, data);
-        dirty_it->second.location = loc << block_order;
-        dirty_it->second.state = ST_D_SUBMITTED;
-#ifdef BLOCKSTORE_DEBUG
-        printf("Allocate block %lu\n", loc);
-#endif
-        data_alloc->set(loc, true);
-        uint64_t stripe_offset = (op->offset % bitmap_granularity);
-        uint64_t stripe_end = (op->offset + op->len) % bitmap_granularity;
-        // Zero fill up to bitmap_granularity
-        int vcnt = 0;
-        if (stripe_offset)
-        {
-            PRIV(op)->iov_zerofill[vcnt++] = (struct iovec){ zero_object, stripe_offset };
-        }
-        PRIV(op)->iov_zerofill[vcnt++] = (struct iovec){ op->buf, op->len };
-        if (stripe_end)
-        {
-            stripe_end = bitmap_granularity - stripe_end;
-            PRIV(op)->iov_zerofill[vcnt++] = (struct iovec){ zero_object, stripe_end };
-        }
-        data->iov.iov_len = op->len + stripe_offset + stripe_end; // to check it in the callback
-        data->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
-        my_uring_prep_writev(
-            sqe, data_fd, PRIV(op)->iov_zerofill, vcnt, data_offset + (loc << block_order) + op->offset - stripe_offset
-        );
-        PRIV(op)->pending_ops = 1;
-        PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 0;
-        if (immediate_commit != IMMEDIATE_ALL)
-        {
-            // Remember big write as unsynced
-            unsynced_big_writes.push_back((obj_ver_id){
-                .oid = op->oid,
-                .version = op->version,
-            });
-            PRIV(op)->op_state = 3;
-        }
-        else
-        {
-            PRIV(op)->op_state = 1;
-        }
-    }
-    else
-    {
-        // Small (journaled) write
-        // First check if the journal has sufficient space
-        blockstore_journal_check_t space_check(this);
-        if (unsynced_big_writes.size() && !space_check.check_available(op, unsynced_big_writes.size(), sizeof(journal_entry_big_write), 0)
-            || !space_check.check_available(op, 1, sizeof(journal_entry_small_write), op->len + JOURNAL_STABILIZE_RESERVATION))
-        {
-            return 0;
-        }
-        // There is sufficient space. Get SQE(s)
-        struct io_uring_sqe *sqe1 = NULL;
-        if (immediate_commit != IMMEDIATE_NONE ||
-            (journal_block_size - journal.in_sector_pos) < sizeof(journal_entry_small_write) &&
-            journal.sector_info[journal.cur_sector].dirty)
-        {
-            // Write current journal sector only if it's dirty and full, or in the immediate_commit mode
-            BS_SUBMIT_GET_SQE_DECL(sqe1);
-        }
-        struct io_uring_sqe *sqe2 = NULL;
-        if (op->len > 0)
-        {
-            BS_SUBMIT_GET_SQE_DECL(sqe2);
-        }
-        // Got SQEs. Prepare previous journal sector write if required
-        auto cb = [this, op](ring_data_t *data) { handle_write_event(data, op); };
-        if (immediate_commit == IMMEDIATE_NONE)
-        {
-            if (sqe1)
-            {
-                prepare_journal_sector_write(journal, journal.cur_sector, sqe1, cb);
-                PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 1 + journal.cur_sector;
-                PRIV(op)->pending_ops++;
-            }
-            else
-            {
-                PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 0;
-            }
-        }
-        // Then pre-fill journal entry
-        journal_entry_small_write *je = (journal_entry_small_write*)
-            prefill_single_journal_entry(journal, JE_SMALL_WRITE, sizeof(journal_entry_small_write));
-        dirty_it->second.journal_sector = journal.sector_info[journal.cur_sector].offset;
-        journal.used_sectors[journal.sector_info[journal.cur_sector].offset]++;
-#ifdef BLOCKSTORE_DEBUG
-        printf(
-            "journal offset %08lx is used by %lu:%lu v%lu (%lu refs)\n",
-            dirty_it->second.journal_sector, dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version,
-            journal.used_sectors[journal.sector_info[journal.cur_sector].offset]
-        );
-#endif
-        // Figure out where data will be
-        journal.next_free = (journal.next_free + op->len) <= journal.len ? journal.next_free : journal_block_size;
-        je->oid = op->oid;
-        je->version = op->version;
-        je->offset = op->offset;
-        je->len = op->len;
-        je->data_offset = journal.next_free;
-        je->crc32_data = crc32c(0, op->buf, op->len);
-        je->crc32 = je_crc32((journal_entry*)je);
-        journal.crc32_last = je->crc32;
-        if (immediate_commit != IMMEDIATE_NONE)
-        {
-            prepare_journal_sector_write(journal, journal.cur_sector, sqe1, cb);
-            PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 1 + journal.cur_sector;
-            PRIV(op)->pending_ops++;
-        }
-        if (op->len > 0)
-        {
-            // Prepare journal data write
-            if (journal.inmemory)
-            {
-                // Copy data
-                memcpy(journal.buffer + journal.next_free, op->buf, op->len);
-            }
-            ring_data_t *data2 = ((ring_data_t*)sqe2->user_data);
-            data2->iov = (struct iovec){ op->buf, op->len };
-            data2->callback = cb;
-            my_uring_prep_writev(
-                sqe2, journal.fd, &data2->iov, 1, journal.offset + journal.next_free
-            );
-            PRIV(op)->pending_ops++;
-        }
-        else
-        {
-            // Zero-length overwrite. Allowed to bump object version in EC placement groups without actually writing data
-        }
-        dirty_it->second.location = journal.next_free;
-        dirty_it->second.state = ST_J_SUBMITTED;
-        journal.next_free += op->len;
-        if (journal.next_free >= journal.len)
-        {
-            journal.next_free = journal_block_size;
-        }
-        if (immediate_commit == IMMEDIATE_NONE)
-        {
-            // Remember small write as unsynced
-            unsynced_small_writes.push_back((obj_ver_id){
-                .oid = op->oid,
-                .version = op->version,
-            });
-        }
-        if (!PRIV(op)->pending_ops)
-        {
-            PRIV(op)->op_state = 4;
-            continue_write(op);
-        }
-        else
-        {
-            PRIV(op)->op_state = 3;
-        }
-    }
-    inflight_writes++;
-    return 1;
-}
-
-int blockstore_impl_t::continue_write(blockstore_op_t *op)
-{
-    io_uring_sqe *sqe = NULL;
-    journal_entry_big_write *je;
-    auto dirty_it = dirty_db.find((obj_ver_id){
-        .oid = op->oid,
-        .version = op->version,
-    });
-    assert(dirty_it != dirty_db.end());
-    if (PRIV(op)->op_state == 2)
-        goto resume_2;
-    else if (PRIV(op)->op_state == 4)
-        goto resume_4;
-    else
-        return 1;
-resume_2:
-    // Only for the immediate_commit mode: prepare and submit big_write journal entry
-    sqe = get_sqe();
-    if (!sqe)
-    {
-        return 0;
-    }
-    je = (journal_entry_big_write*)prefill_single_journal_entry(journal, JE_BIG_WRITE, sizeof(journal_entry_big_write));
-    dirty_it->second.journal_sector = journal.sector_info[journal.cur_sector].offset;
-    journal.sector_info[journal.cur_sector].dirty = false;
-    journal.used_sectors[journal.sector_info[journal.cur_sector].offset]++;
-#ifdef BLOCKSTORE_DEBUG
-    printf(
-        "journal offset %08lx is used by %lu:%lu v%lu (%lu refs)\n",
-        journal.sector_info[journal.cur_sector].offset, op->oid.inode, op->oid.stripe, op->version,
-        journal.used_sectors[journal.sector_info[journal.cur_sector].offset]
-    );
-#endif
-    je->oid = op->oid;
-    je->version = op->version;
-    je->offset = op->offset;
-    je->len = op->len;
-    je->location = dirty_it->second.location;
-    je->crc32 = je_crc32((journal_entry*)je);
-    journal.crc32_last = je->crc32;
-    prepare_journal_sector_write(journal, journal.cur_sector, sqe,
-        [this, op](ring_data_t *data) { handle_write_event(data, op); });
-    PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 1 + journal.cur_sector;
-    PRIV(op)->pending_ops = 1;
-    PRIV(op)->op_state = 3;
-    return 1;
-resume_4:
-    // Switch object state
-#ifdef BLOCKSTORE_DEBUG
-    printf("Ack write %lu:%lu v%lu = %d\n", op->oid.inode, op->oid.stripe, op->version, dirty_it->second.state);
-#endif
-    bool imm = dirty_it->second.state == ST_D_SUBMITTED
-        ? (immediate_commit == IMMEDIATE_ALL)
-        : (immediate_commit != IMMEDIATE_NONE);
-    if (imm)
-    {
-        auto & unstab = unstable_writes[op->oid];
-        unstab = unstab < op->version ? op->version : unstab;
-    }
-    if (dirty_it->second.state == ST_J_SUBMITTED)
-    {
-        dirty_it->second.state = imm ? ST_J_SYNCED : ST_J_WRITTEN;
-    }
-    else if (dirty_it->second.state == ST_D_SUBMITTED)
-    {
-        dirty_it->second.state = imm ? ST_D_SYNCED : ST_D_WRITTEN;
-    }
-    else if (dirty_it->second.state == ST_DEL_SUBMITTED)
-    {
-        dirty_it->second.state = imm ? ST_DEL_SYNCED : ST_DEL_WRITTEN;
-        if (imm)
-        {
-            // Deletions are treated as immediately stable
-            mark_stable(dirty_it->first);
-        }
-    }
-    if (immediate_commit == IMMEDIATE_ALL)
-    {
-        dirty_it++;
-        while (dirty_it != dirty_db.end() && dirty_it->first.oid == op->oid)
-        {
-            if (dirty_it->second.state == ST_J_WAIT_BIG)
-            {
-                dirty_it->second.state = ST_J_IN_FLIGHT;
-            }
-            dirty_it++;
-        }
-    }
-    inflight_writes--;
-    // Acknowledge write
-    op->retval = op->len;
-    FINISH_OP(op);
-    return 1;
-}
-
-void blockstore_impl_t::handle_write_event(ring_data_t *data, blockstore_op_t *op)
-{
-    live = true;
-    if (data->res != data->iov.iov_len)
-    {
-        inflight_writes--;
-        // FIXME: our state becomes corrupted after a write error. maybe do something better than just die
-        throw std::runtime_error(
-            "write operation failed ("+std::to_string(data->res)+" != "+std::to_string(data->iov.iov_len)+
-            "). in-memory state is corrupted. AAAAAAAaaaaaaaaa!!!111"
-        );
-    }
-    PRIV(op)->pending_ops--;
-    if (PRIV(op)->pending_ops == 0)
-    {
-        release_journal_sectors(op);
-        PRIV(op)->op_state++;
-        if (!continue_write(op))
-        {
-            submit_queue.push_front(op);
-        }
-    }
-}
-
-void blockstore_impl_t::release_journal_sectors(blockstore_op_t *op)
-{
-    // Release flushed journal sectors
-    if (PRIV(op)->min_flushed_journal_sector > 0 &&
-        PRIV(op)->max_flushed_journal_sector > 0)
-    {
-        uint64_t s = PRIV(op)->min_flushed_journal_sector;
-        while (1)
-        {
-            journal.sector_info[s-1].usage_count--;
-            if (s != (1+journal.cur_sector) && journal.sector_info[s-1].usage_count == 0)
-            {
-                // We know for sure that we won't write into this sector anymore
-                uint64_t new_ds = journal.sector_info[s-1].offset + journal.block_size;
-                if (new_ds >= journal.len)
-                {
-                    new_ds = journal.block_size;
-                }
-                if ((journal.dirty_start + (journal.dirty_start >= journal.used_start ? 0 : journal.len)) <
-                    (new_ds + (new_ds >= journal.used_start ? 0 : journal.len)))
-                {
-                    journal.dirty_start = new_ds;
-                }
-            }
-            if (s == PRIV(op)->max_flushed_journal_sector)
-                break;
-            s = 1 + s % journal.sector_count;
-        }
-        PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 0;
-    }
-}
-
-int blockstore_impl_t::dequeue_del(blockstore_op_t *op)
-{
-    auto dirty_it = dirty_db.find((obj_ver_id){
-        .oid = op->oid,
-        .version = op->version,
-    });
-    assert(dirty_it != dirty_db.end());
-    blockstore_journal_check_t space_check(this);
-    if (!space_check.check_available(op, 1, sizeof(journal_entry_del), 0))
-    {
-        return 0;
-    }
-    io_uring_sqe *sqe = NULL;
-    if (immediate_commit != IMMEDIATE_NONE ||
-        (journal_block_size - journal.in_sector_pos) < sizeof(journal_entry_del) &&
-        journal.sector_info[journal.cur_sector].dirty)
-    {
-        // Write current journal sector only if it's dirty and full, or in the immediate_commit mode
-        BS_SUBMIT_GET_SQE_DECL(sqe);
-    }
-    auto cb = [this, op](ring_data_t *data) { handle_write_event(data, op); };
-    // Prepare journal sector write
-    if (immediate_commit == IMMEDIATE_NONE)
-    {
-        if (sqe)
-        {
-            prepare_journal_sector_write(journal, journal.cur_sector, sqe, cb);
-            PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 1 + journal.cur_sector;
-            PRIV(op)->pending_ops++;
-        }
-        else
-        {
-            PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 0;
-        }
-    }
-    // Pre-fill journal entry
-    journal_entry_del *je = (journal_entry_del*)
-        prefill_single_journal_entry(journal, JE_DELETE, sizeof(struct journal_entry_del));
-    dirty_it->second.journal_sector = journal.sector_info[journal.cur_sector].offset;
-    journal.used_sectors[journal.sector_info[journal.cur_sector].offset]++;
-#ifdef BLOCKSTORE_DEBUG
-    printf(
-        "journal offset %08lx is used by %lu:%lu v%lu (%lu refs)\n",
-        dirty_it->second.journal_sector, dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version,
-        journal.used_sectors[journal.sector_info[journal.cur_sector].offset]
-    );
-#endif
-    je->oid = op->oid;
-    je->version = op->version;
-    je->crc32 = je_crc32((journal_entry*)je);
-    journal.crc32_last = je->crc32;
-    dirty_it->second.state = ST_DEL_SUBMITTED;
-    if (immediate_commit != IMMEDIATE_NONE)
-    {
-        prepare_journal_sector_write(journal, journal.cur_sector, sqe, cb);
-        PRIV(op)->min_flushed_journal_sector = PRIV(op)->max_flushed_journal_sector = 1 + journal.cur_sector;
-        PRIV(op)->pending_ops++;
-        // Remember small write as unsynced
-        unsynced_small_writes.push_back((obj_ver_id){
-            .oid = op->oid,
-            .version = op->version,
-        });
-    }
-    if (!PRIV(op)->pending_ops)
-    {
-        PRIV(op)->op_state = 4;
-        continue_write(op);
-    }
-    else
-    {
-        PRIV(op)->op_state = 3;
-    }
-    return 1;
-}

cluster_client.cpp

@@ -1,349 +0,0 @@
-#include "cluster_client.h"
-
-cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json & config)
-{
-    this->ringloop = ringloop;
-    this->tfd = tfd;
-
-    msgr.tfd = tfd;
-    msgr.ringloop = ringloop;
-    msgr.repeer_pgs = [this](osd_num_t peer_osd)
-    {
-        // peer_osd just connected or dropped connection
-        if (msgr.osd_peer_fds.find(peer_osd) != msgr.osd_peer_fds.end())
-        {
-            // really connected :)
-            continue_ops();
-        }
-    };
-
-    st_cli.tfd = tfd;
-    st_cli.on_load_config_hook = [this](json11::Json::object & cfg) { on_load_config_hook(cfg); };
-    st_cli.on_change_osd_state_hook = [this](uint64_t peer_osd) { on_change_osd_state_hook(peer_osd); };
-    st_cli.on_change_hook = [this](json11::Json::object & changes) { on_change_hook(changes); };
-    st_cli.on_load_pgs_hook = [this](bool success) { on_load_pgs_hook(success); };
-
-    log_level = config["log_level"].int64_value();
-    st_cli.parse_config(config);
-    st_cli.load_global_config();
-}
-
-void cluster_client_t::continue_ops()
-{
-    for (auto op_it = unsent_ops.begin(); op_it != unsent_ops.end(); )
-    {
-        cluster_op_t *op = *op_it;
-        if (op->needs_reslice && !op->sent_count)
-        {
-            op->parts.clear();
-            op->done_count = 0;
-            op->needs_reslice = false;
-        }
-        if (!op->parts.size())
-        {
-            unsent_ops.erase(op_it++);
-            execute(op);
-            continue;
-        }
-        if (!op->needs_reslice)
-        {
-            for (auto & op_part: op->parts)
-            {
-                if (!op_part.sent && !op_part.done)
-                {
-                    try_send(op, &op_part);
-                }
-            }
-            if (op->sent_count == op->parts.size() - op->done_count)
-            {
-                unsent_ops.erase(op_it++);
-                sent_ops.insert(op);
-            }
-            else
-                op_it++;
-        }
-        else
-            op_it++;
-    }
-}
-
-static uint32_t is_power_of_two(uint64_t value)
-{
-    uint32_t l = 0;
-    while (value > 1)
-    {
-        if (value & 1)
-        {
-            return 64;
-        }
-        value = value >> 1;
-        l++;
-    }
-    return l;
-}
-
-void cluster_client_t::on_load_config_hook(json11::Json::object & config)
-{
-    bs_block_size = config["block_size"].uint64_value();
-    bs_disk_alignment = config["disk_alignment"].uint64_value();
-    bs_bitmap_granularity = config["bitmap_granularity"].uint64_value();
-    if (!bs_block_size)
-        bs_block_size = DEFAULT_BLOCK_SIZE;
-    if (!bs_disk_alignment)
-        bs_disk_alignment = DEFAULT_DISK_ALIGNMENT;
-    if (!bs_bitmap_granularity)
-        bs_bitmap_granularity = DEFAULT_BITMAP_GRANULARITY;
-    {
-        uint32_t block_order;
-        if ((block_order = is_power_of_two(bs_block_size)) >= 64 || bs_block_size < MIN_BLOCK_SIZE || bs_block_size >= MAX_BLOCK_SIZE)
-            throw std::runtime_error("Bad block size");
-    }
-    if (config.find("pg_stripe_size") != config.end())
-    {
-        pg_stripe_size = config["pg_stripe_size"].uint64_value();
-        if (!pg_stripe_size)
-            pg_stripe_size = DEFAULT_PG_STRIPE_SIZE;
-    }
-    if (config["immediate_commit"] == "all")
-    {
-        // Cluster-wide immediate_commit mode
-        immediate_commit = true;
-    }
-    msgr.peer_connect_interval = config["peer_connect_interval"].uint64_value();
-    if (!msgr.peer_connect_interval)
-        msgr.peer_connect_interval = DEFAULT_PEER_CONNECT_INTERVAL;
-    msgr.peer_connect_timeout = config["peer_connect_timeout"].uint64_value();
-    if (!msgr.peer_connect_timeout)
-        msgr.peer_connect_timeout = DEFAULT_PEER_CONNECT_TIMEOUT;
-}
-
-void cluster_client_t::on_load_pgs_hook(bool success)
-{
-    if (success)
-    {
-        pg_count = st_cli.pg_config.size();
-        continue_ops();
-    }
-}
-
-void cluster_client_t::on_change_hook(json11::Json::object & changes)
-{
-    if (pg_count != st_cli.pg_config.size())
-    {
-        // At this point, all operations should be suspended
-        // And they need to be resliced!
-        for (auto op: unsent_ops)
-        {
-            op->needs_reslice = true;
-        }
-        for (auto op: sent_ops)
-        {
-            op->needs_reslice = true;
-        }
-        pg_count = st_cli.pg_config.size();
-    }
-    continue_ops();
-}
-
-void cluster_client_t::on_change_osd_state_hook(uint64_t peer_osd)
-{
-    if (msgr.wanted_peers.find(peer_osd) != msgr.wanted_peers.end())
-    {
-        msgr.connect_peer(peer_osd, st_cli.peer_states[peer_osd]);
-    }
-}
-
-// FIXME: Implement OSD_OP_SYNC for immediate_commit == false
-void cluster_client_t::execute(cluster_op_t *op)
-{
-    if (op->opcode == OSD_OP_SYNC && immediate_commit)
-    {
-        // Syncs are not required in the immediate_commit mode
-        op->retval = 0;
-        std::function<void(cluster_op_t*)>(op->callback)(op);
-        return;
-    }
-    if (op->opcode != OSD_OP_READ && op->opcode != OSD_OP_OUT || !op->inode || !op->len ||
-        op->offset % bs_disk_alignment || op->len % bs_disk_alignment)
-    {
-        op->retval = -EINVAL;
-        std::function<void(cluster_op_t*)>(op->callback)(op);
-        return;
-    }
-    if (!pg_stripe_size)
-    {
-        // Config is not loaded yet
-        unsent_ops.insert(op);
-        return;
-    }
-    if (op->opcode == OSD_OP_WRITE && !immediate_commit)
-    {
-        // Copy operation
-        cluster_op_t *op_copy = new cluster_op_t();
-        op_copy->opcode = op->opcode;
-        op_copy->inode = op->inode;
-        op_copy->offset = op->offset;
-        op_copy->len = op->len;
-        op_copy->buf = malloc(op->len);
-        memcpy(op_copy->buf, op->buf, op->len);
-        unsynced_ops.push_back(op_copy);
-        unsynced_bytes += op->len;
-        if (inmemory_commit)
-        {
-            // Immediately acknowledge write and continue with the copy
-            op->retval = op->len;
-            std::function<void(cluster_op_t*)>(op->callback)(op);
-            op = op_copy;
-        }
-        if (unsynced_bytes >= inmemory_dirty_limit)
-        {
-            // Push an extra SYNC operation
-        }
-    }
-    // Slice the request into individual object stripe requests
-    // Primary OSDs still operate individual stripes, but their size is multiplied by PG minsize in case of EC
-    uint64_t pg_block_size = bs_block_size * pg_part_count;
-    uint64_t first_stripe = (op->offset / pg_block_size) * pg_block_size;
-    uint64_t last_stripe = ((op->offset + op->len + pg_block_size - 1) / pg_block_size - 1) * pg_block_size;
-    int part_count = 0;
-    for (uint64_t stripe = first_stripe; stripe <= last_stripe; stripe += pg_block_size)
-    {
-        if (op->offset < (stripe+pg_block_size) && (op->offset+op->len) > stripe)
-        {
-            part_count++;
-        }
-    }
-    op->parts.resize(part_count);
-    bool resend = false;
-    int i = 0;
-    for (uint64_t stripe = first_stripe; stripe <= last_stripe; stripe += pg_block_size)
-    {
-        uint64_t stripe_end = stripe + pg_block_size;
-        if (op->offset < stripe_end && (op->offset+op->len) > stripe)
-        {
-            pg_num_t pg_num = (op->inode + stripe/pg_stripe_size) % pg_count + 1;
-            op->parts[i] = {
-                .parent = op,
-                .offset = op->offset < stripe ? stripe : op->offset,
-                .len = (uint32_t)((op->offset+op->len) > stripe_end ? pg_block_size : op->offset+op->len-stripe),
-                .pg_num = pg_num,
-                .buf = op->buf + (op->offset < stripe ? stripe-op->offset : 0),
-                .sent = false,
-                .done = false,
-            };
-            if (!try_send(op, &op->parts[i]))
-            {
-                // Part needs to be sent later
-                resend = true;
-            }
-            i++;
-        }
-    }
-    if (resend)
-    {
-        unsent_ops.insert(op);
-    }
-    else
-    {
-        sent_ops.insert(op);
-    }
-}
-
-bool cluster_client_t::try_send(cluster_op_t *op, cluster_op_part_t *part)
-{
-    auto pg_it = st_cli.pg_config.find(part->pg_num);
-    if (pg_it != st_cli.pg_config.end() &&
-        !pg_it->second.pause && pg_it->second.cur_primary)
-    {
-        osd_num_t primary_osd = pg_it->second.cur_primary;
-        auto peer_it = msgr.osd_peer_fds.find(primary_osd);
-        if (peer_it != msgr.osd_peer_fds.end())
-        {
-            int peer_fd = peer_it->second;
-            part->osd_num = primary_osd;
-            part->sent = true;
-            op->sent_count++;
-            part->op = {
-                .op_type = OSD_OP_OUT,
-                .peer_fd = peer_fd,
-                .req = { .rw = {
-                    .header = {
-                        .magic = SECONDARY_OSD_OP_MAGIC,
-                        .id = op_id++,
-                        .opcode = op->opcode,
-                    },
-                    .inode = op->inode,
-                    .offset = part->offset,
-                    .len = part->len,
-                } },
-                .callback = [this, part](osd_op_t *op_part)
-                {
-                    handle_op_part(part);
-                },
-            };
-            part->op.send_list.push_back(part->op.req.buf, OSD_PACKET_SIZE);
-            if (op->opcode == OSD_OP_WRITE)
-            {
-                part->op.send_list.push_back(part->buf, part->len);
-            }
-            else
-            {
-                part->op.buf = part->buf;
-            }
-            msgr.outbox_push(&part->op);
-            return true;
-        }
-        else if (msgr.wanted_peers.find(primary_osd) == msgr.wanted_peers.end())
-        {
-            msgr.connect_peer(primary_osd, st_cli.peer_states[primary_osd]);
-        }
-    }
-    return false;
-}
-
-void cluster_client_t::handle_op_part(cluster_op_part_t *part)
-{
-    cluster_op_t *op = part->parent;
-    part->sent = false;
-    op->sent_count--;
-    part->op.buf = NULL;
-    if (part->op.reply.hdr.retval != part->op.req.rw.len)
-    {
-        // Operation failed, retry
-        printf(
-            "Operation part failed on OSD %lu: retval=%ld (expected %u), reconnecting\n",
-            part->osd_num, part->op.reply.hdr.retval, part->op.req.rw.len
-        );
-        msgr.stop_client(part->op.peer_fd);
-        if (op->sent_count == op->parts.size() - op->done_count - 1)
-        {
-            // Resend later when OSDs come up
-            // FIXME: Check for different types of errors
-            // FIXME: Repeat operations after a small timeout, for the case when OSD is coming up
-            sent_ops.erase(op);
-            unsent_ops.insert(op);
-        }
-        if (op->sent_count == 0 && op->needs_reslice)
-        {
-            // PG count has changed, reslice the operation
-            unsent_ops.erase(op);
-            op->parts.clear();
-            op->done_count = 0;
-            op->needs_reslice = false;
-            execute(op);
-        }
-    }
-    else
-    {
-        // OK
-        part->done = true;
-        op->done_count++;
-        if (op->done_count >= op->parts.size())
-        {
-            // Finished!
-            sent_ops.erase(op);
-            op->retval = op->len;
-            std::function<void(cluster_op_t*)>(op->callback)(op);
-        }
-    }
-}

cluster_client.h

@@ -1,80 +0,0 @@
-#pragma once
-
-#include "messenger.h"
-#include "etcd_state_client.h"
-
-#define MIN_BLOCK_SIZE 4*1024
-#define MAX_BLOCK_SIZE 128*1024*1024
-#define DEFAULT_BLOCK_SIZE 128*1024
-#define DEFAULT_PG_STRIPE_SIZE 4*1024*1024
-#define DEFAULT_DISK_ALIGNMENT 4096
-#define DEFAULT_BITMAP_GRANULARITY 4096
-
-struct cluster_op_t;
-
-struct cluster_op_part_t
-{
-    cluster_op_t *parent;
-    uint64_t offset;
-    uint32_t len;
-    pg_num_t pg_num;
-    osd_num_t osd_num;
-    void *buf;
-    bool sent;
-    bool done;
-    osd_op_t op;
-};
-
-struct cluster_op_t
-{
-    uint64_t opcode; // OSD_OP_READ, OSD_OP_WRITE, OSD_OP_SYNC
-    uint64_t inode;
-    uint64_t offset;
-    uint64_t len;
-    int retval;
-    void *buf;
-    std::function<void(cluster_op_t*)> callback;
-protected:
-    bool needs_reslice = false;
-    int sent_count = 0, done_count = 0;
-    std::vector<cluster_op_part_t> parts;
-    friend class cluster_client_t;
-};
-
-class cluster_client_t
-{
-    timerfd_manager_t *tfd;
-    ring_loop_t *ringloop;
-
-    uint64_t pg_part_count = 2;
-    uint64_t pg_stripe_size = 0;
-    uint64_t bs_block_size = 0;
-    uint64_t bs_disk_alignment = 0;
-    uint64_t bs_bitmap_granularity = 0;
-    uint64_t pg_count = 0;
-    bool immediate_commit = false;
-    bool inmemory_commit = false;
-    uint64_t inmemory_dirty_limit = 32*1024*1024;
-    int log_level;
-
-    uint64_t op_id = 1;
-    etcd_state_client_t st_cli;
-    osd_messenger_t msgr;
-    std::set<cluster_op_t*> sent_ops, unsent_ops;
-    // unsynced operations are copied in memory to allow replay when cluster isn't in the immediate_commit mode
-    std::vector<cluster_op_t*> unsynced_ops;
-    uint64_t unsynced_bytes = 0;
-
-public:
-    cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json & config);
-    void execute(cluster_op_t *op);
-
-protected:
-    void continue_ops();
-    void on_load_config_hook(json11::Json::object & cfg);
-    void on_load_pgs_hook(bool success);
-    void on_change_hook(json11::Json::object & changes);
-    void on_change_osd_state_hook(uint64_t peer_osd);
-    bool try_send(cluster_op_t *op, cluster_op_part_t *part);
-    void handle_op_part(cluster_op_part_t *part);
-};

copy-fio-includes.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+gcc -I. -E -o fio_headers.i src/fio_headers.h
+
+rm -rf fio-copy
+for i in `grep -Po 'fio/[^"]+' fio_headers.i | sort | uniq`; do
+    j=${i##fio/}
+    p=$(dirname $j)
+    mkdir -p fio-copy/$p
+    cp $i fio-copy/$j
+done
+
+rm fio_headers.i

copy-qemu-includes.sh

@@ -0,0 +1,18 @@
+#!/bin/bash
+
+#cd qemu
+#debian/rules b/configure-stamp
+#cd b/qemu; make qapi
+
+gcc -I qemu/b/qemu `pkg-config glib-2.0 --cflags` \
+    -I qemu/include -E -o qemu_driver.i src/qemu_driver.c
+
+rm -rf qemu-copy
+for i in `grep -Po 'qemu/[^"]+' qemu_driver.i | sort | uniq`; do
+    j=${i##qemu/}
+    p=$(dirname $j)
+    mkdir -p qemu-copy/$p
+    cp $i qemu-copy/$j
+done
+
+rm qemu_driver.i

cpp-btree

@@ -0,0 +1 @@
+Subproject commit 5dc108754ad40d3b1d024f9bd7cca0595ef1a1db

csi/.dockerignore

@@ -0,0 +1,3 @@
+vitastor-csi
+go.sum
+Dockerfile

csi/Dockerfile

@@ -0,0 +1,32 @@
+# Compile stage
+FROM golang:buster AS build
+
+ADD go.mod /app/
+RUN cd /app; CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go mod download -x
+ADD . /app
+RUN perl -i -e '$/ = undef; while(<>) { s/\n\s*(\{\s*\n)/$1\n/g; s/\}(\s*\n\s*)else\b/$1} else/g; print; }' `find /app -name '*.go'`
+RUN cd /app; CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go build -o vitastor-csi
+
+# Final stage
+FROM debian:buster
+
+LABEL maintainers="Vitaliy Filippov <vitalif@yourcmc.ru>"
+LABEL description="Vitastor CSI Driver"
+
+ENV NODE_ID=""
+ENV CSI_ENDPOINT=""
+
+RUN apt-get update && \
+    apt-get install -y wget && \
+    wget -q -O /etc/apt/trusted.gpg.d/vitastor.gpg https://vitastor.io/debian/pubkey.gpg && \
+    (echo deb http://vitastor.io/debian buster main > /etc/apt/sources.list.d/vitastor.list) && \
+    (echo deb http://deb.debian.org/debian buster-backports main > /etc/apt/sources.list.d/backports.list) && \
+    (echo "APT::Install-Recommends false;" > /etc/apt/apt.conf) && \
+    apt-get update && \
+    apt-get install -y e2fsprogs xfsprogs vitastor kmod && \
+    apt-get clean && \
+    (echo options nbd nbds_max=128 > /etc/modprobe.d/nbd.conf)
+
+COPY --from=build /app/vitastor-csi /bin/
+
+ENTRYPOINT ["/bin/vitastor-csi"]

csi/Makefile

@@ -0,0 +1,9 @@
+VERSION ?= v0.6.5
+
+all: build push
+
+build:
+	@docker build --rm -t vitalif/vitastor-csi:$(VERSION) .
+
+push:
+	@docker push vitalif/vitastor-csi:$(VERSION)

csi/deploy/000-csi-namespace.yaml

@@ -0,0 +1,5 @@
+---
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: vitastor-system

csi/deploy/001-csi-config-map.yaml

@@ -0,0 +1,9 @@
+---
+apiVersion: v1
+kind: ConfigMap
+data:
+  vitastor.conf: |-
+    {"etcd_address":"http://192.168.7.2:2379","etcd_prefix":"/vitastor"}
+metadata:
+  namespace: vitastor-system
+  name: vitastor-config

csi/deploy/002-csi-nodeplugin-rbac.yaml

@@ -0,0 +1,37 @@
+---
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-nodeplugin
+---
+kind: ClusterRole
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-nodeplugin
+rules:
+  - apiGroups: [""]
+    resources: ["nodes"]
+    verbs: ["get"]
+  # allow to read Vault Token and connection options from the Tenants namespace
+  - apiGroups: [""]
+    resources: ["secrets"]
+    verbs: ["get"]
+  - apiGroups: [""]
+    resources: ["configmaps"]
+    verbs: ["get"]
+---
+kind: ClusterRoleBinding
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-nodeplugin
+subjects:
+  - kind: ServiceAccount
+    name: vitastor-csi-nodeplugin
+    namespace: vitastor-system
+roleRef:
+  kind: ClusterRole
+  name: vitastor-csi-nodeplugin
+  apiGroup: rbac.authorization.k8s.io

csi/deploy/003-csi-nodeplugin-psp.yaml

@@ -0,0 +1,72 @@
+---
+apiVersion: policy/v1beta1
+kind: PodSecurityPolicy
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-nodeplugin-psp
+spec:
+  allowPrivilegeEscalation: true
+  allowedCapabilities:
+    - 'SYS_ADMIN'
+  fsGroup:
+    rule: RunAsAny
+  privileged: true
+  hostNetwork: true
+  hostPID: true
+  runAsUser:
+    rule: RunAsAny
+  seLinux:
+    rule: RunAsAny
+  supplementalGroups:
+    rule: RunAsAny
+  volumes:
+    - 'configMap'
+    - 'emptyDir'
+    - 'projected'
+    - 'secret'
+    - 'downwardAPI'
+    - 'hostPath'
+  allowedHostPaths:
+    - pathPrefix: '/dev'
+      readOnly: false
+    - pathPrefix: '/run/mount'
+      readOnly: false
+    - pathPrefix: '/sys'
+      readOnly: false
+    - pathPrefix: '/lib/modules'
+      readOnly: true
+    - pathPrefix: '/var/lib/kubelet/pods'
+      readOnly: false
+    - pathPrefix: '/var/lib/kubelet/plugins/csi.vitastor.io'
+      readOnly: false
+    - pathPrefix: '/var/lib/kubelet/plugins_registry'
+      readOnly: false
+    - pathPrefix: '/var/lib/kubelet/plugins'
+      readOnly: false
+
+---
+kind: Role
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-nodeplugin-psp
+rules:
+  - apiGroups: ['policy']
+    resources: ['podsecuritypolicies']
+    verbs: ['use']
+    resourceNames: ['vitastor-csi-nodeplugin-psp']
+
+---
+kind: RoleBinding
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-nodeplugin-psp
+subjects:
+  - kind: ServiceAccount
+    name: vitastor-csi-nodeplugin
+    namespace: vitastor-system
+roleRef:
+  kind: Role
+  name: vitastor-csi-nodeplugin-psp
+  apiGroup: rbac.authorization.k8s.io

csi/deploy/004-csi-nodeplugin.yaml

@@ -0,0 +1,140 @@
+---
+kind: DaemonSet
+apiVersion: apps/v1
+metadata:
+  namespace: vitastor-system
+  name: csi-vitastor
+spec:
+  selector:
+    matchLabels:
+      app: csi-vitastor
+  template:
+    metadata:
+      namespace: vitastor-system
+      labels:
+        app: csi-vitastor
+    spec:
+      serviceAccountName: vitastor-csi-nodeplugin
+      hostNetwork: true
+      hostPID: true
+      priorityClassName: system-node-critical
+      # to use e.g. Rook orchestrated cluster, and mons' FQDN is
+      # resolved through k8s service, set dns policy to cluster first
+      dnsPolicy: ClusterFirstWithHostNet
+      containers:
+        - name: driver-registrar
+          # This is necessary only for systems with SELinux, where
+          # non-privileged sidecar containers cannot access unix domain socket
+          # created by privileged CSI driver container.
+          securityContext:
+            privileged: true
+          image: k8s.gcr.io/sig-storage/csi-node-driver-registrar:v2.2.0
+          args:
+            - "--v=5"
+            - "--csi-address=/csi/csi.sock"
+            - "--kubelet-registration-path=/var/lib/kubelet/plugins/csi.vitastor.io/csi.sock"
+          env:
+            - name: KUBE_NODE_NAME
+              valueFrom:
+                fieldRef:
+                  fieldPath: spec.nodeName
+          volumeMounts:
+            - name: socket-dir
+              mountPath: /csi
+            - name: registration-dir
+              mountPath: /registration
+        - name: csi-vitastor
+          securityContext:
+            privileged: true
+            capabilities:
+              add: ["SYS_ADMIN"]
+            allowPrivilegeEscalation: true
+          image: vitalif/vitastor-csi:v0.6.5
+          args:
+            - "--node=$(NODE_ID)"
+            - "--endpoint=$(CSI_ENDPOINT)"
+          env:
+            - name: NODE_ID
+              valueFrom:
+                fieldRef:
+                  fieldPath: spec.nodeName
+            - name: CSI_ENDPOINT
+              value: unix:///csi/csi.sock
+          imagePullPolicy: "IfNotPresent"
+          ports:
+          - containerPort: 9898
+            name: healthz
+            protocol: TCP
+          livenessProbe:
+            failureThreshold: 5
+            httpGet:
+              path: /healthz
+              port: healthz
+            initialDelaySeconds: 10
+            timeoutSeconds: 3
+            periodSeconds: 2
+          volumeMounts:
+            - name: socket-dir
+              mountPath: /csi
+            - mountPath: /dev
+              name: host-dev
+            - mountPath: /sys
+              name: host-sys
+            - mountPath: /run/mount
+              name: host-mount
+            - mountPath: /lib/modules
+              name: lib-modules
+              readOnly: true
+            - name: vitastor-config
+              mountPath: /etc/vitastor
+            - name: plugin-dir
+              mountPath: /var/lib/kubelet/plugins
+              mountPropagation: "Bidirectional"
+            - name: mountpoint-dir
+              mountPath: /var/lib/kubelet/pods
+              mountPropagation: "Bidirectional"
+        - name: liveness-probe
+          securityContext:
+            privileged: true
+          image: quay.io/k8scsi/livenessprobe:v1.1.0
+          args:
+            - "--csi-address=$(CSI_ENDPOINT)"
+            - "--health-port=9898"
+          env:
+            - name: CSI_ENDPOINT
+              value: unix://csi/csi.sock
+          volumeMounts:
+          - mountPath: /csi
+            name: socket-dir
+      volumes:
+        - name: socket-dir
+          hostPath:
+            path: /var/lib/kubelet/plugins/csi.vitastor.io
+            type: DirectoryOrCreate
+        - name: plugin-dir
+          hostPath:
+            path: /var/lib/kubelet/plugins
+            type: Directory
+        - name: mountpoint-dir
+          hostPath:
+            path: /var/lib/kubelet/pods
+            type: DirectoryOrCreate
+        - name: registration-dir
+          hostPath:
+            path: /var/lib/kubelet/plugins_registry/
+            type: Directory
+        - name: host-dev
+          hostPath:
+            path: /dev
+        - name: host-sys
+          hostPath:
+            path: /sys
+        - name: host-mount
+          hostPath:
+            path: /run/mount
+        - name: lib-modules
+          hostPath:
+            path: /lib/modules
+        - name: vitastor-config
+          configMap:
+            name: vitastor-config

csi/deploy/005-csi-provisioner-rbac.yaml

@@ -0,0 +1,102 @@
+---
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-provisioner
+
+---
+kind: ClusterRole
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  namespace: vitastor-system
+  name: vitastor-external-provisioner-runner
+rules:
+  - apiGroups: [""]
+    resources: ["nodes"]
+    verbs: ["get", "list", "watch"]
+  - apiGroups: [""]
+    resources: ["secrets"]
+    verbs: ["get", "list", "watch"]
+  - apiGroups: [""]
+    resources: ["events"]
+    verbs: ["list", "watch", "create", "update", "patch"]
+  - apiGroups: [""]
+    resources: ["persistentvolumes"]
+    verbs: ["get", "list", "watch", "create", "update", "delete", "patch"]
+  - apiGroups: [""]
+    resources: ["persistentvolumeclaims"]
+    verbs: ["get", "list", "watch", "update"]
+  - apiGroups: [""]
+    resources: ["persistentvolumeclaims/status"]
+    verbs: ["update", "patch"]
+  - apiGroups: ["storage.k8s.io"]
+    resources: ["storageclasses"]
+    verbs: ["get", "list", "watch"]
+  - apiGroups: ["snapshot.storage.k8s.io"]
+    resources: ["volumesnapshots"]
+    verbs: ["get", "list"]
+  - apiGroups: ["snapshot.storage.k8s.io"]
+    resources: ["volumesnapshotcontents"]
+    verbs: ["create", "get", "list", "watch", "update", "delete"]
+  - apiGroups: ["snapshot.storage.k8s.io"]
+    resources: ["volumesnapshotclasses"]
+    verbs: ["get", "list", "watch"]
+  - apiGroups: ["storage.k8s.io"]
+    resources: ["volumeattachments"]
+    verbs: ["get", "list", "watch", "update", "patch"]
+  - apiGroups: ["storage.k8s.io"]
+    resources: ["volumeattachments/status"]
+    verbs: ["patch"]
+  - apiGroups: ["storage.k8s.io"]
+    resources: ["csinodes"]
+    verbs: ["get", "list", "watch"]
+  - apiGroups: ["snapshot.storage.k8s.io"]
+    resources: ["volumesnapshotcontents/status"]
+    verbs: ["update"]
+  - apiGroups: [""]
+    resources: ["configmaps"]
+    verbs: ["get"]
+---
+kind: ClusterRoleBinding
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-provisioner-role
+subjects:
+  - kind: ServiceAccount
+    name: vitastor-csi-provisioner
+    namespace: vitastor-system
+roleRef:
+  kind: ClusterRole
+  name: vitastor-external-provisioner-runner
+  apiGroup: rbac.authorization.k8s.io
+
+---
+kind: Role
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  namespace: vitastor-system
+  name: vitastor-external-provisioner-cfg
+rules:
+  - apiGroups: [""]
+    resources: ["configmaps"]
+    verbs: ["get", "list", "watch", "create", "update", "delete"]
+  - apiGroups: ["coordination.k8s.io"]
+    resources: ["leases"]
+    verbs: ["get", "watch", "list", "delete", "update", "create"]
+
+---
+kind: RoleBinding
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  name: vitastor-csi-provisioner-role-cfg
+  namespace: vitastor-system
+subjects:
+  - kind: ServiceAccount
+    name: vitastor-csi-provisioner
+    namespace: vitastor-system
+roleRef:
+  kind: Role
+  name: vitastor-external-provisioner-cfg
+  apiGroup: rbac.authorization.k8s.io

csi/deploy/006-csi-provisioner-psp.yaml

@@ -0,0 +1,60 @@
+---
+apiVersion: policy/v1beta1
+kind: PodSecurityPolicy
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-provisioner-psp
+spec:
+  allowPrivilegeEscalation: true
+  allowedCapabilities:
+    - 'SYS_ADMIN'
+  fsGroup:
+    rule: RunAsAny
+  privileged: true
+  runAsUser:
+    rule: RunAsAny
+  seLinux:
+    rule: RunAsAny
+  supplementalGroups:
+    rule: RunAsAny
+  volumes:
+    - 'configMap'
+    - 'emptyDir'
+    - 'projected'
+    - 'secret'
+    - 'downwardAPI'
+    - 'hostPath'
+  allowedHostPaths:
+    - pathPrefix: '/dev'
+      readOnly: false
+    - pathPrefix: '/sys'
+      readOnly: false
+    - pathPrefix: '/lib/modules'
+      readOnly: true
+
+---
+kind: Role
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  namespace: vitastor-system
+  name: vitastor-csi-provisioner-psp
+rules:
+  - apiGroups: ['policy']
+    resources: ['podsecuritypolicies']
+    verbs: ['use']
+    resourceNames: ['vitastor-csi-provisioner-psp']
+
+---
+kind: RoleBinding
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  name: vitastor-csi-provisioner-psp
+  namespace: vitastor-system
+subjects:
+  - kind: ServiceAccount
+    name: vitastor-csi-provisioner
+    namespace: vitastor-system
+roleRef:
+  kind: Role
+  name: vitastor-csi-provisioner-psp
+  apiGroup: rbac.authorization.k8s.io

csi/deploy/007-csi-provisioner.yaml

@@ -0,0 +1,159 @@
+---
+kind: Service
+apiVersion: v1
+metadata:
+  namespace: vitastor-system
+  name: csi-vitastor-provisioner
+  labels:
+    app: csi-metrics
+spec:
+  selector:
+    app: csi-vitastor-provisioner
+  ports:
+    - name: http-metrics
+      port: 8080
+      protocol: TCP
+      targetPort: 8680
+
+---
+kind: Deployment
+apiVersion: apps/v1
+metadata:
+  namespace: vitastor-system
+  name: csi-vitastor-provisioner
+spec:
+  replicas: 3
+  selector:
+    matchLabels:
+      app: csi-vitastor-provisioner
+  template:
+    metadata:
+      namespace: vitastor-system
+      labels:
+        app: csi-vitastor-provisioner
+    spec:
+      affinity:
+        podAntiAffinity:
+          requiredDuringSchedulingIgnoredDuringExecution:
+            - labelSelector:
+                matchExpressions:
+                  - key: app
+                    operator: In
+                    values:
+                      - csi-vitastor-provisioner
+              topologyKey: "kubernetes.io/hostname"
+      serviceAccountName: vitastor-csi-provisioner
+      priorityClassName: system-cluster-critical
+      containers:
+        - name: csi-provisioner
+          image: k8s.gcr.io/sig-storage/csi-provisioner:v2.2.0
+          args:
+            - "--csi-address=$(ADDRESS)"
+            - "--v=5"
+            - "--timeout=150s"
+            - "--retry-interval-start=500ms"
+            - "--leader-election=true"
+            #  set it to true to use topology based provisioning
+            - "--feature-gates=Topology=false"
+            # if fstype is not specified in storageclass, ext4 is default
+            - "--default-fstype=ext4"
+            - "--extra-create-metadata=true"
+          env:
+            - name: ADDRESS
+              value: unix:///csi/csi-provisioner.sock
+          imagePullPolicy: "IfNotPresent"
+          volumeMounts:
+            - name: socket-dir
+              mountPath: /csi
+        - name: csi-snapshotter
+          image: k8s.gcr.io/sig-storage/csi-snapshotter:v4.0.0
+          args:
+            - "--csi-address=$(ADDRESS)"
+            - "--v=5"
+            - "--timeout=150s"
+            - "--leader-election=true"
+          env:
+            - name: ADDRESS
+              value: unix:///csi/csi-provisioner.sock
+          imagePullPolicy: "IfNotPresent"
+          securityContext:
+            privileged: true
+          volumeMounts:
+            - name: socket-dir
+              mountPath: /csi
+        - name: csi-attacher
+          image: k8s.gcr.io/sig-storage/csi-attacher:v3.1.0
+          args:
+            - "--v=5"
+            - "--csi-address=$(ADDRESS)"
+            - "--leader-election=true"
+            - "--retry-interval-start=500ms"
+          env:
+            - name: ADDRESS
+              value: /csi/csi-provisioner.sock
+          imagePullPolicy: "IfNotPresent"
+          volumeMounts:
+            - name: socket-dir
+              mountPath: /csi
+        - name: csi-resizer
+          image: k8s.gcr.io/sig-storage/csi-resizer:v1.1.0
+          args:
+            - "--csi-address=$(ADDRESS)"
+            - "--v=5"
+            - "--timeout=150s"
+            - "--leader-election"
+            - "--retry-interval-start=500ms"
+            - "--handle-volume-inuse-error=false"
+          env:
+            - name: ADDRESS
+              value: unix:///csi/csi-provisioner.sock
+          imagePullPolicy: "IfNotPresent"
+          volumeMounts:
+            - name: socket-dir
+              mountPath: /csi
+        - name: csi-vitastor
+          securityContext:
+            privileged: true
+            capabilities:
+              add: ["SYS_ADMIN"]
+          image: vitalif/vitastor-csi:v0.6.5
+          args:
+            - "--node=$(NODE_ID)"
+            - "--endpoint=$(CSI_ENDPOINT)"
+          env:
+            - name: NODE_ID
+              valueFrom:
+                fieldRef:
+                  fieldPath: spec.nodeName
+            - name: CSI_ENDPOINT
+              value: unix:///csi/csi-provisioner.sock
+          imagePullPolicy: "IfNotPresent"
+          volumeMounts:
+            - name: socket-dir
+              mountPath: /csi
+            - mountPath: /dev
+              name: host-dev
+            - mountPath: /sys
+              name: host-sys
+            - mountPath: /lib/modules
+              name: lib-modules
+              readOnly: true
+            - name: vitastor-config
+              mountPath: /etc/vitastor
+      volumes:
+        - name: host-dev
+          hostPath:
+            path: /dev
+        - name: host-sys
+          hostPath:
+            path: /sys
+        - name: lib-modules
+          hostPath:
+            path: /lib/modules
+        - name: socket-dir
+          emptyDir: {
+            medium: "Memory"
+          }
+        - name: vitastor-config
+          configMap:
+            name: vitastor-config

csi/deploy/008-csi-driver.yaml

@@ -0,0 +1,11 @@
+---
+# if Kubernetes version is less than 1.18 change
+# apiVersion to storage.k8s.io/v1betav1
+apiVersion: storage.k8s.io/v1
+kind: CSIDriver
+metadata:
+  namespace: vitastor-system
+  name: csi.vitastor.io
+spec:
+  attachRequired: true
+  podInfoOnMount: false

csi/deploy/009-storage-class.yaml

@@ -0,0 +1,19 @@
+---
+apiVersion: storage.k8s.io/v1
+kind: StorageClass
+metadata:
+  namespace: vitastor-system
+  name: vitastor
+  annotations:
+    storageclass.kubernetes.io/is-default-class: "true"
+provisioner: csi.vitastor.io
+volumeBindingMode: Immediate
+parameters:
+  etcdVolumePrefix: ""
+  poolId: "1"
+  # you can choose other configuration file if you have it in the config map
+  #configPath: "/etc/vitastor/vitastor.conf"
+  # you can also specify etcdUrl here, maybe to connect to another Vitastor cluster
+  # multiple etcdUrls may be specified, delimited by comma
+  #etcdUrl: "http://192.168.7.2:2379"
+  #etcdPrefix: "/vitastor"

csi/deploy/example-pvc.yaml

@@ -0,0 +1,12 @@
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: test-vitastor-pvc
+spec:
+  storageClassName: vitastor
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 10Gi

csi/go.mod

@@ -0,0 +1,35 @@
+module vitastor.io/csi
+
+go 1.15
+
+require (
+	github.com/container-storage-interface/spec v1.4.0
+	github.com/coreos/bbolt v0.0.0-00010101000000-000000000000 // indirect
+	github.com/coreos/etcd v3.3.25+incompatible // indirect
+	github.com/coreos/go-semver v0.3.0 // indirect
+	github.com/coreos/go-systemd v0.0.0-20191104093116-d3cd4ed1dbcf // indirect
+	github.com/coreos/pkg v0.0.0-20180928190104-399ea9e2e55f // indirect
+	github.com/dustin/go-humanize v1.0.0 // indirect
+	github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b
+	github.com/gorilla/websocket v1.4.2 // indirect
+	github.com/grpc-ecosystem/go-grpc-middleware v1.3.0 // indirect
+	github.com/grpc-ecosystem/go-grpc-prometheus v1.2.0 // indirect
+	github.com/grpc-ecosystem/grpc-gateway v1.16.0 // indirect
+	github.com/jonboulle/clockwork v0.2.2 // indirect
+	github.com/kubernetes-csi/csi-lib-utils v0.9.1
+	github.com/soheilhy/cmux v0.1.5 // indirect
+	github.com/tmc/grpc-websocket-proxy v0.0.0-20201229170055-e5319fda7802 // indirect
+	github.com/xiang90/probing v0.0.0-20190116061207-43a291ad63a2 // indirect
+	go.etcd.io/bbolt v0.0.0-00010101000000-000000000000 // indirect
+	go.etcd.io/etcd v3.3.25+incompatible
+	golang.org/x/net v0.0.0-20201202161906-c7110b5ffcbb
+	google.golang.org/grpc v1.33.1
+	k8s.io/klog v1.0.0
+	k8s.io/utils v0.0.0-20210305010621-2afb4311ab10
+)
+
+replace github.com/coreos/bbolt => go.etcd.io/bbolt v1.3.5
+
+replace go.etcd.io/bbolt => github.com/coreos/bbolt v1.3.5
+
+replace google.golang.org/grpc => google.golang.org/grpc v1.25.1

csi/src/config.go

@@ -0,0 +1,22 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
+
+package vitastor
+
+const (
+    vitastorCSIDriverName    = "csi.vitastor.io"
+    vitastorCSIDriverVersion = "0.6.5"
+)
+
+// Config struct fills the parameters of request or user input
+type Config struct
+{
+    Endpoint string
+    NodeID   string
+}
+
+// NewConfig returns config struct to initialize new driver
+func NewConfig() *Config
+{
+    return &Config{}
+}

csi/src/controllerserver.go

@@ -0,0 +1,530 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
+
+package vitastor
+
+import (
+    "context"
+    "encoding/json"
+    "strings"
+    "bytes"
+    "strconv"
+    "time"
+    "fmt"
+    "os"
+    "os/exec"
+    "io/ioutil"
+
+    "github.com/kubernetes-csi/csi-lib-utils/protosanitizer"
+    "k8s.io/klog"
+
+    "google.golang.org/grpc/codes"
+    "google.golang.org/grpc/status"
+
+    "go.etcd.io/etcd/clientv3"
+
+    "github.com/container-storage-interface/spec/lib/go/csi"
+)
+
+const (
+    KB int64 = 1024
+    MB int64 = 1024 * KB
+    GB int64 = 1024 * MB
+    TB int64 = 1024 * GB
+    ETCD_TIMEOUT time.Duration = 15*time.Second
+)
+
+type InodeIndex struct
+{
+    Id uint64 `json:"id"`
+    PoolId uint64 `json:"pool_id"`
+}
+
+type InodeConfig struct
+{
+    Name string `json:"name"`
+    Size uint64 `json:"size,omitempty"`
+    ParentPool uint64 `json:"parent_pool,omitempty"`
+    ParentId uint64 `json:"parent_id,omitempty"`
+    Readonly bool `json:"readonly,omitempty"`
+}
+
+type ControllerServer struct
+{
+    *Driver
+}
+
+// NewControllerServer create new instance controller
+func NewControllerServer(driver *Driver) *ControllerServer
+{
+    return &ControllerServer{
+        Driver: driver,
+    }
+}
+
+func GetConnectionParams(params map[string]string) (map[string]string, []string, string)
+{
+    ctxVars := make(map[string]string)
+    configPath := params["configPath"]
+    if (configPath == "")
+    {
+        configPath = "/etc/vitastor/vitastor.conf"
+    }
+    else
+    {
+        ctxVars["configPath"] = configPath
+    }
+    config := make(map[string]interface{})
+    if configFD, err := os.Open(configPath); err == nil
+    {
+        defer configFD.Close()
+        data, _ := ioutil.ReadAll(configFD)
+        json.Unmarshal(data, &config)
+    }
+    // Try to load prefix & etcd URL from the config
+    var etcdUrl []string
+    if (params["etcdUrl"] != "")
+    {
+        ctxVars["etcdUrl"] = params["etcdUrl"]
+        etcdUrl = strings.Split(params["etcdUrl"], ",")
+    }
+    if (len(etcdUrl) == 0)
+    {
+        switch config["etcd_address"].(type)
+        {
+        case string:
+            etcdUrl = strings.Split(config["etcd_address"].(string), ",")
+        case []string:
+            etcdUrl = config["etcd_address"].([]string)
+        }
+    }
+    etcdPrefix := params["etcdPrefix"]
+    if (etcdPrefix == "")
+    {
+        etcdPrefix, _ = config["etcd_prefix"].(string)
+        if (etcdPrefix == "")
+        {
+            etcdPrefix = "/vitastor"
+        }
+    }
+    else
+    {
+        ctxVars["etcdPrefix"] = etcdPrefix
+    }
+    return ctxVars, etcdUrl, etcdPrefix
+}
+
+// Create the volume
+func (cs *ControllerServer) CreateVolume(ctx context.Context, req *csi.CreateVolumeRequest) (*csi.CreateVolumeResponse, error)
+{
+    klog.Infof("received controller create volume request %+v", protosanitizer.StripSecrets(req))
+    if (req == nil)
+    {
+        return nil, status.Errorf(codes.InvalidArgument, "request cannot be empty")
+    }
+    if (req.GetName() == "")
+    {
+        return nil, status.Error(codes.InvalidArgument, "name is a required field")
+    }
+    volumeCapabilities := req.GetVolumeCapabilities()
+    if (volumeCapabilities == nil)
+    {
+        return nil, status.Error(codes.InvalidArgument, "volume capabilities is a required field")
+    }
+
+    etcdVolumePrefix := req.Parameters["etcdVolumePrefix"]
+    poolId, _ := strconv.ParseUint(req.Parameters["poolId"], 10, 64)
+    if (poolId == 0)
+    {
+        return nil, status.Error(codes.InvalidArgument, "poolId is missing in storage class configuration")
+    }
+
+    volName := etcdVolumePrefix + req.GetName()
+    volSize := 1 * GB
+    if capRange := req.GetCapacityRange(); capRange != nil
+    {
+        volSize = ((capRange.GetRequiredBytes() + MB - 1) / MB) * MB
+    }
+
+    // FIXME: The following should PROBABLY be implemented externally in a management tool
+
+    ctxVars, etcdUrl, etcdPrefix := GetConnectionParams(req.Parameters)
+    if (len(etcdUrl) == 0)
+    {
+        return nil, status.Error(codes.InvalidArgument, "no etcdUrl in storage class configuration and no etcd_address in vitastor.conf")
+    }
+
+    // Connect to etcd
+    cli, err := clientv3.New(clientv3.Config{
+        DialTimeout: ETCD_TIMEOUT,
+        Endpoints: etcdUrl,
+    })
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "failed to connect to etcd at "+strings.Join(etcdUrl, ",")+": "+err.Error())
+    }
+    defer cli.Close()
+
+    var imageId uint64 = 0
+    for
+    {
+        // Check if the image exists
+        ctx, cancel := context.WithTimeout(context.Background(), ETCD_TIMEOUT)
+        resp, err := cli.Get(ctx, etcdPrefix+"/index/image/"+volName)
+        cancel()
+        if (err != nil)
+        {
+            return nil, status.Error(codes.Internal, "failed to read key from etcd: "+err.Error())
+        }
+        if (len(resp.Kvs) > 0)
+        {
+            kv := resp.Kvs[0]
+            var v InodeIndex
+            err := json.Unmarshal(kv.Value, &v)
+            if (err != nil)
+            {
+                return nil, status.Error(codes.Internal, "invalid /index/image/"+volName+" key in etcd: "+err.Error())
+            }
+            poolId = v.PoolId
+            imageId = v.Id
+            inodeCfgKey := fmt.Sprintf("/config/inode/%d/%d", poolId, imageId)
+            ctx, cancel := context.WithTimeout(context.Background(), ETCD_TIMEOUT)
+            resp, err := cli.Get(ctx, etcdPrefix+inodeCfgKey)
+            cancel()
+            if (err != nil)
+            {
+                return nil, status.Error(codes.Internal, "failed to read key from etcd: "+err.Error())
+            }
+            if (len(resp.Kvs) == 0)
+            {
+                return nil, status.Error(codes.Internal, "missing "+inodeCfgKey+" key in etcd")
+            }
+            var inodeCfg InodeConfig
+            err = json.Unmarshal(resp.Kvs[0].Value, &inodeCfg)
+            if (err != nil)
+            {
+                return nil, status.Error(codes.Internal, "invalid "+inodeCfgKey+" key in etcd: "+err.Error())
+            }
+            if (inodeCfg.Size < uint64(volSize))
+            {
+                return nil, status.Error(codes.Internal, "image "+volName+" is already created, but size is less than expected")
+            }
+        }
+        else
+        {
+            // Find a free ID
+            // Create image metadata in a transaction verifying that the image doesn't exist yet AND ID is still free
+            maxIdKey := fmt.Sprintf("%s/index/maxid/%d", etcdPrefix, poolId)
+            ctx, cancel := context.WithTimeout(context.Background(), ETCD_TIMEOUT)
+            resp, err := cli.Get(ctx, maxIdKey)
+            cancel()
+            if (err != nil)
+            {
+                return nil, status.Error(codes.Internal, "failed to read key from etcd: "+err.Error())
+            }
+            var modRev int64
+            var nextId uint64
+            if (len(resp.Kvs) > 0)
+            {
+                var err error
+                nextId, err = strconv.ParseUint(string(resp.Kvs[0].Value), 10, 64)
+                if (err != nil)
+                {
+                    return nil, status.Error(codes.Internal, maxIdKey+" contains invalid ID")
+                }
+                modRev = resp.Kvs[0].ModRevision
+                nextId++
+            }
+            else
+            {
+                nextId = 1
+            }
+            inodeIdxJson, _ := json.Marshal(InodeIndex{
+                Id: nextId,
+                PoolId: poolId,
+            })
+            inodeCfgJson, _ := json.Marshal(InodeConfig{
+                Name: volName,
+                Size: uint64(volSize),
+            })
+            ctx, cancel = context.WithTimeout(context.Background(), ETCD_TIMEOUT)
+            txnResp, err := cli.Txn(ctx).If(
+                clientv3.Compare(clientv3.ModRevision(fmt.Sprintf("%s/index/maxid/%d", etcdPrefix, poolId)), "=", modRev),
+                clientv3.Compare(clientv3.CreateRevision(fmt.Sprintf("%s/index/image/%s", etcdPrefix, volName)), "=", 0),
+                clientv3.Compare(clientv3.CreateRevision(fmt.Sprintf("%s/config/inode/%d/%d", etcdPrefix, poolId, nextId)), "=", 0),
+            ).Then(
+                clientv3.OpPut(fmt.Sprintf("%s/index/maxid/%d", etcdPrefix, poolId), fmt.Sprintf("%d", nextId)),
+                clientv3.OpPut(fmt.Sprintf("%s/index/image/%s", etcdPrefix, volName), string(inodeIdxJson)),
+                clientv3.OpPut(fmt.Sprintf("%s/config/inode/%d/%d", etcdPrefix, poolId, nextId), string(inodeCfgJson)),
+            ).Commit()
+            cancel()
+            if (err != nil)
+            {
+                return nil, status.Error(codes.Internal, "failed to commit transaction in etcd: "+err.Error())
+            }
+            if (txnResp.Succeeded)
+            {
+                imageId = nextId
+                break
+            }
+            // Start over if the transaction fails
+        }
+    }
+
+    ctxVars["name"] = volName
+    volumeIdJson, _ := json.Marshal(ctxVars)
+    return &csi.CreateVolumeResponse{
+        Volume: &csi.Volume{
+            // Ugly, but VolumeContext isn't passed to DeleteVolume :-(
+            VolumeId: string(volumeIdJson),
+            CapacityBytes: volSize,
+        },
+    }, nil
+}
+
+// DeleteVolume deletes the given volume
+func (cs *ControllerServer) DeleteVolume(ctx context.Context, req *csi.DeleteVolumeRequest) (*csi.DeleteVolumeResponse, error)
+{
+    klog.Infof("received controller delete volume request %+v", protosanitizer.StripSecrets(req))
+    if (req == nil)
+    {
+        return nil, status.Error(codes.InvalidArgument, "request cannot be empty")
+    }
+
+    ctxVars := make(map[string]string)
+    err := json.Unmarshal([]byte(req.VolumeId), &ctxVars)
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "volume ID not in JSON format")
+    }
+    volName := ctxVars["name"]
+
+    _, etcdUrl, etcdPrefix := GetConnectionParams(ctxVars)
+    if (len(etcdUrl) == 0)
+    {
+        return nil, status.Error(codes.InvalidArgument, "no etcdUrl in storage class configuration and no etcd_address in vitastor.conf")
+    }
+
+    cli, err := clientv3.New(clientv3.Config{
+        DialTimeout: ETCD_TIMEOUT,
+        Endpoints: etcdUrl,
+    })
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "failed to connect to etcd at "+strings.Join(etcdUrl, ",")+": "+err.Error())
+    }
+    defer cli.Close()
+
+    // Find inode by name
+    ctx, cancel := context.WithTimeout(context.Background(), ETCD_TIMEOUT)
+    resp, err := cli.Get(ctx, etcdPrefix+"/index/image/"+volName)
+    cancel()
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "failed to read key from etcd: "+err.Error())
+    }
+    if (len(resp.Kvs) == 0)
+    {
+        return nil, status.Error(codes.NotFound, "volume "+volName+" does not exist")
+    }
+    var idx InodeIndex
+    err = json.Unmarshal(resp.Kvs[0].Value, &idx)
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "invalid /index/image/"+volName+" key in etcd: "+err.Error())
+    }
+
+    // Get inode config
+    inodeCfgKey := fmt.Sprintf("%s/config/inode/%d/%d", etcdPrefix, idx.PoolId, idx.Id)
+    ctx, cancel = context.WithTimeout(context.Background(), ETCD_TIMEOUT)
+    resp, err = cli.Get(ctx, inodeCfgKey)
+    cancel()
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "failed to read key from etcd: "+err.Error())
+    }
+    if (len(resp.Kvs) == 0)
+    {
+        return nil, status.Error(codes.NotFound, "volume "+volName+" does not exist")
+    }
+    var inodeCfg InodeConfig
+    err = json.Unmarshal(resp.Kvs[0].Value, &inodeCfg)
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "invalid "+inodeCfgKey+" key in etcd: "+err.Error())
+    }
+
+    // Delete inode data by invoking vitastor-rm
+    args := []string{
+        "--etcd_address", strings.Join(etcdUrl, ","),
+        "--pool", fmt.Sprintf("%d", idx.PoolId),
+        "--inode", fmt.Sprintf("%d", idx.Id),
+    }
+    if (ctxVars["configPath"] != "")
+    {
+        args = append(args, "--config_path", ctxVars["configPath"])
+    }
+    c := exec.Command("/usr/bin/vitastor-rm", args...)
+    var stderr bytes.Buffer
+    c.Stdout = nil
+    c.Stderr = &stderr
+    err = c.Run()
+    stderrStr := string(stderr.Bytes())
+    if (err != nil)
+    {
+        klog.Errorf("vitastor-rm failed: %s, status %s\n", stderrStr, err)
+        return nil, status.Error(codes.Internal, stderrStr+" (status "+err.Error()+")")
+    }
+
+    // Delete inode config in etcd
+    ctx, cancel = context.WithTimeout(context.Background(), ETCD_TIMEOUT)
+    txnResp, err := cli.Txn(ctx).Then(
+        clientv3.OpDelete(fmt.Sprintf("%s/index/image/%s", etcdPrefix, volName)),
+        clientv3.OpDelete(fmt.Sprintf("%s/config/inode/%d/%d", etcdPrefix, idx.PoolId, idx.Id)),
+    ).Commit()
+    cancel()
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "failed to delete keys in etcd: "+err.Error())
+    }
+    if (!txnResp.Succeeded)
+    {
+        return nil, status.Error(codes.Internal, "failed to delete keys in etcd: transaction failed")
+    }
+
+    return &csi.DeleteVolumeResponse{}, nil
+}
+
+// ControllerPublishVolume return Unimplemented error
+func (cs *ControllerServer) ControllerPublishVolume(ctx context.Context, req *csi.ControllerPublishVolumeRequest) (*csi.ControllerPublishVolumeResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// ControllerUnpublishVolume return Unimplemented error
+func (cs *ControllerServer) ControllerUnpublishVolume(ctx context.Context, req *csi.ControllerUnpublishVolumeRequest) (*csi.ControllerUnpublishVolumeResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// ValidateVolumeCapabilities checks whether the volume capabilities requested are supported.
+func (cs *ControllerServer) ValidateVolumeCapabilities(ctx context.Context, req *csi.ValidateVolumeCapabilitiesRequest) (*csi.ValidateVolumeCapabilitiesResponse, error)
+{
+    klog.Infof("received controller validate volume capability request %+v", protosanitizer.StripSecrets(req))
+    if (req == nil)
+    {
+        return nil, status.Errorf(codes.InvalidArgument, "request is nil")
+    }
+    volumeID := req.GetVolumeId()
+    if (volumeID == "")
+    {
+        return nil, status.Error(codes.InvalidArgument, "volumeId is nil")
+    }
+    volumeCapabilities := req.GetVolumeCapabilities()
+    if (volumeCapabilities == nil)
+    {
+        return nil, status.Error(codes.InvalidArgument, "volumeCapabilities is nil")
+    }
+
+    var volumeCapabilityAccessModes []*csi.VolumeCapability_AccessMode
+    for _, mode := range []csi.VolumeCapability_AccessMode_Mode{
+        csi.VolumeCapability_AccessMode_SINGLE_NODE_WRITER,
+        csi.VolumeCapability_AccessMode_MULTI_NODE_MULTI_WRITER,
+    } {
+        volumeCapabilityAccessModes = append(volumeCapabilityAccessModes, &csi.VolumeCapability_AccessMode{Mode: mode})
+    }
+
+    capabilitySupport := false
+    for _, capability := range volumeCapabilities
+    {
+        for _, volumeCapabilityAccessMode := range volumeCapabilityAccessModes
+        {
+            if (volumeCapabilityAccessMode.Mode == capability.AccessMode.Mode)
+            {
+                capabilitySupport = true
+            }
+        }
+    }
+
+    if (!capabilitySupport)
+    {
+        return nil, status.Errorf(codes.NotFound, "%v not supported", req.GetVolumeCapabilities())
+    }
+
+    return &csi.ValidateVolumeCapabilitiesResponse{
+        Confirmed: &csi.ValidateVolumeCapabilitiesResponse_Confirmed{
+            VolumeCapabilities: req.VolumeCapabilities,
+        },
+    }, nil
+}
+
+// ListVolumes returns a list of volumes
+func (cs *ControllerServer) ListVolumes(ctx context.Context, req *csi.ListVolumesRequest) (*csi.ListVolumesResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// GetCapacity returns the capacity of the storage pool
+func (cs *ControllerServer) GetCapacity(ctx context.Context, req *csi.GetCapacityRequest) (*csi.GetCapacityResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// ControllerGetCapabilities returns the capabilities of the controller service.
+func (cs *ControllerServer) ControllerGetCapabilities(ctx context.Context, req *csi.ControllerGetCapabilitiesRequest) (*csi.ControllerGetCapabilitiesResponse, error)
+{
+    functionControllerServerCapabilities := func(cap csi.ControllerServiceCapability_RPC_Type) *csi.ControllerServiceCapability
+    {
+        return &csi.ControllerServiceCapability{
+            Type: &csi.ControllerServiceCapability_Rpc{
+                Rpc: &csi.ControllerServiceCapability_RPC{
+                    Type: cap,
+                },
+            },
+        }
+    }
+
+    var controllerServerCapabilities []*csi.ControllerServiceCapability
+    for _, capability := range []csi.ControllerServiceCapability_RPC_Type{
+        csi.ControllerServiceCapability_RPC_CREATE_DELETE_VOLUME,
+        csi.ControllerServiceCapability_RPC_LIST_VOLUMES,
+        csi.ControllerServiceCapability_RPC_EXPAND_VOLUME,
+        csi.ControllerServiceCapability_RPC_CREATE_DELETE_SNAPSHOT,
+    } {
+        controllerServerCapabilities = append(controllerServerCapabilities, functionControllerServerCapabilities(capability))
+    }
+
+    return &csi.ControllerGetCapabilitiesResponse{
+        Capabilities: controllerServerCapabilities,
+    }, nil
+}
+
+// CreateSnapshot create snapshot of an existing PV
+func (cs *ControllerServer) CreateSnapshot(ctx context.Context, req *csi.CreateSnapshotRequest) (*csi.CreateSnapshotResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// DeleteSnapshot delete provided snapshot of a PV
+func (cs *ControllerServer) DeleteSnapshot(ctx context.Context, req *csi.DeleteSnapshotRequest) (*csi.DeleteSnapshotResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// ListSnapshots list the snapshots of a PV
+func (cs *ControllerServer) ListSnapshots(ctx context.Context, req *csi.ListSnapshotsRequest) (*csi.ListSnapshotsResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// ControllerExpandVolume resizes a volume
+func (cs *ControllerServer) ControllerExpandVolume(ctx context.Context, req *csi.ControllerExpandVolumeRequest) (*csi.ControllerExpandVolumeResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// ControllerGetVolume get volume info
+func (cs *ControllerServer) ControllerGetVolume(ctx context.Context, req *csi.ControllerGetVolumeRequest) (*csi.ControllerGetVolumeResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}

csi/src/grpc.go

@@ -0,0 +1,137 @@
+/*
+Copyright 2017 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package vitastor
+
+import (
+    "fmt"
+    "net"
+    "os"
+    "strings"
+    "sync"
+
+    "github.com/golang/glog"
+    "golang.org/x/net/context"
+    "google.golang.org/grpc"
+
+    "github.com/container-storage-interface/spec/lib/go/csi"
+    "github.com/kubernetes-csi/csi-lib-utils/protosanitizer"
+)
+
+// Defines Non blocking GRPC server interfaces
+type NonBlockingGRPCServer interface {
+    // Start services at the endpoint
+    Start(endpoint string, ids csi.IdentityServer, cs csi.ControllerServer, ns csi.NodeServer)
+    // Waits for the service to stop
+    Wait()
+    // Stops the service gracefully
+    Stop()
+    // Stops the service forcefully
+    ForceStop()
+}
+
+func NewNonBlockingGRPCServer() NonBlockingGRPCServer {
+    return &nonBlockingGRPCServer{}
+}
+
+// NonBlocking server
+type nonBlockingGRPCServer struct {
+    wg     sync.WaitGroup
+    server *grpc.Server
+}
+
+func (s *nonBlockingGRPCServer) Start(endpoint string, ids csi.IdentityServer, cs csi.ControllerServer, ns csi.NodeServer) {
+
+    s.wg.Add(1)
+
+    go s.serve(endpoint, ids, cs, ns)
+
+    return
+}
+
+func (s *nonBlockingGRPCServer) Wait() {
+    s.wg.Wait()
+}
+
+func (s *nonBlockingGRPCServer) Stop() {
+    s.server.GracefulStop()
+}
+
+func (s *nonBlockingGRPCServer) ForceStop() {
+    s.server.Stop()
+}
+
+func (s *nonBlockingGRPCServer) serve(endpoint string, ids csi.IdentityServer, cs csi.ControllerServer, ns csi.NodeServer) {
+
+    proto, addr, err := ParseEndpoint(endpoint)
+    if err != nil {
+        glog.Fatal(err.Error())
+    }
+
+    if proto == "unix" {
+        addr = "/" + addr
+        if err := os.Remove(addr); err != nil && !os.IsNotExist(err) {
+            glog.Fatalf("Failed to remove %s, error: %s", addr, err.Error())
+        }
+    }
+
+    listener, err := net.Listen(proto, addr)
+    if err != nil {
+        glog.Fatalf("Failed to listen: %v", err)
+    }
+
+    opts := []grpc.ServerOption{
+        grpc.UnaryInterceptor(logGRPC),
+    }
+    server := grpc.NewServer(opts...)
+    s.server = server
+
+    if ids != nil {
+        csi.RegisterIdentityServer(server, ids)
+    }
+    if cs != nil {
+        csi.RegisterControllerServer(server, cs)
+    }
+    if ns != nil {
+        csi.RegisterNodeServer(server, ns)
+    }
+
+    glog.Infof("Listening for connections on address: %#v", listener.Addr())
+
+    server.Serve(listener)
+}
+
+func ParseEndpoint(ep string) (string, string, error) {
+    if strings.HasPrefix(strings.ToLower(ep), "unix://") || strings.HasPrefix(strings.ToLower(ep), "tcp://") {
+        s := strings.SplitN(ep, "://", 2)
+        if s[1] != "" {
+            return s[0], s[1], nil
+        }
+    }
+    return "", "", fmt.Errorf("Invalid endpoint: %v", ep)
+}
+
+func logGRPC(ctx context.Context, req interface{}, info *grpc.UnaryServerInfo, handler grpc.UnaryHandler) (interface{}, error) {
+    glog.V(3).Infof("GRPC call: %s", info.FullMethod)
+    glog.V(5).Infof("GRPC request: %s", protosanitizer.StripSecrets(req))
+    resp, err := handler(ctx, req)
+    if err != nil {
+        glog.Errorf("GRPC error: %v", err)
+    } else {
+        glog.V(5).Infof("GRPC response: %s", protosanitizer.StripSecrets(resp))
+    }
+    return resp, err
+}

csi/src/identityserver.go

@@ -0,0 +1,60 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
+
+package vitastor
+
+import (
+    "context"
+
+    "github.com/kubernetes-csi/csi-lib-utils/protosanitizer"
+    "k8s.io/klog"
+
+    "github.com/container-storage-interface/spec/lib/go/csi"
+)
+
+// IdentityServer struct of Vitastor CSI driver with supported methods of CSI identity server spec.
+type IdentityServer struct
+{
+    *Driver
+}
+
+// NewIdentityServer create new instance identity
+func NewIdentityServer(driver *Driver) *IdentityServer
+{
+    return &IdentityServer{
+        Driver: driver,
+    }
+}
+
+// GetPluginInfo returns metadata of the plugin
+func (is *IdentityServer) GetPluginInfo(ctx context.Context, req *csi.GetPluginInfoRequest) (*csi.GetPluginInfoResponse, error)
+{
+    klog.Infof("received identity plugin info request %+v", protosanitizer.StripSecrets(req))
+    return &csi.GetPluginInfoResponse{
+        Name:          vitastorCSIDriverName,
+        VendorVersion: vitastorCSIDriverVersion,
+    }, nil
+}
+
+// GetPluginCapabilities returns available capabilities of the plugin
+func (is *IdentityServer) GetPluginCapabilities(ctx context.Context, req *csi.GetPluginCapabilitiesRequest) (*csi.GetPluginCapabilitiesResponse, error)
+{
+    klog.Infof("received identity plugin capabilities request %+v", protosanitizer.StripSecrets(req))
+    return &csi.GetPluginCapabilitiesResponse{
+        Capabilities: []*csi.PluginCapability{
+            {
+                Type: &csi.PluginCapability_Service_{
+                    Service: &csi.PluginCapability_Service{
+                        Type: csi.PluginCapability_Service_CONTROLLER_SERVICE,
+                    },
+                },
+            },
+        },
+    }, nil
+}
+
+// Probe returns the health and readiness of the plugin
+func (is *IdentityServer) Probe(ctx context.Context, req *csi.ProbeRequest) (*csi.ProbeResponse, error)
+{
+    return &csi.ProbeResponse{}, nil
+}

csi/src/nodeserver.go

@@ -0,0 +1,279 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
+
+package vitastor
+
+import (
+    "context"
+    "os"
+    "os/exec"
+    "encoding/json"
+    "strings"
+    "bytes"
+
+    "google.golang.org/grpc/codes"
+    "google.golang.org/grpc/status"
+    "k8s.io/utils/mount"
+    utilexec "k8s.io/utils/exec"
+
+    "github.com/container-storage-interface/spec/lib/go/csi"
+    "github.com/kubernetes-csi/csi-lib-utils/protosanitizer"
+    "k8s.io/klog"
+)
+
+// NodeServer struct of Vitastor CSI driver with supported methods of CSI node server spec.
+type NodeServer struct
+{
+    *Driver
+    mounter mount.Interface
+}
+
+// NewNodeServer create new instance node
+func NewNodeServer(driver *Driver) *NodeServer
+{
+    return &NodeServer{
+        Driver: driver,
+        mounter: mount.New(""),
+    }
+}
+
+// NodeStageVolume mounts the volume to a staging path on the node.
+func (ns *NodeServer) NodeStageVolume(ctx context.Context, req *csi.NodeStageVolumeRequest) (*csi.NodeStageVolumeResponse, error)
+{
+    return &csi.NodeStageVolumeResponse{}, nil
+}
+
+// NodeUnstageVolume unstages the volume from the staging path
+func (ns *NodeServer) NodeUnstageVolume(ctx context.Context, req *csi.NodeUnstageVolumeRequest) (*csi.NodeUnstageVolumeResponse, error)
+{
+    return &csi.NodeUnstageVolumeResponse{}, nil
+}
+
+func Contains(list []string, s string) bool
+{
+    for i := 0; i < len(list); i++
+    {
+        if (list[i] == s)
+        {
+            return true
+        }
+    }
+    return false
+}
+
+// NodePublishVolume mounts the volume mounted to the staging path to the target path
+func (ns *NodeServer) NodePublishVolume(ctx context.Context, req *csi.NodePublishVolumeRequest) (*csi.NodePublishVolumeResponse, error)
+{
+    klog.Infof("received node publish volume request %+v", protosanitizer.StripSecrets(req))
+
+    targetPath := req.GetTargetPath()
+
+    // Check that it's not already mounted
+    free, error := mount.IsNotMountPoint(ns.mounter, targetPath)
+    if (error != nil)
+    {
+        if (os.IsNotExist(error))
+        {
+            error := os.MkdirAll(targetPath, 0777)
+            if (error != nil)
+            {
+                return nil, status.Error(codes.Internal, error.Error())
+            }
+            free = true
+        }
+        else
+        {
+            return nil, status.Error(codes.Internal, error.Error())
+        }
+    }
+    if (!free)
+    {
+        return &csi.NodePublishVolumeResponse{}, nil
+    }
+
+    ctxVars := make(map[string]string)
+    err := json.Unmarshal([]byte(req.VolumeId), &ctxVars)
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, "volume ID not in JSON format")
+    }
+    volName := ctxVars["name"]
+
+    _, etcdUrl, etcdPrefix := GetConnectionParams(ctxVars)
+    if (len(etcdUrl) == 0)
+    {
+        return nil, status.Error(codes.InvalidArgument, "no etcdUrl in storage class configuration and no etcd_address in vitastor.conf")
+    }
+
+    // Map NBD device
+    // FIXME: Check if already mapped
+    args := []string{
+        "map", "--etcd_address", strings.Join(etcdUrl, ","),
+        "--etcd_prefix", etcdPrefix,
+        "--image", volName,
+    };
+    if (ctxVars["configPath"] != "")
+    {
+        args = append(args, "--config_path", ctxVars["configPath"])
+    }
+    if (req.GetReadonly())
+    {
+        args = append(args, "--readonly", "1")
+    }
+    c := exec.Command("/usr/bin/vitastor-nbd", args...)
+    var stdout, stderr bytes.Buffer
+    c.Stdout, c.Stderr = &stdout, &stderr
+    err = c.Run()
+    stdoutStr, stderrStr := string(stdout.Bytes()), string(stderr.Bytes())
+    if (err != nil)
+    {
+        klog.Errorf("vitastor-nbd map failed: %s, status %s\n", stdoutStr+stderrStr, err)
+        return nil, status.Error(codes.Internal, stdoutStr+stderrStr+" (status "+err.Error()+")")
+    }
+    devicePath := strings.TrimSpace(stdoutStr)
+
+    // Check existing format
+    diskMounter := &mount.SafeFormatAndMount{Interface: ns.mounter, Exec: utilexec.New()}
+    existingFormat, err := diskMounter.GetDiskFormat(devicePath)
+    if (err != nil)
+    {
+        klog.Errorf("failed to get disk format for path %s, error: %v", err)
+        // unmap NBD device
+        unmapOut, unmapErr := exec.Command("/usr/bin/vitastor-nbd", "unmap", devicePath).CombinedOutput()
+        if (unmapErr != nil)
+        {
+            klog.Errorf("failed to unmap NBD device %s: %s, error: %v", devicePath, unmapOut, unmapErr)
+        }
+        return nil, err
+    }
+
+    // Format the device (ext4 or xfs)
+    fsType := req.GetVolumeCapability().GetMount().GetFsType()
+    isBlock := req.GetVolumeCapability().GetBlock() != nil
+    opt := req.GetVolumeCapability().GetMount().GetMountFlags()
+    opt = append(opt, "_netdev")
+    if ((req.VolumeCapability.AccessMode.Mode == csi.VolumeCapability_AccessMode_MULTI_NODE_READER_ONLY ||
+        req.VolumeCapability.AccessMode.Mode == csi.VolumeCapability_AccessMode_SINGLE_NODE_READER_ONLY) &&
+        !Contains(opt, "ro"))
+    {
+        opt = append(opt, "ro")
+    }
+    if (fsType == "xfs")
+    {
+        opt = append(opt, "nouuid")
+    }
+    readOnly := Contains(opt, "ro")
+    if (existingFormat == "" && !readOnly)
+    {
+        args := []string{}
+        switch fsType
+        {
+            case "ext4":
+                args = []string{"-m0", "-Enodiscard,lazy_itable_init=1,lazy_journal_init=1", devicePath}
+            case "xfs":
+                args = []string{"-K", devicePath}
+        }
+        if (len(args) > 0)
+        {
+            cmdOut, cmdErr := diskMounter.Exec.Command("mkfs."+fsType, args...).CombinedOutput()
+            if (cmdErr != nil)
+            {
+                klog.Errorf("failed to run mkfs error: %v, output: %v", cmdErr, string(cmdOut))
+                // unmap NBD device
+                unmapOut, unmapErr := exec.Command("/usr/bin/vitastor-nbd", "unmap", devicePath).CombinedOutput()
+                if (unmapErr != nil)
+                {
+                    klog.Errorf("failed to unmap NBD device %s: %s, error: %v", devicePath, unmapOut, unmapErr)
+                }
+                return nil, status.Error(codes.Internal, cmdErr.Error())
+            }
+        }
+    }
+    if (isBlock)
+    {
+        opt = append(opt, "bind")
+        err = diskMounter.Mount(devicePath, targetPath, fsType, opt)
+    }
+    else
+    {
+        err = diskMounter.FormatAndMount(devicePath, targetPath, fsType, opt)
+    }
+    if (err != nil)
+    {
+        klog.Errorf(
+            "failed to mount device path (%s) to path (%s) for volume (%s) error: %s",
+            devicePath, targetPath, volName, err,
+        )
+        // unmap NBD device
+        unmapOut, unmapErr := exec.Command("/usr/bin/vitastor-nbd", "unmap", devicePath).CombinedOutput()
+        if (unmapErr != nil)
+        {
+            klog.Errorf("failed to unmap NBD device %s: %s, error: %v", devicePath, unmapOut, unmapErr)
+        }
+        return nil, status.Error(codes.Internal, err.Error())
+    }
+    return &csi.NodePublishVolumeResponse{}, nil
+}
+
+// NodeUnpublishVolume unmounts the volume from the target path
+func (ns *NodeServer) NodeUnpublishVolume(ctx context.Context, req *csi.NodeUnpublishVolumeRequest) (*csi.NodeUnpublishVolumeResponse, error)
+{
+    klog.Infof("received node unpublish volume request %+v", protosanitizer.StripSecrets(req))
+    targetPath := req.GetTargetPath()
+    devicePath, refCount, err := mount.GetDeviceNameFromMount(ns.mounter, targetPath)
+    if (err != nil)
+    {
+        if (os.IsNotExist(err))
+        {
+            return nil, status.Error(codes.NotFound, "Target path not found")
+        }
+        return nil, status.Error(codes.Internal, err.Error())
+    }
+    if (devicePath == "")
+    {
+        return nil, status.Error(codes.NotFound, "Volume not mounted")
+    }
+    // unmount
+    err = mount.CleanupMountPoint(targetPath, ns.mounter, false)
+    if (err != nil)
+    {
+        return nil, status.Error(codes.Internal, err.Error())
+    }
+    // unmap NBD device
+    if (refCount == 1)
+    {
+        unmapOut, unmapErr := exec.Command("/usr/bin/vitastor-nbd", "unmap", devicePath).CombinedOutput()
+        if (unmapErr != nil)
+        {
+            klog.Errorf("failed to unmap NBD device %s: %s, error: %v", devicePath, unmapOut, unmapErr)
+        }
+    }
+    return &csi.NodeUnpublishVolumeResponse{}, nil
+}
+
+// NodeGetVolumeStats returns volume capacity statistics available for the volume
+func (ns *NodeServer) NodeGetVolumeStats(ctx context.Context, req *csi.NodeGetVolumeStatsRequest) (*csi.NodeGetVolumeStatsResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// NodeExpandVolume expanding the file system on the node
+func (ns *NodeServer) NodeExpandVolume(ctx context.Context, req *csi.NodeExpandVolumeRequest) (*csi.NodeExpandVolumeResponse, error)
+{
+    return nil, status.Error(codes.Unimplemented, "")
+}
+
+// NodeGetCapabilities returns the supported capabilities of the node server
+func (ns *NodeServer) NodeGetCapabilities(ctx context.Context, req *csi.NodeGetCapabilitiesRequest) (*csi.NodeGetCapabilitiesResponse, error)
+{
+    return &csi.NodeGetCapabilitiesResponse{}, nil
+}
+
+// NodeGetInfo returns NodeGetInfoResponse for CO.
+func (ns *NodeServer) NodeGetInfo(ctx context.Context, req *csi.NodeGetInfoRequest) (*csi.NodeGetInfoResponse, error)
+{
+    klog.Infof("received node get info request %+v", protosanitizer.StripSecrets(req))
+    return &csi.NodeGetInfoResponse{
+        NodeId: ns.NodeID,
+    }, nil
+}

csi/src/server.go

@@ -0,0 +1,36 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
+
+package vitastor
+
+import (
+    "k8s.io/klog"
+)
+
+type Driver struct
+{
+    *Config
+}
+
+// NewDriver create new instance driver
+func NewDriver(config *Config) (*Driver, error)
+{
+    if (config == nil)
+    {
+        klog.Errorf("Vitastor CSI driver initialization failed")
+        return nil, nil
+    }
+    driver := &Driver{
+        Config: config,
+    }
+    klog.Infof("Vitastor CSI driver initialized")
+    return driver, nil
+}
+
+// Start server
+func (driver *Driver) Run()
+{
+    server := NewNonBlockingGRPCServer()
+    server.Start(driver.Endpoint, NewIdentityServer(driver), NewControllerServer(driver), NewNodeServer(driver))
+    server.Wait()
+}

csi/vitastor-csi.go

@@ -0,0 +1,39 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
+
+package main
+
+import (
+    "flag"
+    "fmt"
+    "os"
+    "k8s.io/klog"
+    "vitastor.io/csi/src"
+)
+
+func main()
+{
+    var config = vitastor.NewConfig()
+    flag.StringVar(&config.Endpoint, "endpoint", "", "CSI endpoint")
+    flag.StringVar(&config.NodeID, "node", "", "Node ID")
+    flag.Parse()
+    if (config.Endpoint == "")
+    {
+        config.Endpoint = os.Getenv("CSI_ENDPOINT")
+    }
+    if (config.NodeID == "")
+    {
+        config.NodeID = os.Getenv("NODE_ID")
+    }
+    if (config.Endpoint == "" && config.NodeID == "")
+    {
+        fmt.Fprintf(os.Stderr, "Please set -endpoint and -node / CSI_ENDPOINT & NODE_ID env vars\n")
+        os.Exit(1)
+    }
+    drv, err := vitastor.NewDriver(config)
+    if (err != nil)
+    {
+        klog.Fatalln(err)
+    }
+    drv.Run()
+}

debian/build-vitastor-bullseye.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+sed 's/$REL/bullseye/g' < vitastor.Dockerfile > ../Dockerfile
+cd ..
+mkdir -p packages
+sudo podman build -v `pwd`/packages:/root/packages -f Dockerfile .
+rm Dockerfile

debian/build-vitastor-buster.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+sed 's/$REL/buster/g' < vitastor.Dockerfile > ../Dockerfile
+cd ..
+mkdir -p packages
+sudo podman build -v `pwd`/packages:/root/packages -f Dockerfile .
+rm Dockerfile

debian/changelog

@@ -0,0 +1,27 @@
+vitastor (0.6.5-1) unstable; urgency=medium
+
+  * RDMA support
+  * Bugfixes
+
+ -- Vitaliy Filippov <vitalif@yourcmc.ru>  Sat, 01 May 2021 18:46:10 +0300
+
+vitastor (0.6.0-1) unstable; urgency=medium
+
+  * Snapshots and Copy-on-Write clones
+  * Image metadata in etcd (name, size)
+  * Image I/O and space statistics in etcd
+  * Write throttling for smoothing random write workloads in SSD+HDD configurations
+
+ -- Vitaliy Filippov <vitalif@yourcmc.ru>  Sun, 11 Apr 2021 00:49:18 +0300
+
+vitastor (0.5.1-1) unstable; urgency=medium
+
+  * Add jerasure support
+
+ -- Vitaliy Filippov <vitalif@yourcmc.ru>  Sat, 05 Dec 2020 17:02:26 +0300
+
+vitastor (0.5-1) unstable; urgency=medium
+
+  * First packaging for Debian
+
+ -- Vitaliy Filippov <vitalif@yourcmc.ru>  Thu, 05 Nov 2020 02:20:59 +0300

debian/compat

@@ -0,0 +1 @@
+13

debian/control

@@ -0,0 +1,17 @@
+Source: vitastor
+Section: admin
+Priority: optional
+Maintainer: Vitaliy Filippov <vitalif@yourcmc.ru>
+Build-Depends: debhelper, liburing-dev (>= 0.6), g++ (>= 8), libstdc++6 (>= 8), linux-libc-dev, libgoogle-perftools-dev, libjerasure-dev, libgf-complete-dev, libibverbs-dev
+Standards-Version: 4.5.0
+Homepage: https://vitastor.io/
+Rules-Requires-Root: no
+
+Package: vitastor
+Architecture: amd64
+Depends: ${shlibs:Depends}, ${misc:Depends}, fio (= ${dep:fio}), qemu (= ${dep:qemu}), nodejs (>= 10), node-sprintf-js, node-ws (>= 7), libjerasure2, lp-solve
+Description: Vitastor, a fast software-defined clustered block storage
+ Vitastor is a small, simple and fast clustered block storage (storage for VM drives),
+ architecturally similar to Ceph which means strong consistency, primary-replication,
+ symmetric clustering and automatic data distribution over any number of drives of any
+ size with configurable redundancy (replication or erasure codes/XOR).

debian/copyright

@@ -0,0 +1,21 @@
+Format: https://www.debian.org/doc/packaging-manuals/copyright-format/1.0/
+Upstream-Name: vitastor
+Upstream-Contact: Vitaliy Filippov <vitalif@yourcmc.ru>
+Source: https://vitastor.io
+
+Files: *
+Copyright: 2019+ Vitaliy Filippov <vitalif@yourcmc.ru>
+License: Multiple licenses VNPL-1.1 and/or GPL-2.0+
+ All server-side code (OSD, Monitor and so on) is licensed under the terms of
+ Vitastor Network Public License 1.1 (VNPL 1.1), a copyleft license based on
+ GNU GPLv3.0 with the additional "Network Interaction" clause which requires
+ opensourcing all programs directly or indirectly interacting with Vitastor
+ through a computer network and expressly designed to be used in conjunction
+ with it ("Proxy Programs"). Proxy Programs may be made public not only under
+ the terms of the same license, but also under the terms of any GPL-Compatible
+ Free Software License, as listed by the Free Software Foundation.
+ This is a stricter copyleft license than the Affero GPL.
+ .
+ Client libraries (cluster_client and so on) are dual-licensed under the same
+ VNPL 1.1 and also GNU GPL 2.0 or later to allow for compatibility with GPLed
+ software like QEMU and fio.

debian/install

@@ -0,0 +1,3 @@
+VNPL-1.1.txt usr/share/doc/vitastor
+GPL-2.0.txt usr/share/doc/vitastor
+mon usr/lib/vitastor

debian/patched-qemu.Dockerfile

@@ -0,0 +1,50 @@
+# Build patched QEMU for Debian Buster or Bullseye/Sid inside a container
+# cd ..; podman build --build-arg REL=bullseye -v `pwd`/packages:/root/packages -f debian/patched-qemu.Dockerfile .
+
+FROM debian:$REL
+
+WORKDIR /root
+
+RUN if [ "$REL" = "buster" ]; then \
+        echo 'deb http://deb.debian.org/debian buster-backports main' >> /etc/apt/sources.list; \
+        echo >> /etc/apt/preferences; \
+        echo 'Package: *' >> /etc/apt/preferences; \
+        echo 'Pin: release a=buster-backports' >> /etc/apt/preferences; \
+        echo 'Pin-Priority: 500' >> /etc/apt/preferences; \
+        echo >> /etc/apt/preferences; \
+        echo 'Package: libglvnd* libgles* libglx* libgl1 libegl* libopengl* mesa*' >> /etc/apt/preferences; \
+        echo 'Pin: release a=buster-backports' >> /etc/apt/preferences; \
+        echo 'Pin-Priority: 50' >> /etc/apt/preferences; \
+    fi; \
+    grep '^deb ' /etc/apt/sources.list | perl -pe 's/^deb/deb-src/' >> /etc/apt/sources.list; \
+    echo 'APT::Install-Recommends false;' >> /etc/apt/apt.conf; \
+    echo 'APT::Install-Suggests false;' >> /etc/apt/apt.conf
+
+RUN apt-get update
+RUN apt-get -y install qemu fio liburing1 liburing-dev libgoogle-perftools-dev devscripts
+RUN apt-get -y build-dep qemu
+RUN apt-get -y build-dep fio
+# To build a custom version
+#RUN cp /root/packages/qemu-orig/* /root
+RUN apt-get --download-only source qemu
+RUN apt-get --download-only source fio
+
+ADD patches/qemu-5.0-vitastor.patch patches/qemu-5.1-vitastor.patch /root/vitastor/patches/
+RUN set -e; \
+    mkdir -p /root/packages/qemu-$REL; \
+    rm -rf /root/packages/qemu-$REL/*; \
+    cd /root/packages/qemu-$REL; \
+    dpkg-source -x /root/qemu*.dsc; \
+    if [ -d /root/packages/qemu-$REL/qemu-5.0 ]; then \
+        cp /root/vitastor/patches/qemu-5.0-vitastor.patch /root/packages/qemu-$REL/qemu-5.0/debian/patches; \
+        echo qemu-5.0-vitastor.patch >> /root/packages/qemu-$REL/qemu-5.0/debian/patches/series; \
+    else \
+        cp /root/vitastor/patches/qemu-5.1-vitastor.patch /root/packages/qemu-$REL/qemu-*/debian/patches; \
+        P=`ls -d /root/packages/qemu-$REL/qemu-*/debian/patches`; \
+        echo qemu-5.1-vitastor.patch >> $P/series; \
+    fi; \
+    cd /root/packages/qemu-$REL/qemu-*/; \
+    V=$(head -n1 debian/changelog | perl -pe 's/^.*\((.*?)(~bpo[\d\+]*)?\).*$/$1/')+vitastor1; \
+    DEBFULLNAME="Vitaliy Filippov <vitalif@yourcmc.ru>" dch -D $REL -v $V 'Plug Vitastor block driver'; \
+    DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage --jobs=auto -sa; \
+    rm -rf /root/packages/qemu-$REL/qemu-*/

debian/rules

@@ -0,0 +1,9 @@
+#!/usr/bin/make -f
+export DH_VERBOSE = 1
+
+%:
+	dh $@
+
+override_dh_installdeb:
+	cat debian/substvars >> debian/vitastor.substvars
+	dh_installdeb

debian/source/format

@@ -0,0 +1 @@
+3.0 (quilt)

debian/substvars

@@ -0,0 +1,2 @@
+dep:fio=3.16-1
+dep:qemu=1:5.1+dfsg-4+vitastor1

debian/vitastor.Dockerfile

@@ -0,0 +1,67 @@
+# Build Vitastor packages for Debian Buster or Bullseye/Sid inside a container
+# cd ..; podman build --build-arg REL=bullseye -v `pwd`/packages:/root/packages -f debian/vitastor.Dockerfile .
+
+FROM debian:$REL
+
+WORKDIR /root
+
+RUN if [ "$REL" = "buster" ]; then \
+        echo 'deb http://deb.debian.org/debian buster-backports main' >> /etc/apt/sources.list; \
+        echo >> /etc/apt/preferences; \
+        echo 'Package: *' >> /etc/apt/preferences; \
+        echo 'Pin: release a=buster-backports' >> /etc/apt/preferences; \
+        echo 'Pin-Priority: 500' >> /etc/apt/preferences; \
+    fi; \
+    grep '^deb ' /etc/apt/sources.list | perl -pe 's/^deb/deb-src/' >> /etc/apt/sources.list; \
+    echo 'APT::Install-Recommends false;' >> /etc/apt/apt.conf; \
+    echo 'APT::Install-Suggests false;' >> /etc/apt/apt.conf
+
+RUN apt-get update
+RUN apt-get -y install qemu fio liburing1 liburing-dev libgoogle-perftools-dev devscripts
+RUN apt-get -y build-dep qemu
+RUN apt-get -y build-dep fio
+RUN apt-get --download-only source qemu
+RUN apt-get --download-only source fio
+RUN apt-get update && apt-get -y install libjerasure-dev cmake libibverbs-dev
+
+ADD . /root/vitastor
+RUN set -e -x; \
+    mkdir -p /root/fio-build/; \
+    cd /root/fio-build/; \
+    rm -rf /root/fio-build/*; \
+    dpkg-source -x /root/fio*.dsc; \
+    cd /root/packages/qemu-$REL/; \
+    rm -rf qemu*/; \
+    dpkg-source -x qemu*.dsc; \
+    cd /root/packages/qemu-$REL/qemu*/; \
+    debian/rules b/configure-stamp; \
+    cd b/qemu; \
+    make -j8 qapi/qapi-builtin-types.h; \
+    mkdir -p /root/packages/vitastor-$REL; \
+    rm -rf /root/packages/vitastor-$REL/*; \
+    cd /root/packages/vitastor-$REL; \
+    cp -r /root/vitastor vitastor-0.6.5; \
+    ln -s /root/packages/qemu-$REL/qemu-*/ vitastor-0.6.5/qemu; \
+    ln -s /root/fio-build/fio-*/ vitastor-0.6.5/fio; \
+    cd vitastor-0.6.5; \
+    FIO=$(head -n1 fio/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
+    QEMU=$(head -n1 qemu/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
+    sh copy-qemu-includes.sh; \
+    sh copy-fio-includes.sh; \
+    rm qemu fio; \
+    mkdir -p a b debian/patches; \
+    mv qemu-copy b/qemu; \
+    mv fio-copy b/fio; \
+    diff -NaurpbB a b > debian/patches/qemu-fio-headers.patch || true; \
+    echo qemu-fio-headers.patch >> debian/patches/series; \
+    rm -rf a b; \
+    rm -rf /root/packages/qemu-$REL/qemu*/; \
+    echo "dep:fio=$FIO" > debian/substvars; \
+    echo "dep:qemu=$QEMU" >> debian/substvars; \
+    cd /root/packages/vitastor-$REL; \
+    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.6.5.orig.tar.xz vitastor-0.6.5; \
+    cd vitastor-0.6.5; \
+    V=$(head -n1 debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
+    DEBFULLNAME="Vitaliy Filippov <vitalif@yourcmc.ru>" dch -D $REL -v "$V""$REL" "Rebuild for $REL"; \
+    DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage --jobs=auto -sa; \
+    rm -rf /root/packages/vitastor-$REL/vitastor-*/

dump_journal.cpp

@@ -1,165 +0,0 @@
-#define _LARGEFILE64_SOURCE
-#include <sys/types.h>
-#include <sys/ioctl.h>
-#include <sys/stat.h>
-#include <sys/time.h>
-#include <fcntl.h>
-#include <unistd.h>
-#include <stdint.h>
-#include <malloc.h>
-#include <linux/fs.h>
-#include <string.h>
-#include <errno.h>
-#include <assert.h>
-#include <stdio.h>
-
-#include "blockstore_impl.h"
-#include "crc32c.h"
-
-struct journal_dump_t
-{
-    char *journal_device;
-    uint32_t journal_block;
-    uint64_t journal_offset;
-    uint64_t journal_len;
-    uint64_t journal_pos;
-    int fd;
-
-    void dump_block(void *buf);
-};
-
-int main(int argc, char *argv[])
-{
-    if (argc < 5)
-    {
-        printf("USAGE: %s <journal_file> <journal_block_size> <offset> <size>\n", argv[0]);
-        return 1;
-    }
-    journal_dump_t self;
-    self.journal_device = argv[1];
-    self.journal_block = strtoul(argv[2], NULL, 10);
-    self.journal_offset = strtoull(argv[3], NULL, 10);
-    self.journal_len = strtoull(argv[4], NULL, 10);
-    if (self.journal_block < MEM_ALIGNMENT || (self.journal_block % MEM_ALIGNMENT) ||
-        self.journal_block > 128*1024)
-    {
-        printf("Invalid journal block size\n");
-        return 1;
-    }
-    self.fd = open(self.journal_device, O_DIRECT|O_RDONLY);
-    if (self.fd == -1)
-    {
-        printf("Failed to open journal\n");
-        return 1;
-    }
-    void *data = memalign(MEM_ALIGNMENT, self.journal_block);
-    self.journal_pos = 0;
-    while (self.journal_pos < self.journal_len)
-    {
-        int r = pread(self.fd, data, self.journal_block, self.journal_offset+self.journal_pos);
-        assert(r == self.journal_block);
-        uint64_t s;
-        for (s = 0; s < self.journal_block; s += 8)
-        {
-            if (*((uint64_t*)(data+s)) != 0)
-                break;
-        }
-        if (s == self.journal_block)
-        {
-            printf("offset %08lx: zeroes\n", self.journal_pos);
-            self.journal_pos += self.journal_block;
-        }
-        else if (((journal_entry*)data)->magic == JOURNAL_MAGIC)
-        {
-            printf("offset %08lx:\n", self.journal_pos);
-            self.dump_block(data);
-        }
-        else
-        {
-            printf("offset %08lx: no magic in the beginning, looks like random data (pattern=%lx)\n", self.journal_pos, *((uint64_t*)data));
-            self.journal_pos += self.journal_block;
-        }
-    }
-    free(data);
-    close(self.fd);
-    return 0;
-}
-
-void journal_dump_t::dump_block(void *buf)
-{
-    uint32_t pos = 0;
-    journal_pos += journal_block;
-    int entry = 0;
-    bool wrapped = false;
-    while (pos < journal_block)
-    {
-        journal_entry *je = (journal_entry*)(buf + pos);
-        if (je->magic != JOURNAL_MAGIC || je->type < JE_START || je->type > JE_DELETE)
-        {
-            break;
-        }
-        const char *crc32_valid = je_crc32(je) == je->crc32 ? "(valid)" : "(invalid)";
-        printf("entry % 3d: crc32=%08x %s prev=%08x ", entry, je->crc32, crc32_valid, je->crc32_prev);
-        if (je->type == JE_START)
-        {
-            printf("je_start start=%08lx\n", je->start.journal_start);
-        }
-        else if (je->type == JE_SMALL_WRITE)
-        {
-            printf(
-                "je_small_write oid=%lu:%lu ver=%lu offset=%u len=%u loc=%08lx",
-                je->small_write.oid.inode, je->small_write.oid.stripe,
-                je->small_write.version, je->small_write.offset, je->small_write.len,
-                je->small_write.data_offset
-            );
-            if (journal_pos + je->small_write.len > journal_len)
-            {
-                // data continues from the beginning of the journal
-                journal_pos = journal_block;
-                wrapped = true;
-            }
-            if (journal_pos != je->small_write.data_offset)
-            {
-                printf(" (mismatched, calculated = %lu)", journal_pos);
-            }
-            journal_pos += je->small_write.len;
-            if (journal_pos >= journal_len)
-            {
-                journal_pos = journal_block;
-                wrapped = true;
-            }
-            uint32_t data_crc32 = 0;
-            void *data = memalign(MEM_ALIGNMENT, je->small_write.len);
-            assert(pread(fd, data, je->small_write.len, journal_offset+je->small_write.data_offset) == je->small_write.len);
-            data_crc32 = crc32c(0, data, je->small_write.len);
-            free(data);
-            printf(
-                " data_crc32=%08x%s", je->small_write.crc32_data,
-                (data_crc32 != je->small_write.crc32_data) ? " (invalid)" : " (valid)"
-            );
-            printf("\n");
-        }
-        else if (je->type == JE_BIG_WRITE)
-        {
-            printf("je_big_write oid=%lu:%lu ver=%lu loc=%08lx\n", je->big_write.oid.inode, je->big_write.oid.stripe, je->big_write.version, je->big_write.location);
-        }
-        else if (je->type == JE_STABLE)
-        {
-            printf("je_stable oid=%lu:%lu ver=%lu\n", je->stable.oid.inode, je->stable.oid.stripe, je->stable.version);
-        }
-        else if (je->type == JE_ROLLBACK)
-        {
-            printf("je_rollback oid=%lu:%lu ver=%lu\n", je->rollback.oid.inode, je->rollback.oid.stripe, je->rollback.version);
-        }
-        else if (je->type == JE_DELETE)
-        {
-            printf("je_delete oid=%lu:%lu ver=%lu\n", je->del.oid.inode, je->del.oid.stripe, je->del.version);
-        }
-        pos += je->size;
-        entry++;
-    }
-    if (wrapped)
-    {
-        journal_pos = journal_len;
-    }
-}

etcd_state_client.cpp

@@ -1,424 +0,0 @@
-#include "osd_ops.h"
-#include "pg_states.h"
-#include "etcd_state_client.h"
-#include "http_client.h"
-#include "base64.h"
-
-json_kv_t etcd_state_client_t::parse_etcd_kv(const json11::Json & kv_json)
-{
-    json_kv_t kv;
-    kv.key = base64_decode(kv_json["key"].string_value());
-    std::string json_err, json_text = base64_decode(kv_json["value"].string_value());
-    kv.value = json_text == "" ? json11::Json() : json11::Json::parse(json_text, json_err);
-    if (json_err != "")
-    {
-        printf("Bad JSON in etcd key %s: %s (value: %s)\n", kv.key.c_str(), json_err.c_str(), json_text.c_str());
-        kv.key = "";
-    }
-    return kv;
-}
-
-void etcd_state_client_t::etcd_txn(json11::Json txn, int timeout, std::function<void(std::string, json11::Json)> callback)
-{
-    etcd_call("/kv/txn", txn, timeout, callback);
-}
-
-void etcd_state_client_t::etcd_call(std::string api, json11::Json payload, int timeout, std::function<void(std::string, json11::Json)> callback)
-{
-    std::string etcd_address = etcd_addresses[rand() % etcd_addresses.size()];
-    std::string etcd_api_path;
-    int pos = etcd_address.find('/');
-    if (pos >= 0)
-    {
-        etcd_api_path = etcd_address.substr(pos);
-        etcd_address = etcd_address.substr(0, pos);
-    }
-    std::string req = payload.dump();
-    req = "POST "+etcd_api_path+api+" HTTP/1.1\r\n"
-        "Host: "+etcd_address+"\r\n"
-        "Content-Type: application/json\r\n"
-        "Content-Length: "+std::to_string(req.size())+"\r\n"
-        "Connection: close\r\n"
-        "\r\n"+req;
-    http_request_json(tfd, etcd_address, req, timeout, callback);
-}
-
-void etcd_state_client_t::parse_config(json11::Json & config)
-{
-    this->etcd_addresses.clear();
-    if (config["etcd_address"].is_string())
-    {
-        std::string ea = config["etcd_address"].string_value();
-        while (1)
-        {
-            int pos = ea.find(',');
-            std::string addr = pos >= 0 ? ea.substr(0, pos) : ea;
-            if (addr.length() > 0)
-            {
-                if (addr.find('/') < 0)
-                    addr += "/v3";
-                this->etcd_addresses.push_back(addr);
-            }
-            if (pos >= 0)
-                ea = ea.substr(pos+1);
-            else
-                break;
-        }
-    }
-    else if (config["etcd_address"].array_items().size())
-    {
-        for (auto & ea: config["etcd_address"].array_items())
-        {
-            std::string addr = ea.string_value();
-            if (addr != "")
-            {
-                if (addr.find('/') < 0)
-                    addr += "/v3";
-                this->etcd_addresses.push_back(addr);
-            }
-        }
-    }
-    this->etcd_prefix = config["etcd_prefix"].string_value();
-    if (this->etcd_prefix == "")
-    {
-        this->etcd_prefix = "/microceph";
-    }
-    else if (this->etcd_prefix[0] != '/')
-    {
-        this->etcd_prefix = "/"+this->etcd_prefix;
-    }
-    this->log_level = config["log_level"].int64_value();
-}
-
-void etcd_state_client_t::start_etcd_watcher()
-{
-    std::string etcd_address = etcd_addresses[rand() % etcd_addresses.size()];
-    std::string etcd_api_path;
-    int pos = etcd_address.find('/');
-    if (pos >= 0)
-    {
-        etcd_api_path = etcd_address.substr(pos);
-        etcd_address = etcd_address.substr(0, pos);
-    }
-    etcd_watches_initialised = 0;
-    etcd_watch_ws = open_websocket(tfd, etcd_address, etcd_api_path+"/watch", ETCD_SLOW_TIMEOUT, [this](const http_response_t *msg)
-    {
-        if (msg->body.length())
-        {
-            std::string json_err;
-            json11::Json data = json11::Json::parse(msg->body, json_err);
-            if (json_err != "")
-            {
-                printf("Bad JSON in etcd event: %s, ignoring event\n", json_err.c_str());
-            }
-            else
-            {
-                if (data["result"]["created"].bool_value())
-                {
-                    etcd_watches_initialised++;
-                }
-                if (etcd_watches_initialised == 4)
-                {
-                    etcd_watch_revision = data["result"]["header"]["revision"].uint64_value();
-                }
-                // First gather all changes into a hash to remove multiple overwrites
-                json11::Json::object changes;
-                for (auto & ev: data["result"]["events"].array_items())
-                {
-                    auto kv = parse_etcd_kv(ev["kv"]);
-                    if (kv.key != "")
-                    {
-                        changes[kv.key] = kv.value;
-                    }
-                }
-                for (auto & kv: changes)
-                {
-                    if (this->log_level > 0)
-                    {
-                        printf("Incoming event: %s -> %s\n", kv.first.c_str(), kv.second.dump().c_str());
-                    }
-                    parse_state(kv.first, kv.second);
-                }
-                // React to changes
-                if (on_change_hook != NULL)
-                {
-                    on_change_hook(changes);
-                }
-            }
-        }
-        if (msg->eof)
-        {
-            etcd_watch_ws = NULL;
-            if (etcd_watches_initialised == 0)
-            {
-                // Connection not established, retry in <ETCD_SLOW_TIMEOUT>
-                tfd->set_timer(ETCD_SLOW_TIMEOUT, false, [this](int)
-                {
-                    start_etcd_watcher();
-                });
-            }
-            else
-            {
-                // Connection was live, retry immediately
-                start_etcd_watcher();
-            }
-        }
-    });
-    etcd_watch_ws->post_message(WS_TEXT, json11::Json(json11::Json::object {
-        { "create_request", json11::Json::object {
-            { "key", base64_encode(etcd_prefix+"/config/") },
-            { "range_end", base64_encode(etcd_prefix+"/config0") },
-            { "start_revision", etcd_watch_revision+1 },
-            { "watch_id", ETCD_CONFIG_WATCH_ID },
-        } }
-    }).dump());
-    etcd_watch_ws->post_message(WS_TEXT, json11::Json(json11::Json::object {
-        { "create_request", json11::Json::object {
-            { "key", base64_encode(etcd_prefix+"/osd/state/") },
-            { "range_end", base64_encode(etcd_prefix+"/osd/state0") },
-            { "start_revision", etcd_watch_revision+1 },
-            { "watch_id", ETCD_OSD_STATE_WATCH_ID },
-        } }
-    }).dump());
-    etcd_watch_ws->post_message(WS_TEXT, json11::Json(json11::Json::object {
-        { "create_request", json11::Json::object {
-            { "key", base64_encode(etcd_prefix+"/pg/state/") },
-            { "range_end", base64_encode(etcd_prefix+"/pg/state0") },
-            { "start_revision", etcd_watch_revision+1 },
-            { "watch_id", ETCD_PG_STATE_WATCH_ID },
-        } }
-    }).dump());
-    etcd_watch_ws->post_message(WS_TEXT, json11::Json(json11::Json::object {
-        { "create_request", json11::Json::object {
-            { "key", base64_encode(etcd_prefix+"/pg/history/") },
-            { "range_end", base64_encode(etcd_prefix+"/pg/history0") },
-            { "start_revision", etcd_watch_revision+1 },
-            { "watch_id", ETCD_PG_HISTORY_WATCH_ID },
-        } }
-    }).dump());
-}
-
-void etcd_state_client_t::load_global_config()
-{
-    etcd_call("/kv/range", json11::Json::object {
-        { "key", base64_encode(etcd_prefix+"/config/global") }
-    }, ETCD_SLOW_TIMEOUT, [this](std::string err, json11::Json data)
-    {
-        if (err != "")
-        {
-            printf("Error reading OSD configuration from etcd: %s\n", err.c_str());
-            tfd->set_timer(ETCD_SLOW_TIMEOUT, false, [this](int timer_id)
-            {
-                load_global_config();
-            });
-            return;
-        }
-        if (!etcd_watch_revision)
-        {
-            etcd_watch_revision = data["header"]["revision"].uint64_value();
-        }
-        json11::Json::object global_config;
-        if (data["kvs"].array_items().size() > 0)
-        {
-            auto kv = parse_etcd_kv(data["kvs"][0]);
-            if (kv.value.is_object())
-            {
-                global_config = kv.value.object_items();
-            }
-        }
-        on_load_config_hook(global_config);
-    });
-}
-
-void etcd_state_client_t::load_pgs()
-{
-    json11::Json::array txn = {
-        json11::Json::object {
-            { "request_range", json11::Json::object {
-                { "key", base64_encode(etcd_prefix+"/config/pgs") },
-            } }
-        },
-        json11::Json::object {
-            { "request_range", json11::Json::object {
-                { "key", base64_encode(etcd_prefix+"/pg/history/") },
-                { "range_end", base64_encode(etcd_prefix+"/pg/history0") },
-            } }
-        },
-        json11::Json::object {
-            { "request_range", json11::Json::object {
-                { "key", base64_encode(etcd_prefix+"/pg/state/") },
-                { "range_end", base64_encode(etcd_prefix+"/pg/state0") },
-            } }
-        },
-        json11::Json::object {
-            { "request_range", json11::Json::object {
-                { "key", base64_encode(etcd_prefix+"/osd/state/") },
-                { "range_end", base64_encode(etcd_prefix+"/osd/state0") },
-            } }
-        },
-    };
-    json11::Json::object req = { { "success", txn } };
-    json11::Json checks = load_pgs_checks_hook != NULL ? load_pgs_checks_hook() : json11::Json();
-    if (checks.array_items().size() > 0)
-    {
-        req["compare"] = checks;
-    }
-    etcd_txn(req, ETCD_SLOW_TIMEOUT, [this](std::string err, json11::Json data)
-    {
-        if (err != "")
-        {
-            printf("Error loading PGs from etcd: %s\n", err.c_str());
-            tfd->set_timer(ETCD_SLOW_TIMEOUT, false, [this](int timer_id)
-            {
-                load_pgs();
-            });
-            return;
-        }
-        if (!data["succeeded"].bool_value())
-        {
-            on_load_pgs_hook(false);
-            return;
-        }
-        for (auto & res: data["responses"].array_items())
-        {
-            for (auto & kv_json: res["response_range"]["kvs"].array_items())
-            {
-                auto kv = parse_etcd_kv(kv_json);
-                parse_state(kv.key, kv.value);
-            }
-        }
-        on_load_pgs_hook(true);
-    });
-}
-
-void etcd_state_client_t::parse_state(const std::string & key, const json11::Json & value)
-{
-    if (key == etcd_prefix+"/config/pgs")
-    {
-        for (auto & pg_item: this->pg_config)
-        {
-            pg_item.second.exists = false;
-        }
-        for (auto & pg_item: value["items"].object_items())
-        {
-            pg_num_t pg_num = stoull_full(pg_item.first);
-            if (!pg_num)
-            {
-                printf("Bad key in PG configuration: %s (must be a number), skipped\n", pg_item.first.c_str());
-                continue;
-            }
-            this->pg_config[pg_num].exists = true;
-            this->pg_config[pg_num].pause = pg_item.second["pause"].bool_value();
-            this->pg_config[pg_num].primary = pg_item.second["primary"].uint64_value();
-            this->pg_config[pg_num].target_set.clear();
-            for (auto pg_osd: pg_item.second["osd_set"].array_items())
-            {
-                this->pg_config[pg_num].target_set.push_back(pg_osd.uint64_value());
-            }
-            if (this->pg_config[pg_num].target_set.size() != 3)
-            {
-                printf("Bad PG %u config format: incorrect osd_set = %s\n", pg_num, pg_item.second["osd_set"].dump().c_str());
-                this->pg_config[pg_num].target_set.resize(3);
-                this->pg_config[pg_num].pause = true;
-            }
-        }
-    }
-    else if (key.substr(0, etcd_prefix.length()+12) == etcd_prefix+"/pg/history/")
-    {
-        // <etcd_prefix>/pg/history/%d
-        pg_num_t pg_num = stoull_full(key.substr(etcd_prefix.length()+12));
-        if (!pg_num)
-        {
-            printf("Bad etcd key %s, ignoring\n", key.c_str());
-        }
-        else
-        {
-            auto & pg_cfg = this->pg_config[pg_num];
-            pg_cfg.target_history.clear();
-            pg_cfg.all_peers.clear();
-            // Refuse to start PG if any set of the <osd_sets> has no live OSDs
-            for (auto hist_item: value["osd_sets"].array_items())
-            {
-                std::vector<osd_num_t> history_set;
-                for (auto pg_osd: hist_item.array_items())
-                {
-                    history_set.push_back(pg_osd.uint64_value());
-                }
-                pg_cfg.target_history.push_back(history_set);
-            }
-            // Include these additional OSDs when peering the PG
-            for (auto pg_osd: value["all_peers"].array_items())
-            {
-                pg_cfg.all_peers.push_back(pg_osd.uint64_value());
-            }
-        }
-    }
-    else if (key.substr(0, etcd_prefix.length()+10) == etcd_prefix+"/pg/state/")
-    {
-        // <etcd_prefix>/pg/state/%d
-        pg_num_t pg_num = stoull_full(key.substr(etcd_prefix.length()+10));
-        if (!pg_num)
-        {
-            printf("Bad etcd key %s, ignoring\n", key.c_str());
-        }
-        else if (value.is_null())
-        {
-            this->pg_config[pg_num].cur_primary = 0;
-            this->pg_config[pg_num].cur_state = 0;
-        }
-        else
-        {
-            osd_num_t cur_primary = value["primary"].uint64_value();
-            int state = 0;
-            for (auto & e: value["state"].array_items())
-            {
-                int i;
-                for (i = 0; i < pg_state_bit_count; i++)
-                {
-                    if (e.string_value() == pg_state_names[i])
-                    {
-                        state = state | pg_state_bits[i];
-                        break;
-                    }
-                }
-                if (i >= pg_state_bit_count)
-                {
-                    printf("Unexpected PG %u state keyword in etcd: %s\n", pg_num, e.dump().c_str());
-                    return;
-                }
-            }
-            if (!cur_primary || !value["state"].is_array() || !state ||
-                (state & PG_OFFLINE) && state != PG_OFFLINE ||
-                (state & PG_PEERING) && state != PG_PEERING ||
-                (state & PG_INCOMPLETE) && state != PG_INCOMPLETE)
-            {
-                printf("Unexpected PG %u state in etcd: primary=%lu, state=%s\n", pg_num, cur_primary, value["state"].dump().c_str());
-                return;
-            }
-            this->pg_config[pg_num].cur_primary = cur_primary;
-            this->pg_config[pg_num].cur_state = state;
-        }
-    }
-    else if (key.substr(0, etcd_prefix.length()+11) == etcd_prefix+"/osd/state/")
-    {
-        // <etcd_prefix>/osd/state/%d
-        osd_num_t peer_osd = std::stoull(key.substr(etcd_prefix.length()+11));
-        if (peer_osd > 0)
-        {
-            if (value.is_object() && value["state"] == "up" &&
-                value["addresses"].is_array() &&
-                value["port"].int64_value() > 0 && value["port"].int64_value() < 65536)
-            {
-                this->peer_states[peer_osd] = value;
-            }
-            else
-            {
-                this->peer_states.erase(peer_osd);
-            }
-            if (on_change_osd_state_hook != NULL)
-            {
-                on_change_osd_state_hook(peer_osd);
-            }
-        }
-    }
-}

etcd_state_client.h

@@ -1,61 +0,0 @@
-#pragma once
-
-#include "osd_id.h"
-#include "http_client.h"
-#include "timerfd_manager.h"
-
-#define ETCD_CONFIG_WATCH_ID 1
-#define ETCD_PG_STATE_WATCH_ID 2
-#define ETCD_PG_HISTORY_WATCH_ID 3
-#define ETCD_OSD_STATE_WATCH_ID 4
-
-#define MAX_ETCD_ATTEMPTS 5
-#define ETCD_SLOW_TIMEOUT 5000
-#define ETCD_QUICK_TIMEOUT 1000
-
-struct pg_config_t
-{
-    bool exists;
-    osd_num_t primary;
-    std::vector<osd_num_t> target_set;
-    std::vector<std::vector<osd_num_t>> target_history;
-    std::vector<osd_num_t> all_peers;
-    bool pause;
-    osd_num_t cur_primary;
-    int cur_state;
-};
-
-struct json_kv_t
-{
-    std::string key;
-    json11::Json value;
-};
-
-struct etcd_state_client_t
-{
-    std::vector<std::string> etcd_addresses;
-    std::string etcd_prefix;
-    int log_level = 0;
-    timerfd_manager_t *tfd = NULL;
-
-    int etcd_watches_initialised = 0;
-    uint64_t etcd_watch_revision = 0;
-    websocket_t *etcd_watch_ws = NULL;
-    std::map<pg_num_t, pg_config_t> pg_config;
-    std::map<osd_num_t, json11::Json> peer_states;
-
-    std::function<void(json11::Json::object &)> on_change_hook;
-    std::function<void(json11::Json::object &)> on_load_config_hook;
-    std::function<json11::Json()> load_pgs_checks_hook;
-    std::function<void(bool)> on_load_pgs_hook;
-    std::function<void(uint64_t)> on_change_osd_state_hook;
-
-    json_kv_t parse_etcd_kv(const json11::Json & kv_json);
-    void etcd_call(std::string api, json11::Json payload, int timeout, std::function<void(std::string, json11::Json)> callback);
-    void etcd_txn(json11::Json txn, int timeout, std::function<void(std::string, json11::Json)> callback);
-    void start_etcd_watcher();
-    void load_global_config();
-    void load_pgs();
-    void parse_state(const std::string & key, const json11::Json & value);
-    void parse_config(json11::Json & config);
-};

fio_cluster.cpp

@@ -1,298 +0,0 @@
-// FIO engine to test cluster I/O
-//
-// Random write:
-//
-// fio -thread -ioengine=./libfio_cluster.so -name=test -bs=4k -direct=1 -fsync=16 -iodepth=16 -rw=randwrite \
-//     -etcd=127.0.0.1:2379 [-etcd_prefix=/microceph] -size=1000M
-//
-// Linear write:
-//
-// fio -thread -ioengine=./libfio_cluster.so -name=test -bs=128k -direct=1 -fsync=32 -iodepth=32 -rw=write \
-//     -etcd=127.0.0.1:2379 [-etcd_prefix=/microceph] -size=1000M
-//
-// Random read (run with -iodepth=32 or -iodepth=1):
-//
-// fio -thread -ioengine=./libfio_cluster.so -name=test -bs=4k -direct=1 -iodepth=32 -rw=randread \
-//     -etcd=127.0.0.1:2379 [-etcd_prefix=/microceph] -size=1000M
-
-#include <sys/types.h>
-#include <sys/socket.h>
-#include <netinet/in.h>
-#include <netinet/tcp.h>
-
-#include <vector>
-#include <unordered_map>
-
-#include "epoll_manager.h"
-#include "cluster_client.h"
-extern "C" {
-#define CONFIG_HAVE_GETTID
-#define CONFIG_PWRITEV2
-#include "fio/fio.h"
-#include "fio/optgroup.h"
-}
-
-struct sec_data
-{
-    ring_loop_t *ringloop = NULL;
-    epoll_manager_t *epmgr = NULL;
-    cluster_client_t *cli = NULL;
-    bool last_sync = false;
-    /* The list of completed io_u structs. */
-    std::vector<io_u*> completed;
-    uint64_t op_n = 0, inflight = 0;
-    bool trace = false;
-};
-
-struct sec_options
-{
-    int __pad;
-    char *etcd_host = NULL;
-    char *etcd_prefix = NULL;
-    int inode = 0;
-    int trace = 0;
-};
-
-static struct fio_option options[] = {
-    {
-        .name   = "etcd",
-        .lname  = "etcd address",
-        .type   = FIO_OPT_STR_STORE,
-        .off1   = offsetof(struct sec_options, etcd_host),
-        .help   = "etcd address in the form HOST:PORT[/PATH]",
-        .category = FIO_OPT_C_ENGINE,
-        .group  = FIO_OPT_G_FILENAME,
-    },
-    {
-        .name   = "etcd",
-        .lname  = "etcd key prefix",
-        .type   = FIO_OPT_STR_STORE,
-        .off1   = offsetof(struct sec_options, etcd_prefix),
-        .help   = "etcd key prefix, by default /microceph",
-        .category = FIO_OPT_C_ENGINE,
-        .group  = FIO_OPT_G_FILENAME,
-    },
-    {
-        .name   = "inode",
-        .lname  = "inode to run tests on",
-        .type   = FIO_OPT_INT,
-        .off1   = offsetof(struct sec_options, inode),
-        .help   = "inode to run tests on (1 by default)",
-        .category = FIO_OPT_C_ENGINE,
-        .group  = FIO_OPT_G_FILENAME,
-    },
-    {
-        .name   = "osd_trace",
-        .lname  = "OSD trace",
-        .type   = FIO_OPT_BOOL,
-        .off1   = offsetof(struct sec_options, trace),
-        .help   = "Trace OSD operations",
-        .def    = "0",
-        .category = FIO_OPT_C_ENGINE,
-        .group  = FIO_OPT_G_FILENAME,
-    },
-    {
-        .name = NULL,
-    },
-};
-
-static int sec_setup(struct thread_data *td)
-{
-    sec_data *bsd;
-
-    bsd = new sec_data;
-    if (!bsd)
-    {
-        td_verror(td, errno, "calloc");
-        return 1;
-    }
-    td->io_ops_data = bsd;
-
-    if (!td->files_index)
-    {
-        add_file(td, "osd_cluster", 0, 0);
-        td->o.nr_files = td->o.nr_files ? : 1;
-        td->o.open_files++;
-    }
-
-    return 0;
-}
-
-static void sec_cleanup(struct thread_data *td)
-{
-    sec_data *bsd = (sec_data*)td->io_ops_data;
-    if (bsd)
-    {
-        delete bsd->cli;
-        delete bsd->epmgr;
-        delete bsd->ringloop;
-        bsd->cli = NULL;
-        bsd->epmgr = NULL;
-        bsd->ringloop = NULL;
-    }
-}
-
-/* Connect to the server from each thread. */
-static int sec_init(struct thread_data *td)
-{
-    sec_options *o = (sec_options*)td->eo;
-    sec_data *bsd = (sec_data*)td->io_ops_data;
-
-    json11::Json cfg = json11::Json::object {
-        { "etcd_address", std::string(o->etcd_host) },
-        { "etcd_prefix", std::string(o->etcd_prefix ? o->etcd_prefix : "/microceph") },
-    };
-
-    bsd->ringloop = new ring_loop_t(512);
-    bsd->epmgr = new epoll_manager_t(bsd->ringloop);
-    bsd->cli = new cluster_client_t(bsd->ringloop, bsd->epmgr->tfd, cfg);
-
-    bsd->trace = o->trace ? true : false;
-
-    return 0;
-}
-
-/* Begin read or write request. */
-static enum fio_q_status sec_queue(struct thread_data *td, struct io_u *io)
-{
-    sec_options *opt = (sec_options*)td->eo;
-    sec_data *bsd = (sec_data*)td->io_ops_data;
-    int n = bsd->op_n;
-
-    fio_ro_check(td, io);
-    if (io->ddir == DDIR_SYNC && bsd->last_sync)
-    {
-        return FIO_Q_COMPLETED;
-    }
-
-    io->engine_data = bsd;
-    cluster_op_t *op = new cluster_op_t;
-
-    switch (io->ddir)
-    {
-    case DDIR_READ:
-        op->opcode = OSD_OP_READ;
-        op->inode = opt->inode;
-        op->offset = io->offset;
-        op->len = io->xfer_buflen;
-        op->buf = io->xfer_buf;
-        bsd->last_sync = false;
-        break;
-    case DDIR_WRITE:
-        op->opcode = OSD_OP_WRITE;
-        op->inode = opt->inode;
-        op->offset = io->offset;
-        op->len = io->xfer_buflen;
-        op->buf = io->xfer_buf;
-        bsd->last_sync = false;
-        break;
-    case DDIR_SYNC:
-        op->opcode = OSD_OP_SYNC;
-        bsd->last_sync = true;
-        break;
-    default:
-        io->error = EINVAL;
-        return FIO_Q_COMPLETED;
-    }
-
-    op->callback = [io, n](cluster_op_t *op)
-    {
-        io->error = op->retval < 0 ? -op->retval : 0;
-        sec_data *bsd = (sec_data*)io->engine_data;
-        bsd->inflight--;
-        bsd->completed.push_back(io);
-        if (bsd->trace)
-        {
-            printf("--- %s n=%d retval=%d\n", io->ddir == DDIR_READ ? "READ" :
-                (io->ddir == DDIR_WRITE ? "WRITE" : "SYNC"), n, op->retval);
-        }
-        delete op;
-    };
-
-    if (opt->trace)
-    {
-        printf("+++ %s # %d\n", io->ddir == DDIR_READ ? "READ" :
-            (io->ddir == DDIR_WRITE ? "WRITE" : "SYNC"), n);
-    }
-
-    io->error = 0;
-    bsd->inflight++;
-    bsd->op_n++;
-    bsd->cli->execute(op);
-
-    if (io->error != 0)
-        return FIO_Q_COMPLETED;
-    return FIO_Q_QUEUED;
-}
-
-static int sec_getevents(struct thread_data *td, unsigned int min, unsigned int max, const struct timespec *t)
-{
-    sec_data *bsd = (sec_data*)td->io_ops_data;
-    while (true)
-    {
-        bsd->ringloop->loop();
-        if (bsd->completed.size() >= min)
-            break;
-        bsd->ringloop->wait();
-    }
-    return bsd->completed.size();
-}
-
-static struct io_u *sec_event(struct thread_data *td, int event)
-{
-    sec_data *bsd = (sec_data*)td->io_ops_data;
-    if (bsd->completed.size() == 0)
-        return NULL;
-    /* FIXME We ignore the event number and assume fio calls us exactly once for [0..nr_events-1] */
-    struct io_u *ev = bsd->completed.back();
-    bsd->completed.pop_back();
-    return ev;
-}
-
-static int sec_io_u_init(struct thread_data *td, struct io_u *io)
-{
-    io->engine_data = NULL;
-    return 0;
-}
-
-static void sec_io_u_free(struct thread_data *td, struct io_u *io)
-{
-}
-
-static int sec_open_file(struct thread_data *td, struct fio_file *f)
-{
-    return 0;
-}
-
-static int sec_invalidate(struct thread_data *td, struct fio_file *f)
-{
-    return 0;
-}
-
-struct ioengine_ops ioengine = {
-    .name               = "microceph_cluster",
-    .version            = FIO_IOOPS_VERSION,
-    .flags              = FIO_MEMALIGN | FIO_DISKLESSIO | FIO_NOEXTEND,
-    .setup              = sec_setup,
-    .init               = sec_init,
-    .queue              = sec_queue,
-    .getevents          = sec_getevents,
-    .event              = sec_event,
-    .cleanup            = sec_cleanup,
-    .open_file          = sec_open_file,
-    .invalidate         = sec_invalidate,
-    .io_u_init          = sec_io_u_init,
-    .io_u_free          = sec_io_u_free,
-    .option_struct_size = sizeof(struct sec_options),
-    .options            = options,
-};
-
-static void fio_init fio_sec_register(void)
-{
-    register_ioengine(&ioengine);
-}
-
-static void fio_exit fio_sec_unregister(void)
-{
-    unregister_ioengine(&ioengine);
-}

json11

@@ -0,0 +1 @@
+Subproject commit 97f06cb20c1e136fd37d58fb40f57dd8f8a3a4a7

lambda_size.cpp

@@ -1,48 +0,0 @@
-#include <iostream>
-#include <functional>
-#include <array>
-#include <cstdlib> // for malloc() and free()
-using namespace std;
-
-// replace operator new and delete to log allocations
-void* operator new(std::size_t n)
-{
-    cout << "Allocating " << n << " bytes" << endl;
-    return malloc(n);
-}
-
-void operator delete(void* p) throw()
-{
-    free(p);
-}
-
-class test
-{
-public:
-    std::string s;
-    void a(std::function<void()> & f, const char *str)
-    {
-        auto l = [this, str]() { cout << str << " ? " << s << " from this\n"; };
-        cout << "Assigning lambda3 of size " << sizeof(l) << endl;
-        f = l;
-    }
-};
-
-int main()
-{
-    std::array<char, 16> arr1;
-    auto lambda1 = [arr1](){};
-    cout << "Assigning lambda1 of size " << sizeof(lambda1) << endl;
-    std::function<void()> f1 = lambda1;
-
-    std::array<char, 17> arr2;
-    auto lambda2 = [arr2](){};
-    cout << "Assigning lambda2 of size " << sizeof(lambda2) << endl;
-    std::function<void()> f2 = lambda2;
-
-    test t;
-    std::function<void()> f3;
-    t.s = "str";
-    t.a(f3, "huyambda");
-    f3();
-}

lp/mon.js

@@ -1,858 +0,0 @@
-const http = require('http');
-const os = require('os');
-const WebSocket = require('ws');
-const LPOptimizer = require('./lp-optimizer.js');
-const stableStringify = require('./stable-stringify.js');
-
-class Mon
-{
-    static etcd_tree = {
-        config: {
-            global: null,
-            /* placement_tree = {
-                levels: { datacenter: 1, rack: 2, host: 3, osd: 4, ... },
-                nodes: { host1: { level: 'host', parent: 'rack1' }, ... },
-                failure_domain: 'host',
-            } */
-            placement_tree: null,
-            osd: {},
-            pgs: {},
-        },
-        osd: {
-            state: {},
-            stats: {},
-        },
-        mon: {
-            master: null,
-        },
-        pg: {
-            change_stamp: null,
-            state: {},
-            stats: {},
-            history: {},
-        },
-    }
-
-    constructor(config)
-    {
-        // FIXME: Maybe prefer local etcd
-        this.etcd_urls = [];
-        for (let url of config.etcd_url.split(/,/))
-        {
-            let scheme = 'http';
-            url = url.trim().replace(/^(https?):\/\//, (m, m1) => { scheme = m1; return ''; });
-            if (!/\/[^\/]/.exec(url))
-                url += '/v3';
-            this.etcd_urls.push(scheme+'://'+url);
-        }
-        this.etcd_prefix = config.etcd_prefix || '/rage';
-        this.etcd_prefix = this.etcd_prefix.replace(/\/\/+/g, '/').replace(/^\/?(.*[^\/])\/?$/, '/$1');
-        this.etcd_start_timeout = (config.etcd_start_timeout || 5) * 1000;
-        this.state = JSON.parse(JSON.stringify(Mon.etcd_tree));
-    }
-
-    async start()
-    {
-        await this.load_config();
-        await this.get_lease();
-        await this.become_master();
-        await this.load_cluster_state();
-        await this.start_watcher();
-        await this.recheck_pgs();
-    }
-
-    async load_config()
-    {
-        const res = await this.etcd_call('/txn', { success: [
-            { requestRange: { key: b64(this.etcd_prefix+'/config/global') } }
-        ] }, this.etcd_start_timeout, -1);
-        this.parse_kv(res.responses[0].response_range.kvs[0]);
-        this.check_config();
-    }
-
-    check_config()
-    {
-        this.config.etcd_mon_timeout = Number(this.config.etcd_mon_timeout) || 0;
-        if (this.config.etcd_mon_timeout <= 0)
-        {
-            this.config.etcd_mon_timeout = 1000;
-        }
-        this.config.etcd_mon_retries = Number(this.config.etcd_mon_retries) || 5;
-        if (this.config.etcd_mon_retries < 0)
-        {
-            this.config.etcd_mon_retries = 0;
-        }
-        this.config.mon_change_timeout = Number(this.config.mon_change_timeout) || 1000;
-        if (this.config.mon_change_timeout < 100)
-        {
-            this.config.mon_change_timeout = 100;
-        }
-        this.config.mon_stats_timeout = Number(this.config.mon_stats_timeout) || 1000;
-        if (this.config.mon_stats_timeout < 100)
-        {
-            this.config.mon_stats_timeout = 100;
-        }
-        // After this number of seconds, a dead OSD will be removed from PG distribution
-        this.config.osd_out_time = Number(this.config.osd_out_time) || 0;
-        if (!this.config.osd_out_time)
-        {
-            this.config.osd_out_time = 30*60; // 30 minutes by default
-        }
-        this.config.max_osd_combinations = Number(this.config.max_osd_combinations) || 10000;
-        if (this.config.max_osd_combinations < 100)
-        {
-            this.config.max_osd_combinations = 100;
-        }
-    }
-
-    async start_watcher(retries)
-    {
-        let retry = 0;
-        if (retries >= 0 && retries < 1)
-        {
-            retries = 1;
-        }
-        while (retries < 0 || retry < retries)
-        {
-            const base = 'ws'+this.etcd_urls[Math.floor(Math.random()*this.etcd_urls.length)].substr(4);
-            const ok = await new Promise((ok, no) =>
-            {
-                const timer_id = setTimeout(() =>
-                {
-                    this.ws.close();
-                    ok(false);
-                }, timeout);
-                this.ws = new WebSocket(base+'/watch');
-                this.ws.on('open', () =>
-                {
-                    if (timer_id)
-                        clearTimeout(timer_id);
-                    ok(true);
-                });
-            });
-            if (!ok)
-            {
-                this.ws = null;
-            }
-            retry++;
-        }
-        if (!this.ws)
-        {
-            this.die('Failed to open etcd watch websocket');
-        }
-        this.ws.send(JSON.stringify({
-            create_request: {
-                key: b64(this.etcd_prefix+'/'),
-                range_end: b64(this.etcd_prefix+'0'),
-                start_revision: ''+this.etcd_watch_revision,
-                watch_id: 1,
-            },
-        }));
-        this.ws.on('message', (msg) =>
-        {
-            let data;
-            try
-            {
-                data = JSON.parse(msg);
-            }
-            catch (e)
-            {
-            }
-            if (!data || !data.result || !data.result.events)
-            {
-                console.error('Garbage received from watch websocket: '+msg);
-            }
-            else
-            {
-                let stats_changed = false, changed = false;
-                console.log('Revision '+data.result.header.revision+' events: ');
-                for (const e of data.result.events)
-                {
-                    this.parse_kv(e.kv);
-                    const key = e.kv.key.substr(this.etcd_prefix.length);
-                    if (key.substr(0, 11) == '/osd/stats/' || key.substr(0, 10) == '/pg/stats/')
-                    {
-                        stats_changed = true;
-                    }
-                    else if (key != '/stats')
-                    {
-                        changed = true;
-                    }
-                    console.log(e);
-                }
-                if (stats_changed)
-                {
-                    this.schedule_update_stats();
-                }
-                if (changed)
-                {
-                    this.schedule_recheck();
-                }
-            }
-        });
-    }
-
-    async get_lease()
-    {
-        const max_ttl = this.config.etcd_mon_ttl + this.config.etcd_mon_timeout/1000*this.config.etcd_mon_retries;
-        const res = await this.etcd_call('/lease/grant', { TTL: max_ttl }, this.config.etcd_mon_timeout, this.config.etcd_mon_retries);
-        this.etcd_lease_id = res.ID;
-        setInterval(async () =>
-        {
-            const res = await this.etcd_call('/lease/keepalive', { ID: this.etcd_lease_id }, this.config.etcd_mon_timeout, this.config.etcd_mon_retries);
-            if (!res.result.TTL)
-            {
-                this.die('Lease expired');
-            }
-        }, config.etcd_mon_timeout);
-    }
-
-    async become_master()
-    {
-        const state = { ip: this.local_ips() };
-        while (1)
-        {
-            const res = await this.etcd_call('/txn', {
-                compare: [ { target: 'CREATE', create_revision: 0, key: b64(this.etcd_prefix+'/mon/master') } ],
-                success: [ { key: b64(this.etcd_prefix+'/mon/master'), value: b64(JSON.stringify(state)), lease: ''+this.etcd_lease_id } ],
-            }, this.etcd_start_timeout, 0);
-            if (!res.succeeded)
-            {
-                await new Promise(ok => setTimeout(ok, this.etcd_start_timeout));
-            }
-        }
-    }
-
-    async load_cluster_state()
-    {
-        const res = await this.etcd_call('/txn', { success: [
-            { requestRange: { key: b64(this.etcd_prefix+'/'), range_end: b64(this.etcd_prefix+'0') } },
-        ] }, this.etcd_start_timeout, -1);
-        this.etcd_watch_revision = BigInt(res.header.revision)+BigInt(1);
-        const data = JSON.parse(JSON.stringify(Mon.etcd_tree));
-        for (const response of res.responses)
-        {
-            for (const kv of response.response_range.kvs)
-            {
-                this.parse_kv(kv);
-            }
-        }
-        this.state = data;
-    }
-
-    all_osds()
-    {
-        return Object.keys(this.state.osd.stats);
-    }
-
-    get_osd_tree()
-    {
-        this.state.config.placement_tree = this.state.config.placement_tree||{};
-        const levels = this.state.config.placement_tree.levels||{};
-        levels.host = levels.host || 100;
-        levels.osd = levels.osd || 101;
-        const tree = { '': { children: [] } };
-        for (const node_id in this.state.config.placement_tree.nodes||{})
-        {
-            const node_cfg = this.state.config.placement_tree.nodes[node_id];
-            if (!node_id || /^\d/.exec(node_id) ||
-                !node_cfg.level || !levels[node_cfg.level])
-            {
-                // All nodes must have non-empty non-numeric IDs and valid levels
-                continue;
-            }
-            tree[node_id] = { id: node_id, level: node_cfg.level, parent: node_cfg.parent, children: [] };
-        }
-        // This requires monitor system time to be in sync with OSD system times (at least to some extent)
-        const down_time = Date.now()/1000 - this.config.osd_out_time;
-        for (const osd_num of this.all_osds().sort((a, b) => a - b))
-        {
-            const stat = this.state.osd.stats[osd_num];
-            if (stat.size && (this.state.osd.state[osd_num] || Number(stat.time) >= down_time))
-            {
-                // Numeric IDs are reserved for OSDs
-                const reweight = this.state.config.osd[osd_num] && Number(this.state.config.osd[osd_num].reweight) || 1;
-                tree[osd_num] = tree[osd_num] || { id: osd_num, parent: stat.host };
-                tree[osd_num].level = 'osd';
-                tree[osd_num].size = reweight * stat.size / 1024 / 1024 / 1024 / 1024; // terabytes
-                delete tree[osd_num].children;
-            }
-        }
-        for (const node_id in tree)
-        {
-            if (node_id === '')
-            {
-                continue;
-            }
-            const node_cfg = tree[node_id];
-            const node_level = levels[node_cfg.level] || node_cfg.level;
-            let parent_level = node_cfg.parent && tree[node_cfg.parent] && tree[node_cfg.parent].children
-                && tree[node_cfg.parent].level;
-            parent_level = parent_level ? (levels[parent_level] || parent_level) : null;
-            // Parent's level must be less than child's; OSDs must be leaves
-            const parent = parent_level && parent_level < node_level ? tree[node_cfg.parent] : '';
-            tree[parent].children.push(tree[node_id]);
-            delete node_cfg.parent;
-        }
-        return LPOptimizer.flatten_tree(tree[''].children, levels, this.state.config.failure_domain, 'osd');
-    }
-
-    async stop_all_pgs()
-    {
-        let has_online = false, paused = true;
-        for (const pg in this.state.config.pgs.items||{})
-        {
-            const cur_state = ((this.state.pg.state[pg]||{}).state||[]).join(',');
-            if (cur_state != '' && cur_state != 'offline')
-            {
-                has_online = true;
-            }
-            if (!this.state.config.pgs.items[pg].pause)
-            {
-                paused = false;
-            }
-        }
-        if (!paused)
-        {
-            console.log('Stopping all PGs before changing PG count');
-            const new_cfg = JSON.parse(JSON.stringify(this.state.config.pgs));
-            for (const pg in new_cfg.items)
-            {
-                new_cfg.items[pg].pause = true;
-            }
-            // Check that no OSDs change their state before we pause PGs
-            // Doing this we make sure that OSDs don't wake up in the middle of our "transaction"
-            // and can't see the old PG configuration
-            const checks = [];
-            for (const osd_num of this.all_osds())
-            {
-                const key = b64(this.etcd_prefix+'/osd/state/'+osd_num);
-                checks.push({ key, target: 'MOD', result: 'LESS', mod_revision: ''+this.etcd_watch_revision });
-            }
-            const res = await this.etcd_call('/txn', {
-                compare: [
-                    { key: b64(this.etcd_prefix+'/mon/master'), target: 'LEASE', lease: ''+this.etcd_lease_id },
-                    { key: b64(this.etcd_prefix+'/config/pgs'), target: 'MOD', mod_revision: ''+this.etcd_watch_revision, result: 'LESS' },
-                    ...checks,
-                ],
-                success: [
-                    { requestPut: { key: b64(this.etcd_prefix+'/config/pgs'), value: b64(JSON.stringify(new_cfg)) } },
-                ],
-            }, this.config.etcd_mon_timeout, 0);
-            if (!res.succeeded)
-            {
-                return false;
-            }
-            this.state.config.pgs = new_cfg;
-        }
-        return !has_online;
-    }
-
-    scale_pg_count(prev_pgs, pg_history, new_pg_count)
-    {
-        const old_pg_count = prev_pgs.length;
-        // Add all possibly intersecting PGs into the history of new PGs
-        if (!(new_pg_count % old_pg_count))
-        {
-            // New PG count is a multiple of the old PG count
-            const mul = (new_pg_count / old_pg_count);
-            for (let i = 0; i < new_pg_count; i++)
-            {
-                const old_i = Math.floor(new_pg_count / mul);
-                pg_history[i] = JSON.parse(JSON.stringify(this.state.pg.history[1+old_i]));
-            }
-        }
-        else if (!(old_pg_count % new_pg_count))
-        {
-            // Old PG count is a multiple of the new PG count
-            const mul = (old_pg_count / new_pg_count);
-            for (let i = 0; i < new_pg_count; i++)
-            {
-                pg_history[i] = {
-                    osd_sets: [],
-                    all_peers: [],
-                };
-                for (let j = 0; j < mul; j++)
-                {
-                    pg_history[i].osd_sets.push(prev_pgs[i*mul]);
-                    const hist = this.state.pg.history[1+i*mul+j];
-                    if (hist && hist.osd_sets && hist.osd_sets.length)
-                    {
-                        Array.prototype.push.apply(pg_history[i].osd_sets, hist.osd_sets);
-                    }
-                    if (hist && hist.all_peers && hist.all_peers.length)
-                    {
-                        Array.prototype.push.apply(pg_history[i].all_peers, hist.all_peers);
-                    }
-                }
-            }
-        }
-        else
-        {
-            // Any PG may intersect with any PG after non-multiple PG count change
-            // So, merge ALL PGs history
-            let all_sets = {};
-            let all_peers = {};
-            for (const pg of prev_pgs)
-            {
-                all_sets[pg.join(' ')] = pg;
-            }
-            for (const pg in this.state.pg.history)
-            {
-                const hist = this.state.pg.history[pg];
-                if (hist && hist.osd_sets)
-                {
-                    for (const pg of hist.osd_sets)
-                    {
-                        all_sets[pg.join(' ')] = pg;
-                    }
-                }
-                if (hist && hist.all_peers)
-                {
-                    for (const osd_num of hist.all_peers)
-                    {
-                        all_peers[osd_num] = Number(osd_num);
-                    }
-                }
-            }
-            all_sets = Object.values(all_sets);
-            all_peers = Object.values(all_peers);
-            for (let i = 0; i < new_pg_count; i++)
-            {
-                pg_history[i] = { osd_sets: all_sets, all_peers };
-            }
-        }
-        // Mark history keys for removed PGs as removed
-        for (let i = new_pg_count; i < old_pg_count; i++)
-        {
-            pg_history[i] = null;
-        }
-        if (old_pg_count < new_pg_count)
-        {
-            for (let i = new_pg_count-1; i >= 0; i--)
-            {
-                prev_pgs[i] = prev_pgs[Math.floor(i/new_pg_count*old_pg_count)];
-            }
-        }
-        else if (old_pg_count > new_pg_count)
-        {
-            for (let i = 0; i < new_pg_count; i++)
-            {
-                prev_pgs[i] = prev_pgs[Math.round(i/new_pg_count*old_pg_count)];
-            }
-            prev_pgs.splice(new_pg_count, old_pg_count-new_pg_count);
-        }
-    }
-
-    async save_new_pgs(prev_pgs, new_pgs, pg_history, tree_hash)
-    {
-        const txn = [], checks = [];
-        const pg_items = {};
-        new_pgs.map((osd_set, i) =>
-        {
-            osd_set = osd_set.map(osd_num => osd_num === LPOptimizer.NO_OSD ? 0 : osd_num);
-            const alive_set = osd_set.filter(osd_num => osd_num);
-            pg_items[i+1] = {
-                osd_set,
-                primary: alive_set.length ? alive_set[Math.floor(Math.random()*alive_set.length)] : 0,
-            };
-            if (prev_pgs[i] && prev_pgs[i].join(' ') != osd_set.join(' '))
-            {
-                pg_history[i] = pg_history[i] || {};
-                pg_history[i].osd_sets = pg_history[i].osd_sets || [];
-                pg_history[i].osd_sets.push(prev_pgs[i]);
-            }
-        });
-        for (let i = 0; i < new_pgs.length || i < prev_pgs.length; i++)
-        {
-            checks.push({
-                key: b64(this.etcd_prefix+'/pg/history/'+(i+1)),
-                target: 'MOD',
-                mod_revision: ''+this.etcd_watch_revision,
-                result: 'LESS',
-            });
-            if (pg_history[i])
-            {
-                txn.push({
-                    requestPut: {
-                        key: b64(this.etcd_prefix+'/pg/history/'+(i+1)),
-                        value: b64(JSON.stringify(pg_history[i])),
-                    },
-                });
-            }
-            else
-            {
-                txn.push({
-                    requestDeleteRange: {
-                        key: b64(this.etcd_prefix+'/pg/history/'+(i+1)),
-                    },
-                });
-            }
-        }
-        this.state.config.pgs = {
-            hash: tree_hash,
-            items: pg_items,
-        };
-        const res = await this.etcd_call('/txn', {
-            compare: [
-                { key: b64(this.etcd_prefix+'/mon/master'), target: 'LEASE', lease: ''+this.etcd_lease_id },
-                { key: b64(this.etcd_prefix+'/config/pgs'), target: 'MOD', mod_revision: ''+this.etcd_watch_revision, result: 'LESS' },
-                ...checks,
-            ],
-            success: [
-                { requestPut: { key: b64(this.etcd_prefix+'/config/pgs'), value: b64(JSON.stringify(this.state.config.pgs)) } },
-                ...txn,
-            ],
-        }, this.config.etcd_mon_timeout, 0);
-        return res.succeeded;
-    }
-
-    async recheck_pgs()
-    {
-        // Take configuration and state, check it against the stored configuration hash
-        // Recalculate PGs and save them to etcd if the configuration is changed
-        const tree_cfg = {
-            osd_tree: this.get_osd_tree(),
-            pg_count: this.config.pg_count || Object.keys(this.state.config.pgs.items||{}).length || 128,
-            max_osd_combinations: this.config.max_osd_combinations,
-        };
-        const tree_hash = sha1hex(stableStringify(tree_cfg));
-        if (this.state.config.pgs.hash != tree_hash)
-        {
-            // Something has changed
-            const prev_pgs = [];
-            for (const pg in this.state.config.pgs.items||{})
-            {
-                prev_pgs[pg-1] = this.state.config.pgs.items[pg].osd_set;
-            }
-            const pg_history = [];
-            const old_pg_count = prev_pgs.length;
-            let optimize_result;
-            if (old_pg_count > 0)
-            {
-                if (old_pg_count != tree_cfg.pg_count)
-                {
-                    // PG count changed. Need to bring all PGs down.
-                    if (!await this.stop_all_pgs())
-                    {
-                        this.schedule_recheck();
-                        return;
-                    }
-                    this.scale_pg_count(prev_pgs, pg_history, new_pg_count);
-                }
-                optimize_result = await LPOptimizer.optimize_change(prev_pgs, tree_cfg.osd_tree, tree_cfg.max_osd_combinations);
-            }
-            else
-            {
-                optimize_result = await LPOptimizer.optimize_initial(tree_cfg.osd_tree, tree_cfg.pg_count, tree_cfg.max_osd_combinations);
-            }
-            if (!await this.save_new_pgs(prev_pgs, optimize_result.int_pgs, pg_history, tree_hash))
-            {
-                console.log('Someone changed PG configuration while we also tried to change it. Retrying in '+this.config.mon_change_timeout+' ms');
-                this.schedule_recheck();
-                return;
-            }
-            console.log('PG configuration successfully changed');
-            if (old_pg_count != optimize_result.int_pgs.length)
-            {
-                console.log(`PG count changed from: ${old_pg_count} to ${optimize_result.int_pgs.length}`);
-            }
-            LPOptimizer.print_change_stats(optimize_result);
-        }
-    }
-
-    schedule_recheck()
-    {
-        if (this.recheck_timer)
-        {
-            clearTimeout(this.recheck_timer);
-            this.recheck_timer = null;
-        }
-        this.recheck_timer = setTimeout(() =>
-        {
-            this.recheck_timer = null;
-            this.recheck_pgs().catch(console.error);
-        }, this.config.mon_change_timeout || 1000);
-    }
-
-    sum_stats()
-    {
-        let overflow = false;
-        this.prev_stats = this.prev_stats || { op_stats: {}, subop_stats: {}, recovery_stats: {} };
-        const op_stats = {}, subop_stats = {}, recovery_stats = {};
-        for (const osd in this.state.osd.stats)
-        {
-            const st = this.state.osd.stats[osd];
-            for (const op in st.op_stats||{})
-            {
-                op_stats[op] = op_stats[op] || { count: 0n, usec: 0n, bytes: 0n };
-                op_stats[op].count += BigInt(st.op_stats.count||0);
-                op_stats[op].usec += BigInt(st.op_stats.usec||0);
-                op_stats[op].bytes += BigInt(st.op_stats.bytes||0);
-            }
-            for (const op in st.subop_stats||{})
-            {
-                subop_stats[op] = subop_stats[op] || { count: 0n, usec: 0n };
-                subop_stats[op].count += BigInt(st.subop_stats.count||0);
-                subop_stats[op].usec += BigInt(st.subop_stats.usec||0);
-            }
-            for (const op in st.recovery_stats||{})
-            {
-                recovery_stats[op] = recovery_stats[op] || { count: 0n, bytes: 0n };
-                recovery_stats[op].count += BigInt(st.recovery_stats.count||0);
-                recovery_stats[op].bytes += BigInt(st.recovery_stats.bytes||0);
-            }
-        }
-        for (const op in op_stats)
-        {
-            if (op_stats[op].count >= 0x10000000000000000n)
-            {
-                if (!this.prev_stats.op_stats[op])
-                {
-                    overflow = true;
-                }
-                else
-                {
-                    op_stats[op].count -= this.prev_stats.op_stats[op].count;
-                    op_stats[op].usec -= this.prev_stats.op_stats[op].usec;
-                    op_stats[op].bytes -= this.prev_stats.op_stats[op].bytes;
-                }
-            }
-        }
-        for (const op in subop_stats)
-        {
-            if (subop_stats[op].count >= 0x10000000000000000n)
-            {
-                if (!this.prev_stats.subop_stats[op])
-                {
-                    overflow = true;
-                }
-                else
-                {
-                    subop_stats[op].count -= this.prev_stats.subop_stats[op].count;
-                    subop_stats[op].usec -= this.prev_stats.subop_stats[op].usec;
-                }
-            }
-        }
-        for (const op in recovery_stats)
-        {
-            if (recovery_stats[op].count >= 0x10000000000000000n)
-            {
-                if (!this.prev_stats.recovery_stats[op])
-                {
-                    overflow = true;
-                }
-                else
-                {
-                    recovery_stats[op].count -= this.prev_stats.recovery_stats[op].count;
-                    recovery_stats[op].bytes -= this.prev_stats.recovery_stats[op].bytes;
-                }
-            }
-        }
-        const object_counts = { object: 0n, clean: 0n, misplaced: 0n, degraded: 0n, incomplete: 0n };
-        for (const pg_num in this.state.pg.stats)
-        {
-            const st = this.state.pg.stats[pg_num];
-            for (const k in object_counts)
-            {
-                if (st[k+'_count'])
-                {
-                    object_counts[k] += BigInt(st[k+'_count']);
-                }
-            }
-        }
-        return (this.prev_stats = { overflow, op_stats, subop_stats, recovery_stats, object_counts });
-    }
-
-    async update_total_stats()
-    {
-        const stats = this.sum_stats();
-        if (!stats.overflow)
-        {
-            // Convert to strings, serialize and save
-            const ser = {};
-            for (const st of [ 'op_stats', 'subop_stats', 'recovery_stats' ])
-            {
-                ser[st] = {};
-                for (const op in stats[st])
-                {
-                    ser[st][op] = {};
-                    for (const k in stats[st][op])
-                    {
-                        ser[st][op][k] = ''+stats[st][op][k];
-                    }
-                }
-            }
-            ser.object_counts = {};
-            for (const k in stats.object_counts)
-            {
-                ser.object_counts[k] = ''+stats.object_counts[k];
-            }
-            await this.etcd_call('/txn', {
-                success: [ { requestPut: { key: b64(this.etcd_prefix+'/stats'), value: b64(JSON.stringify(ser)) } } ],
-            }, this.config.etcd_mon_timeout, 0);
-        }
-    }
-
-    schedule_update_stats()
-    {
-        if (this.stats_timer)
-        {
-            clearTimeout(this.stats_timer);
-            this.stats_timer = null;
-        }
-        this.stats_timer = setTimeout(() =>
-        {
-            this.stats_timer = null;
-            this.update_total_stats().catch(console.error);
-        }, this.config.mon_stats_timeout || 1000);
-    }
-
-    parse_kv(kv)
-    {
-        if (!kv || !kv.key)
-        {
-            return;
-        }
-        kv.key = de64(kv.key);
-        kv.value = kv.value ? JSON.parse(de64(kv.value)) : null;
-        const key = kv.key.substr(this.etcd_prefix.length).replace(/^\/+/, '').split('/');
-        const cur = this.state, orig = Mon.etcd_tree;
-        for (let i = 0; i < key.length-1; i++)
-        {
-            if (!orig[key[i]])
-            {
-                console.log('Bad key in etcd: '+kv.key+' = '+kv.value);
-                return;
-            }
-            orig = orig[key[i]];
-            cur = (cur[key[i]] = cur[key[i]] || {});
-        }
-        if (orig[key.length-1])
-        {
-            console.log('Bad key in etcd: '+kv.key+' = '+kv.value);
-            return;
-        }
-        cur[key[key.length-1]] = kv.value;
-        if (key.join('/') === 'config/global')
-        {
-            this.state.config.global = this.state.config.global || {};
-            this.config = this.state.config.global;
-            this.check_config();
-        }
-    }
-
-    async etcd_call(path, body, timeout, retries)
-    {
-        let retry = 0;
-        if (retries >= 0 && retries < 1)
-        {
-            retries = 1;
-        }
-        while (retries < 0 || retry < retries)
-        {
-            const base = this.etcd_urls[Math.floor(Math.random()*this.etcd_urls.length)];
-            const res = await POST(base+path, body, timeout);
-            if (res.json)
-            {
-                if (res.json.error)
-                {
-                    console.log('etcd returned error: '+res.json.error);
-                    break;
-                }
-                return res.json;
-            }
-            retry++;
-        }
-        this.die();
-    }
-
-    die(err)
-    {
-        // In fact we can just try to rejoin
-        console.fatal(err || 'Cluster connection failed');
-        process.exit(1);
-    }
-
-    local_ips()
-    {
-        const ips = [];
-        const ifaces = os.networkInterfaces();
-        for (const ifname in ifaces)
-        {
-            for (const iface of ifaces[ifname])
-            {
-                if (iface.family == 'IPv4' && !iface.internal)
-                {
-                    ips.push(iface.address);
-                }
-            }
-        }
-        return ips;
-    }
-}
-
-function POST(url, body, timeout)
-{
-    return new Promise((ok, no) =>
-    {
-        const body_text = Buffer.from(JSON.stringify(body));
-        let timer_id = timeout > 0 ? setTimeout(() =>
-        {
-            if (req)
-                req.abort();
-            req = null;
-            ok({ error: 'timeout' });
-        }, timeout) : null;
-        let req = http.request(url, { method: 'POST', headers: {
-            'Content-Type': 'application/json',
-            'Content-Length': body_text,
-        } }, (res) =>
-        {
-            if (!req)
-            {
-                return;
-            }
-            clearTimeout(timer_id);
-            if (res.statusCode != 200)
-            {
-                ok({ error: res.statusCode, response: res });
-                return;
-            }
-            let res_body = '';
-            res.setEncoding('utf8');
-            res.on('data', chunk => { res_body += chunk });
-            res.on('end', () =>
-            {
-                try
-                {
-                    res_body = JSON.parse(res_body);
-                    ok({ response: res, json: res_body });
-                }
-                catch (e)
-                {
-                    ok({ error: e, response: res, body: res_body });
-                }
-            });
-        });
-        req.write(body_text);
-        req.end();
-    });
-}
-
-function b64(str)
-{
-    return Buffer.from(str).toString('base64');
-}
-
-function de64(str)
-{
-    return Buffer.from(str, 'base64').toString();
-}
-
-function sha1hex(str)
-{
-    const hash = crypto.createHash('sha1');
-    hash.update(str);
-    return hash.digest('hex');
-}

messenger.cpp

@@ -1,398 +0,0 @@
-#include <unistd.h>
-#include <fcntl.h>
-#include <sys/socket.h>
-#include <sys/epoll.h>
-#include <netinet/tcp.h>
-
-#include "messenger.h"
-
-osd_op_t::~osd_op_t()
-{
-    assert(!bs_op);
-    assert(!op_data);
-    if (rmw_buf)
-    {
-        free(rmw_buf);
-    }
-    if (buf)
-    {
-        // Note: reusing osd_op_t WILL currently lead to memory leaks
-        // So we don't reuse it, but free it every time
-        free(buf);
-    }
-}
-
-void osd_messenger_t::connect_peer(uint64_t peer_osd, json11::Json peer_state)
-{
-    if (wanted_peers.find(peer_osd) == wanted_peers.end())
-    {
-        wanted_peers[peer_osd] = (osd_wanted_peer_t){
-            .address_list = peer_state["addresses"],
-            .port = (int)peer_state["port"].int64_value(),
-        };
-    }
-    else
-    {
-        wanted_peers[peer_osd].address_list = peer_state["addresses"];
-        wanted_peers[peer_osd].port = (int)peer_state["port"].int64_value();
-    }
-    wanted_peers[peer_osd].address_changed = true;
-    if (!wanted_peers[peer_osd].connecting &&
-        (time(NULL) - wanted_peers[peer_osd].last_connect_attempt) >= peer_connect_interval)
-    {
-        try_connect_peer(peer_osd);
-    }
-}
-
-void osd_messenger_t::try_connect_peer(uint64_t peer_osd)
-{
-    auto wp_it = wanted_peers.find(peer_osd);
-    if (wp_it == wanted_peers.end())
-    {
-        return;
-    }
-    if (osd_peer_fds.find(peer_osd) != osd_peer_fds.end())
-    {
-        wanted_peers.erase(peer_osd);
-        return;
-    }
-    auto & wp = wp_it->second;
-    if (wp.address_index >= wp.address_list.array_items().size())
-    {
-        return;
-    }
-    wp.cur_addr = wp.address_list[wp.address_index].string_value();
-    wp.cur_port = wp.port;
-    try_connect_peer_addr(peer_osd, wp.cur_addr.c_str(), wp.cur_port);
-}
-
-void osd_messenger_t::try_connect_peer_addr(osd_num_t peer_osd, const char *peer_host, int peer_port)
-{
-    struct sockaddr_in addr;
-    int r;
-    if ((r = inet_pton(AF_INET, peer_host, &addr.sin_addr)) != 1)
-    {
-        on_connect_peer(peer_osd, -EINVAL);
-        return;
-    }
-    addr.sin_family = AF_INET;
-    addr.sin_port = htons(peer_port ? peer_port : 11203);
-    int peer_fd = socket(AF_INET, SOCK_STREAM, 0);
-    if (peer_fd < 0)
-    {
-        on_connect_peer(peer_osd, -errno);
-        return;
-    }
-    fcntl(peer_fd, F_SETFL, fcntl(peer_fd, F_GETFL, 0) | O_NONBLOCK);
-    int timeout_id = -1;
-    if (peer_connect_timeout > 0)
-    {
-        timeout_id = tfd->set_timer(1000*peer_connect_timeout, false, [this, peer_fd](int timer_id)
-        {
-            osd_num_t peer_osd = clients[peer_fd].osd_num;
-            stop_client(peer_fd);
-            on_connect_peer(peer_osd, -EIO);
-            return;
-        });
-    }
-    r = connect(peer_fd, (sockaddr*)&addr, sizeof(addr));
-    if (r < 0 && errno != EINPROGRESS)
-    {
-        close(peer_fd);
-        on_connect_peer(peer_osd, -errno);
-        return;
-    }
-    assert(peer_osd != this->osd_num);
-    clients[peer_fd] = (osd_client_t){
-        .peer_addr = addr,
-        .peer_port = peer_port,
-        .peer_fd = peer_fd,
-        .peer_state = PEER_CONNECTING,
-        .connect_timeout_id = timeout_id,
-        .osd_num = peer_osd,
-        .in_buf = malloc(receive_buffer_size),
-    };
-    tfd->set_fd_handler(peer_fd, [this](int peer_fd, int epoll_events)
-    {
-        // Either OUT (connected) or HUP
-        handle_connect_epoll(peer_fd);
-    });
-}
-
-void osd_messenger_t::handle_connect_epoll(int peer_fd)
-{
-    auto & cl = clients[peer_fd];
-    if (cl.connect_timeout_id >= 0)
-    {
-        tfd->clear_timer(cl.connect_timeout_id);
-        cl.connect_timeout_id = -1;
-    }
-    osd_num_t peer_osd = cl.osd_num;
-    int result = 0;
-    socklen_t result_len = sizeof(result);
-    if (getsockopt(peer_fd, SOL_SOCKET, SO_ERROR, &result, &result_len) < 0)
-    {
-        result = errno;
-    }
-    if (result != 0)
-    {
-        stop_client(peer_fd);
-        on_connect_peer(peer_osd, -result);
-        return;
-    }
-    int one = 1;
-    setsockopt(peer_fd, SOL_TCP, TCP_NODELAY, &one, sizeof(one));
-    cl.peer_state = PEER_CONNECTED;
-    // FIXME Disable EPOLLOUT on this fd
-    tfd->set_fd_handler(peer_fd, [this](int peer_fd, int epoll_events)
-    {
-        handle_peer_epoll(peer_fd, epoll_events);
-    });
-    // Check OSD number
-    check_peer_config(cl);
-}
-
-void osd_messenger_t::handle_peer_epoll(int peer_fd, int epoll_events)
-{
-    // Mark client as ready (i.e. some data is available)
-    if (epoll_events & EPOLLRDHUP)
-    {
-        // Stop client
-        printf("[OSD %lu] client %d disconnected\n", this->osd_num, peer_fd);
-        stop_client(peer_fd);
-    }
-    else if (epoll_events & EPOLLIN)
-    {
-        // Mark client as ready (i.e. some data is available)
-        auto & cl = clients[peer_fd];
-        cl.read_ready++;
-        if (cl.read_ready == 1)
-        {
-            read_ready_clients.push_back(cl.peer_fd);
-            ringloop->wakeup();
-        }
-    }
-}
-
-void osd_messenger_t::on_connect_peer(osd_num_t peer_osd, int peer_fd)
-{
-    auto & wp = wanted_peers.at(peer_osd);
-    wp.connecting = false;
-    if (peer_fd < 0)
-    {
-        printf("Failed to connect to peer OSD %lu address %s port %d: %s\n", peer_osd, wp.cur_addr.c_str(), wp.cur_port, strerror(-peer_fd));
-        if (wp.address_changed)
-        {
-            wp.address_changed = false;
-            wp.address_index = 0;
-            try_connect_peer(peer_osd);
-        }
-        else if (wp.address_index < wp.address_list.array_items().size()-1)
-        {
-            // Try other addresses
-            wp.address_index++;
-            try_connect_peer(peer_osd);
-        }
-        else
-        {
-            // Retry again in <peer_connect_interval> seconds
-            wp.last_connect_attempt = time(NULL);
-            wp.address_index = 0;
-            tfd->set_timer(1000*peer_connect_interval, false, [this, peer_osd](int)
-            {
-                try_connect_peer(peer_osd);
-            });
-        }
-        return;
-    }
-    printf("Connected with peer OSD %lu (fd %d)\n", peer_osd, peer_fd);
-    wanted_peers.erase(peer_osd);
-    repeer_pgs(peer_osd);
-}
-
-void osd_messenger_t::check_peer_config(osd_client_t & cl)
-{
-    osd_op_t *op = new osd_op_t();
-    op->op_type = OSD_OP_OUT;
-    op->send_list.push_back(op->req.buf, OSD_PACKET_SIZE);
-    op->peer_fd = cl.peer_fd;
-    op->req = {
-        .show_conf = {
-            .header = {
-                .magic = SECONDARY_OSD_OP_MAGIC,
-                .id = this->next_subop_id++,
-                .opcode = OSD_OP_SHOW_CONFIG,
-            },
-        },
-    };
-    op->callback = [this](osd_op_t *op)
-    {
-        osd_client_t & cl = clients[op->peer_fd];
-        std::string json_err;
-        json11::Json config;
-        bool err = false;
-        if (op->reply.hdr.retval < 0)
-        {
-            err = true;
-            printf("Failed to get config from OSD %lu (retval=%ld), disconnecting peer\n", cl.osd_num, op->reply.hdr.retval);
-        }
-        else
-        {
-            config = json11::Json::parse(std::string((char*)op->buf), json_err);
-            if (json_err != "")
-            {
-                err = true;
-                printf("Failed to get config from OSD %lu: bad JSON: %s, disconnecting peer\n", cl.osd_num, json_err.c_str());
-            }
-            else if (config["osd_num"].uint64_value() != cl.osd_num)
-            {
-                err = true;
-                printf("Connected to OSD %lu instead of OSD %lu, peer state is outdated, disconnecting peer\n", config["osd_num"].uint64_value(), cl.osd_num);
-                on_connect_peer(cl.osd_num, -1);
-            }
-        }
-        if (err)
-        {
-            stop_client(op->peer_fd);
-            delete op;
-            return;
-        }
-        osd_peer_fds[cl.osd_num] = cl.peer_fd;
-        on_connect_peer(cl.osd_num, cl.peer_fd);
-        delete op;
-    };
-    outbox_push(op);
-}
-
-void osd_messenger_t::cancel_osd_ops(osd_client_t & cl)
-{
-    for (auto p: cl.sent_ops)
-    {
-        cancel_op(p.second);
-    }
-    cl.sent_ops.clear();
-    for (auto op: cl.outbox)
-    {
-        cancel_op(op);
-    }
-    cl.outbox.clear();
-    if (cl.write_op)
-    {
-        cancel_op(cl.write_op);
-        cl.write_op = NULL;
-    }
-}
-
-void osd_messenger_t::cancel_op(osd_op_t *op)
-{
-    if (op->op_type == OSD_OP_OUT)
-    {
-        op->reply.hdr.magic = SECONDARY_OSD_REPLY_MAGIC;
-        op->reply.hdr.id = op->req.hdr.id;
-        op->reply.hdr.opcode = op->req.hdr.opcode;
-        op->reply.hdr.retval = -EPIPE;
-        // Copy lambda to be unaffected by `delete op`
-        std::function<void(osd_op_t*)>(op->callback)(op);
-    }
-    else
-    {
-        // This function is only called in stop_client(), so it's fine to destroy the operation
-        delete op;
-    }
-}
-
-void osd_messenger_t::stop_client(int peer_fd)
-{
-    assert(peer_fd != 0);
-    auto it = clients.find(peer_fd);
-    if (it == clients.end())
-    {
-        return;
-    }
-    uint64_t repeer_osd = 0;
-    osd_client_t cl = it->second;
-    if (cl.peer_state == PEER_CONNECTED)
-    {
-        if (cl.osd_num)
-        {
-            // Reload configuration from etcd when the connection is dropped
-            printf("[OSD %lu] Stopping client %d (OSD peer %lu)\n", osd_num, peer_fd, cl.osd_num);
-            repeer_osd = cl.osd_num;
-        }
-        else
-        {
-            printf("[OSD %lu] Stopping client %d (regular client)\n", osd_num, peer_fd);
-        }
-    }
-    clients.erase(it);
-    tfd->set_fd_handler(peer_fd, NULL);
-    if (cl.osd_num)
-    {
-        osd_peer_fds.erase(cl.osd_num);
-        // Cancel outbound operations
-        cancel_osd_ops(cl);
-    }
-    if (cl.read_op)
-    {
-        delete cl.read_op;
-        cl.read_op = NULL;
-    }
-    for (auto rit = read_ready_clients.begin(); rit != read_ready_clients.end(); rit++)
-    {
-        if (*rit == peer_fd)
-        {
-            read_ready_clients.erase(rit);
-            break;
-        }
-    }
-    for (auto wit = write_ready_clients.begin(); wit != write_ready_clients.end(); wit++)
-    {
-        if (*wit == peer_fd)
-        {
-            write_ready_clients.erase(wit);
-            break;
-        }
-    }
-    free(cl.in_buf);
-    close(peer_fd);
-    if (repeer_osd)
-    {
-        repeer_pgs(repeer_osd);
-    }
-}
-
-void osd_messenger_t::accept_connections(int listen_fd)
-{
-    // Accept new connections
-    sockaddr_in addr;
-    socklen_t peer_addr_size = sizeof(addr);
-    int peer_fd;
-    while ((peer_fd = accept(listen_fd, (sockaddr*)&addr, &peer_addr_size)) >= 0)
-    {
-        assert(peer_fd != 0);
-        char peer_str[256];
-        printf("[OSD %lu] new client %d: connection from %s port %d\n", this->osd_num, peer_fd,
-            inet_ntop(AF_INET, &addr.sin_addr, peer_str, 256), ntohs(addr.sin_port));
-        fcntl(peer_fd, F_SETFL, fcntl(peer_fd, F_GETFL, 0) | O_NONBLOCK);
-        int one = 1;
-        setsockopt(peer_fd, SOL_TCP, TCP_NODELAY, &one, sizeof(one));
-        clients[peer_fd] = {
-            .peer_addr = addr,
-            .peer_port = ntohs(addr.sin_port),
-            .peer_fd = peer_fd,
-            .peer_state = PEER_CONNECTED,
-            .in_buf = malloc(receive_buffer_size),
-        };
-        // Add FD to epoll
-        tfd->set_fd_handler(peer_fd, [this](int peer_fd, int epoll_events)
-        {
-            handle_peer_epoll(peer_fd, epoll_events);
-        });
-        // Try to accept next connection
-        peer_addr_size = sizeof(addr);
-    }
-    if (peer_fd == -1 && errno != EAGAIN)
-    {
-        throw std::runtime_error(std::string("accept: ") + strerror(errno));
-    }
-}

messenger.h

@@ -1,213 +0,0 @@
-#pragma once
-
-#include <sys/types.h>
-#include <stdint.h>
-#include <arpa/inet.h>
-#include <malloc.h>
-
-#include <set>
-#include <map>
-#include <deque>
-#include <vector>
-
-#include "json11/json11.hpp"
-#include "osd_ops.h"
-#include "timerfd_manager.h"
-#include "ringloop.h"
-
-#define OSD_OP_IN 0
-#define OSD_OP_OUT 1
-
-#define CL_READ_HDR 1
-#define CL_READ_DATA 2
-#define CL_READ_REPLY_DATA 3
-#define CL_WRITE_READY 1
-#define CL_WRITE_REPLY 2
-#define OSD_OP_INLINE_BUF_COUNT 16
-
-#define PEER_CONNECTING 1
-#define PEER_CONNECTED 2
-
-#define DEFAULT_PEER_CONNECT_INTERVAL 5
-#define DEFAULT_PEER_CONNECT_TIMEOUT 5
-
-struct osd_op_buf_list_t
-{
-    int count = 0, alloc = 0, sent = 0;
-    iovec *buf = NULL;
-    iovec inline_buf[OSD_OP_INLINE_BUF_COUNT];
-
-    ~osd_op_buf_list_t()
-    {
-        if (buf && buf != inline_buf)
-        {
-            free(buf);
-        }
-    }
-
-    inline iovec* get_iovec()
-    {
-        return (buf ? buf : inline_buf) + sent;
-    }
-
-    inline int get_size()
-    {
-        return count - sent;
-    }
-
-    inline void push_back(void *nbuf, size_t len)
-    {
-        if (count >= alloc)
-        {
-            if (!alloc)
-            {
-                alloc = OSD_OP_INLINE_BUF_COUNT;
-                buf = inline_buf;
-            }
-            else if (buf == inline_buf)
-            {
-                int old = alloc;
-                alloc = ((alloc/16)*16 + 1);
-                buf = (iovec*)malloc(sizeof(iovec) * alloc);
-                memcpy(buf, inline_buf, sizeof(iovec)*old);
-            }
-            else
-            {
-                alloc = ((alloc/16)*16 + 1);
-                buf = (iovec*)realloc(buf, sizeof(iovec) * alloc);
-            }
-        }
-        buf[count++] = { .iov_base = nbuf, .iov_len = len };
-    }
-};
-
-struct blockstore_op_t;
-
-struct osd_primary_op_data_t;
-
-struct osd_op_t
-{
-    timespec tv_begin;
-    uint64_t op_type = OSD_OP_IN;
-    int peer_fd;
-    osd_any_op_t req;
-    osd_any_reply_t reply;
-    blockstore_op_t *bs_op = NULL;
-    void *buf = NULL;
-    void *rmw_buf = NULL;
-    osd_primary_op_data_t* op_data = NULL;
-    std::function<void(osd_op_t*)> callback;
-
-    osd_op_buf_list_t send_list;
-
-    ~osd_op_t();
-};
-
-struct osd_client_t
-{
-    sockaddr_in peer_addr;
-    int peer_port;
-    int peer_fd;
-    int peer_state;
-    int connect_timeout_id = -1;
-    osd_num_t osd_num = 0;
-
-    void *in_buf = NULL;
-
-    // Read state
-    int read_ready = 0;
-    osd_op_t *read_op = NULL;
-    int read_reply_id = 0;
-    iovec read_iov;
-    msghdr read_msg;
-    void *read_buf = NULL;
-    int read_remaining = 0;
-    int read_state = 0;
-
-    // Incoming operations
-    std::vector<osd_op_t*> received_ops;
-
-    // Outbound operations
-    std::deque<osd_op_t*> outbox;
-    std::map<int, osd_op_t*> sent_ops;
-
-    // PGs dirtied by this client's primary-writes (FIXME to drop the connection)
-    std::set<pg_num_t> dirty_pgs;
-
-    // Write state
-    osd_op_t *write_op = NULL;
-    msghdr write_msg;
-    int write_state = 0;
-};
-
-struct osd_wanted_peer_t
-{
-    json11::Json address_list;
-    int port;
-    time_t last_connect_attempt;
-    bool connecting, address_changed;
-    int address_index;
-    std::string cur_addr;
-    int cur_port;
-};
-
-struct osd_op_stats_t
-{
-    uint64_t op_stat_sum[OSD_OP_MAX+1] = { 0 };
-    uint64_t op_stat_count[OSD_OP_MAX+1] = { 0 };
-    uint64_t op_stat_bytes[OSD_OP_MAX+1] = { 0 };
-    uint64_t subop_stat_sum[OSD_OP_MAX+1] = { 0 };
-    uint64_t subop_stat_count[OSD_OP_MAX+1] = { 0 };
-};
-
-struct osd_messenger_t
-{
-    timerfd_manager_t *tfd;
-    ring_loop_t *ringloop;
-
-    // osd_num_t is only for logging and asserts
-    osd_num_t osd_num;
-    int receive_buffer_size = 9000;
-    int peer_connect_interval = DEFAULT_PEER_CONNECT_INTERVAL;
-    int peer_connect_timeout = DEFAULT_PEER_CONNECT_TIMEOUT;
-    int log_level = 0;
-
-    std::map<osd_num_t, osd_wanted_peer_t> wanted_peers;
-    std::map<uint64_t, int> osd_peer_fds;
-    uint64_t next_subop_id = 1;
-
-    std::map<int, osd_client_t> clients;
-    std::vector<int> read_ready_clients;
-    std::vector<int> write_ready_clients;
-
-    // op statistics
-    osd_op_stats_t stats;
-
-public:
-    void connect_peer(uint64_t osd_num, json11::Json peer_state);
-    void stop_client(int peer_fd);
-    void outbox_push(osd_op_t *cur_op);
-    std::function<void(osd_op_t*)> exec_op;
-    std::function<void(osd_num_t)> repeer_pgs;
-    void handle_peer_epoll(int peer_fd, int epoll_events);
-    void read_requests();
-    void send_replies();
-    void accept_connections(int listen_fd);
-
-protected:
-    void try_connect_peer(uint64_t osd_num);
-    void try_connect_peer_addr(osd_num_t peer_osd, const char *peer_host, int peer_port);
-    void handle_connect_epoll(int peer_fd);
-    void on_connect_peer(osd_num_t peer_osd, int peer_fd);
-    void check_peer_config(osd_client_t & cl);
-    void cancel_osd_ops(osd_client_t & cl);
-    void cancel_op(osd_op_t *op);
-
-    bool try_send(osd_client_t & cl);
-    void handle_send(int result, int peer_fd);
-
-    bool handle_read(int result, int peer_fd);
-    void handle_finished_read(osd_client_t & cl);
-    void handle_op_hdr(osd_client_t *cl);
-    void handle_reply_hdr(osd_client_t *cl);
-};

mon/PGUtil.js

@@ -0,0 +1,104 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
+module.exports = {
+    scale_pg_count,
+};
+
+function add_pg_history(new_pg_history, new_pg, prev_pgs, prev_pg_history, old_pg)
+{
+    if (!new_pg_history[new_pg])
+    {
+        new_pg_history[new_pg] = {
+            osd_sets: {},
+            all_peers: {},
+            epoch: 0,
+        };
+    }
+    const nh = new_pg_history[new_pg], oh = prev_pg_history[old_pg];
+    nh.osd_sets[prev_pgs[old_pg].join(' ')] = prev_pgs[old_pg];
+    if (oh && oh.osd_sets && oh.osd_sets.length)
+    {
+        for (const pg of oh.osd_sets)
+        {
+            nh.osd_sets[pg.join(' ')] = pg;
+        }
+    }
+    if (oh && oh.all_peers && oh.all_peers.length)
+    {
+        for (const osd_num of oh.all_peers)
+        {
+            nh.all_peers[osd_num] = Number(osd_num);
+        }
+    }
+    if (oh && oh.epoch)
+    {
+        nh.epoch = nh.epoch < oh.epoch ? oh.epoch : nh.epoch;
+    }
+}
+
+function finish_pg_history(merged_history)
+{
+    merged_history.osd_sets = Object.values(merged_history.osd_sets);
+    merged_history.all_peers = Object.values(merged_history.all_peers);
+}
+
+function scale_pg_count(prev_pgs, prev_pg_history, new_pg_history, new_pg_count)
+{
+    const old_pg_count = prev_pgs.length;
+    // Add all possibly intersecting PGs to the history of new PGs
+    if (!(new_pg_count % old_pg_count))
+    {
+        // New PG count is a multiple of old PG count
+        for (let i = 0; i < new_pg_count; i++)
+        {
+            add_pg_history(new_pg_history, i, prev_pgs, prev_pg_history, i % old_pg_count);
+            finish_pg_history(new_pg_history[i]);
+        }
+    }
+    else if (!(old_pg_count % new_pg_count))
+    {
+        // Old PG count is a multiple of the new PG count
+        const mul = (old_pg_count / new_pg_count);
+        for (let i = 0; i < new_pg_count; i++)
+        {
+            for (let j = 0; j < mul; j++)
+            {
+                add_pg_history(new_pg_history, i, prev_pgs, prev_pg_history, i+j*new_pg_count);
+            }
+            finish_pg_history(new_pg_history[i]);
+        }
+    }
+    else
+    {
+        // Any PG may intersect with any PG after non-multiple PG count change
+        // So, merge ALL PGs history
+        let merged_history = {};
+        for (let i = 0; i < old_pg_count; i++)
+        {
+            add_pg_history(merged_history, 1, prev_pgs, prev_pg_history, i);
+        }
+        finish_pg_history(merged_history[1]);
+        for (let i = 0; i < new_pg_count; i++)
+        {
+            new_pg_history[i] = { ...merged_history[1] };
+        }
+    }
+    // Mark history keys for removed PGs as removed
+    for (let i = new_pg_count; i < old_pg_count; i++)
+    {
+        new_pg_history[i] = null;
+    }
+    // Just for the lp_solve optimizer - pick a "previous" PG for each "new" one
+    if (old_pg_count < new_pg_count)
+    {
+        for (let i = old_pg_count; i < new_pg_count; i++)
+        {
+            prev_pgs[i] = prev_pgs[i % old_pg_count];
+        }
+    }
+    else if (old_pg_count > new_pg_count)
+    {
+        prev_pgs.splice(new_pg_count, old_pg_count-new_pg_count);
+    }
+}

mon/afr.js

@@ -0,0 +1,89 @@
+// Functions to calculate Annualized Failure Rate of your cluster
+// if you know AFR of your drives, number of drives, expected rebalance time
+// and replication factor
+// License: VNPL-1.1 (see https://yourcmc.ru/git/vitalif/vitastor/src/branch/master/README.md for details) or AGPL-3.0
+// Author: Vitaliy Filippov, 2020+
+
+module.exports = {
+    cluster_afr_fullmesh,
+    failure_rate_fullmesh,
+    cluster_afr,
+    c_n_k,
+};
+
+/******** "FULL MESH": ASSUME EACH OSD COMMUNICATES WITH ALL OTHER OSDS ********/
+
+// Estimate AFR of the cluster
+// n - number of drives
+// afr - annualized failure rate of a single drive
+// l - expected rebalance time in days after a single drive failure
+// k - replication factor / number of drives that must fail at the same time for the cluster to fail
+function cluster_afr_fullmesh(n, afr, l, k)
+{
+    return 1 - (1 - afr * failure_rate_fullmesh(n-(k-1), afr*l/365, k-1)) ** (n-(k-1));
+}
+
+// Probability of at least <f> failures in a cluster with <n> drives with AFR=<a>
+function failure_rate_fullmesh(n, a, f)
+{
+    if (f <= 0)
+    {
+        return (1-a)**n;
+    }
+    let p = 1;
+    for (let i = 0; i < f; i++)
+    {
+        p -= c_n_k(n, i) * (1-a)**(n-i) * a**i;
+    }
+    return p;
+}
+
+/******** PGS: EACH OSD ONLY COMMUNICATES WITH <pgs> OTHER OSDs ********/
+
+// <n> hosts of <m> drives of <capacity> GB, each able to backfill at <speed> GB/s,
+// <k> replicas, <pgs> unique peer PGs per OSD (~50 for 100 PG-per-OSD in a big cluster)
+//
+// For each of n*m drives: P(drive fails in a year) * P(any of its peers fail in <l*365> next days).
+// More peers per OSD increase rebalance speed (more drives work together to resilver) if you
+// let them finish rebalance BEFORE replacing the failed drive (degraded_replacement=false).
+// At the same time, more peers per OSD increase probability of any of them to fail!
+// osd_rm=true means that failed OSDs' data is rebalanced over all other hosts,
+// not over the same host as it's in Ceph by default (dead OSDs are marked 'out').
+//
+// Probability of all except one drives in a replica group to fail is (AFR^(k-1)).
+// So with <x> PGs it becomes ~ (x * (AFR*L/365)^(k-1)). Interesting but reasonable consequence
+// is that, with k=2, total failure rate doesn't depend on number of peers per OSD,
+// because it gets increased linearly by increased number of peers to fail
+// and decreased linearly by reduced rebalance time.
+function cluster_afr({ n_hosts, n_drives, afr_drive, afr_host, capacity, speed, ec, ec_data, ec_parity, replicas, pgs = 1, osd_rm, degraded_replacement, down_out_interval = 600 })
+{
+    const pg_size = (ec ? ec_data+ec_parity : replicas);
+    pgs = Math.min(pgs, (n_hosts-1)*n_drives/(pg_size-1));
+    const host_pgs = Math.min(pgs*n_drives, (n_hosts-1)*n_drives/(pg_size-1));
+    const resilver_disk = n_drives == 1 || osd_rm ? pgs : (n_drives-1);
+    const disk_heal_time = (down_out_interval + capacity/(degraded_replacement ? 1 : resilver_disk)/speed)/86400/365;
+    const host_heal_time = (down_out_interval + n_drives*capacity/pgs/speed)/86400/365;
+    const disk_heal_fail = ((afr_drive+afr_host/n_drives)*disk_heal_time);
+    const host_heal_fail = ((afr_drive+afr_host/n_drives)*host_heal_time);
+    const disk_pg_fail = ec
+        ? failure_rate_fullmesh(ec_data+ec_parity-1, disk_heal_fail, ec_parity)
+        : disk_heal_fail**(replicas-1);
+    const host_pg_fail = ec
+        ? failure_rate_fullmesh(ec_data+ec_parity-1, host_heal_fail, ec_parity)
+        : host_heal_fail**(replicas-1);
+    return 1 - ((1 - afr_drive * (1-(1-disk_pg_fail)**pgs)) ** (n_hosts*n_drives))
+        * ((1 - afr_host * (1-(1-host_pg_fail)**host_pgs)) ** n_hosts);
+}
+
+/******** UTILITY ********/
+
+// Combination count
+function c_n_k(n, k)
+{
+    let r = 1;
+    for (let i = 0; i < k; i++)
+    {
+        r *= (n-i) / (i+1);
+    }
+    return r;
+}

mon/afr_test.js

@@ -0,0 +1,28 @@
+const { sprintf } = require('sprintf-js');
+const { cluster_afr } = require('./afr.js');
+
+print_cluster_afr({ n_hosts: 4, n_drives: 6, afr_drive: 0.03, afr_host: 0.05, capacity: 4000, speed: 0.1, replicas: 2 });
+print_cluster_afr({ n_hosts: 4, n_drives: 3, afr_drive: 0.03, afr_host: 0, capacity: 4000, speed: 0.1, replicas: 2 });
+print_cluster_afr({ n_hosts: 4, n_drives: 3, afr_drive: 0.03, afr_host: 0.05, capacity: 4000, speed: 0.1, replicas: 2 });
+print_cluster_afr({ n_hosts: 4, n_drives: 3, afr_drive: 0.03, afr_host: 0, capacity: 4000, speed: 0.1, ec: true, ec_data: 2, ec_parity: 1 });
+print_cluster_afr({ n_hosts: 4, n_drives: 3, afr_drive: 0.03, afr_host: 0.05, capacity: 4000, speed: 0.1, ec: true, ec_data: 2, ec_parity: 1 });
+print_cluster_afr({ n_hosts: 10, n_drives: 10, afr_drive: 0.1, afr_host: 0, capacity: 8000, speed: 0.02, replicas: 2 });
+print_cluster_afr({ n_hosts: 10, n_drives: 10, afr_drive: 0.1, afr_host: 0.05, capacity: 8000, speed: 0.02, replicas: 2 });
+print_cluster_afr({ n_hosts: 10, n_drives: 10, afr_drive: 0.1, afr_host: 0, capacity: 8000, speed: 0.02, replicas: 3 });
+print_cluster_afr({ n_hosts: 10, n_drives: 10, afr_drive: 0.1, afr_host: 0.05, capacity: 8000, speed: 0.02, replicas: 3 });
+print_cluster_afr({ n_hosts: 10, n_drives: 10, afr_drive: 0.1, afr_host: 0, capacity: 8000, speed: 0.02, replicas: 3, pgs: 100 });
+print_cluster_afr({ n_hosts: 10, n_drives: 10, afr_drive: 0.1, afr_host: 0.05, capacity: 8000, speed: 0.02, replicas: 3, pgs: 100 });
+print_cluster_afr({ n_hosts: 10, n_drives: 10, afr_drive: 0.1, afr_host: 0.05, capacity: 8000, speed: 0.02, replicas: 3, pgs: 100, degraded_replacement: 1 });
+
+function print_cluster_afr(config)
+{
+    console.log(
+        `${config.n_hosts} nodes with ${config.n_drives} ${sprintf("%.1f", config.capacity/1000)}TB drives`+
+        `, capable to backfill at ${sprintf("%.1f", config.speed*1000)} MB/s, drive AFR ${sprintf("%.1f", config.afr_drive*100)}%`+
+        (config.afr_host ? `, host AFR ${sprintf("%.1f", config.afr_host*100)}%` : '')+
+        (config.ec ? `, EC ${config.ec_data}+${config.ec_parity}` : `, ${config.replicas} replicas`)+
+        `, ${config.pgs||1} PG per OSD`+
+        (config.degraded_replacement ? `\n...and you don't let the rebalance finish before replacing drives` : '')
+    );
+    console.log('-> '+sprintf("%.7f%%", 100*cluster_afr(config))+'\n');
+}

mon/lp-optimizer.js

@@ -1,3 +1,6 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
 // Data distribution optimizer using linear programming (lp_solve)
 
 const child_process = require('child_process');
@@ -25,7 +28,7 @@ async function lp_solve(text)
     let vars = {};
     for (const line of stdout.split(/\n/))
     {
-        let m = /^(^Value of objective function: ([\d\.]+)|Actual values of the variables:)\s*$/.exec(line);
+        let m = /^(^Value of objective function: (-?[\d\.]+)|Actual values of the variables:)\s*$/.exec(line);
         if (m)
         {
             if (m[2])
@@ -47,34 +50,34 @@ async function lp_solve(text)
     return { score, vars };
 }
 
-async function optimize_initial(osd_tree, pg_count, max_combinations)
+async function optimize_initial({ osd_tree, pg_count, pg_size = 3, pg_minsize = 2, max_combinations = 10000, parity_space = 1 })
 {
-    max_combinations = max_combinations || 10000;
+    if (!pg_count || !osd_tree)
+    {
+        return null;
+    }
     const all_weights = Object.assign({}, ...Object.values(osd_tree));
     const total_weight = Object.values(all_weights).reduce((a, c) => Number(a) + Number(c), 0);
-    let all_pgs = all_combinations(osd_tree, null, true);
-    if (all_pgs.length > max_combinations)
-    {
-        const prob = max_combinations/all_pgs.length;
-        all_pgs = all_pgs.filter(pg => Math.random() < prob);
-    }
+    const all_pgs = Object.values(random_combinations(osd_tree, pg_size, max_combinations, parity_space > 1));
     const pg_per_osd = {};
     for (const pg of all_pgs)
     {
-        for (const osd of pg)
+        for (let i = 0; i < pg.length; i++)
         {
+            const osd = pg[i];
             pg_per_osd[osd] = pg_per_osd[osd] || [];
-            pg_per_osd[osd].push("pg_"+pg.join("_"));
+            pg_per_osd[osd].push((i >= pg_minsize ? parity_space+'*' : '')+"pg_"+pg.join("_"));
         }
     }
-    const pg_size = Math.min(Object.keys(osd_tree).length, 3);
+    const pg_effsize = Math.min(pg_minsize, Object.keys(osd_tree).length)
+        + Math.max(0, Math.min(pg_size, Object.keys(osd_tree).length) - pg_minsize) * parity_space;
     let lp = '';
     lp += "max: "+all_pgs.map(pg => 'pg_'+pg.join('_')).join(' + ')+";\n";
     for (const osd in pg_per_osd)
     {
         if (osd !== NO_OSD)
         {
-            let osd_pg_count = all_weights[osd]/total_weight*pg_size*pg_count;
+            let osd_pg_count = all_weights[osd]/total_weight*pg_effsize*pg_count;
             lp += pg_per_osd[osd].join(' + ')+' <= '+osd_pg_count+';\n';
         }
     }
@@ -86,11 +89,30 @@ async function optimize_initial(osd_tree, pg_count, max_combinations)
     const lp_result = await lp_solve(lp);
     if (!lp_result)
     {
+        console.log(lp);
         throw new Error('Problem is infeasible or unbounded - is it a bug?');
     }
     const int_pgs = make_int_pgs(lp_result.vars, pg_count);
-    const eff = pg_list_space_efficiency(int_pgs, all_weights);
-    return { score: lp_result.score, weights: lp_result.vars, int_pgs, space: eff*pg_size, total_space: total_weight };
+    const eff = pg_list_space_efficiency(int_pgs, all_weights, pg_minsize, parity_space);
+    const res = {
+        score: lp_result.score,
+        weights: lp_result.vars,
+        int_pgs,
+        space: eff * pg_effsize,
+        total_space: total_weight,
+    };
+    return res;
+}
+
+function shuffle(array)
+{
+    for (let i = array.length - 1, j, x; i > 0; i--)
+    {
+        j = Math.floor(Math.random() * (i + 1));
+        x = array[i];
+        array[i] = array[j];
+        array[j] = x;
+    }
 }
 
 function make_int_pgs(weights, pg_count)
@@ -109,14 +131,122 @@ function make_int_pgs(weights, pg_count)
         weight_left -= weights[pg_name];
         pg_left -= n;
     }
+    shuffle(int_pgs);
     return int_pgs;
 }
 
-// Try to minimize data movement
-async function optimize_change(prev_int_pgs, osd_tree, max_combinations)
+function calc_intersect_weights(pg_size, pg_count, prev_weights, all_pgs)
 {
-    max_combinations = max_combinations || 10000;
-    const pg_size = Math.min(Object.keys(osd_tree).length, 3);
+    const move_weights = {};
+    if ((1 << pg_size) < pg_count)
+    {
+        const intersect = {};
+        for (const pg_name in prev_weights)
+        {
+            const pg = pg_name.substr(3).split(/_/);
+            for (let omit = 1; omit < (1 << pg_size); omit++)
+            {
+                let pg_omit = [ ...pg ];
+                let intersect_count = pg_size;
+                for (let i = 0; i < pg_size; i++)
+                {
+                    if (omit & (1 << i))
+                    {
+                        pg_omit[i] = '';
+                        intersect_count--;
+                    }
+                }
+                pg_omit = pg_omit.join(':');
+                intersect[pg_omit] = Math.max(intersect[pg_omit] || 0, intersect_count);
+            }
+        }
+        for (const pg of all_pgs)
+        {
+            let max_int = 0;
+            for (let omit = 1; omit < (1 << pg_size); omit++)
+            {
+                let pg_omit = [ ...pg ];
+                for (let i = 0; i < pg_size; i++)
+                {
+                    if (omit & (1 << i))
+                    {
+                        pg_omit[i] = '';
+                    }
+                }
+                pg_omit = pg_omit.join(':');
+                max_int = Math.max(max_int, intersect[pg_omit] || 0);
+            }
+            move_weights['pg_'+pg.join('_')] = pg_size-max_int;
+        }
+    }
+    else
+    {
+        const prev_pg_hashed = Object.keys(prev_weights).map(pg_name => pg_name.substr(3).split(/_/).reduce((a, c) => { a[c] = 1; return a; }, {}));
+        for (const pg of all_pgs)
+        {
+            if (!prev_weights['pg_'+pg.join('_')])
+            {
+                let max_int = 0;
+                for (const prev_hash in prev_pg_hashed)
+                {
+                    const intersect_count = pg.reduce((a, osd) => a + (prev_hash[osd] ? 1 : 0), 0);
+                    if (max_int < intersect_count)
+                    {
+                        max_int = intersect_count;
+                        if (max_int >= pg_size)
+                        {
+                            break;
+                        }
+                    }
+                }
+                move_weights['pg_'+pg.join('_')] = pg_size-max_int;
+            }
+        }
+    }
+    return move_weights;
+}
+
+function add_valid_previous(osd_tree, prev_weights, all_pgs)
+{
+    // Add previous combinations that are still valid
+    const hosts = Object.keys(osd_tree).sort();
+    const host_per_osd = {};
+    for (const host in osd_tree)
+    {
+        for (const osd in osd_tree[host])
+        {
+            host_per_osd[osd] = host;
+        }
+    }
+    skip_pg: for (const pg_name in prev_weights)
+    {
+        const seen_hosts = {};
+        const pg = pg_name.substr(3).split(/_/);
+        for (const osd of pg)
+        {
+            if (!host_per_osd[osd] || seen_hosts[host_per_osd[osd]])
+            {
+                continue skip_pg;
+            }
+            seen_hosts[host_per_osd[osd]] = true;
+        }
+        if (!all_pgs[pg_name])
+        {
+            all_pgs[pg_name] = pg;
+        }
+    }
+}
+
+// Try to minimize data movement
+async function optimize_change({ prev_pgs: prev_int_pgs, osd_tree, pg_size = 3, pg_minsize = 2, max_combinations = 10000, parity_space = 1 })
+{
+    if (!osd_tree)
+    {
+        return null;
+    }
+    // FIXME: use parity_chunks with parity_space instead of pg_minsize
+    const pg_effsize = Math.min(pg_minsize, Object.keys(osd_tree).length)
+        + Math.max(0, Math.min(pg_size, Object.keys(osd_tree).length) - pg_minsize) * parity_space;
     const pg_count = prev_int_pgs.length;
     const prev_weights = {};
     const prev_pg_per_osd = {};
@@ -124,70 +254,55 @@ async function optimize_change(prev_int_pgs, osd_tree, max_combinations)
     {
         const pg_name = 'pg_'+pg.join('_');
         prev_weights[pg_name] = (prev_weights[pg_name]||0) + 1;
-        for (const osd of pg)
+        for (let i = 0; i < pg.length; i++)
         {
+            const osd = pg[i];
             prev_pg_per_osd[osd] = prev_pg_per_osd[osd] || [];
-            prev_pg_per_osd[osd].push(pg_name);
+            prev_pg_per_osd[osd].push([ pg_name, (i >= pg_minsize ? parity_space : 1) ]);
         }
     }
     // Get all combinations
-    let all_pgs = all_combinations(osd_tree, null, true);
-    if (all_pgs.length > max_combinations)
-    {
-        const intersecting = all_pgs.filter(pg => prev_weights['pg_'+pg.join('_')]);
-        if (intersecting.length > max_combinations)
-        {
-            const prob = max_combinations/intersecting.length;
-            all_pgs = intersecting.filter(pg => Math.random() < prob);
-        }
-        else
-        {
-            const prob = (max_combinations-intersecting.length)/all_pgs.length;
-            all_pgs = all_pgs.filter(pg => Math.random() < prob || prev_weights['pg_'+pg.join('_')]);
-        }
-    }
+    let all_pgs = random_combinations(osd_tree, pg_size, max_combinations, parity_space > 1);
+    add_valid_previous(osd_tree, prev_weights, all_pgs);
+    all_pgs = Object.values(all_pgs);
     const pg_per_osd = {};
     for (const pg of all_pgs)
     {
         const pg_name = 'pg_'+pg.join('_');
-        for (const osd of pg)
+        for (let i = 0; i < pg.length; i++)
         {
+            const osd = pg[i];
             pg_per_osd[osd] = pg_per_osd[osd] || [];
-            pg_per_osd[osd].push(pg_name);
+            pg_per_osd[osd].push([ pg_name, (i >= pg_minsize ? parity_space : 1) ]);
         }
     }
     // Penalize PGs based on their similarity to old PGs
-    const intersect = {};
-    for (const pg_name in prev_weights)
-    {
-        const pg = pg_name.substr(3).split(/_/);
-        intersect[pg[0]+'::'] = intersect[':'+pg[1]+':'] = intersect['::'+pg[2]] = 2;
-        intersect[pg[0]+'::'+pg[2]] = intersect[':'+pg[1]+':'+pg[2]] = intersect[pg[0]+':'+pg[1]+':'] = 1;
-    }
-    const move_weights = {};
-    for (const pg of all_pgs)
-    {
-        move_weights['pg_'+pg.join('_')] =
-            intersect[pg[0]+'::'+pg[2]] || intersect[':'+pg[1]+':'+pg[2]] || intersect[pg[0]+':'+pg[1]+':'] ||
-            intersect[pg[0]+'::'] || intersect[':'+pg[1]+':'] || intersect['::'+pg[2]] ||
-            3;
-    }
+    const move_weights = calc_intersect_weights(pg_size, pg_count, prev_weights, all_pgs);
     // Calculate total weight - old PG weights
     const all_pg_names = all_pgs.map(pg => 'pg_'+pg.join('_'));
+    const all_pgs_hash = all_pg_names.reduce((a, c) => { a[c] = true; return a; }, {});
     const all_weights = Object.assign({}, ...Object.values(osd_tree));
     const total_weight = Object.values(all_weights).reduce((a, c) => Number(a) + Number(c), 0);
     // Generate the LP problem
     let lp = '';
     lp += 'max: '+all_pg_names.map(pg_name => (
-        prev_weights[pg_name] ? `${4-move_weights[pg_name]}*add_${pg_name} - 4*del_${pg_name}` : `${4-move_weights[pg_name]}*${pg_name}`
+        prev_weights[pg_name] ? `${pg_size+1}*add_${pg_name} - ${pg_size+1}*del_${pg_name}` : `${pg_size+1-move_weights[pg_name]}*${pg_name}`
     )).join(' + ')+';\n';
+    lp += all_pg_names
+        .map(pg_name => (prev_weights[pg_name] ? `add_${pg_name} - del_${pg_name}` : `${pg_name}`))
+        .join(' + ')+' = '+(pg_count
+            - Object.keys(prev_weights).reduce((a, old_pg_name) => (a + (all_pgs_hash[old_pg_name] ? prev_weights[old_pg_name] : 0)), 0)
+        )+';\n';
     for (const osd in pg_per_osd)
     {
         if (osd !== NO_OSD)
         {
-            const osd_sum = (pg_per_osd[osd]||[]).map(pg_name => prev_weights[pg_name] ? `add_${pg_name} - del_${pg_name}` : pg_name).join(' + ');
-            const rm_osd_pg_count = (prev_pg_per_osd[osd]||[]).filter(old_pg_name => move_weights[old_pg_name]).length;
-            let osd_pg_count = all_weights[osd]*3/total_weight*pg_count - rm_osd_pg_count;
+            const osd_sum = (pg_per_osd[osd]||[]).map(([ pg_name, space ]) => (
+                prev_weights[pg_name] ? `${space} * add_${pg_name} - ${space} * del_${pg_name}` : `${space} * ${pg_name}`
+            )).join(' + ');
+            const rm_osd_pg_count = (prev_pg_per_osd[osd]||[])
+                .reduce((a, [ old_pg_name, space ]) => (a + (all_pgs_hash[old_pg_name] ? space : 0)), 0);
+            const osd_pg_count = all_weights[osd]*pg_effsize/total_weight*pg_count - rm_osd_pg_count;
             lp += osd_sum + ' <= ' + osd_pg_count + ';\n';
         }
     }
@@ -221,7 +336,7 @@ async function optimize_change(prev_int_pgs, osd_tree, max_combinations)
     const weights = { ...prev_weights };
     for (const k in prev_weights)
     {
-        if (!move_weights[k])
+        if (!all_pgs_hash[k])
         {
             delete weights[k];
         }
@@ -236,7 +351,7 @@ async function optimize_change(prev_int_pgs, osd_tree, max_combinations)
         {
             weights[k.substr(4)] = (weights[k.substr(4)] || 0) - Number(lp_result.vars[k]);
         }
-        else
+        else if (k.substr(0, 3) === 'pg_')
         {
             weights[k] = Number(lp_result.vars[k]);
         }
@@ -258,7 +373,7 @@ async function optimize_change(prev_int_pgs, osd_tree, max_combinations)
         {
             differs++;
         }
-        for (let j = 0; j < 3; j++)
+        for (let j = 0; j < pg_size; j++)
         {
             if (new_pgs[i][j] != prev_int_pgs[i][j])
             {
@@ -273,7 +388,7 @@ async function optimize_change(prev_int_pgs, osd_tree, max_combinations)
         int_pgs: new_pgs,
         differs,
         osd_differs,
-        space: pg_size * pg_list_space_efficiency(new_pgs, all_weights),
+        space: pg_effsize * pg_list_space_efficiency(new_pgs, all_weights, pg_minsize, parity_space),
         total_space: total_weight,
     };
 }
@@ -391,64 +506,155 @@ function extract_osds(osd_tree, levels, osd_level, osds = {})
     return osds;
 }
 
-// FIXME: support different pg_sizes, not just 3
-// osd_tree = { failure_domain1: { osd1: size1, ... }, ... }
-function all_combinations(osd_tree, count, ordered)
+// ordered = don't treat (x,y) and (y,x) as equal
+function random_combinations(osd_tree, pg_size, count, ordered)
 {
+    let seed = 0x5f020e43;
+    let rng = () =>
+    {
+        seed ^= seed << 13;
+        seed ^= seed >> 17;
+        seed ^= seed << 5;
+        return seed + 2147483648;
+    };
     const hosts = Object.keys(osd_tree).sort();
     const osds = Object.keys(osd_tree).reduce((a, c) => { a[c] = Object.keys(osd_tree[c]).sort(); return a; }, {});
-    while (hosts.length < 3)
+    const r = {};
+    // Generate random combinations including each OSD at least once
+    for (let h = 0; h < hosts.length; h++)
     {
-        osds[NO_OSD] = [ NO_OSD ];
-        hosts.push(NO_OSD);
-    }
-    let host_idx = [ 0, 1, 2 ];
-    let osd_idx = [ 0, 0, 0 ];
-    const r = [];
-    while (!count || count < 0 || r.length < count)
-    {
-        let inc;
-        if (host_idx[2] != host_idx[1] && host_idx[2] != host_idx[0] && host_idx[1] != host_idx[0])
+        for (let o = 0; o < osds[hosts[h]].length; o++)
         {
-            r.push(host_idx.map((hi, i) => osds[hosts[hi]][osd_idx[i]]));
-            inc = 2;
-            while (inc >= 0)
+            const pg = [ osds[hosts[h]][o] ];
+            const cur_hosts = [ ...hosts ];
+            cur_hosts.splice(h, 1);
+            for (let i = 1; i < pg_size && i < hosts.length; i++)
             {
-                osd_idx[inc]++;
-                if (osd_idx[inc] >= osds[hosts[host_idx[inc]]].length)
+                const next_host = rng() % cur_hosts.length;
+                const next_osd = rng() % osds[cur_hosts[next_host]].length;
+                pg.push(osds[cur_hosts[next_host]][next_osd]);
+                cur_hosts.splice(next_host, 1);
+            }
+            const cyclic_pgs = [ pg ];
+            if (ordered)
+            {
+                for (let i = 1; i < pg.size; i++)
                 {
-                    osd_idx[inc] = 0;
-                    inc--;
+                    cyclic_pgs.push([ ...pg.slice(i), ...pg.slice(0, i) ]);
                 }
-                else
+            }
+            for (const pg of cyclic_pgs)
+            {
+                while (pg.length < pg_size)
                 {
-                    break;
+                    pg.push(NO_OSD);
                 }
+                r['pg_'+pg.join('_')] = pg;
+            }
+        }
+    }
+    // Generate purely random combinations
+    while (count > 0)
+    {
+        let host_idx = [];
+        const cur_hosts = [ ...hosts.map((h, i) => i) ];
+        const max_hosts = pg_size < hosts.length ? pg_size : hosts.length;
+        if (ordered)
+        {
+            for (let i = 0; i < max_hosts; i++)
+            {
+                const r = rng() % cur_hosts.length;
+                host_idx[i] = cur_hosts[r];
+                cur_hosts.splice(r, 1);
             }
         }
         else
         {
-            inc = -1;
+            for (let i = 0; i < max_hosts; i++)
+            {
+                const r = rng() % (cur_hosts.length - (max_hosts - i - 1));
+                host_idx[i] = cur_hosts[r];
+                cur_hosts.splice(0, r+1);
+            }
+        }
+        let pg = host_idx.map(h => osds[hosts[h]][rng() % osds[hosts[h]].length]);
+        while (pg.length < pg_size)
+        {
+            pg.push(NO_OSD);
+        }
+        r['pg_'+pg.join('_')] = pg;
+        count--;
+    }
+    return r;
+}
+
+// Super-stupid algorithm. Given the current OSD tree, generate all possible OSD combinations
+// osd_tree = { failure_domain1: { osd1: size1, ... }, ... }
+// ordered = return combinations without duplicates having different order
+function all_combinations(osd_tree, pg_size, ordered, count)
+{
+    const hosts = Object.keys(osd_tree).sort();
+    const osds = Object.keys(osd_tree).reduce((a, c) => { a[c] = Object.keys(osd_tree[c]).sort(); return a; }, {});
+    while (hosts.length < pg_size)
+    {
+        osds[NO_OSD] = [ NO_OSD ];
+        hosts.push(NO_OSD);
+    }
+    let host_idx = [];
+    let osd_idx = [];
+    for (let i = 0; i < pg_size; i++)
+    {
+        host_idx.push(i);
+        osd_idx.push(0);
+    }
+    const r = [];
+    while (!count || count < 0 || r.length < count)
+    {
+        r.push(host_idx.map((hi, i) => osds[hosts[hi]][osd_idx[i]]));
+        let inc = pg_size-1;
+        while (inc >= 0)
+        {
+            osd_idx[inc]++;
+            if (osd_idx[inc] >= osds[hosts[host_idx[inc]]].length)
+            {
+                osd_idx[inc] = 0;
+                inc--;
+            }
+            else
+            {
+                break;
+            }
         }
         if (inc < 0)
         {
-            // no osds left in current host combination, select the next one
-            osd_idx = [ 0, 0, 0 ];
-            host_idx[2]++;
-            if (host_idx[2] >= hosts.length)
+            // no osds left in the current host combination, select the next one
+            inc = pg_size-1;
+            same_again: while (inc >= 0)
             {
-                host_idx[1]++;
-                host_idx[2] = ordered ? host_idx[1]+1 : 0;
-                if ((ordered ? host_idx[2] : host_idx[1]) >= hosts.length)
+                host_idx[inc]++;
+                for (let prev_host = 0; prev_host < inc; prev_host++)
                 {
-                    host_idx[0]++;
-                    host_idx[1] = ordered ? host_idx[0]+1 : 0;
-                    host_idx[2] = ordered ? host_idx[1]+1 : 0;
-                    if ((ordered ? host_idx[2] : host_idx[0]) >= hosts.length)
+                    if (host_idx[prev_host] == host_idx[inc])
                     {
-                        break;
+                        continue same_again;
                     }
                 }
+                if (host_idx[inc] < (ordered ? hosts.length-(pg_size-1-inc) : hosts.length))
+                {
+                    while ((++inc) < pg_size)
+                    {
+                        host_idx[inc] = (ordered ? host_idx[inc-1]+1 : 0);
+                    }
+                    break;
+                }
+                else
+                {
+                    inc--;
+                }
+            }
+            if (inc < 0)
+            {
+                break;
             }
         }
     }
@@ -468,14 +674,15 @@ function pg_weights_space_efficiency(weights, pg_count, osd_sizes)
     return pg_per_osd_space_efficiency(per_osd, pg_count, osd_sizes);
 }
 
-function pg_list_space_efficiency(pgs, osd_sizes)
+function pg_list_space_efficiency(pgs, osd_sizes, pg_minsize, parity_space)
 {
     const per_osd = {};
     for (const pg of pgs)
     {
-        for (const osd of pg)
+        for (let i = 0; i < pg.length; i++)
         {
-            per_osd[osd] = (per_osd[osd]||0) + 1;
+            const osd = pg[i];
+            per_osd[osd] = (per_osd[osd]||0) + (i >= pg_minsize ? (parity_space||1) : 1);
         }
     }
     return pg_per_osd_space_efficiency(per_osd, pgs.length, osd_sizes);
@@ -517,5 +724,6 @@ module.exports = {
     lp_solve,
     make_int_pgs,
     align_pgs,
+    random_combinations,
     all_combinations,
 };

mon/make-osd.sh

@@ -0,0 +1,75 @@
+#!/bin/bash
+# Very simple systemd unit generator for vitastor-osd services
+# Not the final solution yet, mostly for tests
+# Copyright (c) Vitaliy Filippov, 2019+
+# License: MIT
+
+# USAGE: ./make-osd.sh /dev/disk/by-partuuid/xxx [ /dev/disk/by-partuuid/yyy]...
+
+IP_SUBSTR="10.200.1."
+ETCD_HOSTS="etcd0=http://10.200.1.10:2380,etcd1=http://10.200.1.11:2380,etcd2=http://10.200.1.12:2380"
+
+set -e -x
+
+IP=`ip -json a s | jq -r '.[].addr_info[] | select(.local | startswith("'$IP_SUBSTR'")) | .local'`
+[ "$IP" != "" ] || exit 1
+ETCD_MON=$(echo $ETCD_HOSTS | perl -pe 's/:2380/:2379/g; s/etcd\d*=//g;')
+D=`dirname $0`
+
+# Create OSDs on all passed devices
+OSD_NUM=1
+for DEV in $*; do
+
+# Ugly :) -> node.js rework pending
+while true; do
+    ST=$(etcdctl --endpoints="$ETCD_MON" get --print-value-only /vitastor/osd/stats/$OSD_NUM)
+    if [ "$ST" = "" ]; then
+        break
+    fi
+    OSD_NUM=$((OSD_NUM+1))
+done
+etcdctl --endpoints="$ETCD_MON" put /vitastor/osd/stats/$OSD_NUM '{}'
+
+echo Creating OSD $OSD_NUM on $DEV
+
+OPT=`node $D/simple-offsets.js --device $DEV --format options | tr '\n' ' '`
+META=`echo $OPT | grep -Po '(?<=data_offset )\d+'`
+dd if=/dev/zero of=$DEV bs=1048576 count=$(((META+1048575)/1048576)) oflag=direct
+
+cat >/etc/systemd/system/vitastor-osd$OSD_NUM.service <<EOF
+[Unit]
+Description=Vitastor object storage daemon osd.$OSD_NUM
+After=network-online.target local-fs.target time-sync.target
+Wants=network-online.target local-fs.target time-sync.target
+PartOf=vitastor.target
+
+[Service]
+LimitNOFILE=1048576
+LimitNPROC=1048576
+LimitMEMLOCK=infinity
+ExecStart=/usr/bin/vitastor-osd \\
+    --etcd_address $IP:2379/v3 \\
+    --bind_address $IP \\
+    --osd_num $OSD_NUM \\
+    --disable_data_fsync 1 \\
+    --immediate_commit all \\
+    --disk_alignment 4096 --journal_block_size 4096 --meta_block_size 4096 \\
+    --journal_no_same_sector_overwrites true \\
+    --journal_sector_buffer_count 1024 \\
+    $OPT
+WorkingDirectory=/
+ExecStartPre=+chown vitastor:vitastor $DEV
+User=vitastor
+PrivateTmp=false
+TasksMax=infinity
+Restart=always
+StartLimitInterval=0
+RestartSec=10
+
+[Install]
+WantedBy=vitastor.target
+EOF
+
+systemctl enable vitastor-osd$OSD_NUM
+
+done

mon/make-units.sh

@@ -0,0 +1,86 @@
+#!/bin/bash
+# Very simple systemd unit generator for etcd & vitastor-mon services
+# Not the final solution yet, mostly for tests
+# Copyright (c) Vitaliy Filippov, 2019+
+# License: MIT
+
+# USAGE: ./make-units.sh
+
+IP_SUBSTR="10.200.1."
+ETCD_HOSTS="etcd0=http://10.200.1.10:2380,etcd1=http://10.200.1.11:2380,etcd2=http://10.200.1.12:2380"
+
+# determine IP
+IP=`ip -json a s | jq -r '.[].addr_info[] | select(.local | startswith("'$IP_SUBSTR'")) | .local'`
+[ "$IP" != "" ] || exit 1
+ETCD_NUM=${ETCD_HOSTS/$IP*/}
+[ "$ETCD_NUM" != "$ETCD_HOSTS" ] || exit 1
+ETCD_NUM=$(echo $ETCD_NUM | tr -d -c , | wc -c)
+
+# etcd
+useradd etcd
+
+mkdir -p /var/lib/etcd$ETCD_NUM.etcd
+cat >/etc/systemd/system/etcd.service <<EOF
+[Unit]
+Description=etcd for vitastor
+After=network-online.target local-fs.target time-sync.target
+Wants=network-online.target local-fs.target time-sync.target
+
+[Service]
+Restart=always
+ExecStart=/usr/local/bin/etcd -name etcd$ETCD_NUM --data-dir /var/lib/etcd$ETCD_NUM.etcd \\
+    --advertise-client-urls http://$IP:2379 --listen-client-urls http://$IP:2379 \\
+    --initial-advertise-peer-urls http://$IP:2380 --listen-peer-urls http://$IP:2380 \\
+    --initial-cluster-token vitastor-etcd-1 --initial-cluster $ETCD_HOSTS \\
+    --initial-cluster-state new --max-txn-ops=100000 --max-request-bytes=104857600 \\
+    --auto-compaction-retention=10 --auto-compaction-mode=revision
+WorkingDirectory=/var/lib/etcd$ETCD_NUM.etcd
+ExecStartPre=+chown -R etcd /var/lib/etcd$ETCD_NUM.etcd
+User=etcd
+PrivateTmp=false
+TasksMax=infinity
+Restart=always
+StartLimitInterval=0
+RestartSec=10
+
+[Install]
+WantedBy=local.target
+EOF
+
+systemctl daemon-reload
+systemctl enable etcd
+systemctl start etcd
+
+useradd vitastor
+chmod 755 /root
+
+# Vitastor target
+cat >/etc/systemd/system/vitastor.target <<EOF
+[Unit]
+Description=vitastor target
+[Install]
+WantedBy=multi-user.target
+EOF
+
+# Monitor unit
+ETCD_MON=$(echo $ETCD_HOSTS | perl -pe 's/:2380/:2379/g; s/etcd\d*=//g;')
+cat >/etc/systemd/system/vitastor-mon.service <<EOF
+[Unit]
+Description=Vitastor monitor
+After=network-online.target local-fs.target time-sync.target
+Wants=network-online.target local-fs.target time-sync.target
+
+[Service]
+Restart=always
+ExecStart=node /usr/lib/vitastor/mon/mon-main.js --etcd_url '$ETCD_MON' --etcd_prefix '/vitastor' --etcd_start_timeout 5
+WorkingDirectory=/
+User=vitastor
+PrivateTmp=false
+TasksMax=infinity
+Restart=always
+StartLimitInterval=0
+RestartSec=10
+
+[Install]
+WantedBy=vitastor.target
+EOF

mon/merge.js

@@ -0,0 +1,23 @@
+const fsp = require('fs').promises;
+
+async function merge(file1, file2, out)
+{
+    if (!out)
+    {
+        console.error('USAGE: nodejs merge.js layer1 layer2 output');
+        process.exit();
+    }
+    const layer1 = await fsp.readFile(file1);
+    const layer2 = await fsp.readFile(file2);
+    const zero = Buffer.alloc(4096);
+    for (let i = 0; i < layer2.length; i += 4096)
+    {
+        if (zero.compare(layer2, i, i+4096) != 0)
+        {
+            layer2.copy(layer1, i, i, i+4096);
+        }
+    }
+    await fsp.writeFile(out, layer1);
+}
+
+merge(process.argv[2], process.argv[3], process.argv[4]);

mon/mon-main.js

@@ -1,5 +1,8 @@
 #!/usr/bin/node
 
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
 const Mon = require('./mon.js');
 
 const options = {};
@@ -15,8 +18,8 @@ for (let i = 2; i < process.argv.length; i++)
 
 if (!options.etcd_url)
 {
-    console.error('USAGE: '+process.argv[0]+' '+process.argv[1]+' --etcd_url "http://127.0.0.1:2379,..." --etcd_prefix "/rage" --etcd_start_timeout 5');
+    console.error('USAGE: '+process.argv[0]+' '+process.argv[1]+' --etcd_url "http://127.0.0.1:2379,..." --etcd_prefix "/vitastor" --etcd_start_timeout 5 [--verbose 1]');
     process.exit();
 }
 
-new Mon(options).start();
+new Mon(options).start().catch(e => { console.error(e); process.exit(); });

mon/package.json

@@ -1,14 +1,15 @@
 {
-  "name": "rage-mon",
+  "name": "vitastor-mon",
   "version": "1.0.0",
-  "description": "RAGE storage monitor service",
-  "main": "mon.js",
+  "description": "Vitastor SDS monitor service",
+  "main": "mon-main.js",
   "scripts": {
     "test": "echo \"Error: no test specified\" && exit 1"
   },
   "author": "Vitaliy Filippov",
   "license": "UNLICENSED",
   "dependencies": {
+    "sprintf-js": "^1.1.2",
     "ws": "^7.2.5"
   }
 }

mon/simple-offsets.js

@@ -0,0 +1,96 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: MIT
+
+// Simple tool to calculate journal and metadata offsets for a single device
+// Will be replaced by smarter tools in the future
+
+const fs = require('fs').promises;
+const child_process = require('child_process');
+
+async function run()
+{
+    const options = {
+        object_size: 128*1024,
+        bitmap_granularity: 4096,
+        journal_size: 16*1024*1024,
+        device_block_size: 4096,
+        journal_offset: 0,
+        device_size: 0,
+        format: 'text',
+    };
+    for (let i = 2; i < process.argv.length; i++)
+    {
+        if (process.argv[i].substr(0, 2) == '--')
+        {
+            options[process.argv[i].substr(2)] = process.argv[i+1];
+            i++;
+        }
+    }
+    if (!options.device)
+    {
+        process.stderr.write('USAGE: nodejs '+process.argv[1]+' --device /dev/sdXXX\n');
+        process.exit(1);
+    }
+    options.device_size = Number(options.device_size);
+    let device_size = options.device_size;
+    if (!device_size)
+    {
+        const st = await fs.stat(options.device);
+        options.device_block_size = st.blksize;
+        if (st.isBlockDevice())
+            device_size = Number(await system("/sbin/blockdev --getsize64 "+options.device))
+        else
+            device_size = st.size;
+    }
+    if (!device_size)
+    {
+        process.stderr.write('Failed to get device size\n');
+        process.exit(1);
+    }
+    options.journal_offset = Math.ceil(options.journal_offset/options.device_block_size)*options.device_block_size;
+    const meta_offset = options.journal_offset + Math.ceil(options.journal_size/options.device_block_size)*options.device_block_size;
+    const entries_per_block = Math.floor(options.device_block_size / (24 + 2*options.object_size/options.bitmap_granularity/8));
+    const object_count = Math.floor((device_size-meta_offset)/options.object_size);
+    const meta_size = Math.ceil(1 + object_count / entries_per_block) * options.device_block_size;
+    const data_offset = meta_offset + meta_size;
+    const meta_size_fmt = (meta_size > 1024*1024*1024 ? Math.round(meta_size/1024/1024/1024*100)/100+" GB"
+        : Math.round(meta_size/1024/1024*100)/100+" MB");
+    if (options.format == 'text' || options.format == 'options')
+    {
+        if (options.format == 'text')
+        {
+            process.stderr.write(
+                `Metadata size: ${meta_size_fmt}\n`+
+                `Options for the OSD:\n`
+            );
+        }
+        process.stdout.write(
+            (options.device_block_size != 4096 ?
+                `    --meta_block_size ${options.device}\n`+
+                `    --journal_block-size ${options.device}\n` : '')+
+            `    --data_device ${options.device}\n`+
+            `    --journal_offset ${options.journal_offset}\n`+
+            `    --meta_offset ${meta_offset}\n`+
+            `    --data_offset ${data_offset}\n`+
+            (options.device_size ? `    --data_size ${device_size-data_offset}\n` : '')
+        );
+    }
+    else if (options.format == 'env')
+    {
+        process.stdout.write(
+            `journal_offset=${options.journal_offset}\n`+
+            `meta_offset=${meta_offset}\n`+
+            `data_offset=${data_offset}\n`+
+            `data_size=${device_size-data_offset}\n`
+        );
+    }
+    else
+        process.stdout.write('Unknown format: '+options.format);
+}
+
+function system(cmd)
+{
+    return new Promise((ok, no) => child_process.exec(cmd, { maxBuffer: 64*1024*1024 }, (err, stdout, stderr) => (err ? no(err.message) : ok(stdout))));
+}
+
+run().catch(err => { console.error(err); process.exit(1); });

mon/stable-stringify.js

@@ -0,0 +1,78 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: MIT
+
+function stableStringify(obj, opts)
+{
+    if (!opts)
+        opts = {};
+    if (typeof opts === 'function')
+        opts = { cmp: opts };
+    let space = opts.space || '';
+    if (typeof space === 'number')
+        space = Array(space+1).join(' ');
+    const cycles = (typeof opts.cycles === 'boolean') ? opts.cycles : false;
+    const cmp = opts.cmp && (function (f)
+    {
+        return function (node)
+        {
+            return function (a, b)
+            {
+                let aobj = { key: a, value: node[a] };
+                let bobj = { key: b, value: node[b] };
+                return f(aobj, bobj);
+            };
+        };
+    })(opts.cmp);
+    const seen = new Map();
+    return (function stringify (parent, key, node, level)
+    {
+        const indent = space ? ('\n' + new Array(level + 1).join(space)) : '';
+        const colonSeparator = space ? ': ' : ':';
+        if (node === undefined)
+        {
+            return;
+        }
+        if (typeof node !== 'object' || node === null)
+        {
+            return JSON.stringify(node);
+        }
+        if (node instanceof Array)
+        {
+            const out = [];
+            for (let i = 0; i < node.length; i++)
+            {
+                const item = stringify(node, i, node[i], level+1) || JSON.stringify(null);
+                out.push(indent + space + item);
+            }
+            return '[' + out.join(',') + indent + ']';
+        }
+        else
+        {
+            if (seen.has(node))
+            {
+                if (cycles)
+                    return JSON.stringify('__cycle__');
+                throw new TypeError('Converting circular structure to JSON');
+            }
+            else
+                seen.set(node, true);
+            const keys = Object.keys(node).sort(cmp && cmp(node));
+            const out = [];
+            for (let i = 0; i < keys.length; i++)
+            {
+                const key = keys[i];
+                const value = stringify(node, key, node[key], level+1);
+                if (!value)
+                    continue;
+                const keyValue = JSON.stringify(key)
+                    + colonSeparator
+                    + value;
+                out.push(indent + space + keyValue);
+            }
+            seen.delete(node);
+            return '{' + out.join(',') + indent + '}';
+        }
+    })({ '': obj }, '', obj, 0);
+}
+
+module.exports = stableStringify;

mon/test-nonuniform.js

@@ -0,0 +1,130 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
+// Interesting real-world example coming from Ceph with EC and compression enabled.
+// EC parity chunks can't be compressed as efficiently as data chunks,
+// thus they occupy more space (2.26x more space) in OSD object stores.
+// This leads to really uneven OSD fill ratio in Ceph even when PGs are perfectly balanced.
+// But we support this case with the "parity_space" parameter in optimize_initial()/optimize_change().
+
+const LPOptimizer = require('./lp-optimizer.js');
+
+const osd_tree = {
+    ripper5: {
+        osd0: 3.493144989013672,
+        osd1: 3.493144989013672,
+        osd2: 3.454082489013672,
+        osd12: 3.461894989013672,
+    },
+    ripper7: {
+        osd4: 3.638690948486328,
+        osd5: 3.638690948486328,
+        osd6: 3.638690948486328,
+    },
+    ripper4: {
+        osd9: 3.4609375,
+        osd10: 3.4609375,
+        osd11: 3.4609375,
+    },
+    ripper6: {
+        osd3: 3.5849609375,
+        osd7: 3.5859336853027344,
+        osd8: 3.638690948486328,
+        osd13: 3.461894989013672
+    },
+};
+
+const prev_pgs = [[12,7,5],[6,11,12],[3,6,9],[10,0,5],[2,5,13],[9,8,6],[3,4,12],[7,4,12],[12,11,13],[13,6,0],[4,13,10],[9,7,6],[7,10,0],[10,8,0],[3,10,2],[3,0,4],[6,13,0],[13,10,0],[13,10,5],[8,11,6],[3,9,2],[2,8,5],[8,9,5],[3,12,11],[0,7,4],[13,11,1],[11,3,12],[12,8,10],[7,5,12],[2,13,5],[7,11,0],[13,2,6],[0,6,8],[13,1,6],[0,13,4],[0,8,10],[4,10,0],[8,12,4],[8,12,9],[12,7,4],[13,9,5],[3,2,11],[1,9,7],[1,8,5],[5,12,9],[3,5,12],[2,8,10],[0,8,4],[1,4,11],[7,10,2],[12,13,5],[3,1,11],[7,1,4],[4,12,8],[7,0,9],[11,1,8],[3,0,5],[11,13,0],[1,13,5],[12,7,10],[12,8,4],[11,13,5],[0,11,6],[2,11,3],[13,1,11],[2,7,10],[7,10,12],[7,12,10],[12,11,5],[13,12,10],[2,3,9],[4,3,9],[13,2,5],[7,12,6],[12,10,13],[9,8,1],[13,1,5],[9,5,12],[5,11,7],[6,2,9],[8,11,6],[12,5,8],[6,13,1],[7,6,11],[2,3,6],[8,5,9],[1,13,6],[9,3,2],[7,11,1],[3,10,1],[0,11,7],[3,0,5],[1,3,6],[6,0,9],[3,11,4],[8,10,2],[13,1,9],[12,6,9],[3,12,9],[12,8,9],[7,5,0],[8,12,5],[0,11,3],[12,11,13],[0,7,11],[0,3,10],[1,3,11],[2,7,11],[13,2,6],[9,12,13],[8,2,4],[0,7,4],[5,13,0],[13,12,9],[1,9,8],[0,10,3],[3,5,10],[7,12,9],[2,13,4],[12,7,5],[9,2,7],[3,2,9],[6,2,7],[3,1,9],[4,3,2],[5,3,11],[0,7,6],[1,6,13],[7,10,2],[12,4,8],[13,12,6],[7,5,11],[6,2,3],[2,7,6],[2,3,10],[2,7,10],[11,12,6],[0,13,5],[10,2,4],[13,0,11],[7,0,6],[8,9,4],[8,4,11],[7,11,2],[3,4,2],[6,1,3],[7,2,11],[8,9,4],[11,4,8],[10,3,1],[2,10,13],[1,7,11],[13,11,12],[2,6,9],[10,0,13],[7,10,4],[0,11,13],[13,10,1],[7,5,0],[7,12,10],[3,1,4],[7,1,5],[3,11,5],[7,5,0],[1,3,5],[10,5,12],[0,3,9],[7,1,11],[11,8,12],[3,6,2],[7,12,9],[7,11,12],[4,11,3],[0,11,13],[13,2,5],[1,5,8],[0,11,8],[3,5,1],[11,0,6],[3,11,2],[11,8,12],[4,1,3],[10,13,4],[13,9,6],[2,3,10],[12,7,9],[10,0,4],[10,13,2],[3,11,1],[7,2,9],[1,7,4],[13,1,4],[7,0,6],[5,3,9],[10,0,7],[0,7,10],[3,6,10],[13,0,5],[8,4,1],[3,1,10],[2,10,13],[13,0,5],[13,10,2],[12,7,9],[6,8,10],[6,1,8],[10,8,1],[13,5,0],[5,11,3],[7,6,1],[8,5,9],[2,13,11],[10,12,4],[13,4,1],[2,13,4],[11,7,0],[2,9,7],[1,7,6],[8,0,4],[8,1,9],[7,10,12],[13,9,6],[7,6,11],[13,0,4],[1,8,4],[3,12,5],[10,3,1],[10,2,13],[2,4,8],[6,2,3],[3,0,10],[6,7,12],[8,12,5],[3,0,6],[13,12,10],[11,3,6],[9,0,13],[10,0,6],[7,5,2],[1,3,11],[7,10,2],[2,9,8],[11,13,12],[0,8,4],[8,12,11],[6,0,3],[1,13,4],[11,8,2],[12,3,6],[4,7,1],[7,6,12],[3,10,6],[0,10,7],[8,9,1],[0,10,6],[8,10,1]]
+    .map(pg => pg.map(n => 'osd'+n));
+
+const by_osd = {};
+
+for (let i = 0; i < prev_pgs.length; i++)
+{
+    for (let j = 0; j < prev_pgs[i].length; j++)
+    {
+        by_osd[prev_pgs[i][j]] = by_osd[prev_pgs[i][j]] || [];
+        by_osd[prev_pgs[i][j]][j] = (by_osd[prev_pgs[i][j]][j] || 0) + 1;
+    }
+}
+
+/*
+
+This set of PGs was balanced by hand, by heavily tuning OSD weights in Ceph:
+
+{
+  osd0: 4.2,
+  osd1: 3.5,
+  osd2: 3.45409,
+  osd3: 4.5,
+  osd4: 1.4,
+  osd5: 1.4,
+  osd6: 1.75,
+  osd7: 4.5,
+  osd8: 4.4,
+  osd9: 2.2,
+  osd10: 2.7,
+  osd11: 2,
+  osd12: 3.4,
+  osd13: 3.4,
+}
+
+EC+compression is a nightmare in Ceph, yeah :))
+
+To calculate the average ratio between data chunks and parity chunks we
+calculate the number of PG chunks for each chunk role for each OSD:
+
+{
+  osd12: [ 18, 22, 17 ],
+  osd7: [ 35, 22, 8 ],
+  osd5: [ 6, 17, 27 ],
+  osd6: [ 13, 12, 28 ],
+  osd11: [ 13, 26, 20 ],
+  osd3: [ 30, 20, 10 ],
+  osd9: [ 8, 12, 26 ],
+  osd10: [ 15, 23, 20 ],
+  osd0: [ 22, 22, 14 ],
+  osd2: [ 22, 16, 16 ],
+  osd13: [ 29, 19, 13 ],
+  osd8: [ 20, 18, 12 ],
+  osd4: [ 8, 10, 28 ],
+  osd1: [ 17, 17, 17 ]
+}
+
+And now we can pick a pair of OSDs and determine the ratio by solving the following:
+
+osd5 = 23*X + 27*Y = 3249728140
+osd13 = 48*X + 13*Y = 2991675992
+
+=>
+
+osd5 - 27/13*osd13 = 23*X - 27/13*48*X = -76.6923076923077*X = -2963752766.46154
+
+=>
+
+X = 38644720.1243731
+Y = (osd5-23*X)/27 = 87440725.0792377
+Y/X = 2.26268232239284 ~= 2.26
+
+Which means that parity chunks are compressed ~2.26 times worse than data chunks.
+
+Fine, let's try to optimize for it.
+
+*/
+
+async function run()
+{
+    const all_weights = Object.assign({}, ...Object.values(osd_tree));
+    const total_weight = Object.values(all_weights).reduce((a, c) => Number(a) + Number(c), 0);
+    const eff = LPOptimizer.pg_list_space_efficiency(prev_pgs, all_weights, 2, 2.26);
+    const orig = eff*4.26 / total_weight;
+    console.log('Original efficiency was: '+Math.round(orig*10000)/100+' %');
+
+    let prev = await LPOptimizer.optimize_initial({ osd_tree, pg_size: 3, pg_count: 256, parity_space: 2.26 });
+    LPOptimizer.print_change_stats(prev);
+
+    let next = await LPOptimizer.optimize_change({ prev_pgs, osd_tree, pg_size: 3, max_combinations: 10000, parity_space: 2.26 });
+    LPOptimizer.print_change_stats(next);
+}
+
+run().catch(console.error);

mon/test-optimize-simple.js

@@ -0,0 +1,25 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
+const LPOptimizer = require('./lp-optimizer.js');
+
+async function run()
+{
+    const osd_tree = { a: { 1: 1 }, b: { 2: 1 }, c: { 3: 1 } };
+    let res;
+
+    console.log('16 PGs, size=3');
+    res = await LPOptimizer.optimize_initial({ osd_tree, pg_size: 3, pg_count: 16 });
+    LPOptimizer.print_change_stats(res, false);
+
+    console.log('\nReduce PG size to 2');
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs.map(pg => pg.slice(0, 2)), osd_tree, pg_size: 2 });
+    LPOptimizer.print_change_stats(res, false);
+
+    console.log('\nRemove OSD 3');
+    delete osd_tree['c'];
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree, pg_size: 2 });
+    LPOptimizer.print_change_stats(res, false);
+}
+
+run().catch(console.error);

mon/test-optimize-undersized.js

@@ -1,3 +1,6 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
 const LPOptimizer = require('./lp-optimizer.js');
 
 const crush_tree = [
@@ -40,31 +43,31 @@ async function run()
 {
     const cur_tree = {};
     console.log('Empty tree:');
-    let res = await LPOptimizer.optimize_initial(cur_tree, 256);
+    let res = await LPOptimizer.optimize_initial({ osd_tree: cur_tree, pg_size: 3, pg_count: 256 });
     LPOptimizer.print_change_stats(res, false);
     console.log('\nAdding 1st failure domain:');
     cur_tree['dom1'] = osd_tree['dom1'];
-    res = await LPOptimizer.optimize_change(res.int_pgs, cur_tree);
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree: cur_tree, pg_size: 3 });
     LPOptimizer.print_change_stats(res, false);
     console.log('\nAdding 2nd failure domain:');
     cur_tree['dom2'] = osd_tree['dom2'];
-    res = await LPOptimizer.optimize_change(res.int_pgs, cur_tree);
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree: cur_tree, pg_size: 3 });
     LPOptimizer.print_change_stats(res, false);
     console.log('\nAdding 3rd failure domain:');
     cur_tree['dom3'] = osd_tree['dom3'];
-    res = await LPOptimizer.optimize_change(res.int_pgs, cur_tree);
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree: cur_tree, pg_size: 3 });
     LPOptimizer.print_change_stats(res, false);
     console.log('\nRemoving 3rd failure domain:');
     delete cur_tree['dom3'];
-    res = await LPOptimizer.optimize_change(res.int_pgs, cur_tree);
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree: cur_tree, pg_size: 3 });
     LPOptimizer.print_change_stats(res, false);
     console.log('\nRemoving 2nd failure domain:');
     delete cur_tree['dom2'];
-    res = await LPOptimizer.optimize_change(res.int_pgs, cur_tree);
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree: cur_tree, pg_size: 3 });
     LPOptimizer.print_change_stats(res, false);
     console.log('\nRemoving 1st failure domain:');
     delete cur_tree['dom1'];
-    res = await LPOptimizer.optimize_change(res.int_pgs, cur_tree);
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree: cur_tree, pg_size: 3 });
     LPOptimizer.print_change_stats(res, false);
 }
 

mon/test-optimize.js

@@ -1,3 +1,6 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
 const LPOptimizer = require('./lp-optimizer.js');
 
 const osd_tree = {
@@ -75,19 +78,37 @@ const crush_tree = [
 
 async function run()
 {
+    let res;
+
     // Test: add 1 OSD of almost the same size. Ideal data movement could be 1/12 = 8.33%. Actual is ~13%
-    // Space efficiency is ~99.5% in both cases.
-    let res = await LPOptimizer.optimize_initial(osd_tree, 256);
+    // Space efficiency is ~99% in all cases.
+
+    console.log('256 PGs, size=2');
+    res = await LPOptimizer.optimize_initial({ osd_tree, pg_size: 2, pg_count: 256 });
     LPOptimizer.print_change_stats(res, false);
-    console.log('adding osd.8');
+    console.log('\nAdding osd.8');
     osd_tree[500][8] = 3.58589;
-    res = await LPOptimizer.optimize_change(res.int_pgs, osd_tree);
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree, pg_size: 2 });
     LPOptimizer.print_change_stats(res, false);
-    console.log('removing osd.8');
+    console.log('\nRemoving osd.8');
     delete osd_tree[500][8];
-    res = await LPOptimizer.optimize_change(res.int_pgs, osd_tree);
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree, pg_size: 2 });
     LPOptimizer.print_change_stats(res, false);
-    res = await LPOptimizer.optimize_initial(LPOptimizer.flatten_tree(crush_tree, {}, 1, 3), 256);
+
+    console.log('\n256 PGs, size=3');
+    res = await LPOptimizer.optimize_initial({ osd_tree, pg_size: 3, pg_count: 256 });
+    LPOptimizer.print_change_stats(res, false);
+    console.log('\nAdding osd.8');
+    osd_tree[500][8] = 3.58589;
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree, pg_size: 3 });
+    LPOptimizer.print_change_stats(res, false);
+    console.log('\nRemoving osd.8');
+    delete osd_tree[500][8];
+    res = await LPOptimizer.optimize_change({ prev_pgs: res.int_pgs, osd_tree, pg_size: 3 });
+    LPOptimizer.print_change_stats(res, false);
+
+    console.log('\n256 PGs, size=3, failure domain=rack');
+    res = await LPOptimizer.optimize_initial({ osd_tree: LPOptimizer.flatten_tree(crush_tree, {}, 1, 3), pg_size: 3, pg_count: 256 });
     LPOptimizer.print_change_stats(res, false);
 }
 

msgr_receive.cpp

@@ -1,273 +0,0 @@
-#include "messenger.h"
-
-void osd_messenger_t::read_requests()
-{
-    for (int i = 0; i < read_ready_clients.size(); i++)
-    {
-        int peer_fd = read_ready_clients[i];
-        auto & cl = clients[peer_fd];
-        io_uring_sqe* sqe = ringloop->get_sqe();
-        if (!sqe)
-        {
-            read_ready_clients.erase(read_ready_clients.begin(), read_ready_clients.begin() + i);
-            return;
-        }
-        ring_data_t* data = ((ring_data_t*)sqe->user_data);
-        if (!cl.read_op || cl.read_remaining < receive_buffer_size)
-        {
-            cl.read_iov.iov_base = cl.in_buf;
-            cl.read_iov.iov_len = receive_buffer_size;
-        }
-        else
-        {
-            cl.read_iov.iov_base = cl.read_buf;
-            cl.read_iov.iov_len = cl.read_remaining;
-        }
-        cl.read_msg.msg_iov = &cl.read_iov;
-        cl.read_msg.msg_iovlen = 1;
-        data->callback = [this, peer_fd](ring_data_t *data) { handle_read(data->res, peer_fd); };
-        my_uring_prep_recvmsg(sqe, peer_fd, &cl.read_msg, 0);
-    }
-    read_ready_clients.clear();
-}
-
-bool osd_messenger_t::handle_read(int result, int peer_fd)
-{
-    auto cl_it = clients.find(peer_fd);
-    if (cl_it != clients.end())
-    {
-        auto & cl = cl_it->second;
-        if (result < 0 && result != -EAGAIN)
-        {
-            // this is a client socket, so don't panic. just disconnect it
-            printf("Client %d socket read error: %d (%s). Disconnecting client\n", peer_fd, -result, strerror(-result));
-            stop_client(peer_fd);
-            return false;
-        }
-        if (result == -EAGAIN || result < cl.read_iov.iov_len)
-        {
-            cl.read_ready--;
-            if (cl.read_ready > 0)
-                read_ready_clients.push_back(peer_fd);
-        }
-        else
-        {
-            read_ready_clients.push_back(peer_fd);
-        }
-        if (result > 0)
-        {
-            if (cl.read_iov.iov_base == cl.in_buf)
-            {
-                // Compose operation(s) from the buffer
-                int remain = result;
-                void *curbuf = cl.in_buf;
-                while (remain > 0)
-                {
-                    if (!cl.read_op)
-                    {
-                        cl.read_op = new osd_op_t;
-                        cl.read_op->peer_fd = peer_fd;
-                        cl.read_op->op_type = OSD_OP_IN;
-                        cl.read_buf = cl.read_op->req.buf;
-                        cl.read_remaining = OSD_PACKET_SIZE;
-                        cl.read_state = CL_READ_HDR;
-                    }
-                    if (cl.read_remaining > remain)
-                    {
-                        memcpy(cl.read_buf, curbuf, remain);
-                        cl.read_remaining -= remain;
-                        cl.read_buf += remain;
-                        remain = 0;
-                        if (cl.read_remaining <= 0)
-                            handle_finished_read(cl);
-                    }
-                    else
-                    {
-                        memcpy(cl.read_buf, curbuf, cl.read_remaining);
-                        curbuf += cl.read_remaining;
-                        remain -= cl.read_remaining;
-                        cl.read_remaining = 0;
-                        cl.read_buf = NULL;
-                        handle_finished_read(cl);
-                    }
-                }
-            }
-            else
-            {
-                // Long data
-                cl.read_remaining -= result;
-                cl.read_buf += result;
-                if (cl.read_remaining <= 0)
-                {
-                    handle_finished_read(cl);
-                }
-            }
-            if (result >= cl.read_iov.iov_len)
-            {
-                return true;
-            }
-        }
-    }
-    return false;
-}
-
-void osd_messenger_t::handle_finished_read(osd_client_t & cl)
-{
-    if (cl.read_state == CL_READ_HDR)
-    {
-        if (cl.read_op->req.hdr.magic == SECONDARY_OSD_REPLY_MAGIC)
-            handle_reply_hdr(&cl);
-        else
-            handle_op_hdr(&cl);
-    }
-    else if (cl.read_state == CL_READ_DATA)
-    {
-        // Operation is ready
-        cl.received_ops.push_back(cl.read_op);
-        exec_op(cl.read_op);
-        cl.read_op = NULL;
-        cl.read_state = 0;
-    }
-    else if (cl.read_state == CL_READ_REPLY_DATA)
-    {
-        // Reply is ready
-        auto req_it = cl.sent_ops.find(cl.read_reply_id);
-        osd_op_t *request = req_it->second;
-        cl.sent_ops.erase(req_it);
-        cl.read_reply_id = 0;
-        delete cl.read_op;
-        cl.read_op = NULL;
-        cl.read_state = 0;
-        // Measure subop latency
-        timespec tv_end;
-        clock_gettime(CLOCK_REALTIME, &tv_end);
-        stats.subop_stat_count[request->req.hdr.opcode]++;
-        if (!stats.subop_stat_count[request->req.hdr.opcode])
-        {
-            stats.subop_stat_count[request->req.hdr.opcode]++;
-            stats.subop_stat_sum[request->req.hdr.opcode] = 0;
-        }
-        stats.subop_stat_sum[request->req.hdr.opcode] += (
-            (tv_end.tv_sec - request->tv_begin.tv_sec)*1000000 +
-            (tv_end.tv_nsec - request->tv_begin.tv_nsec)/1000
-        );
-        request->callback(request);
-    }
-    else
-    {
-        assert(0);
-    }
-}
-
-void osd_messenger_t::handle_op_hdr(osd_client_t *cl)
-{
-    osd_op_t *cur_op = cl->read_op;
-    if (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_READ)
-    {
-        if (cur_op->req.sec_rw.len > 0)
-            cur_op->buf = memalign(MEM_ALIGNMENT, cur_op->req.sec_rw.len);
-        cl->read_remaining = 0;
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_WRITE)
-    {
-        if (cur_op->req.sec_rw.len > 0)
-            cur_op->buf = memalign(MEM_ALIGNMENT, cur_op->req.sec_rw.len);
-        cl->read_remaining = cur_op->req.sec_rw.len;
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_STABILIZE ||
-        cur_op->req.hdr.opcode == OSD_OP_SECONDARY_ROLLBACK)
-    {
-        if (cur_op->req.sec_stab.len > 0)
-            cur_op->buf = memalign(MEM_ALIGNMENT, cur_op->req.sec_stab.len);
-        cl->read_remaining = cur_op->req.sec_stab.len;
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_READ)
-    {
-        if (cur_op->req.rw.len > 0)
-            cur_op->buf = memalign(MEM_ALIGNMENT, cur_op->req.rw.len);
-        cl->read_remaining = 0;
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_WRITE)
-    {
-        if (cur_op->req.rw.len > 0)
-            cur_op->buf = memalign(MEM_ALIGNMENT, cur_op->req.rw.len);
-        cl->read_remaining = cur_op->req.rw.len;
-    }
-    if (cl->read_remaining > 0)
-    {
-        // Read data
-        cl->read_buf = cur_op->buf;
-        cl->read_state = CL_READ_DATA;
-    }
-    else
-    {
-        // Operation is ready
-        cl->read_op = NULL;
-        cl->read_state = 0;
-        cl->received_ops.push_back(cur_op);
-        exec_op(cur_op);
-    }
-}
-
-void osd_messenger_t::handle_reply_hdr(osd_client_t *cl)
-{
-    osd_op_t *cur_op = cl->read_op;
-    auto req_it = cl->sent_ops.find(cur_op->req.hdr.id);
-    if (req_it == cl->sent_ops.end())
-    {
-        // Command out of sync. Drop connection
-        printf("Client %d command out of sync: id %lu\n", cl->peer_fd, cur_op->req.hdr.id);
-        stop_client(cl->peer_fd);
-        return;
-    }
-    osd_op_t *op = req_it->second;
-    memcpy(op->reply.buf, cur_op->req.buf, OSD_PACKET_SIZE);
-    if ((op->reply.hdr.opcode == OSD_OP_SECONDARY_READ || op->reply.hdr.opcode == OSD_OP_READ) &&
-        op->reply.hdr.retval > 0)
-    {
-        // Read data. In this case we assume that the buffer is preallocated by the caller (!)
-        assert(op->buf);
-        cl->read_state = CL_READ_REPLY_DATA;
-        cl->read_reply_id = op->req.hdr.id;
-        cl->read_buf = op->buf;
-        cl->read_remaining = op->reply.hdr.retval;
-    }
-    else if (op->reply.hdr.opcode == OSD_OP_SECONDARY_LIST && op->reply.hdr.retval > 0)
-    {
-        op->buf = memalign(MEM_ALIGNMENT, sizeof(obj_ver_id) * op->reply.hdr.retval);
-        cl->read_state = CL_READ_REPLY_DATA;
-        cl->read_reply_id = op->req.hdr.id;
-        cl->read_buf = op->buf;
-        cl->read_remaining = sizeof(obj_ver_id) * op->reply.hdr.retval;
-    }
-    else if (op->reply.hdr.opcode == OSD_OP_SHOW_CONFIG && op->reply.hdr.retval > 0)
-    {
-        op->buf = malloc(op->reply.hdr.retval);
-        cl->read_state = CL_READ_REPLY_DATA;
-        cl->read_reply_id = op->req.hdr.id;
-        cl->read_buf = op->buf;
-        cl->read_remaining = op->reply.hdr.retval;
-    }
-    else
-    {
-        delete cl->read_op;
-        cl->read_state = 0;
-        cl->read_op = NULL;
-        cl->sent_ops.erase(req_it);
-        // Measure subop latency
-        timespec tv_end;
-        clock_gettime(CLOCK_REALTIME, &tv_end);
-        stats.subop_stat_count[op->req.hdr.opcode]++;
-        if (!stats.subop_stat_count[op->req.hdr.opcode])
-        {
-            stats.subop_stat_count[op->req.hdr.opcode]++;
-            stats.subop_stat_sum[op->req.hdr.opcode] = 0;
-        }
-        stats.subop_stat_sum[op->req.hdr.opcode] += (
-            (tv_end.tv_sec - op->tv_begin.tv_sec)*1000000 +
-            (tv_end.tv_nsec - op->tv_begin.tv_nsec)/1000
-        );
-        // Copy lambda to be unaffected by `delete op`
-        std::function<void(osd_op_t*)>(op->callback)(op);
-    }
-}

msgr_send.cpp

@@ -1,157 +0,0 @@
-#include "messenger.h"
-
-void osd_messenger_t::outbox_push(osd_op_t *cur_op)
-{
-    assert(cur_op->peer_fd);
-    auto & cl = clients.at(cur_op->peer_fd);
-    if (cur_op->op_type == OSD_OP_OUT)
-    {
-        clock_gettime(CLOCK_REALTIME, &cur_op->tv_begin);
-    }
-    else
-    {
-        // Check that operation actually belongs to this client
-        bool found = false;
-        for (auto it = cl.received_ops.begin(); it != cl.received_ops.end(); it++)
-        {
-            if (*it == cur_op)
-            {
-                found = true;
-                cl.received_ops.erase(it, it+1);
-                break;
-            }
-        }
-        if (!found)
-        {
-            delete cur_op;
-            return;
-        }
-    }
-    cl.outbox.push_back(cur_op);
-    if (cl.write_op || cl.outbox.size() > 1 || !try_send(cl))
-    {
-        if (cl.write_state == 0)
-        {
-            cl.write_state = CL_WRITE_READY;
-            write_ready_clients.push_back(cur_op->peer_fd);
-        }
-        ringloop->wakeup();
-    }
-}
-
-bool osd_messenger_t::try_send(osd_client_t & cl)
-{
-    int peer_fd = cl.peer_fd;
-    io_uring_sqe* sqe = ringloop->get_sqe();
-    if (!sqe)
-    {
-        return false;
-    }
-    ring_data_t* data = ((ring_data_t*)sqe->user_data);
-    if (!cl.write_op)
-    {
-        // pick next command
-        cl.write_op = cl.outbox.front();
-        cl.outbox.pop_front();
-        cl.write_state = CL_WRITE_REPLY;
-        if (cl.write_op->op_type == OSD_OP_IN)
-        {
-            // Measure execution latency
-            timespec tv_end;
-            clock_gettime(CLOCK_REALTIME, &tv_end);
-            stats.op_stat_count[cl.write_op->req.hdr.opcode]++;
-            if (!stats.op_stat_count[cl.write_op->req.hdr.opcode])
-            {
-                stats.op_stat_count[cl.write_op->req.hdr.opcode]++;
-                stats.op_stat_sum[cl.write_op->req.hdr.opcode] = 0;
-                stats.op_stat_bytes[cl.write_op->req.hdr.opcode] = 0;
-            }
-            stats.op_stat_sum[cl.write_op->req.hdr.opcode] += (
-                (tv_end.tv_sec - cl.write_op->tv_begin.tv_sec)*1000000 +
-                (tv_end.tv_nsec - cl.write_op->tv_begin.tv_nsec)/1000
-            );
-            if (cl.write_op->req.hdr.opcode == OSD_OP_READ ||
-                cl.write_op->req.hdr.opcode == OSD_OP_WRITE)
-            {
-                stats.op_stat_bytes[cl.write_op->req.hdr.opcode] += cl.write_op->req.rw.len;
-            }
-            else if (cl.write_op->req.hdr.opcode == OSD_OP_SECONDARY_READ ||
-                cl.write_op->req.hdr.opcode == OSD_OP_SECONDARY_WRITE)
-            {
-                stats.op_stat_bytes[cl.write_op->req.hdr.opcode] += cl.write_op->req.sec_rw.len;
-            }
-        }
-    }
-    cl.write_msg.msg_iov = cl.write_op->send_list.get_iovec();
-    cl.write_msg.msg_iovlen = cl.write_op->send_list.get_size();
-    data->callback = [this, peer_fd](ring_data_t *data) { handle_send(data->res, peer_fd); };
-    my_uring_prep_sendmsg(sqe, peer_fd, &cl.write_msg, 0);
-    return true;
-}
-
-void osd_messenger_t::send_replies()
-{
-    for (int i = 0; i < write_ready_clients.size(); i++)
-    {
-        int peer_fd = write_ready_clients[i];
-        if (!try_send(clients[peer_fd]))
-        {
-            write_ready_clients.erase(write_ready_clients.begin(), write_ready_clients.begin() + i);
-            return;
-        }
-    }
-    write_ready_clients.clear();
-}
-
-void osd_messenger_t::handle_send(int result, int peer_fd)
-{
-    auto cl_it = clients.find(peer_fd);
-    if (cl_it != clients.end())
-    {
-        auto & cl = cl_it->second;
-        if (result < 0 && result != -EAGAIN)
-        {
-            // this is a client socket, so don't panic. just disconnect it
-            printf("Client %d socket write error: %d (%s). Disconnecting client\n", peer_fd, -result, strerror(-result));
-            stop_client(peer_fd);
-            return;
-        }
-        if (result >= 0)
-        {
-            osd_op_t *cur_op = cl.write_op;
-            while (result > 0 && cur_op->send_list.sent < cur_op->send_list.count)
-            {
-                iovec & iov = cur_op->send_list.buf[cur_op->send_list.sent];
-                if (iov.iov_len <= result)
-                {
-                    result -= iov.iov_len;
-                    cur_op->send_list.sent++;
-                }
-                else
-                {
-                    iov.iov_len -= result;
-                    iov.iov_base += result;
-                    break;
-                }
-            }
-            if (cur_op->send_list.sent >= cur_op->send_list.count)
-            {
-                // Done
-                if (cur_op->op_type == OSD_OP_IN)
-                {
-                    delete cur_op;
-                }
-                else
-                {
-                    cl.sent_ops[cl.write_op->req.hdr.id] = cl.write_op;
-                }
-                cl.write_op = NULL;
-                cl.write_state = cl.outbox.size() > 0 ? CL_WRITE_READY : 0;
-            }
-        }
-        if (cl.write_state != 0)
-        {
-            write_ready_clients.push_back(peer_fd);
-        }
-    }
-}

osd.cpp

@@ -1,414 +0,0 @@
-#include <sys/socket.h>
-#include <sys/epoll.h>
-#include <sys/poll.h>
-#include <netinet/in.h>
-#include <netinet/tcp.h>
-#include <arpa/inet.h>
-
-#include "osd.h"
-
-#define MAX_EPOLL_EVENTS 64
-
-const char* osd_op_names[] = {
-    "",
-    "read",
-    "write",
-    "sync",
-    "stabilize",
-    "rollback",
-    "delete",
-    "sync_stab_all",
-    "list",
-    "show_config",
-    "primary_read",
-    "primary_write",
-    "primary_sync",
-    "primary_delete",
-};
-
-osd_t::osd_t(blockstore_config_t & config, blockstore_t *bs, ring_loop_t *ringloop)
-{
-    this->config = config;
-    this->bs = bs;
-    this->ringloop = ringloop;
-
-    this->bs_block_size = bs->get_block_size();
-    // FIXME: use bitmap granularity instead
-    this->bs_disk_alignment = bs->get_disk_alignment();
-
-    parse_config(config);
-
-    epoll_fd = epoll_create(1);
-    if (epoll_fd < 0)
-    {
-        throw std::runtime_error(std::string("epoll_create: ") + strerror(errno));
-    }
-
-    this->tfd = new timerfd_manager_t([this](int fd, std::function<void(int, int)> handler) { set_fd_handler(fd, handler); });
-    this->tfd->set_timer(print_stats_interval*1000, true, [this](int timer_id)
-    {
-        print_stats();
-    });
-
-    c_cli.tfd = this->tfd;
-    c_cli.ringloop = this->ringloop;
-    c_cli.exec_op = [this](osd_op_t *op) { exec_op(op); };
-    c_cli.repeer_pgs = [this](osd_num_t peer_osd) { repeer_pgs(peer_osd); };
-
-    init_cluster();
-
-    consumer.loop = [this]() { loop(); };
-    ringloop->register_consumer(&consumer);
-}
-
-osd_t::~osd_t()
-{
-    if (tfd)
-    {
-        delete tfd;
-        tfd = NULL;
-    }
-    ringloop->unregister_consumer(&consumer);
-    close(epoll_fd);
-    close(listen_fd);
-}
-
-void osd_t::parse_config(blockstore_config_t & config)
-{
-    // Initial startup configuration
-    json11::Json json_config = json11::Json(config);
-    st_cli.parse_config(json_config);
-    etcd_report_interval = strtoull(config["etcd_report_interval"].c_str(), NULL, 10);
-    if (etcd_report_interval <= 0)
-        etcd_report_interval = 30;
-    osd_num = strtoull(config["osd_num"].c_str(), NULL, 10);
-    if (!osd_num)
-        throw std::runtime_error("osd_num is required in the configuration");
-    c_cli.osd_num = osd_num;
-    run_primary = config["run_primary"] != "false" && config["run_primary"] != "0" && config["run_primary"] != "no";
-    // Cluster configuration
-    bind_address = config["bind_address"];
-    if (bind_address == "")
-        bind_address = "0.0.0.0";
-    bind_port = stoull_full(config["bind_port"]);
-    if (bind_port <= 0 || bind_port > 65535)
-        bind_port = 0;
-    if (config["immediate_commit"] == "all")
-        immediate_commit = IMMEDIATE_ALL;
-    else if (config["immediate_commit"] == "small")
-        immediate_commit = IMMEDIATE_SMALL;
-    if (config.find("autosync_interval") != config.end())
-    {
-        autosync_interval = strtoull(config["autosync_interval"].c_str(), NULL, 10);
-        if (autosync_interval > MAX_AUTOSYNC_INTERVAL)
-            autosync_interval = DEFAULT_AUTOSYNC_INTERVAL;
-    }
-    if (config.find("client_queue_depth") != config.end())
-    {
-        client_queue_depth = strtoull(config["client_queue_depth"].c_str(), NULL, 10);
-        if (client_queue_depth < 128)
-            client_queue_depth = 128;
-    }
-    if (config.find("pg_stripe_size") != config.end())
-    {
-        pg_stripe_size = strtoull(config["pg_stripe_size"].c_str(), NULL, 10);
-        if (!pg_stripe_size || !bs_block_size || pg_stripe_size < bs_block_size || (pg_stripe_size % bs_block_size) != 0)
-            pg_stripe_size = DEFAULT_PG_STRIPE_SIZE;
-    }
-    recovery_queue_depth = strtoull(config["recovery_queue_depth"].c_str(), NULL, 10);
-    if (recovery_queue_depth < 1 || recovery_queue_depth > MAX_RECOVERY_QUEUE)
-        recovery_queue_depth = DEFAULT_RECOVERY_QUEUE;
-    if (config["readonly"] == "true" || config["readonly"] == "1" || config["readonly"] == "yes")
-        readonly = true;
-    print_stats_interval = strtoull(config["print_stats_interval"].c_str(), NULL, 10);
-    if (!print_stats_interval)
-        print_stats_interval = 3;
-    c_cli.peer_connect_interval = strtoull(config["peer_connect_interval"].c_str(), NULL, 10);
-    if (!c_cli.peer_connect_interval)
-        c_cli.peer_connect_interval = DEFAULT_PEER_CONNECT_INTERVAL;
-    c_cli.peer_connect_timeout = strtoull(config["peer_connect_timeout"].c_str(), NULL, 10);
-    if (!c_cli.peer_connect_timeout)
-        c_cli.peer_connect_timeout = DEFAULT_PEER_CONNECT_TIMEOUT;
-    log_level = strtoull(config["log_level"].c_str(), NULL, 10);
-    c_cli.log_level = log_level;
-}
-
-void osd_t::bind_socket()
-{
-    listen_fd = socket(AF_INET, SOCK_STREAM, 0);
-    if (listen_fd < 0)
-    {
-        throw std::runtime_error(std::string("socket: ") + strerror(errno));
-    }
-    int enable = 1;
-    setsockopt(listen_fd, SOL_SOCKET, SO_REUSEADDR, &enable, sizeof(enable));
-
-    sockaddr_in addr;
-    int r;
-    if ((r = inet_pton(AF_INET, bind_address.c_str(), &addr.sin_addr)) != 1)
-    {
-        close(listen_fd);
-        throw std::runtime_error("bind address "+bind_address+(r == 0 ? " is not valid" : ": no ipv4 support"));
-    }
-    addr.sin_family = AF_INET;
-
-    addr.sin_port = htons(bind_port);
-    if (bind(listen_fd, (sockaddr*)&addr, sizeof(addr)) < 0)
-    {
-        close(listen_fd);
-        throw std::runtime_error(std::string("bind: ") + strerror(errno));
-    }
-    if (bind_port == 0)
-    {
-        socklen_t len = sizeof(addr);
-        if (getsockname(listen_fd, (sockaddr *)&addr, &len) == -1)
-        {
-            close(listen_fd);
-            throw std::runtime_error(std::string("getsockname: ") + strerror(errno));
-        }
-        listening_port = ntohs(addr.sin_port);
-    }
-    else
-    {
-        listening_port = bind_port;
-    }
-
-    if (listen(listen_fd, listen_backlog) < 0)
-    {
-        close(listen_fd);
-        throw std::runtime_error(std::string("listen: ") + strerror(errno));
-    }
-
-    fcntl(listen_fd, F_SETFL, fcntl(listen_fd, F_GETFL, 0) | O_NONBLOCK);
-
-    epoll_event ev;
-    ev.data.fd = listen_fd;
-    ev.events = EPOLLIN | EPOLLET;
-    if (epoll_ctl(epoll_fd, EPOLL_CTL_ADD, listen_fd, &ev) < 0)
-    {
-        close(listen_fd);
-        close(epoll_fd);
-        throw std::runtime_error(std::string("epoll_ctl: ") + strerror(errno));
-    }
-}
-
-bool osd_t::shutdown()
-{
-    stopping = true;
-    if (inflight_ops > 0)
-    {
-        return false;
-    }
-    return bs->is_safe_to_stop();
-}
-
-void osd_t::loop()
-{
-    if (!wait_state)
-    {
-        handle_epoll_events();
-        wait_state = 1;
-    }
-    handle_peers();
-    c_cli.read_requests();
-    c_cli.send_replies();
-    ringloop->submit();
-}
-
-void osd_t::set_fd_handler(int fd, std::function<void(int, int)> handler)
-{
-    if (handler != NULL)
-    {
-        bool exists = epoll_handlers.find(fd) != epoll_handlers.end();
-        epoll_event ev;
-        ev.data.fd = fd;
-        ev.events = EPOLLOUT | EPOLLIN | EPOLLRDHUP | EPOLLET;
-        if (epoll_ctl(epoll_fd, exists ? EPOLL_CTL_MOD : EPOLL_CTL_ADD, fd, &ev) < 0)
-        {
-            throw std::runtime_error(std::string("epoll_ctl: ") + strerror(errno));
-        }
-        epoll_handlers[fd] = handler;
-    }
-    else
-    {
-        if (epoll_ctl(epoll_fd, EPOLL_CTL_DEL, fd, NULL) < 0 && errno != ENOENT)
-        {
-            throw std::runtime_error(std::string("epoll_ctl: ") + strerror(errno));
-        }
-        epoll_handlers.erase(fd);
-    }
-}
-
-void osd_t::handle_epoll_events()
-{
-    io_uring_sqe *sqe = ringloop->get_sqe();
-    if (!sqe)
-    {
-        throw std::runtime_error("can't get SQE, will fall out of sync with EPOLLET");
-    }
-    ring_data_t *data = ((ring_data_t*)sqe->user_data);
-    my_uring_prep_poll_add(sqe, epoll_fd, POLLIN);
-    data->callback = [this](ring_data_t *data)
-    {
-        if (data->res < 0)
-        {
-            throw std::runtime_error(std::string("epoll failed: ") + strerror(-data->res));
-        }
-        handle_epoll_events();
-    };
-    ringloop->submit();
-    int nfds;
-    epoll_event events[MAX_EPOLL_EVENTS];
-restart:
-    nfds = epoll_wait(epoll_fd, events, MAX_EPOLL_EVENTS, 0);
-    for (int i = 0; i < nfds; i++)
-    {
-        if (events[i].data.fd == listen_fd)
-        {
-            c_cli.accept_connections(listen_fd);
-        }
-        else
-        {
-            auto & cb = epoll_handlers[events[i].data.fd];
-            cb(events[i].data.fd, events[i].events);
-        }
-    }
-    if (nfds == MAX_EPOLL_EVENTS)
-    {
-        goto restart;
-    }
-}
-
-void osd_t::exec_op(osd_op_t *cur_op)
-{
-    clock_gettime(CLOCK_REALTIME, &cur_op->tv_begin);
-    if (stopping)
-    {
-        // Throw operation away
-        delete cur_op;
-        return;
-    }
-    inflight_ops++;
-    cur_op->send_list.push_back(cur_op->reply.buf, OSD_PACKET_SIZE);
-    if (cur_op->req.hdr.magic != SECONDARY_OSD_OP_MAGIC ||
-        cur_op->req.hdr.opcode < OSD_OP_MIN || cur_op->req.hdr.opcode > OSD_OP_MAX ||
-        (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_READ || cur_op->req.hdr.opcode == OSD_OP_SECONDARY_WRITE) &&
-        (cur_op->req.sec_rw.len > OSD_RW_MAX || cur_op->req.sec_rw.len % bs_disk_alignment || cur_op->req.sec_rw.offset % bs_disk_alignment) ||
-        (cur_op->req.hdr.opcode == OSD_OP_READ || cur_op->req.hdr.opcode == OSD_OP_WRITE || cur_op->req.hdr.opcode == OSD_OP_DELETE) &&
-        (cur_op->req.rw.len > OSD_RW_MAX || cur_op->req.rw.len % bs_disk_alignment || cur_op->req.rw.offset % bs_disk_alignment))
-    {
-        // Bad command
-        finish_op(cur_op, -EINVAL);
-        return;
-    }
-    if (readonly &&
-        cur_op->req.hdr.opcode != OSD_OP_SECONDARY_READ &&
-        cur_op->req.hdr.opcode != OSD_OP_SECONDARY_LIST &&
-        cur_op->req.hdr.opcode != OSD_OP_READ &&
-        cur_op->req.hdr.opcode != OSD_OP_SHOW_CONFIG)
-    {
-        // Readonly mode
-        finish_op(cur_op, -EROFS);
-        return;
-    }
-    if (cur_op->req.hdr.opcode == OSD_OP_TEST_SYNC_STAB_ALL)
-    {
-        exec_sync_stab_all(cur_op);
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_SHOW_CONFIG)
-    {
-        exec_show_config(cur_op);
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_READ)
-    {
-        continue_primary_read(cur_op);
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_WRITE)
-    {
-        continue_primary_write(cur_op);
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_SYNC)
-    {
-        continue_primary_sync(cur_op);
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_DELETE)
-    {
-        continue_primary_del(cur_op);
-    }
-    else
-    {
-        exec_secondary(cur_op);
-    }
-}
-
-void osd_t::reset_stats()
-{
-    c_cli.stats = { 0 };
-    prev_stats = { 0 };
-    memset(recovery_stat_count, 0, sizeof(recovery_stat_count));
-    memset(recovery_stat_bytes, 0, sizeof(recovery_stat_bytes));
-}
-
-void osd_t::print_stats()
-{
-    for (int i = 0; i <= OSD_OP_MAX; i++)
-    {
-        if (c_cli.stats.op_stat_count[i] != prev_stats.op_stat_count[i])
-        {
-            uint64_t avg = (c_cli.stats.op_stat_sum[i] - prev_stats.op_stat_sum[i])/(c_cli.stats.op_stat_count[i] - prev_stats.op_stat_count[i]);
-            uint64_t bw = (c_cli.stats.op_stat_bytes[i] - prev_stats.op_stat_bytes[i]) / print_stats_interval;
-            if (c_cli.stats.op_stat_bytes[i] != 0)
-            {
-                printf(
-                    "[OSD %lu] avg latency for op %d (%s): %lu us, B/W: %.2f %s\n", osd_num, i, osd_op_names[i], avg,
-                    (bw > 1024*1024*1024 ? bw/1024.0/1024/1024 : (bw > 1024*1024 ? bw/1024.0/1024 : bw/1024.0)),
-                    (bw > 1024*1024*1024 ? "GB/s" : (bw > 1024*1024 ? "MB/s" : "KB/s"))
-                );
-            }
-            else
-            {
-                printf("[OSD %lu] avg latency for op %d (%s): %lu us\n", osd_num, i, osd_op_names[i], avg);
-            }
-            prev_stats.op_stat_count[i] = c_cli.stats.op_stat_count[i];
-            prev_stats.op_stat_sum[i] = c_cli.stats.op_stat_sum[i];
-            prev_stats.op_stat_bytes[i] = c_cli.stats.op_stat_bytes[i];
-        }
-    }
-    for (int i = 0; i <= OSD_OP_MAX; i++)
-    {
-        if (c_cli.stats.subop_stat_count[i] != prev_stats.subop_stat_count[i])
-        {
-            uint64_t avg = (c_cli.stats.subop_stat_sum[i] - prev_stats.subop_stat_sum[i])/(c_cli.stats.subop_stat_count[i] - prev_stats.subop_stat_count[i]);
-            printf("[OSD %lu] avg latency for subop %d (%s): %ld us\n", osd_num, i, osd_op_names[i], avg);
-            prev_stats.subop_stat_count[i] = c_cli.stats.subop_stat_count[i];
-            prev_stats.subop_stat_sum[i] = c_cli.stats.subop_stat_sum[i];
-        }
-    }
-    for (int i = 0; i < 2; i++)
-    {
-        if (recovery_stat_count[0][i] != recovery_stat_count[1][i])
-        {
-            uint64_t bw = (recovery_stat_bytes[0][i] - recovery_stat_bytes[1][i]) / print_stats_interval;
-            printf(
-                "[OSD %lu] %s recovery: %.1f op/s, B/W: %.2f %s\n", osd_num, recovery_stat_names[i],
-                (recovery_stat_count[0][i] - recovery_stat_count[1][i]) * 1.0 / print_stats_interval,
-                (bw > 1024*1024*1024 ? bw/1024.0/1024/1024 : (bw > 1024*1024 ? bw/1024.0/1024 : bw/1024.0)),
-                (bw > 1024*1024*1024 ? "GB/s" : (bw > 1024*1024 ? "MB/s" : "KB/s"))
-            );
-            recovery_stat_count[1][i] = recovery_stat_count[0][i];
-            recovery_stat_bytes[1][i] = recovery_stat_bytes[0][i];
-        }
-    }
-    if (incomplete_objects > 0)
-    {
-        printf("[OSD %lu] %lu object(s) incomplete\n", osd_num, incomplete_objects);
-    }
-    if (degraded_objects > 0)
-    {
-        printf("[OSD %lu] %lu object(s) degraded\n", osd_num, degraded_objects);
-    }
-    if (misplaced_objects > 0)
-    {
-        printf("[OSD %lu] %lu object(s) misplaced\n", osd_num, misplaced_objects);
-    }
-}

osd_id.h

@@ -1,4 +0,0 @@
-#pragma once
-
-typedef uint64_t osd_num_t;
-typedef uint32_t pg_num_t;

osd_primary.cpp

@@ -1,666 +0,0 @@
-#include "osd_primary.h"
-
-// read: read directly or read paired stripe(s), reconstruct, return
-// write: read paired stripe(s), reconstruct, modify, calculate parity, write
-//
-// nuance: take care to read the same version from paired stripes!
-// to do so, we remember "last readable" version until a write request completes
-// and we postpone other write requests to the same stripe until completion of previous ones
-//
-// sync: sync peers, get unstable versions, stabilize them
-
-bool osd_t::prepare_primary_rw(osd_op_t *cur_op)
-{
-    // PG number is calculated from the offset
-    // Our EC scheme stores data in fixed chunks equal to (K*block size)
-    // K = pg_minsize and will be a property of the inode. Not it's hardcoded (FIXME)
-    uint64_t pg_block_size = bs_block_size * 2;
-    object_id oid = {
-        .inode = cur_op->req.rw.inode,
-        // oid.stripe = starting offset of the parity stripe
-        .stripe = (cur_op->req.rw.offset/pg_block_size)*pg_block_size,
-    };
-    pg_num_t pg_num = (cur_op->req.rw.inode + oid.stripe/pg_stripe_size) % pg_count + 1;
-    auto pg_it = pgs.find(pg_num);
-    if (pg_it == pgs.end() || !(pg_it->second.state & PG_ACTIVE))
-    {
-        // This OSD is not primary for this PG or the PG is inactive
-        finish_op(cur_op, -EPIPE);
-        return false;
-    }
-    if ((cur_op->req.rw.offset + cur_op->req.rw.len) > (oid.stripe + pg_block_size) ||
-        (cur_op->req.rw.offset % bs_disk_alignment) != 0 ||
-        (cur_op->req.rw.len % bs_disk_alignment) != 0)
-    {
-        finish_op(cur_op, -EINVAL);
-        return false;
-    }
-    osd_primary_op_data_t *op_data = (osd_primary_op_data_t*)calloc(
-        sizeof(osd_primary_op_data_t) + sizeof(osd_rmw_stripe_t) * pg_it->second.pg_size, 1
-    );
-    op_data->pg_num = pg_num;
-    op_data->oid = oid;
-    op_data->stripes = ((osd_rmw_stripe_t*)(op_data+1));
-    cur_op->op_data = op_data;
-    split_stripes(pg_it->second.pg_minsize, bs_block_size, (uint32_t)(cur_op->req.rw.offset - oid.stripe), cur_op->req.rw.len, op_data->stripes);
-    pg_it->second.inflight++;
-    return true;
-}
-
-static uint64_t* get_object_osd_set(pg_t &pg, object_id &oid, uint64_t *def, pg_osd_set_state_t **object_state)
-{
-    if (!(pg.state & (PG_HAS_INCOMPLETE | PG_HAS_DEGRADED | PG_HAS_MISPLACED)))
-    {
-        *object_state = NULL;
-        return def;
-    }
-    auto st_it = pg.incomplete_objects.find(oid);
-    if (st_it != pg.incomplete_objects.end())
-    {
-        *object_state = st_it->second;
-        return st_it->second->read_target.data();
-    }
-    st_it = pg.degraded_objects.find(oid);
-    if (st_it != pg.degraded_objects.end())
-    {
-        *object_state = st_it->second;
-        return st_it->second->read_target.data();
-    }
-    st_it = pg.misplaced_objects.find(oid);
-    if (st_it != pg.misplaced_objects.end())
-    {
-        *object_state = st_it->second;
-        return st_it->second->read_target.data();
-    }
-    *object_state = NULL;
-    return def;
-}
-
-void osd_t::continue_primary_read(osd_op_t *cur_op)
-{
-    if (!cur_op->op_data && !prepare_primary_rw(cur_op))
-    {
-        return;
-    }
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    if (op_data->st == 1)      goto resume_1;
-    else if (op_data->st == 2) goto resume_2;
-    {
-        auto & pg = pgs[op_data->pg_num];
-        for (int role = 0; role < pg.pg_minsize; role++)
-        {
-            op_data->stripes[role].read_start = op_data->stripes[role].req_start;
-            op_data->stripes[role].read_end = op_data->stripes[role].req_end;
-        }
-        // Determine version
-        auto vo_it = pg.ver_override.find(op_data->oid);
-        op_data->target_ver = vo_it != pg.ver_override.end() ? vo_it->second : UINT64_MAX;
-        if (pg.state == PG_ACTIVE)
-        {
-            // Fast happy-path
-            cur_op->buf = alloc_read_buffer(op_data->stripes, pg.pg_minsize, 0);
-            submit_primary_subops(SUBMIT_READ, pg.pg_minsize, pg.cur_set.data(), cur_op);
-            cur_op->send_list.push_back(cur_op->buf, cur_op->req.rw.len);
-            op_data->st = 1;
-        }
-        else
-        {
-            // PG may be degraded or have misplaced objects
-            uint64_t* cur_set = get_object_osd_set(pg, op_data->oid, pg.cur_set.data(), &op_data->object_state);
-            if (extend_missing_stripes(op_data->stripes, cur_set, pg.pg_minsize, pg.pg_size) < 0)
-            {
-                finish_op(cur_op, -EIO);
-                return;
-            }
-            // Submit reads
-            op_data->pg_minsize = pg.pg_minsize;
-            op_data->pg_size = pg.pg_size;
-            op_data->degraded = 1;
-            cur_op->buf = alloc_read_buffer(op_data->stripes, pg.pg_size, 0);
-            submit_primary_subops(SUBMIT_READ, pg.pg_size, cur_set, cur_op);
-            op_data->st = 1;
-        }
-    }
-resume_1:
-    return;
-resume_2:
-    if (op_data->errors > 0)
-    {
-        finish_op(cur_op, op_data->epipe > 0 ? -EPIPE : -EIO);
-        return;
-    }
-    if (op_data->degraded)
-    {
-        // Reconstruct missing stripes
-        // FIXME: Always EC(k+1) by now. Add different coding schemes
-        osd_rmw_stripe_t *stripes = op_data->stripes;
-        for (int role = 0; role < op_data->pg_minsize; role++)
-        {
-            if (stripes[role].read_end != 0 && stripes[role].missing)
-            {
-                reconstruct_stripe(stripes, op_data->pg_size, role);
-            }
-            if (stripes[role].req_end != 0)
-            {
-                // Send buffer in parts to avoid copying
-                cur_op->send_list.push_back(
-                    stripes[role].read_buf + (stripes[role].req_start - stripes[role].read_start),
-                    stripes[role].req_end - stripes[role].req_start
-                );
-            }
-        }
-    }
-    finish_op(cur_op, cur_op->req.rw.len);
-}
-
-bool osd_t::check_write_queue(osd_op_t *cur_op, pg_t & pg)
-{
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    // Check if actions are pending for this object
-    auto act_it = pg.flush_actions.lower_bound((obj_piece_id_t){
-        .oid = op_data->oid,
-        .osd_num = 0,
-    });
-    if (act_it != pg.flush_actions.end() &&
-        act_it->first.oid.inode == op_data->oid.inode &&
-        (act_it->first.oid.stripe & ~STRIPE_MASK) == op_data->oid.stripe)
-    {
-        pg.write_queue.emplace(op_data->oid, cur_op);
-        return false;
-    }
-    // Check if there are other write requests to the same object
-    auto vo_it = pg.write_queue.find(op_data->oid);
-    if (vo_it != pg.write_queue.end())
-    {
-        op_data->st = 1;
-        pg.write_queue.emplace(op_data->oid, cur_op);
-        return false;
-    }
-    pg.write_queue.emplace(op_data->oid, cur_op);
-    return true;
-}
-
-void osd_t::continue_primary_write(osd_op_t *cur_op)
-{
-    if (!cur_op->op_data && !prepare_primary_rw(cur_op))
-    {
-        return;
-    }
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    auto & pg = pgs[op_data->pg_num];
-    if (op_data->st == 1)      goto resume_1;
-    else if (op_data->st == 2) goto resume_2;
-    else if (op_data->st == 3) goto resume_3;
-    else if (op_data->st == 4) goto resume_4;
-    else if (op_data->st == 5) goto resume_5;
-    else if (op_data->st == 6) goto resume_6;
-    else if (op_data->st == 7) goto resume_7;
-    else if (op_data->st == 8) goto resume_8;
-    else if (op_data->st == 9) goto resume_9;
-    assert(op_data->st == 0);
-    if (!check_write_queue(cur_op, pg))
-    {
-        return;
-    }
-resume_1:
-    // Determine blocks to read and write
-    // Missing chunks are allowed to be overwritten even in incomplete objects
-    // FIXME: Allow to do small writes to the old (degraded/misplaced) OSD set for the lower performance impact
-    op_data->prev_set = get_object_osd_set(pg, op_data->oid, pg.cur_set.data(), &op_data->object_state);
-    cur_op->rmw_buf = calc_rmw(cur_op->buf, op_data->stripes, op_data->prev_set,
-        pg.pg_size, pg.pg_minsize, pg.pg_cursize, pg.cur_set.data(), bs_block_size);
-    // Read required blocks
-    submit_primary_subops(SUBMIT_RMW_READ, pg.pg_size, op_data->prev_set, cur_op);
-resume_2:
-    op_data->st = 2;
-    return;
-resume_3:
-    if (op_data->errors > 0)
-    {
-        pg_cancel_write_queue(pg, cur_op, op_data->oid, op_data->epipe > 0 ? -EPIPE : -EIO);
-        return;
-    }
-    // Save version override for parallel reads
-    pg.ver_override[op_data->oid] = op_data->fact_ver;
-    // Recover missing stripes, calculate parity
-    calc_rmw_parity(op_data->stripes, pg.pg_size, op_data->prev_set, pg.cur_set.data(), bs_block_size);
-    // Send writes
-    submit_primary_subops(SUBMIT_WRITE, pg.pg_size, pg.cur_set.data(), cur_op);
-resume_4:
-    op_data->st = 4;
-    return;
-resume_5:
-    if (op_data->errors > 0)
-    {
-        pg_cancel_write_queue(pg, cur_op, op_data->oid, op_data->epipe > 0 ? -EPIPE : -EIO);
-        return;
-    }
-    if (op_data->fact_ver == 1)
-    {
-        // Object is created
-        pg.clean_count++;
-        pg.total_count++;
-    }
-    if (op_data->object_state)
-    {
-        {
-            int recovery_type = op_data->object_state->state & (OBJ_DEGRADED|OBJ_INCOMPLETE) ? 0 : 1;
-            recovery_stat_count[0][recovery_type]++;
-            if (!recovery_stat_count[0][recovery_type])
-            {
-                recovery_stat_count[0][recovery_type]++;
-                recovery_stat_bytes[0][recovery_type] = 0;
-            }
-            for (int role = 0; role < pg.pg_size; role++)
-            {
-                recovery_stat_bytes[0][recovery_type] += op_data->stripes[role].write_end - op_data->stripes[role].write_start;
-            }
-        }
-        if (op_data->object_state->state & OBJ_MISPLACED)
-        {
-            // Remove extra chunks
-            submit_primary_del_subops(cur_op, pg.cur_set.data(), op_data->object_state->osd_set);
-            if (op_data->n_subops > 0)
-            {
-resume_8:
-                op_data->st = 8;
-                return;
-resume_9:
-                if (op_data->errors > 0)
-                {
-                    pg_cancel_write_queue(pg, cur_op, op_data->oid, op_data->epipe > 0 ? -EPIPE : -EIO);
-                    return;
-                }
-            }
-        }
-        // Clear object state
-        remove_object_from_state(op_data->oid, op_data->object_state, pg);
-        pg.clean_count++;
-    }
-    // Remove version override
-    pg.ver_override.erase(op_data->oid);
-    // FIXME: Check for immediate_commit == IMMEDIATE_SMALL
-resume_6:
-resume_7:
-    if (!remember_unstable_write(cur_op, pg, pg.cur_loc_set, 6))
-    {
-        return;
-    }
-    object_id oid = op_data->oid;
-    finish_op(cur_op, cur_op->req.rw.len);
-    // Continue other write operations to the same object
-    auto next_it = pg.write_queue.find(oid);
-    auto this_it = next_it;
-    if (this_it != pg.write_queue.end() && this_it->second == cur_op)
-    {
-        next_it++;
-        pg.write_queue.erase(this_it);
-        if (next_it != pg.write_queue.end() && next_it->first == oid)
-        {
-            osd_op_t *next_op = next_it->second;
-            continue_primary_write(next_op);
-        }
-    }
-}
-
-bool osd_t::remember_unstable_write(osd_op_t *cur_op, pg_t & pg, pg_osd_set_t & loc_set, int base_state)
-{
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    if (op_data->st == base_state)
-    {
-        goto resume_6;
-    }
-    else if (op_data->st == base_state+1)
-    {
-        goto resume_7;
-    }
-    if (immediate_commit == IMMEDIATE_ALL)
-    {
-        op_data->unstable_write_osds = new std::vector<unstable_osd_num_t>();
-        op_data->unstable_writes = new obj_ver_id[loc_set.size()];
-        {
-            int last_start = 0;
-            for (auto & chunk: loc_set)
-            {
-                op_data->unstable_writes[last_start] = (obj_ver_id){
-                    .oid = {
-                        .inode = op_data->oid.inode,
-                        .stripe = op_data->oid.stripe | chunk.role,
-                    },
-                    .version = op_data->fact_ver,
-                };
-                op_data->unstable_write_osds->push_back((unstable_osd_num_t){
-                    .osd_num = chunk.osd_num,
-                    .start = last_start,
-                    .len = 1,
-                });
-                last_start++;
-            }
-        }
-        submit_primary_stab_subops(cur_op);
-resume_6:
-        op_data->st = 6;
-        return false;
-resume_7:
-        // FIXME: Free those in the destructor?
-        delete op_data->unstable_write_osds;
-        delete[] op_data->unstable_writes;
-        op_data->unstable_writes = NULL;
-        op_data->unstable_write_osds = NULL;
-        if (op_data->errors > 0)
-        {
-            pg_cancel_write_queue(pg, cur_op, op_data->oid, op_data->epipe > 0 ? -EPIPE : -EIO);
-            return false;
-        }
-    }
-    else
-    {
-        // Remember version as unstable
-        for (auto & chunk: loc_set)
-        {
-            this->unstable_writes[(osd_object_id_t){
-                .osd_num = chunk.osd_num,
-                .oid = {
-                    .inode = op_data->oid.inode,
-                    .stripe = op_data->oid.stripe | chunk.role,
-                },
-            }] = op_data->fact_ver;
-        }
-        // Remember PG as dirty to drop the connection when PG goes offline
-        // (this is required because of the "lazy sync")
-        c_cli.clients[cur_op->peer_fd].dirty_pgs.insert(op_data->pg_num);
-        dirty_pgs.insert(op_data->pg_num);
-    }
-    return true;
-}
-
-// Save and clear unstable_writes -> SYNC all -> STABLE all
-void osd_t::continue_primary_sync(osd_op_t *cur_op)
-{
-    if (!cur_op->op_data)
-    {
-        cur_op->op_data = (osd_primary_op_data_t*)calloc(sizeof(osd_primary_op_data_t), 1);
-    }
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    if (op_data->st == 1)      goto resume_1;
-    else if (op_data->st == 2) goto resume_2;
-    else if (op_data->st == 3) goto resume_3;
-    else if (op_data->st == 4) goto resume_4;
-    else if (op_data->st == 5) goto resume_5;
-    else if (op_data->st == 6) goto resume_6;
-    assert(op_data->st == 0);
-    if (syncs_in_progress.size() > 0)
-    {
-        // Wait for previous syncs, if any
-        // FIXME: We may try to execute the current one in parallel, like in Blockstore, but I'm not sure if it matters at all
-        syncs_in_progress.push_back(cur_op);
-        op_data->st = 1;
-resume_1:
-        return;
-    }
-    else
-    {
-        syncs_in_progress.push_back(cur_op);
-    }
-resume_2:
-    if (unstable_writes.size() == 0)
-    {
-        // Nothing to sync
-        goto finish;
-    }
-    // Save and clear unstable_writes
-    // In theory it is possible to do in on a per-client basis, but this seems to be an unnecessary complication
-    // It would be cool not to copy these here at all, but someone has to deduplicate them by object IDs anyway
-    {
-        op_data->unstable_write_osds = new std::vector<unstable_osd_num_t>();
-        op_data->unstable_writes = new obj_ver_id[this->unstable_writes.size()];
-        op_data->dirty_pgs = new pg_num_t[dirty_pgs.size()];
-        op_data->dirty_pg_count = dirty_pgs.size();
-        osd_num_t last_osd = 0;
-        int last_start = 0, last_end = 0;
-        for (auto it = this->unstable_writes.begin(); it != this->unstable_writes.end(); it++)
-        {
-            if (last_osd != it->first.osd_num)
-            {
-                if (last_osd != 0)
-                {
-                    op_data->unstable_write_osds->push_back((unstable_osd_num_t){
-                        .osd_num = last_osd,
-                        .start = last_start,
-                        .len = last_end - last_start,
-                    });
-                }
-                last_osd = it->first.osd_num;
-                last_start = last_end;
-            }
-            op_data->unstable_writes[last_end] = (obj_ver_id){
-                .oid = it->first.oid,
-                .version = it->second,
-            };
-            last_end++;
-        }
-        if (last_osd != 0)
-        {
-            op_data->unstable_write_osds->push_back((unstable_osd_num_t){
-                .osd_num = last_osd,
-                .start = last_start,
-                .len = last_end - last_start,
-            });
-        }
-        int dpg = 0;
-        for (auto dirty_pg_num: dirty_pgs)
-        {
-            pgs[dirty_pg_num].inflight++;
-            op_data->dirty_pgs[dpg++] = dirty_pg_num;
-        }
-        dirty_pgs.clear();
-        this->unstable_writes.clear();
-    }
-    if (immediate_commit != IMMEDIATE_ALL)
-    {
-        // SYNC
-        submit_primary_sync_subops(cur_op);
-resume_3:
-        op_data->st = 3;
-        return;
-resume_4:
-        if (op_data->errors > 0)
-        {
-            goto resume_6;
-        }
-    }
-    // Stabilize version sets
-    submit_primary_stab_subops(cur_op);
-resume_5:
-    op_data->st = 5;
-    return;
-resume_6:
-    if (op_data->errors > 0)
-    {
-        // Return objects back into the unstable write set
-        for (auto unstable_osd: *(op_data->unstable_write_osds))
-        {
-            for (int i = 0; i < unstable_osd.len; i++)
-            {
-                // Except those from peered PGs
-                auto & w = op_data->unstable_writes[i];
-                pg_num_t wpg = map_to_pg(w.oid);
-                if (pgs[wpg].state & PG_ACTIVE)
-                {
-                    uint64_t & dest = this->unstable_writes[(osd_object_id_t){
-                        .osd_num = unstable_osd.osd_num,
-                        .oid = w.oid,
-                    }];
-                    dest = dest < w.version ? w.version : dest;
-                    dirty_pgs.insert(wpg);
-                }
-            }
-        }
-    }
-    for (int i = 0; i < op_data->dirty_pg_count; i++)
-    {
-        auto & pg = pgs.at(op_data->dirty_pgs[i]);
-        pg.inflight--;
-        if ((pg.state & PG_STOPPING) && pg.inflight == 0 && !pg.flush_batch)
-        {
-            finish_stop_pg(pg);
-        }
-    }
-    // FIXME: Free those in the destructor?
-    delete op_data->dirty_pgs;
-    delete op_data->unstable_write_osds;
-    delete[] op_data->unstable_writes;
-    op_data->unstable_writes = NULL;
-    op_data->unstable_write_osds = NULL;
-    if (op_data->errors > 0)
-    {
-        finish_op(cur_op, op_data->epipe > 0 ? -EPIPE : -EIO);
-    }
-    else
-    {
-finish:
-        if (cur_op->peer_fd)
-        {
-            auto it = c_cli.clients.find(cur_op->peer_fd);
-            if (it != c_cli.clients.end())
-                it->second.dirty_pgs.clear();
-        }
-        finish_op(cur_op, 0);
-    }
-    assert(syncs_in_progress.front() == cur_op);
-    syncs_in_progress.pop_front();
-    if (syncs_in_progress.size() > 0)
-    {
-        cur_op = syncs_in_progress.front();
-        op_data = cur_op->op_data;
-        op_data->st++;
-        goto resume_2;
-    }
-}
-
-// Decrement pg_osd_set_state_t's object_count and change PG state accordingly
-void osd_t::remove_object_from_state(object_id & oid, pg_osd_set_state_t *object_state, pg_t & pg)
-{
-    if (object_state->state & OBJ_INCOMPLETE)
-    {
-        // Successful write means that object is not incomplete anymore
-        this->incomplete_objects--;
-        pg.incomplete_objects.erase(oid);
-        if (!pg.incomplete_objects.size())
-        {
-            pg.state = pg.state & ~PG_HAS_INCOMPLETE;
-            report_pg_state(pg);
-        }
-    }
-    else if (object_state->state & OBJ_DEGRADED)
-    {
-        this->degraded_objects--;
-        pg.degraded_objects.erase(oid);
-        if (!pg.degraded_objects.size())
-        {
-            pg.state = pg.state & ~PG_HAS_DEGRADED;
-            report_pg_state(pg);
-        }
-    }
-    else if (object_state->state & OBJ_MISPLACED)
-    {
-        this->misplaced_objects--;
-        pg.misplaced_objects.erase(oid);
-        if (!pg.misplaced_objects.size())
-        {
-            pg.state = pg.state & ~PG_HAS_MISPLACED;
-            report_pg_state(pg);
-        }
-    }
-    else
-    {
-        throw std::runtime_error("BUG: Invalid object state: "+std::to_string(object_state->state));
-    }
-    object_state->object_count--;
-    if (!object_state->object_count)
-    {
-        pg.state_dict.erase(object_state->osd_set);
-    }
-}
-
-void osd_t::continue_primary_del(osd_op_t *cur_op)
-{
-    if (!cur_op->op_data && !prepare_primary_rw(cur_op))
-    {
-        return;
-    }
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    auto & pg = pgs[op_data->pg_num];
-    if (op_data->st == 1)      goto resume_1;
-    else if (op_data->st == 2) goto resume_2;
-    else if (op_data->st == 3) goto resume_3;
-    else if (op_data->st == 4) goto resume_4;
-    else if (op_data->st == 5) goto resume_5;
-    assert(op_data->st == 0);
-    // Delete is forbidden even in active PGs if they're also degraded or have previous dead OSDs
-    if (pg.state & (PG_DEGRADED | PG_LEFT_ON_DEAD))
-    {
-        finish_op(cur_op, -EBUSY);
-        return;
-    }
-    if (!check_write_queue(cur_op, pg))
-    {
-        return;
-    }
-resume_1:
-    // Determine which OSDs contain this object and delete it
-    op_data->prev_set = get_object_osd_set(pg, op_data->oid, pg.cur_set.data(), &op_data->object_state);
-    // Submit 1 read to determine the actual version number
-    submit_primary_subops(SUBMIT_RMW_READ, pg.pg_size, op_data->prev_set, cur_op);
-resume_2:
-    op_data->st = 2;
-    return;
-resume_3:
-    if (op_data->errors > 0)
-    {
-        pg_cancel_write_queue(pg, cur_op, op_data->oid, op_data->epipe > 0 ? -EPIPE : -EIO);
-        return;
-    }
-    // Save version override for parallel reads
-    pg.ver_override[op_data->oid] = op_data->fact_ver;
-    // Submit deletes
-    op_data->fact_ver++;
-    submit_primary_del_subops(cur_op, NULL, op_data->object_state ? op_data->object_state->osd_set : pg.cur_loc_set);
-resume_4:
-    op_data->st = 4;
-    return;
-resume_5:
-    if (op_data->errors > 0)
-    {
-        pg_cancel_write_queue(pg, cur_op, op_data->oid, op_data->epipe > 0 ? -EPIPE : -EIO);
-        return;
-    }
-    // Remove version override
-    pg.ver_override.erase(op_data->oid);
-    // Adjust PG stats after "instant stabilize", because we need object_state above
-    if (!op_data->object_state)
-    {
-        pg.clean_count--;
-    }
-    else
-    {
-        remove_object_from_state(op_data->oid, op_data->object_state, pg);
-    }
-    pg.total_count--;
-    object_id oid = op_data->oid;
-    finish_op(cur_op, cur_op->req.rw.len);
-    // Continue other write operations to the same object
-    auto next_it = pg.write_queue.find(oid);
-    auto this_it = next_it;
-    if (this_it != pg.write_queue.end() && this_it->second == cur_op)
-    {
-        next_it++;
-        pg.write_queue.erase(this_it);
-        if (next_it != pg.write_queue.end() &&
-            next_it->first == oid)
-        {
-            osd_op_t *next_op = next_it->second;
-            continue_primary_write(next_op);
-        }
-    }
-}

osd_primary.h

@@ -1,35 +0,0 @@
-#pragma once
-
-#include "osd.h"
-#include "osd_rmw.h"
-
-#define SUBMIT_READ 0
-#define SUBMIT_RMW_READ 1
-#define SUBMIT_WRITE 2
-
-struct unstable_osd_num_t
-{
-    osd_num_t osd_num;
-    int start, len;
-};
-
-struct osd_primary_op_data_t
-{
-    int st = 0;
-    pg_num_t pg_num;
-    object_id oid;
-    uint64_t target_ver;
-    uint64_t fact_ver = 0;
-    int n_subops = 0, done = 0, errors = 0, epipe = 0;
-    int degraded = 0, pg_size, pg_minsize;
-    osd_rmw_stripe_t *stripes;
-    osd_op_t *subops = NULL;
-    uint64_t *prev_set = NULL;
-    pg_osd_set_state_t *object_state = NULL;
-
-    // for sync. oops, requires freeing
-    std::vector<unstable_osd_num_t> *unstable_write_osds = NULL;
-    pg_num_t *dirty_pgs = NULL;
-    int dirty_pg_count = 0;
-    obj_ver_id *unstable_writes = NULL;
-};

osd_primary_subops.cpp

@@ -1,551 +0,0 @@
-#include "osd_primary.h"
-
-void osd_t::autosync()
-{
-    // FIXME Autosync based on the number of unstable writes to prevent
-    // "journal_sector_buffer_count is too low for this batch" errors
-    if (immediate_commit != IMMEDIATE_ALL && !autosync_op)
-    {
-        autosync_op = new osd_op_t();
-        autosync_op->op_type = OSD_OP_IN;
-        autosync_op->req = {
-            .sync = {
-                .header = {
-                    .magic = SECONDARY_OSD_OP_MAGIC,
-                    .id = 1,
-                    .opcode = OSD_OP_SYNC,
-                },
-            },
-        };
-        autosync_op->callback = [this](osd_op_t *op)
-        {
-            if (op->reply.hdr.retval < 0)
-            {
-                printf("Warning: automatic sync resulted in an error: %ld (%s)\n", -op->reply.hdr.retval, strerror(-op->reply.hdr.retval));
-            }
-            delete autosync_op;
-            autosync_op = NULL;
-        };
-        exec_op(autosync_op);
-    }
-}
-
-void osd_t::finish_op(osd_op_t *cur_op, int retval)
-{
-    inflight_ops--;
-    if (cur_op->op_data)
-    {
-        if (cur_op->op_data->pg_num > 0)
-        {
-            auto & pg = pgs[cur_op->op_data->pg_num];
-            pg.inflight--;
-            assert(pg.inflight >= 0);
-            if ((pg.state & PG_STOPPING) && pg.inflight == 0 && !pg.flush_batch)
-            {
-                finish_stop_pg(pg);
-            }
-        }
-        assert(!cur_op->op_data->subops);
-        assert(!cur_op->op_data->unstable_write_osds);
-        assert(!cur_op->op_data->unstable_writes);
-        assert(!cur_op->op_data->dirty_pgs);
-        free(cur_op->op_data);
-        cur_op->op_data = NULL;
-    }
-    if (!cur_op->peer_fd)
-    {
-        // Copy lambda to be unaffected by `delete op`
-        std::function<void(osd_op_t*)>(cur_op->callback)(cur_op);
-    }
-    else
-    {
-        // FIXME add separate magic number
-        auto cl_it = c_cli.clients.find(cur_op->peer_fd);
-        if (cl_it != c_cli.clients.end())
-        {
-            cur_op->reply.hdr.magic = SECONDARY_OSD_REPLY_MAGIC;
-            cur_op->reply.hdr.id = cur_op->req.hdr.id;
-            cur_op->reply.hdr.opcode = cur_op->req.hdr.opcode;
-            cur_op->reply.hdr.retval = retval;
-            c_cli.outbox_push(cur_op);
-        }
-        else
-        {
-            delete cur_op;
-        }
-    }
-}
-
-void osd_t::submit_primary_subops(int submit_type, int pg_size, const uint64_t* osd_set, osd_op_t *cur_op)
-{
-    bool w = submit_type == SUBMIT_WRITE;
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    osd_rmw_stripe_t *stripes = op_data->stripes;
-    // Allocate subops
-    int n_subops = 0, zero_read = -1;
-    for (int role = 0; role < pg_size; role++)
-    {
-        if (osd_set[role] == this->osd_num || osd_set[role] != 0 && zero_read == -1)
-        {
-            zero_read = role;
-        }
-        if (osd_set[role] != 0 && (w || stripes[role].read_end != 0))
-        {
-            n_subops++;
-        }
-    }
-    if (!n_subops && submit_type == SUBMIT_RMW_READ)
-    {
-        n_subops = 1;
-    }
-    else
-    {
-        zero_read = -1;
-    }
-    uint64_t op_version = w ? op_data->fact_ver+1 : (submit_type == SUBMIT_RMW_READ ? UINT64_MAX : op_data->target_ver);
-    osd_op_t *subops = new osd_op_t[n_subops];
-    op_data->fact_ver = 0;
-    op_data->done = op_data->errors = 0;
-    op_data->n_subops = n_subops;
-    op_data->subops = subops;
-    int i = 0;
-    for (int role = 0; role < pg_size; role++)
-    {
-        // We always submit zero-length writes to all replicas, even if the stripe is not modified
-        if (!(w || stripes[role].read_end != 0 || zero_read == role))
-        {
-            continue;
-        }
-        osd_num_t role_osd_num = osd_set[role];
-        if (role_osd_num != 0)
-        {
-            if (role_osd_num == this->osd_num)
-            {
-                clock_gettime(CLOCK_REALTIME, &subops[i].tv_begin);
-                subops[i].op_type = (uint64_t)cur_op;
-                subops[i].bs_op = new blockstore_op_t({
-                    .opcode = (uint64_t)(w ? BS_OP_WRITE : BS_OP_READ),
-                    .callback = [subop = &subops[i], this](blockstore_op_t *bs_subop)
-                    {
-                        handle_primary_bs_subop(subop);
-                    },
-                    .oid = {
-                        .inode = op_data->oid.inode,
-                        .stripe = op_data->oid.stripe | role,
-                    },
-                    .version = op_version,
-                    .offset = w ? stripes[role].write_start : stripes[role].read_start,
-                    .len = w ? stripes[role].write_end - stripes[role].write_start : stripes[role].read_end - stripes[role].read_start,
-                    .buf = w ? stripes[role].write_buf : stripes[role].read_buf,
-                });
-#ifdef OSD_DEBUG
-                printf(
-                    "Submit %s to local: %lu:%lu v%lu %u-%u\n", w ? "write" : "read",
-                    op_data->oid.inode, op_data->oid.stripe | role, op_version,
-                    subops[i].bs_op->offset, subops[i].bs_op->len
-                );
-#endif
-                bs->enqueue_op(subops[i].bs_op);
-            }
-            else
-            {
-                subops[i].op_type = OSD_OP_OUT;
-                subops[i].send_list.push_back(subops[i].req.buf, OSD_PACKET_SIZE);
-                subops[i].peer_fd = c_cli.osd_peer_fds.at(role_osd_num);
-                subops[i].req.sec_rw = {
-                    .header = {
-                        .magic = SECONDARY_OSD_OP_MAGIC,
-                        .id = c_cli.next_subop_id++,
-                        .opcode = (uint64_t)(w ? OSD_OP_SECONDARY_WRITE : OSD_OP_SECONDARY_READ),
-                    },
-                    .oid = {
-                        .inode = op_data->oid.inode,
-                        .stripe = op_data->oid.stripe | role,
-                    },
-                    .version = op_version,
-                    .offset = w ? stripes[role].write_start : stripes[role].read_start,
-                    .len = w ? stripes[role].write_end - stripes[role].write_start : stripes[role].read_end - stripes[role].read_start,
-                };
-#ifdef OSD_DEBUG
-                printf(
-                    "Submit %s to osd %lu: %lu:%lu v%lu %u-%u\n", w ? "write" : "read", role_osd_num,
-                    op_data->oid.inode, op_data->oid.stripe | role, op_version,
-                    subops[i].req.sec_rw.offset, subops[i].req.sec_rw.len
-                );
-#endif
-                subops[i].buf = w ? stripes[role].write_buf : stripes[role].read_buf;
-                if (w && stripes[role].write_end > 0)
-                {
-                    subops[i].send_list.push_back(stripes[role].write_buf, stripes[role].write_end - stripes[role].write_start);
-                }
-                subops[i].callback = [cur_op, this](osd_op_t *subop)
-                {
-                    int fail_fd = subop->req.hdr.opcode == OSD_OP_SECONDARY_WRITE &&
-                        subop->reply.hdr.retval != subop->req.sec_rw.len ? subop->peer_fd : -1;
-                    // so it doesn't get freed
-                    subop->buf = NULL;
-                    handle_primary_subop(subop, cur_op);
-                    if (fail_fd >= 0)
-                    {
-                        // write operation failed, drop the connection
-                        c_cli.stop_client(fail_fd);
-                    }
-                };
-                c_cli.outbox_push(&subops[i]);
-            }
-            i++;
-        }
-    }
-}
-
-static uint64_t bs_op_to_osd_op[] = {
-    0,
-    OSD_OP_SECONDARY_READ,      // BS_OP_READ
-    OSD_OP_SECONDARY_WRITE,     // BS_OP_WRITE
-    OSD_OP_SECONDARY_SYNC,      // BS_OP_SYNC
-    OSD_OP_SECONDARY_STABILIZE, // BS_OP_STABLE
-    OSD_OP_SECONDARY_DELETE,    // BS_OP_DELETE
-    OSD_OP_SECONDARY_LIST,      // BS_OP_LIST
-    OSD_OP_SECONDARY_ROLLBACK,  // BS_OP_ROLLBACK
-    OSD_OP_TEST_SYNC_STAB_ALL,  // BS_OP_SYNC_STAB_ALL
-};
-
-void osd_t::handle_primary_bs_subop(osd_op_t *subop)
-{
-    osd_op_t *cur_op = (osd_op_t*)subop->op_type;
-    blockstore_op_t *bs_op = subop->bs_op;
-    int expected = bs_op->opcode == BS_OP_READ || bs_op->opcode == BS_OP_WRITE ? bs_op->len : 0;
-    if (bs_op->retval != expected && bs_op->opcode != BS_OP_READ)
-    {
-        // die
-        throw std::runtime_error(
-            "local blockstore modification failed (opcode = "+std::to_string(bs_op->opcode)+
-            " retval = "+std::to_string(bs_op->retval)+")"
-        );
-    }
-    add_bs_subop_stats(subop);
-    subop->req.hdr.opcode = bs_op_to_osd_op[bs_op->opcode];
-    subop->reply.hdr.retval = bs_op->retval;
-    if (bs_op->opcode == BS_OP_READ || bs_op->opcode == BS_OP_WRITE)
-    {
-        subop->req.sec_rw.len = bs_op->len;
-        subop->reply.sec_rw.version = bs_op->version;
-    }
-    delete bs_op;
-    subop->bs_op = NULL;
-    handle_primary_subop(subop, cur_op);
-}
-
-void osd_t::add_bs_subop_stats(osd_op_t *subop)
-{
-    // Include local blockstore ops in statistics
-    uint64_t opcode = bs_op_to_osd_op[subop->bs_op->opcode];
-    timespec tv_end;
-    clock_gettime(CLOCK_REALTIME, &tv_end);
-    c_cli.stats.op_stat_count[opcode]++;
-    if (!c_cli.stats.op_stat_count[opcode])
-    {
-        c_cli.stats.op_stat_count[opcode] = 1;
-        c_cli.stats.op_stat_sum[opcode] = 0;
-        c_cli.stats.op_stat_bytes[opcode] = 0;
-    }
-    c_cli.stats.op_stat_sum[opcode] += (
-        (tv_end.tv_sec - subop->tv_begin.tv_sec)*1000000 +
-        (tv_end.tv_nsec - subop->tv_begin.tv_nsec)/1000
-    );
-    if (opcode == OSD_OP_SECONDARY_READ || opcode == OSD_OP_SECONDARY_WRITE)
-    {
-        c_cli.stats.op_stat_bytes[opcode] += subop->bs_op->len;
-    }
-}
-
-void osd_t::handle_primary_subop(osd_op_t *subop, osd_op_t *cur_op)
-{
-    uint64_t opcode = subop->req.hdr.opcode;
-    int retval = subop->reply.hdr.retval;
-    int expected = opcode == OSD_OP_SECONDARY_READ || opcode == OSD_OP_SECONDARY_WRITE
-        ? subop->req.sec_rw.len : 0;
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    if (retval != expected)
-    {
-        printf("%s subop failed: retval = %d (expected %d)\n", osd_op_names[opcode], retval, expected);
-        if (retval == -EPIPE)
-        {
-            op_data->epipe++;
-        }
-        op_data->errors++;
-    }
-    else
-    {
-        op_data->done++;
-        if (opcode == OSD_OP_SECONDARY_READ || opcode == OSD_OP_SECONDARY_WRITE)
-        {
-            uint64_t version = subop->reply.sec_rw.version;
-#ifdef OSD_DEBUG
-            uint64_t peer_osd = c_cli.clients.find(subop->peer_fd) != c_cli.clients.end()
-                ? c_cli.clients[subop->peer_fd].osd_num : osd_num;
-            printf("subop %lu from osd %lu: version = %lu\n", opcode, peer_osd, version);
-#endif
-            if (op_data->fact_ver != 0 && op_data->fact_ver != version)
-            {
-                throw std::runtime_error(
-                    "different fact_versions returned from "+std::string(osd_op_names[opcode])+
-                    " subops: "+std::to_string(version)+" vs "+std::to_string(op_data->fact_ver)
-                );
-            }
-            op_data->fact_ver = version;
-        }
-    }
-    if ((op_data->errors + op_data->done) >= op_data->n_subops)
-    {
-        delete[] op_data->subops;
-        op_data->subops = NULL;
-        op_data->st++;
-        if (cur_op->req.hdr.opcode == OSD_OP_READ)
-        {
-            continue_primary_read(cur_op);
-        }
-        else if (cur_op->req.hdr.opcode == OSD_OP_WRITE)
-        {
-            continue_primary_write(cur_op);
-        }
-        else if (cur_op->req.hdr.opcode == OSD_OP_SYNC)
-        {
-            continue_primary_sync(cur_op);
-        }
-        else if (cur_op->req.hdr.opcode == OSD_OP_DELETE)
-        {
-            continue_primary_del(cur_op);
-        }
-        else
-        {
-            throw std::runtime_error("BUG: unknown opcode");
-        }
-    }
-}
-
-void osd_t::cancel_primary_write(osd_op_t *cur_op)
-{
-    if (cur_op->op_data && cur_op->op_data->subops)
-    {
-        // Primary-write operation is waiting for subops, subops
-        // are sent to peer OSDs, so we can't just throw them away.
-        // Mark them with an extra EPIPE.
-        cur_op->op_data->errors++;
-        cur_op->op_data->epipe++;
-        cur_op->op_data->done--; // Caution: `done` must be signed because may become -1 here
-    }
-    else
-    {
-        finish_op(cur_op, -EPIPE);
-    }
-}
-
-void osd_t::submit_primary_del_subops(osd_op_t *cur_op, uint64_t *cur_set, pg_osd_set_t & loc_set)
-{
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    int extra_chunks = 0;
-    for (auto & chunk: loc_set)
-    {
-        if (!cur_set || chunk.osd_num != cur_set[chunk.role])
-        {
-            extra_chunks++;
-        }
-    }
-    op_data->n_subops = extra_chunks;
-    op_data->done = op_data->errors = 0;
-    if (!extra_chunks)
-    {
-        return;
-    }
-    osd_op_t *subops = new osd_op_t[extra_chunks];
-    op_data->subops = subops;
-    int i = 0;
-    for (auto & chunk: loc_set)
-    {
-        if (!cur_set || chunk.osd_num != cur_set[chunk.role])
-        {
-            if (chunk.osd_num == this->osd_num)
-            {
-                clock_gettime(CLOCK_REALTIME, &subops[i].tv_begin);
-                subops[i].op_type = (uint64_t)cur_op;
-                subops[i].bs_op = new blockstore_op_t({
-                    .opcode = BS_OP_DELETE,
-                    .callback = [subop = &subops[i], this](blockstore_op_t *bs_subop)
-                    {
-                        handle_primary_bs_subop(subop);
-                    },
-                    .oid = {
-                        .inode = op_data->oid.inode,
-                        .stripe = op_data->oid.stripe | chunk.role,
-                    },
-                    // Same version as write
-                    .version = op_data->fact_ver,
-                });
-                bs->enqueue_op(subops[i].bs_op);
-            }
-            else
-            {
-                subops[i].op_type = OSD_OP_OUT;
-                subops[i].send_list.push_back(subops[i].req.buf, OSD_PACKET_SIZE);
-                subops[i].peer_fd = c_cli.osd_peer_fds.at(chunk.osd_num);
-                subops[i].req.sec_del = {
-                    .header = {
-                        .magic = SECONDARY_OSD_OP_MAGIC,
-                        .id = c_cli.next_subop_id++,
-                        .opcode = OSD_OP_SECONDARY_DELETE,
-                    },
-                    .oid = {
-                        .inode = op_data->oid.inode,
-                        .stripe = op_data->oid.stripe | chunk.role,
-                    },
-                    // Same version as write
-                    .version = op_data->fact_ver,
-                };
-                subops[i].callback = [cur_op, this](osd_op_t *subop)
-                {
-                    int fail_fd = subop->reply.hdr.retval != 0 ? subop->peer_fd : -1;
-                    handle_primary_subop(subop, cur_op);
-                    if (fail_fd >= 0)
-                    {
-                        // delete operation failed, drop the connection
-                        c_cli.stop_client(fail_fd);
-                    }
-                };
-                c_cli.outbox_push(&subops[i]);
-            }
-            i++;
-        }
-    }
-}
-
-void osd_t::submit_primary_sync_subops(osd_op_t *cur_op)
-{
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    int n_osds = op_data->unstable_write_osds->size();
-    osd_op_t *subops = new osd_op_t[n_osds];
-    op_data->done = op_data->errors = 0;
-    op_data->n_subops = n_osds;
-    op_data->subops = subops;
-    for (int i = 0; i < n_osds; i++)
-    {
-        osd_num_t sync_osd = (*(op_data->unstable_write_osds))[i].osd_num;
-        if (sync_osd == this->osd_num)
-        {
-            clock_gettime(CLOCK_REALTIME, &subops[i].tv_begin);
-            subops[i].op_type = (uint64_t)cur_op;
-            subops[i].bs_op = new blockstore_op_t({
-                .opcode = BS_OP_SYNC,
-                .callback = [subop = &subops[i], this](blockstore_op_t *bs_subop)
-                {
-                    handle_primary_bs_subop(subop);
-                },
-            });
-            bs->enqueue_op(subops[i].bs_op);
-        }
-        else
-        {
-            subops[i].op_type = OSD_OP_OUT;
-            subops[i].send_list.push_back(subops[i].req.buf, OSD_PACKET_SIZE);
-            subops[i].peer_fd = c_cli.osd_peer_fds.at(sync_osd);
-            subops[i].req.sec_sync = {
-                .header = {
-                    .magic = SECONDARY_OSD_OP_MAGIC,
-                    .id = c_cli.next_subop_id++,
-                    .opcode = OSD_OP_SECONDARY_SYNC,
-                },
-            };
-            subops[i].callback = [cur_op, this](osd_op_t *subop)
-            {
-                int fail_fd = subop->reply.hdr.retval != 0 ? subop->peer_fd : -1;
-                handle_primary_subop(subop, cur_op);
-                if (fail_fd >= 0)
-                {
-                    // sync operation failed, drop the connection
-                    c_cli.stop_client(fail_fd);
-                }
-            };
-            c_cli.outbox_push(&subops[i]);
-        }
-    }
-}
-
-void osd_t::submit_primary_stab_subops(osd_op_t *cur_op)
-{
-    osd_primary_op_data_t *op_data = cur_op->op_data;
-    int n_osds = op_data->unstable_write_osds->size();
-    osd_op_t *subops = new osd_op_t[n_osds];
-    op_data->done = op_data->errors = 0;
-    op_data->n_subops = n_osds;
-    op_data->subops = subops;
-    for (int i = 0; i < n_osds; i++)
-    {
-        auto & stab_osd = (*(op_data->unstable_write_osds))[i];
-        if (stab_osd.osd_num == this->osd_num)
-        {
-            clock_gettime(CLOCK_REALTIME, &subops[i].tv_begin);
-            subops[i].op_type = (uint64_t)cur_op;
-            subops[i].bs_op = new blockstore_op_t({
-                .opcode = BS_OP_STABLE,
-                .callback = [subop = &subops[i], this](blockstore_op_t *bs_subop)
-                {
-                    handle_primary_bs_subop(subop);
-                },
-                .len = (uint32_t)stab_osd.len,
-                .buf = (void*)(op_data->unstable_writes + stab_osd.start),
-            });
-            bs->enqueue_op(subops[i].bs_op);
-        }
-        else
-        {
-            subops[i].op_type = OSD_OP_OUT;
-            subops[i].send_list.push_back(subops[i].req.buf, OSD_PACKET_SIZE);
-            subops[i].peer_fd = c_cli.osd_peer_fds.at(stab_osd.osd_num);
-            subops[i].req.sec_stab = {
-                .header = {
-                    .magic = SECONDARY_OSD_OP_MAGIC,
-                    .id = c_cli.next_subop_id++,
-                    .opcode = OSD_OP_SECONDARY_STABILIZE,
-                },
-                .len = (uint64_t)(stab_osd.len * sizeof(obj_ver_id)),
-            };
-            subops[i].send_list.push_back(op_data->unstable_writes + stab_osd.start, stab_osd.len * sizeof(obj_ver_id));
-            subops[i].callback = [cur_op, this](osd_op_t *subop)
-            {
-                int fail_fd = subop->reply.hdr.retval != 0 ? subop->peer_fd : -1;
-                handle_primary_subop(subop, cur_op);
-                if (fail_fd >= 0)
-                {
-                    // sync operation failed, drop the connection
-                    c_cli.stop_client(fail_fd);
-                }
-            };
-            c_cli.outbox_push(&subops[i]);
-        }
-    }
-}
-
-void osd_t::pg_cancel_write_queue(pg_t & pg, osd_op_t *first_op, object_id oid, int retval)
-{
-    auto st_it = pg.write_queue.find(oid), it = st_it;
-    finish_op(first_op, retval);
-    if (it != pg.write_queue.end() && it->second == first_op)
-    {
-        it++;
-    }
-    else
-    {
-        // Write queue doesn't match the first operation.
-        // first_op is a leftover operation from the previous peering of the same PG.
-        return;
-    }
-    while (it != pg.write_queue.end() && it->first == oid)
-    {
-        finish_op(it->second, retval);
-        it++;
-    }
-    if (st_it != it)
-    {
-        pg.write_queue.erase(st_it, it);
-    }
-}

osd_rmw.cpp

@@ -1,450 +0,0 @@
-#include <malloc.h>
-#include <string.h>
-#include <assert.h>
-#include "xor.h"
-#include "osd_rmw.h"
-
-static inline void extend_read(uint32_t start, uint32_t end, osd_rmw_stripe_t & stripe)
-{
-    if (stripe.read_end == 0)
-    {
-        stripe.read_start = start;
-        stripe.read_end = end;
-    }
-    else
-    {
-        if (stripe.read_end < end)
-            stripe.read_end = end;
-        if (stripe.read_start > start)
-            stripe.read_start = start;
-    }
-}
-
-static inline void cover_read(uint32_t start, uint32_t end, osd_rmw_stripe_t & stripe)
-{
-    // Subtract <to> write request from <from> request
-    if (start >= stripe.req_start &&
-        end <= stripe.req_end)
-    {
-        return;
-    }
-    if (start <= stripe.req_start &&
-        end >= stripe.req_start &&
-        end <= stripe.req_end)
-    {
-        end = stripe.req_start;
-    }
-    else if (start >= stripe.req_start &&
-        start <= stripe.req_end &&
-        end >= stripe.req_end)
-    {
-        start = stripe.req_end;
-    }
-    if (stripe.read_end == 0)
-    {
-        stripe.read_start = start;
-        stripe.read_end = end;
-    }
-    else
-    {
-        if (stripe.read_end < end)
-            stripe.read_end = end;
-        if (stripe.read_start > start)
-            stripe.read_start = start;
-    }
-}
-
-void split_stripes(uint64_t pg_minsize, uint32_t bs_block_size, uint32_t start, uint32_t end, osd_rmw_stripe_t *stripes)
-{
-    if (end == 0)
-    {
-        // Zero length request - offset doesn't matter
-        return;
-    }
-    end = start+end;
-    for (int role = 0; role < pg_minsize; role++)
-    {
-        if (start < (1+role)*bs_block_size && end > role*bs_block_size)
-        {
-            stripes[role].req_start = start < role*bs_block_size ? 0 : start-role*bs_block_size;
-            stripes[role].req_end = end > (role+1)*bs_block_size ? bs_block_size : end-role*bs_block_size;
-        }
-    }
-}
-
-void reconstruct_stripe(osd_rmw_stripe_t *stripes, int pg_size, int role)
-{
-    int prev = -2;
-    for (int other = 0; other < pg_size; other++)
-    {
-        if (other != role)
-        {
-            if (prev == -2)
-            {
-                prev = other;
-            }
-            else if (prev >= 0)
-            {
-                assert(stripes[role].read_start >= stripes[prev].read_start &&
-                    stripes[role].read_start >= stripes[other].read_start);
-                memxor(
-                    stripes[prev].read_buf + (stripes[role].read_start - stripes[prev].read_start),
-                    stripes[other].read_buf + (stripes[role].read_start - stripes[other].read_start),
-                    stripes[role].read_buf, stripes[role].read_end - stripes[role].read_start
-                );
-                prev = -1;
-            }
-            else
-            {
-                assert(stripes[role].read_start >= stripes[other].read_start);
-                memxor(
-                    stripes[role].read_buf,
-                    stripes[other].read_buf + (stripes[role].read_start - stripes[other].read_start),
-                    stripes[role].read_buf, stripes[role].read_end - stripes[role].read_start
-                );
-            }
-        }
-    }
-}
-
-int extend_missing_stripes(osd_rmw_stripe_t *stripes, osd_num_t *osd_set, int minsize, int size)
-{
-    for (int role = 0; role < minsize; role++)
-    {
-        if (stripes[role].read_end != 0 && osd_set[role] == 0)
-        {
-            stripes[role].missing = true;
-            // Stripe is missing. Extend read to other stripes.
-            // We need at least pg_minsize stripes to recover the lost part.
-            // FIXME: LRC EC and similar don't require to read all other stripes.
-            int exist = 0;
-            for (int j = 0; j < size; j++)
-            {
-                if (osd_set[j] != 0)
-                {
-                    extend_read(stripes[role].read_start, stripes[role].read_end, stripes[j]);
-                    exist++;
-                    if (exist >= minsize)
-                    {
-                        break;
-                    }
-                }
-            }
-            if (exist < minsize)
-            {
-                // Less than minsize stripes are available for this object
-                return -1;
-            }
-        }
-    }
-    return 0;
-}
-
-void* alloc_read_buffer(osd_rmw_stripe_t *stripes, int read_pg_size, uint64_t add_size)
-{
-    // Calculate buffer size
-    uint64_t buf_size = add_size;
-    for (int role = 0; role < read_pg_size; role++)
-    {
-        if (stripes[role].read_end != 0)
-        {
-            buf_size += stripes[role].read_end - stripes[role].read_start;
-        }
-    }
-    // Allocate buffer
-    void *buf = memalign(MEM_ALIGNMENT, buf_size);
-    uint64_t buf_pos = add_size;
-    for (int role = 0; role < read_pg_size; role++)
-    {
-        if (stripes[role].read_end != 0)
-        {
-            stripes[role].read_buf = buf + buf_pos;
-            buf_pos += stripes[role].read_end - stripes[role].read_start;
-        }
-    }
-    return buf;
-}
-
-void* calc_rmw(void *request_buf, osd_rmw_stripe_t *stripes, uint64_t *read_osd_set,
-    uint64_t pg_size, uint64_t pg_minsize, uint64_t pg_cursize, uint64_t *write_osd_set, uint64_t chunk_size)
-{
-    // Generic parity modification (read-modify-write) algorithm
-    // Read -> Reconstruct missing chunks -> Calc parity chunks -> Write
-    // Now we always read continuous ranges. This means that an update of the beginning
-    // of one data stripe and the end of another will lead to a read of full paired stripes.
-    // FIXME: (Maybe) read small individual ranges in that case instead.
-    uint32_t start = 0, end = 0;
-    for (int role = 0; role < pg_minsize; role++)
-    {
-        if (stripes[role].req_end != 0)
-        {
-            start = !end || stripes[role].req_start < start ? stripes[role].req_start : start;
-            end = std::max(stripes[role].req_end, end);
-            stripes[role].write_start = stripes[role].req_start;
-            stripes[role].write_end = stripes[role].req_end;
-        }
-    }
-    int write_parity = 0;
-    for (int role = pg_minsize; role < pg_size; role++)
-    {
-        if (write_osd_set[role] != 0)
-        {
-            write_parity = 1;
-            stripes[role].write_start = start;
-            stripes[role].write_end = end;
-        }
-    }
-    if (write_parity)
-    {
-        for (int role = 0; role < pg_minsize; role++)
-        {
-            cover_read(start, end, stripes[role]);
-        }
-    }
-    if (write_osd_set != read_osd_set)
-    {
-        pg_cursize = 0;
-        // Object is degraded/misplaced and will be moved to <write_osd_set>
-        for (int role = 0; role < pg_size; role++)
-        {
-            if (write_osd_set[role] != read_osd_set[role])
-            {
-                // FIXME: For EC more than 2+1: handle case when write_osd_set == 0 and read_osd_set != 0
-                // We need to get data for any moved / recovered chunk
-                // And we need a continuous write buffer so we'll only optimize
-                // for the case when the whole chunk is ovewritten in the request
-                if (stripes[role].req_start != 0 ||
-                    stripes[role].req_end != chunk_size)
-                {
-                    stripes[role].read_start = 0;
-                    stripes[role].read_end = chunk_size;
-                    // Warning: We don't modify write_start/write_end here, we do it in calc_rmw_parity()
-                }
-            }
-            if (read_osd_set[role] != 0)
-            {
-                pg_cursize++;
-            }
-        }
-    }
-    if (pg_cursize < pg_size)
-    {
-        // Some stripe(s) are missing, so we need to read parity
-        for (int role = 0; role < pg_size; role++)
-        {
-            if (read_osd_set[role] == 0)
-            {
-                stripes[role].missing = true;
-                if (stripes[role].read_end != 0)
-                {
-                    int found = 0;
-                    for (int r2 = 0; r2 < pg_size && found < pg_minsize; r2++)
-                    {
-                        // Read the non-covered range of <role> from at least <minsize> other stripes to reconstruct it
-                        if (read_osd_set[r2] != 0)
-                        {
-                            extend_read(stripes[role].read_start, stripes[role].read_end, stripes[r2]);
-                            found++;
-                        }
-                    }
-                    if (found < pg_minsize)
-                    {
-                        // FIXME Object is incomplete - refuse partial overwrite
-                        assert(0);
-                    }
-                }
-            }
-        }
-    }
-    // Allocate read buffers
-    void *rmw_buf = alloc_read_buffer(stripes, pg_size, (write_parity ? pg_size-pg_minsize : 0) * (end - start));
-    // Position write buffers
-    uint64_t buf_pos = 0, in_pos = 0;
-    for (int role = 0; role < pg_size; role++)
-    {
-        if (stripes[role].req_end != 0)
-        {
-            stripes[role].write_buf = request_buf + in_pos;
-            in_pos += stripes[role].req_end - stripes[role].req_start;
-        }
-        else if (role >= pg_minsize && write_osd_set[role] != 0 && end != 0)
-        {
-            stripes[role].write_buf = rmw_buf + buf_pos;
-            buf_pos += end - start;
-        }
-    }
-    return rmw_buf;
-}
-
-static void get_old_new_buffers(osd_rmw_stripe_t & stripe, uint32_t wr_start, uint32_t wr_end, buf_len_t *bufs, int & nbufs)
-{
-    uint32_t ns = 0, ne = 0, os = 0, oe = 0;
-    if (stripe.req_end > wr_start &&
-        stripe.req_start < wr_end)
-    {
-        ns = std::max(stripe.req_start, wr_start);
-        ne = std::min(stripe.req_end, wr_end);
-    }
-    if (stripe.read_end > wr_start &&
-        stripe.read_start < wr_end)
-    {
-        os = std::max(stripe.read_start, wr_start);
-        oe = std::min(stripe.read_end, wr_end);
-    }
-    if (ne && (!oe || ns <= os))
-    {
-        // NEW or NEW->OLD
-        bufs[nbufs++] = { .buf = stripe.write_buf + ns - stripe.req_start, .len = ne-ns };
-        if (os < ne)
-            os = ne;
-        if (oe > os)
-        {
-            // NEW->OLD
-            bufs[nbufs++] = { .buf = stripe.read_buf + os - stripe.read_start, .len = oe-os };
-        }
-    }
-    else if (oe)
-    {
-        // OLD or OLD->NEW or OLD->NEW->OLD
-        if (ne)
-        {
-            // OLD->NEW or OLD->NEW->OLD
-            bufs[nbufs++] = { .buf = stripe.read_buf + os - stripe.read_start, .len = ns-os };
-            bufs[nbufs++] = { .buf = stripe.write_buf + ns - stripe.req_start, .len = ne-ns };
-            if (oe > ne)
-            {
-                // OLD->NEW->OLD
-                bufs[nbufs++] = { .buf = stripe.read_buf + ne - stripe.read_start, .len = oe-ne };
-            }
-        }
-        else
-        {
-            // OLD
-            bufs[nbufs++] = { .buf = stripe.read_buf + os - stripe.read_start, .len = oe-os };
-        }
-    }
-}
-
-static void xor_multiple_buffers(buf_len_t *xor1, int n1, buf_len_t *xor2, int n2, void *dest, uint32_t len)
-{
-    assert(n1 > 0 && n2 > 0);
-    int i1 = 0, i2 = 0;
-    uint32_t start1 = 0, start2 = 0, end1 = xor1[0].len, end2 = xor2[0].len;
-    uint32_t pos = 0;
-    while (pos < len)
-    {
-        // We know for sure that ranges overlap
-        uint32_t end = std::min(end1, end2);
-        memxor(xor1[i1].buf + pos-start1, xor2[i2].buf + pos-start2, dest+pos, end-pos);
-        pos = end;
-        if (pos >= end1)
-        {
-            i1++;
-            if (i1 >= n1)
-            {
-                assert(pos >= end2);
-                return;
-            }
-            start1 = end1;
-            end1 += xor1[i1].len;
-        }
-        if (pos >= end2)
-        {
-            i2++;
-            start2 = end2;
-            end2 += xor2[i2].len;
-        }
-    }
-}
-
-void calc_rmw_parity(osd_rmw_stripe_t *stripes, int pg_size, uint64_t *read_osd_set, uint64_t *write_osd_set, uint32_t chunk_size)
-{
-    int pg_minsize = pg_size-1;
-    for (int role = 0; role < pg_size; role++)
-    {
-        if (stripes[role].read_end != 0 && stripes[role].missing)
-        {
-            // Reconstruct missing stripe (EC k+1)
-            reconstruct_stripe(stripes, pg_size, role);
-            break;
-        }
-    }
-    uint32_t start = 0, end = 0;
-    if (!stripes[pg_minsize].missing || write_osd_set != read_osd_set)
-    {
-        for (int role = 0; role < pg_minsize; role++)
-        {
-            if (stripes[role].req_end != 0)
-            {
-                start = !end || stripes[role].req_start < start ? stripes[role].req_start : start;
-                end = std::max(stripes[role].req_end, end);
-            }
-        }
-    }
-    if (write_osd_set != read_osd_set)
-    {
-        for (int role = 0; role < pg_minsize; role++)
-        {
-            if (write_osd_set[role] != read_osd_set[role] &&
-                (stripes[role].req_start != 0 || stripes[role].req_end != chunk_size))
-            {
-                // FIXME again, handle case when write_osd_set[role] is 0
-                // Copy modified chunk into the read buffer to write it back
-                memcpy(
-                    stripes[role].read_buf + stripes[role].req_start,
-                    stripes[role].write_buf,
-                    stripes[role].req_end - stripes[role].req_start
-                );
-                stripes[role].write_buf = stripes[role].read_buf;
-                stripes[role].write_start = 0;
-                stripes[role].write_end = chunk_size;
-            }
-        }
-    }
-    if (!stripes[pg_minsize].missing && end != 0)
-    {
-        // Calculate new parity (EC k+1)
-        int parity = pg_minsize, prev = -2;
-        for (int other = 0; other < pg_minsize; other++)
-        {
-            if (prev == -2)
-            {
-                prev = other;
-            }
-            else
-            {
-                int n1 = 0, n2 = 0;
-                buf_len_t xor1[3], xor2[3];
-                if (prev == -1)
-                {
-                    xor1[n1++] = { .buf = stripes[parity].write_buf, .len = end-start };
-                }
-                else
-                {
-                    get_old_new_buffers(stripes[prev], start, end, xor1, n1);
-                    prev = -1;
-                }
-                get_old_new_buffers(stripes[other], start, end, xor2, n2);
-                xor_multiple_buffers(xor1, n1, xor2, n2, stripes[parity].write_buf, end-start);
-            }
-        }
-    }
-    if (write_osd_set != read_osd_set)
-    {
-        for (int role = pg_minsize; role < pg_size; role++)
-        {
-            if (write_osd_set[role] != read_osd_set[role] && (start != 0 || end != chunk_size))
-            {
-                // Copy new parity into the read buffer to write it back
-                memcpy(
-                    stripes[role].read_buf + start,
-                    stripes[role].write_buf,
-                    end - start
-                );
-                stripes[role].write_buf = stripes[role].read_buf;
-                stripes[role].write_start = 0;
-                stripes[role].write_end = chunk_size;
-            }
-        }
-    }
-}

osd_rmw.h

@@ -1,37 +0,0 @@
-#pragma once
-
-#include <stdint.h>
-#include "object_id.h"
-#include "osd_id.h"
-
-#ifndef MEM_ALIGNMENT
-#define MEM_ALIGNMENT 512
-#endif
-
-struct buf_len_t
-{
-    void *buf;
-    uint64_t len;
-};
-
-struct osd_rmw_stripe_t
-{
-    void *read_buf, *write_buf;
-    uint32_t req_start, req_end;
-    uint32_t read_start, read_end;
-    uint32_t write_start, write_end;
-    bool missing;
-};
-
-void split_stripes(uint64_t pg_minsize, uint32_t bs_block_size, uint32_t start, uint32_t len, osd_rmw_stripe_t *stripes);
-
-void reconstruct_stripe(osd_rmw_stripe_t *stripes, int pg_size, int role);
-
-int extend_missing_stripes(osd_rmw_stripe_t *stripes, osd_num_t *osd_set, int minsize, int size);
-
-void* alloc_read_buffer(osd_rmw_stripe_t *stripes, int read_pg_size, uint64_t add_size);
-
-void* calc_rmw(void *request_buf, osd_rmw_stripe_t *stripes, uint64_t *read_osd_set,
-    uint64_t pg_size, uint64_t pg_minsize, uint64_t pg_cursize, uint64_t *write_osd_set, uint64_t chunk_size);
-
-void calc_rmw_parity(osd_rmw_stripe_t *stripes, int pg_size, uint64_t *read_osd_set, uint64_t *write_osd_set, uint32_t chunk_size);

osd_rmw_test.cpp

@@ -1,360 +0,0 @@
-#include <string.h>
-#include "osd_rmw.cpp"
-#include "test_pattern.h"
-
-void dump_stripes(osd_rmw_stripe_t *stripes, int pg_size);
-void test1();
-void test4();
-void test5();
-void test6();
-void test7();
-void test8();
-void test9();
-
-/***
-
-Cases:
-
-1. split(offset=128K-4K, len=8K)
-   = [ [ 128K-4K, 128K ], [ 0, 4K ], [ 0, 0 ] ]
-
-2. read(offset=128K-4K, len=8K, osd_set=[1,0,3])
-   = { read: [ [ 0, 128K ], [ 0, 4K ], [ 0, 4K ] ] }
-
-3. cover_read(0, 128K, { req: [ 128K-4K, 4K ] })
-   = { read: [ 0, 128K-4K ] }
-
-4. write(offset=128K-4K, len=8K, osd_set=[1,0,3])
-   = {
-     read: [ [ 0, 128K ], [ 4K, 128K ], [ 4K, 128K ] ],
-     write: [ [ 128K-4K, 128K ], [ 0, 4K ], [ 0, 128K ] ],
-     input buffer: [ write0, write1 ],
-     rmw buffer: [ write2, read0, read1, read2 ],
-   }
-   + check write2 buffer
-
-5. write(offset=0, len=128K+64K, osd_set=[1,0,3])
-   = {
-     req: [ [ 0, 128K ], [ 0, 64K ], [ 0, 0 ] ],
-     read: [ [ 64K, 128K ], [ 64K, 128K ], [ 64K, 128K ] ],
-     write: [ [ 0, 128K ], [ 0, 64K ], [ 0, 128K ] ],
-     input buffer: [ write0, write1 ],
-     rmw buffer: [ write2, read0, read1, read2 ],
-   }
-
-6. write(offset=0, len=128K+64K, osd_set=[1,2,3])
-   = {
-     req: [ [ 0, 128K ], [ 0, 64K ], [ 0, 0 ] ],
-     read: [ [ 0, 0 ], [ 64K, 128K ], [ 0, 0 ] ],
-     write: [ [ 0, 128K ], [ 0, 64K ], [ 0, 128K ] ],
-     input buffer: [ write0, write1 ],
-     rmw buffer: [ write2, read1 ],
-   }
-
-7. calc_rmw(offset=128K-4K, len=8K, osd_set=[1,0,3], write_set=[1,2,3])
-   = {
-     read: [ [ 0, 128K ], [ 0, 128K ], [ 0, 128K ] ],
-     write: [ [ 128K-4K, 128K ], [ 0, 4K ], [ 0, 128K ] ],
-     input buffer: [ write0, write1 ],
-     rmw buffer: [ write2, read0, read1, read2 ],
-   }
-   then, after calc_rmw_parity(): {
-     write: [ [ 128K-4K, 128K ], [ 0, 128K ], [ 0, 128K ] ],
-     write1==read1,
-   }
-   + check write1 buffer
-   + check write2 buffer
-
-8. calc_rmw(offset=0, len=128K+4K, osd_set=[0,2,3], write_set=[1,2,3])
-   = {
-     read: [ [ 0, 0 ], [ 4K, 128K ], [ 0, 0 ] ],
-     write: [ [ 0, 128K ], [ 0, 4K ], [ 0, 128K ] ],
-     input buffer: [ write0, write1 ],
-     rmw buffer: [ write2, read1 ],
-   }
-   + check write2 buffer
-
-9. object recovery case:
-   calc_rmw(offset=0, len=0, read_osd_set=[0,2,3], write_osd_set=[1,2,3])
-   = {
-     read: [ [ 0, 128K ], [ 0, 128K ], [ 0, 128K ] ],
-     write: [ [ 0, 0 ], [ 0, 0 ], [ 0, 0 ] ],
-     input buffer: NULL,
-     rmw buffer: [ read0, read1, read2 ],
-   }
-   then, after calc_rmw_parity(): {
-     write: [ [ 0, 128K ], [ 0, 0 ], [ 0, 0 ] ],
-     write0==read0,
-   }
-   + check write0 buffer
-
-***/
-
-int main(int narg, char *args[])
-{
-    // Test 1
-    test1();
-    // Test 4
-    test4();
-    // Test 5
-    test5();
-    // Test 6
-    test6();
-    // Test 7
-    test7();
-    // Test 8
-    test8();
-    // Test 9
-    test9();
-    // End
-    printf("all ok\n");
-    return 0;
-}
-
-void dump_stripes(osd_rmw_stripe_t *stripes, int pg_size)
-{
-    printf("request");
-    for (int i = 0; i < pg_size; i++)
-    {
-        printf(" {%uK-%uK}", stripes[i].req_start/1024, stripes[i].req_end/1024);
-    }
-    printf("\n");
-    printf("read");
-    for (int i = 0; i < pg_size; i++)
-    {
-        printf(" {%uK-%uK}", stripes[i].read_start/1024, stripes[i].read_end/1024);
-    }
-    printf("\n");
-    printf("write");
-    for (int i = 0; i < pg_size; i++)
-    {
-        printf(" {%uK-%uK}", stripes[i].write_start/1024, stripes[i].write_end/1024);
-    }
-    printf("\n");
-}
-
-void test1()
-{
-    osd_num_t osd_set[3] = { 1, 0, 3 };
-    osd_rmw_stripe_t stripes[3] = { 0 };
-    // Test 1.1
-    split_stripes(2, 128*1024, 128*1024-4096, 8192, stripes);
-    assert(stripes[0].req_start == 128*1024-4096 && stripes[0].req_end == 128*1024);
-    assert(stripes[1].req_start == 0 && stripes[1].req_end == 4096);
-    assert(stripes[2].req_end == 0);
-    // Test 1.2
-    for (int i = 0; i < 3; i++)
-    {
-        stripes[i].read_start = stripes[i].req_start;
-        stripes[i].read_end = stripes[i].req_end;
-    }
-    assert(extend_missing_stripes(stripes, osd_set, 2, 3) == 0);
-    assert(stripes[0].read_start == 0 && stripes[0].read_end == 128*1024);
-    assert(stripes[2].read_start == 0 && stripes[2].read_end == 4096);
-    // Test 1.3
-    stripes[0] = { .req_start = 128*1024-4096, .req_end = 128*1024 };
-    cover_read(0, 128*1024, stripes[0]);
-    assert(stripes[0].read_start == 0 && stripes[0].read_end == 128*1024-4096);
-}
-
-void test4()
-{
-    osd_num_t osd_set[3] = { 1, 0, 3 };
-    osd_rmw_stripe_t stripes[3] = { 0 };
-    // Test 4.1
-    split_stripes(2, 128*1024, 128*1024-4096, 8192, stripes);
-    void* write_buf = malloc(8192);
-    void* rmw_buf = calc_rmw(write_buf, stripes, osd_set, 3, 2, 2, osd_set, 128*1024);
-    assert(stripes[0].read_start == 0 && stripes[0].read_end == 128*1024);
-    assert(stripes[1].read_start == 4096 && stripes[1].read_end == 128*1024);
-    assert(stripes[2].read_start == 4096 && stripes[2].read_end == 128*1024);
-    assert(stripes[0].write_start == 128*1024-4096 && stripes[0].write_end == 128*1024);
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 4096);
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 128*1024);
-    assert(stripes[0].read_buf == rmw_buf+128*1024);
-    assert(stripes[1].read_buf == rmw_buf+128*1024*2);
-    assert(stripes[2].read_buf == rmw_buf+128*1024*3-4096);
-    assert(stripes[0].write_buf == write_buf);
-    assert(stripes[1].write_buf == write_buf+4096);
-    assert(stripes[2].write_buf == rmw_buf);
-    // Test 4.2
-    set_pattern(write_buf, 8192, PATTERN0);
-    set_pattern(stripes[0].read_buf, 128*1024, PATTERN1); // old data
-    set_pattern(stripes[1].read_buf, 128*1024-4096, UINT64_MAX); // didn't read it, it's missing
-    set_pattern(stripes[2].read_buf, 128*1024-4096, 0); // old parity = 0
-    calc_rmw_parity(stripes, 3, osd_set, osd_set, 128*1024);
-    check_pattern(stripes[2].write_buf, 4096, PATTERN0^PATTERN1); // new parity
-    check_pattern(stripes[2].write_buf+4096, 128*1024-4096*2, 0); // new parity
-    check_pattern(stripes[2].write_buf+128*1024-4096, 4096, PATTERN0^PATTERN1); // new parity
-    free(rmw_buf);
-    free(write_buf);
-}
-
-void test5()
-{
-    osd_num_t osd_set[3] = { 1, 0, 3 };
-    osd_rmw_stripe_t stripes[3] = { 0 };
-    // Test 5.1
-    split_stripes(2, 128*1024, 0, 64*1024*3, stripes);
-    assert(stripes[0].req_start == 0 && stripes[0].req_end == 128*1024);
-    assert(stripes[1].req_start == 0 && stripes[1].req_end == 64*1024);
-    assert(stripes[2].req_end == 0);
-    // Test 5.2
-    void *write_buf = malloc(64*1024*3);
-    void *rmw_buf = calc_rmw(write_buf, stripes, osd_set, 3, 2, 2, osd_set, 128*1024);
-    assert(stripes[0].read_start == 64*1024 && stripes[0].read_end == 128*1024);
-    assert(stripes[1].read_start == 64*1024 && stripes[1].read_end == 128*1024);
-    assert(stripes[2].read_start == 64*1024 && stripes[2].read_end == 128*1024);
-    assert(stripes[0].write_start == 0 && stripes[0].write_end == 128*1024);
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 64*1024);
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 128*1024);
-    assert(stripes[0].read_buf == rmw_buf+128*1024);
-    assert(stripes[1].read_buf == rmw_buf+64*3*1024);
-    assert(stripes[2].read_buf == rmw_buf+64*4*1024);
-    assert(stripes[0].write_buf == write_buf);
-    assert(stripes[1].write_buf == write_buf+128*1024);
-    assert(stripes[2].write_buf == rmw_buf);
-    free(rmw_buf);
-    free(write_buf);
-}
-
-void test6()
-{
-    osd_num_t osd_set[3] = { 1, 2, 3 };
-    osd_rmw_stripe_t stripes[3] = { 0 };
-    // Test 6.1
-    split_stripes(2, 128*1024, 0, 64*1024*3, stripes);
-    void *write_buf = malloc(64*1024*3);
-    void *rmw_buf = calc_rmw(write_buf, stripes, osd_set, 3, 2, 3, osd_set, 128*1024);
-    assert(stripes[0].read_end == 0);
-    assert(stripes[1].read_start == 64*1024 && stripes[1].read_end == 128*1024);
-    assert(stripes[2].read_end == 0);
-    assert(stripes[0].write_start == 0 && stripes[0].write_end == 128*1024);
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 64*1024);
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 128*1024);
-    assert(stripes[0].read_buf == 0);
-    assert(stripes[1].read_buf == rmw_buf+128*1024);
-    assert(stripes[2].read_buf == 0);
-    assert(stripes[0].write_buf == write_buf);
-    assert(stripes[1].write_buf == write_buf+128*1024);
-    assert(stripes[2].write_buf == rmw_buf);
-    free(rmw_buf);
-    free(write_buf);
-}
-
-void test7()
-{
-    osd_num_t osd_set[3] = { 1, 0, 3 };
-    osd_num_t write_osd_set[3] = { 1, 2, 3 };
-    osd_rmw_stripe_t stripes[3] = { 0 };
-    // Test 7.1
-    split_stripes(2, 128*1024, 128*1024-4096, 8192, stripes);
-    void *write_buf = malloc(8192);
-    void *rmw_buf = calc_rmw(write_buf, stripes, osd_set, 3, 2, 2, write_osd_set, 128*1024);
-    assert(stripes[0].read_start == 0 && stripes[0].read_end == 128*1024);
-    assert(stripes[1].read_start == 0 && stripes[1].read_end == 128*1024);
-    assert(stripes[2].read_start == 0 && stripes[2].read_end == 128*1024);
-    assert(stripes[0].write_start == 128*1024-4096 && stripes[0].write_end == 128*1024);
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 4096);
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 128*1024);
-    assert(stripes[0].read_buf == rmw_buf+128*1024);
-    assert(stripes[1].read_buf == rmw_buf+128*1024*2);
-    assert(stripes[2].read_buf == rmw_buf+128*1024*3);
-    assert(stripes[0].write_buf == write_buf);
-    assert(stripes[1].write_buf == write_buf+4096);
-    assert(stripes[2].write_buf == rmw_buf);
-    // Test 7.2
-    set_pattern(write_buf, 8192, PATTERN0);
-    set_pattern(stripes[0].read_buf, 128*1024, PATTERN1); // old data
-    set_pattern(stripes[1].read_buf, 128*1024, UINT64_MAX); // didn't read it, it's missing
-    set_pattern(stripes[2].read_buf, 128*1024, 0); // old parity = 0
-    calc_rmw_parity(stripes, 3, osd_set, write_osd_set, 128*1024);
-    assert(stripes[0].write_start == 128*1024-4096 && stripes[0].write_end == 128*1024);
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 128*1024);
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 128*1024);
-    assert(stripes[1].write_buf == stripes[1].read_buf);
-    check_pattern(stripes[1].write_buf, 4096, PATTERN0);
-    check_pattern(stripes[1].write_buf+4096, 128*1024-4096, PATTERN1);
-    check_pattern(stripes[2].write_buf, 4096, PATTERN0^PATTERN1); // new parity
-    check_pattern(stripes[2].write_buf+4096, 128*1024-4096*2, 0); // new parity
-    check_pattern(stripes[2].write_buf+128*1024-4096, 4096, PATTERN0^PATTERN1); // new parity
-    free(rmw_buf);
-    free(write_buf);
-}
-
-void test8()
-{
-    osd_num_t osd_set[3] = { 0, 2, 3 };
-    osd_num_t write_osd_set[3] = { 1, 2, 3 };
-    osd_rmw_stripe_t stripes[3] = { 0 };
-    // Test 8.1
-    split_stripes(2, 128*1024, 0, 128*1024+4096, stripes);
-    void *write_buf = malloc(128*1024+4096);
-    void *rmw_buf = calc_rmw(write_buf, stripes, osd_set, 3, 2, 2, write_osd_set, 128*1024);
-    assert(stripes[0].read_start == 0 && stripes[0].read_end == 0);
-    assert(stripes[1].read_start == 4096 && stripes[1].read_end == 128*1024);
-    assert(stripes[2].read_start == 0 && stripes[2].read_end == 0);
-    assert(stripes[0].write_start == 0 && stripes[0].write_end == 128*1024);
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 4096);
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 128*1024);
-    assert(stripes[0].read_buf == NULL);
-    assert(stripes[1].read_buf == rmw_buf+128*1024);
-    assert(stripes[2].read_buf == NULL);
-    assert(stripes[0].write_buf == write_buf);
-    assert(stripes[1].write_buf == write_buf+128*1024);
-    assert(stripes[2].write_buf == rmw_buf);
-    // Test 8.2
-    set_pattern(write_buf, 128*1024+4096, PATTERN0);
-    set_pattern(stripes[1].read_buf, 128*1024-4096, PATTERN1);
-    calc_rmw_parity(stripes, 3, osd_set, write_osd_set, 128*1024);
-    assert(stripes[0].write_start == 0 && stripes[0].write_end == 128*1024); // recheck again
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 4096);     // recheck again
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 128*1024); // recheck again
-    assert(stripes[0].write_buf == write_buf);                               // recheck again
-    assert(stripes[1].write_buf == write_buf+128*1024);                      // recheck again
-    assert(stripes[2].write_buf == rmw_buf);                                 // recheck again
-    check_pattern(stripes[2].write_buf, 4096, 0); // new parity
-    check_pattern(stripes[2].write_buf+4096, 128*1024-4096, PATTERN0^PATTERN1); // new parity
-    free(rmw_buf);
-    free(write_buf);
-}
-
-void test9()
-{
-    osd_num_t osd_set[3] = { 0, 2, 3 };
-    osd_num_t write_osd_set[3] = { 1, 2, 3 };
-    osd_rmw_stripe_t stripes[3] = { 0 };
-    // Test 9.0
-    split_stripes(2, 128*1024, 64*1024, 0, stripes);
-    assert(stripes[0].req_start == 0 && stripes[0].req_end == 0);
-    assert(stripes[1].req_start == 0 && stripes[1].req_end == 0);
-    assert(stripes[2].req_start == 0 && stripes[2].req_end == 0);
-    // Test 9.1
-    void *write_buf = NULL;
-    void *rmw_buf = calc_rmw(write_buf, stripes, osd_set, 3, 2, 3, write_osd_set, 128*1024);
-    assert(stripes[0].read_start == 0 && stripes[0].read_end == 128*1024);
-    assert(stripes[1].read_start == 0 && stripes[1].read_end == 128*1024);
-    assert(stripes[2].read_start == 0 && stripes[2].read_end == 128*1024);
-    assert(stripes[0].write_start == 0 && stripes[0].write_end == 0);
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 0);
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 0);
-    assert(stripes[0].read_buf == rmw_buf);
-    assert(stripes[1].read_buf == rmw_buf+128*1024);
-    assert(stripes[2].read_buf == rmw_buf+128*1024*2);
-    assert(stripes[0].write_buf == NULL);
-    assert(stripes[1].write_buf == NULL);
-    assert(stripes[2].write_buf == NULL);
-    // Test 8.2
-    set_pattern(stripes[1].read_buf, 128*1024, 0);
-    set_pattern(stripes[2].read_buf, 128*1024, PATTERN1);
-    calc_rmw_parity(stripes, 3, osd_set, write_osd_set, 128*1024);
-    assert(stripes[0].write_start == 0 && stripes[0].write_end == 128*1024);
-    assert(stripes[1].write_start == 0 && stripes[1].write_end == 0);
-    assert(stripes[2].write_start == 0 && stripes[2].write_end == 0);
-    assert(stripes[0].write_buf == rmw_buf);
-    assert(stripes[1].write_buf == NULL);
-    assert(stripes[2].write_buf == NULL);
-    check_pattern(stripes[0].read_buf, 128*1024, PATTERN1);
-    check_pattern(stripes[0].write_buf, 128*1024, PATTERN1);
-    free(rmw_buf);
-}

osd_secondary.cpp

@@ -1,134 +0,0 @@
-#include "osd.h"
-
-#include "json11/json11.hpp"
-
-void osd_t::secondary_op_callback(osd_op_t *op)
-{
-    if (op->req.hdr.opcode == OSD_OP_SECONDARY_READ ||
-        op->req.hdr.opcode == OSD_OP_SECONDARY_WRITE)
-    {
-        op->reply.sec_rw.version = op->bs_op->version;
-    }
-    else if (op->req.hdr.opcode == OSD_OP_SECONDARY_DELETE)
-    {
-        op->reply.sec_del.version = op->bs_op->version;
-    }
-    if (op->req.hdr.opcode == OSD_OP_SECONDARY_READ &&
-        op->bs_op->retval > 0)
-    {
-        op->send_list.push_back(op->buf, op->bs_op->retval);
-    }
-    else if (op->req.hdr.opcode == OSD_OP_SECONDARY_LIST)
-    {
-        // allocated by blockstore
-        op->buf = op->bs_op->buf;
-        if (op->bs_op->retval > 0)
-        {
-            op->send_list.push_back(op->buf, op->bs_op->retval * sizeof(obj_ver_id));
-        }
-        op->reply.sec_list.stable_count = op->bs_op->version;
-    }
-    int retval = op->bs_op->retval;
-    delete op->bs_op;
-    op->bs_op = NULL;
-    finish_op(op, retval);
-}
-
-void osd_t::exec_secondary(osd_op_t *cur_op)
-{
-    cur_op->bs_op = new blockstore_op_t();
-    cur_op->bs_op->callback = [this, cur_op](blockstore_op_t* bs_op) { secondary_op_callback(cur_op); };
-    cur_op->bs_op->opcode = (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_READ ? BS_OP_READ
-        : (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_WRITE ? BS_OP_WRITE
-        : (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_SYNC ? BS_OP_SYNC
-        : (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_STABILIZE ? BS_OP_STABLE
-        : (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_ROLLBACK ? BS_OP_ROLLBACK
-        : (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_DELETE ? BS_OP_DELETE
-        : (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_LIST ? BS_OP_LIST
-        : -1)))))));
-    if (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_READ ||
-        cur_op->req.hdr.opcode == OSD_OP_SECONDARY_WRITE)
-    {
-        cur_op->bs_op->oid = cur_op->req.sec_rw.oid;
-        cur_op->bs_op->version = cur_op->req.sec_rw.version;
-        cur_op->bs_op->offset = cur_op->req.sec_rw.offset;
-        cur_op->bs_op->len = cur_op->req.sec_rw.len;
-        cur_op->bs_op->buf = cur_op->buf;
-#ifdef OSD_STUB
-        cur_op->bs_op->retval = cur_op->bs_op->len;
-#endif
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_DELETE)
-    {
-        cur_op->bs_op->oid = cur_op->req.sec_del.oid;
-        cur_op->bs_op->version = cur_op->req.sec_del.version;
-#ifdef OSD_STUB
-        cur_op->bs_op->retval = 0;
-#endif
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_STABILIZE ||
-        cur_op->req.hdr.opcode == OSD_OP_SECONDARY_ROLLBACK)
-    {
-        cur_op->bs_op->len = cur_op->req.sec_stab.len/sizeof(obj_ver_id);
-        cur_op->bs_op->buf = cur_op->buf;
-#ifdef OSD_STUB
-        cur_op->bs_op->retval = 0;
-#endif
-    }
-    else if (cur_op->req.hdr.opcode == OSD_OP_SECONDARY_LIST)
-    {
-        if (cur_op->req.sec_list.pg_count < cur_op->req.sec_list.list_pg)
-        {
-            // requested pg number is greater than total pg count
-            cur_op->bs_op->retval = -EINVAL;
-            secondary_op_callback(cur_op);
-            return;
-        }
-        cur_op->bs_op->oid.stripe = cur_op->req.sec_list.pg_stripe_size;
-        cur_op->bs_op->len = cur_op->req.sec_list.pg_count;
-        cur_op->bs_op->offset = cur_op->req.sec_list.list_pg - 1;
-#ifdef OSD_STUB
-        cur_op->bs_op->retval = 0;
-        cur_op->bs_op->buf = NULL;
-#endif
-    }
-#ifdef OSD_STUB
-    secondary_op_callback(cur_op);
-#else
-    bs->enqueue_op(cur_op->bs_op);
-#endif
-}
-
-void osd_t::exec_show_config(osd_op_t *cur_op)
-{
-    // FIXME: Send the real config, not its source
-    std::string cfg_str = json11::Json(config).dump();
-    cur_op->buf = malloc(cfg_str.size()+1);
-    memcpy(cur_op->buf, cfg_str.c_str(), cfg_str.size()+1);
-    cur_op->send_list.push_back(cur_op->buf, cfg_str.size()+1);
-    finish_op(cur_op, cfg_str.size()+1);
-}
-
-void osd_t::exec_sync_stab_all(osd_op_t *cur_op)
-{
-    // Sync and stabilize all objects
-    // This command is only valid for tests
-    cur_op->bs_op = new blockstore_op_t();
-    if (!allow_test_ops)
-    {
-        cur_op->bs_op->retval = -EINVAL;
-        secondary_op_callback(cur_op);
-        return;
-    }
-    cur_op->bs_op->opcode = BS_OP_SYNC_STAB_ALL;
-    cur_op->bs_op->callback = [this, cur_op](blockstore_op_t *bs_op)
-    {
-        secondary_op_callback(cur_op);
-    };
-#ifdef OSD_STUB
-    cur_op->bs_op->retval = 0;
-    secondary_op_callback(cur_op);
-#else
-    bs->enqueue_op(cur_op->bs_op);
-#endif
-}

patches/cinder-vitastor.py

@@ -0,0 +1,948 @@
+# Vitastor Driver for OpenStack Cinder
+#
+# --------------------------------------------
+# Install as cinder/volume/drivers/vitastor.py
+# --------------------------------------------
+#
+# Copyright 2020 Vitaliy Filippov
+#
+# Licensed under the Apache License, Version 2.0 (the "License"); you may
+# not use this file except in compliance with the License. You may obtain
+# a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+# License for the specific language governing permissions and limitations
+# under the License.
+"""Cinder Vitastor Driver"""
+
+import binascii
+import base64
+import errno
+import json
+import math
+import os
+import tempfile
+
+from castellan import key_manager
+from oslo_config import cfg
+from oslo_log import log as logging
+from oslo_service import loopingcall
+from oslo_concurrency import processutils
+from oslo_utils import encodeutils
+from oslo_utils import excutils
+from oslo_utils import fileutils
+from oslo_utils import units
+import six
+from six.moves.urllib import request
+
+from cinder import exception
+from cinder.i18n import _
+from cinder.image import image_utils
+from cinder import interface
+from cinder import objects
+from cinder.objects import fields
+from cinder import utils
+from cinder.volume import configuration
+from cinder.volume import driver
+from cinder.volume import volume_utils
+
+VERSION = '0.6.5'
+
+LOG = logging.getLogger(__name__)
+
+VITASTOR_OPTS = [
+    cfg.StrOpt(
+        'vitastor_config_path',
+        default='/etc/vitastor/vitastor.conf',
+        help='Vitastor configuration file path'
+    ),
+    cfg.StrOpt(
+        'vitastor_etcd_address',
+        default='',
+        help='Vitastor etcd address(es)'),
+    cfg.StrOpt(
+        'vitastor_etcd_prefix',
+        default='/vitastor',
+        help='Vitastor etcd prefix'
+    ),
+    cfg.StrOpt(
+        'vitastor_pool_id',
+        default='',
+        help='Vitastor pool ID to use for volumes'
+    ),
+    # FIXME exclusive_cinder_pool ?
+]
+
+CONF = cfg.CONF
+CONF.register_opts(VITASTOR_OPTS, group = configuration.SHARED_CONF_GROUP)
+
+class VitastorDriverException(exception.VolumeDriverException):
+    message = _("Vitastor Cinder driver failure: %(reason)s")
+
+@interface.volumedriver
+class VitastorDriver(driver.CloneableImageVD,
+    driver.ManageableVD, driver.ManageableSnapshotsVD,
+    driver.BaseVD):
+    """Implements Vitastor volume commands."""
+
+    cfg = {}
+    _etcd_urls = []
+
+    def __init__(self, active_backend_id = None, *args, **kwargs):
+        super(VitastorDriver, self).__init__(*args, **kwargs)
+        self.configuration.append_config_values(VITASTOR_OPTS)
+
+    @classmethod
+    def get_driver_options(cls):
+        additional_opts = cls._get_oslo_driver_opts(
+            'reserved_percentage',
+            'max_over_subscription_ratio',
+            'volume_dd_blocksize'
+        )
+        return VITASTOR_OPTS + additional_opts
+
+    def do_setup(self, context):
+        """Performs initialization steps that could raise exceptions."""
+        super(VitastorDriver, self).do_setup(context)
+        # Make sure configuration is in UTF-8
+        for attr in [ 'config_path', 'etcd_address', 'etcd_prefix', 'pool_id' ]:
+            val = self.configuration.safe_get('vitastor_'+attr)
+            if val is not None:
+                self.cfg[attr] = utils.convert_str(val)
+        self.cfg = self._load_config(self.cfg)
+
+    def _load_config(self, cfg):
+        # Try to load configuration file
+        try:
+            f = open(cfg['config_path'] or '/etc/vitastor/vitastor.conf')
+            conf = json.loads(f.read())
+            f.close()
+            for k in conf:
+                cfg[k] = cfg.get(k, conf[k])
+        except:
+            pass
+        if isinstance(cfg['etcd_address'], str):
+            cfg['etcd_address'] = cfg['etcd_address'].split(',')
+        # Sanitize etcd URLs
+        for i, etcd_url in enumerate(cfg['etcd_address']):
+            ssl = False
+            if etcd_url.lower().startswith('http://'):
+                etcd_url = etcd_url[7:]
+            elif etcd_url.lower().startswith('https://'):
+                etcd_url = etcd_url[8:]
+                ssl = True
+            if etcd_url.find('/') < 0:
+                etcd_url += '/v3'
+            if ssl:
+                etcd_url = 'https://'+etcd_url
+            else:
+                etcd_url = 'http://'+etcd_url
+            cfg['etcd_address'][i] = etcd_url
+        return cfg
+
+    def check_for_setup_error(self):
+        """Returns an error if prerequisites aren't met."""
+
+    def _encode_etcd_key(self, key):
+        if not isinstance(key, bytes):
+            key = str(key).encode('utf-8')
+        return base64.b64encode(self.cfg['etcd_prefix'].encode('utf-8')+b'/'+key).decode('utf-8')
+
+    def _encode_etcd_value(self, value):
+        if not isinstance(value, bytes):
+            value = str(value).encode('utf-8')
+        return base64.b64encode(value).decode('utf-8')
+
+    def _encode_etcd_requests(self, obj):
+        for v in obj:
+            for rt in v:
+                if 'key' in v[rt]:
+                    v[rt]['key'] = self._encode_etcd_key(v[rt]['key'])
+                if 'range_end' in v[rt]:
+                    v[rt]['range_end'] = self._encode_etcd_key(v[rt]['range_end'])
+                if 'value' in v[rt]:
+                    v[rt]['value'] = self._encode_etcd_value(v[rt]['value'])
+
+    def _etcd_txn(self, params):
+        if 'compare' in params:
+            for v in params['compare']:
+                if 'key' in v:
+                    v['key'] = self._encode_etcd_key(v['key'])
+        if 'failure' in params:
+            self._encode_etcd_requests(params['failure'])
+        if 'success' in params:
+            self._encode_etcd_requests(params['success'])
+        body = json.dumps(params).encode('utf-8')
+        headers = {
+            'Content-Type': 'application/json'
+        }
+        err = None
+        for etcd_url in self.cfg['etcd_address']:
+            try:
+                resp = request.urlopen(request.Request(etcd_url+'/kv/txn', body, headers), timeout = 5)
+                data = json.loads(resp.read())
+                if 'responses' not in data:
+                    data['responses'] = []
+                for i, resp in enumerate(data['responses']):
+                    if 'response_range' in resp:
+                        if 'kvs' not in resp['response_range']:
+                            resp['response_range']['kvs'] = []
+                        for kv in resp['response_range']['kvs']:
+                            kv['key'] = base64.b64decode(kv['key'].encode('utf-8')).decode('utf-8')
+                            if kv['key'].startswith(self.cfg['etcd_prefix']+'/'):
+                                kv['key'] = kv['key'][len(self.cfg['etcd_prefix'])+1 : ]
+                            kv['value'] = json.loads(base64.b64decode(kv['value'].encode('utf-8')))
+                    if len(resp.keys()) != 1:
+                        LOG.exception('unknown responses['+str(i)+'] format: '+json.dumps(resp))
+                    else:
+                        resp = data['responses'][i] = resp[list(resp.keys())[0]]
+                return data
+            except Exception as e:
+                LOG.exception('error calling etcd transaction: '+body.decode('utf-8')+'\nerror: '+str(e))
+                err = e
+        raise err
+
+    def _etcd_foreach(self, prefix, add_fn):
+        total = 0
+        batch = 1000
+        begin = prefix+'/'
+        while True:
+            resp = self._etcd_txn({ 'success': [
+                { 'request_range': {
+                    'key': begin,
+                    'range_end': prefix+'0',
+                    'limit': batch+1,
+                } },
+            ] })
+            i = 0
+            while i < batch and i < len(resp['responses'][0]['kvs']):
+                kv = resp['responses'][0]['kvs'][i]
+                add_fn(kv)
+                i += 1
+            if len(resp['responses'][0]['kvs']) <= batch:
+                break
+            begin = resp['responses'][0]['kvs'][batch]['key']
+        return total
+
+    def _update_volume_stats(self):
+        location_info = json.dumps({
+            'config': self.configuration.vitastor_config_path,
+            'etcd_address': self.configuration.vitastor_etcd_address,
+            'etcd_prefix': self.configuration.vitastor_etcd_prefix,
+            'pool_id': self.configuration.vitastor_pool_id,
+        })
+
+        stats = {
+            'vendor_name': 'Vitastor',
+            'driver_version': self.VERSION,
+            'storage_protocol': 'vitastor',
+            'total_capacity_gb': 'unknown',
+            'free_capacity_gb': 'unknown',
+            # FIXME check if safe_get is required
+            'reserved_percentage': self.configuration.safe_get('reserved_percentage'),
+            'multiattach': True,
+            'thin_provisioning_support': True,
+            'max_over_subscription_ratio': self.configuration.safe_get('max_over_subscription_ratio'),
+            'location_info': location_info,
+            'backend_state': 'down',
+            'volume_backend_name': self.configuration.safe_get('volume_backend_name') or 'vitastor',
+            'replication_enabled': False,
+        }
+
+        try:
+            pool_stats = self._etcd_txn({ 'success': [
+                { 'request_range': { 'key': 'pool/stats/'+str(self.cfg['pool_id']) } }
+            ] })
+            total_provisioned = 0
+            def add_total(kv):
+                nonlocal total_provisioned
+                if kv['key'].find('@') >= 0:
+                    total_provisioned += kv['value']['size']
+            self._etcd_foreach('config/inode/'+str(self.cfg['pool_id']), lambda kv: add_total(kv))
+            stats['provisioned_capacity_gb'] = round(total_provisioned/1024.0/1024.0/1024.0, 2)
+            pool_stats = pool_stats['responses'][0]['kvs']
+            if len(pool_stats):
+                pool_stats = pool_stats[0]
+                stats['free_capacity_gb'] = round(1024.0*(pool_stats['total_raw_tb']-pool_stats['used_raw_tb'])/pool_stats['raw_to_usable'], 2)
+                stats['total_capacity_gb'] = round(1024.0*pool_stats['total_raw_tb'], 2)
+            stats['backend_state'] = 'up'
+        except Exception as e:
+            # just log and return unknown capacities
+            LOG.exception('error getting vitastor pool stats: '+str(e))
+
+        self._stats = stats
+
+    def _next_id(self, resp):
+        if len(resp['kvs']) == 0:
+            return (1, 0)
+        else:
+            return (1 + resp['kvs'][0]['value'], resp['kvs'][0]['mod_revision'])
+
+    def create_volume(self, volume):
+        """Creates a logical volume."""
+
+        size = int(volume.size) * units.Gi
+        # FIXME: Check if convert_str is really required
+        vol_name = utils.convert_str(volume.name)
+        if vol_name.find('@') >= 0 or vol_name.find('/') >= 0:
+            raise exception.VolumeBackendAPIException(data = '@ and / are forbidden in volume and snapshot names')
+
+        LOG.debug("creating volume '%s'", vol_name)
+
+        self._create_image(vol_name, { 'size': size })
+
+        if volume.encryption_key_id:
+            self._create_encrypted_volume(volume, volume.obj_context)
+
+        volume_update = {}
+        return volume_update
+
+    def _create_encrypted_volume(self, volume, context):
+        """Create a new LUKS encrypted image directly in Vitastor."""
+        vol_name = utils.convert_str(volume.name)
+        f, opts = self._encrypt_opts(volume, context)
+        # FIXME: Check if it works at all :-)
+        self._execute(
+            'qemu-img', 'convert', '-f', 'luks', *opts,
+            'vitastor:image='+vol_name.replace(':', '\\:')+self._qemu_args(),
+            '%sM' % (volume.size * 1024)
+        )
+        f.close()
+
+    def _encrypt_opts(self, volume, context):
+        encryption = volume_utils.check_encryption_provider(self.db, volume, context)
+        # Fetch the key associated with the volume and decode the passphrase
+        keymgr = key_manager.API(CONF)
+        key = keymgr.get(context, encryption['encryption_key_id'])
+        passphrase = binascii.hexlify(key.get_encoded()).decode('utf-8')
+        # Decode the dm-crypt style cipher spec into something qemu-img can use
+        cipher_spec = image_utils.decode_cipher(encryption['cipher'], encryption['key_size'])
+        tmp_dir = volume_utils.image_conversion_dir()
+        f = tempfile.NamedTemporaryFile(prefix = 'luks_', dir = tmp_dir)
+        f.write(passphrase)
+        f.flush()
+        return (f, [
+            '--object', 'secret,id=luks_sec,format=raw,file=%(passfile)s' % {'passfile': f.name},
+            '-o', 'key-secret=luks_sec,cipher-alg=%(cipher_alg)s,cipher-mode=%(cipher_mode)s,ivgen-alg=%(ivgen_alg)s' % cipher_spec,
+        ])
+
+    def create_snapshot(self, snapshot):
+        """Creates a volume snapshot."""
+
+        vol_name = utils.convert_str(snapshot.volume_name)
+        snap_name = utils.convert_str(snapshot.name)
+        if snap_name.find('@') >= 0 or snap_name.find('/') >= 0:
+            raise exception.VolumeBackendAPIException(data = '@ and / are forbidden in volume and snapshot names')
+        self._create_snapshot(vol_name, vol_name+'@'+snap_name)
+
+    def snapshot_revert_use_temp_snapshot(self):
+        """Disable the use of a temporary snapshot on revert."""
+        return False
+
+    def revert_to_snapshot(self, context, volume, snapshot):
+        """Revert a volume to a given snapshot."""
+
+        # FIXME Delete the image, then recreate it from the snapshot
+
+    def delete_snapshot(self, snapshot):
+        """Deletes a snapshot."""
+
+        vol_name = utils.convert_str(snapshot.volume_name)
+        snap_name = utils.convert_str(snapshot.name)
+
+        # Find the snapshot
+        resp = self._etcd_txn({ 'success': [
+            { 'request_range': { 'key': 'index/image/'+vol_name+'@'+snap_name } },
+        ] })
+        if len(resp['responses'][0]['kvs']) == 0:
+            raise exception.SnapshotNotFound(snapshot_id = snap_name)
+        inode_id = int(resp['responses'][0]['kvs'][0]['value']['id'])
+        pool_id = int(resp['responses'][0]['kvs'][0]['value']['pool_id'])
+        parents = {}
+        parents[(pool_id << 48) | (inode_id & 0xffffffffffff)] = True
+
+        # Check if there are child volumes
+        children = self._child_count(parents)
+        if children > 0:
+            raise exception.SnapshotIsBusy(snapshot_name = snap_name)
+
+        # FIXME: We can't delete snapshots because we can't merge layers yet
+        raise exception.VolumeBackendAPIException(data = 'Snapshot delete (layer merge) is not implemented yet')
+
+    def _child_count(self, parents):
+        children = 0
+        def add_child(kv):
+            nonlocal children
+            children += self._check_parent(kv, parents)
+        self._etcd_foreach('config/inode', lambda kv: add_child(kv))
+        return children
+
+    def _check_parent(self, kv, parents):
+        if 'parent_id' not in kv['value']:
+            return 0
+        parent_id = kv['value']['parent_id']
+        _, _, pool_id, inode_id = kv['key'].split('/')
+        parent_pool_id = pool_id
+        if 'parent_pool_id' in kv['value'] and kv['value']['parent_pool_id']:
+            parent_pool_id = kv['value']['parent_pool_id']
+        inode = (int(pool_id) << 48) | (int(inode_id) & 0xffffffffffff)
+        parent = (int(parent_pool_id) << 48) | (int(parent_id) & 0xffffffffffff)
+        if parent in parents and inode not in parents:
+            return 1
+        return 0
+
+    def create_cloned_volume(self, volume, src_vref):
+        """Create a cloned volume from another volume."""
+
+        size = int(volume.size) * units.Gi
+        src_name = utils.convert_str(src_vref.name)
+        dest_name = utils.convert_str(volume.name)
+        if dest_name.find('@') >= 0 or dest_name.find('/') >= 0:
+            raise exception.VolumeBackendAPIException(data = '@ and / are forbidden in volume and snapshot names')
+
+        # FIXME Do full copy if requested (cfg.disable_clone)
+
+        if src_vref.admin_metadata.get('readonly') == 'True':
+            # source volume is a volume-image cache entry or other readonly volume
+            # clone without intermediate snapshot
+            src = self._get_image(src_name)
+            LOG.debug("creating image '%s' from '%s'", dest_name, src_name)
+            new_cfg = self._create_image(dest_name, {
+                'size': size,
+                'parent_id': src['idx']['id'],
+                'parent_pool_id': src['idx']['pool_id'],
+            })
+            return {}
+
+        clone_snap = "%s@%s.clone_snap" % (src_name, dest_name)
+        make_img = True
+        if (volume.display_name and
+            volume.display_name.startswith('image-') and
+            src_vref.project_id != volume.project_id):
+            # idiotic openstack creates image-volume cache entries
+            # as clones of normal VM volumes... :-X prevent it :-D
+            clone_snap = dest_name
+            make_img = False
+
+        LOG.debug("creating layer '%s' under '%s'", clone_snap, src_name)
+        new_cfg = self._create_snapshot(src_name, clone_snap, True)
+        if make_img:
+            # Then create a clone from it
+            new_cfg = self._create_image(dest_name, {
+                'size': size,
+                'parent_id': new_cfg['parent_id'],
+                'parent_pool_id': new_cfg['parent_pool_id'],
+            })
+
+        return {}
+
+    def create_volume_from_snapshot(self, volume, snapshot):
+        """Creates a cloned volume from an existing snapshot."""
+
+        vol_name = utils.convert_str(volume.name)
+        snap_name = utils.convert_str(snapshot.name)
+
+        snap = self._get_image(vol_name+'@'+snap_name)
+        if not snap:
+            raise exception.SnapshotNotFound(snapshot_id = snap_name)
+        snap_inode_id = int(resp['responses'][0]['kvs'][0]['value']['id'])
+        snap_pool_id = int(resp['responses'][0]['kvs'][0]['value']['pool_id'])
+
+        size = snap['cfg']['size']
+        if int(volume.size):
+            size = int(volume.size) * units.Gi
+        new_cfg = self._create_image(vol_name, {
+            'size': size,
+            'parent_id': snap['idx']['id'],
+            'parent_pool_id': snap['idx']['pool_id'],
+        })
+
+        return {}
+
+    def _vitastor_args(self):
+        args = []
+        for k in [ 'config_path', 'etcd_address', 'etcd_prefix' ]:
+            v = self.configuration.safe_get('vitastor_'+k)
+            if v:
+                args.extend(['--'+k, v])
+        return args
+
+    def _qemu_args(self):
+        args = ''
+        for k in [ 'config_path', 'etcd_address', 'etcd_prefix' ]:
+            v = self.configuration.safe_get('vitastor_'+k)
+            kk = k
+            if kk == 'etcd_address':
+                # FIXME use etcd_address in qemu driver
+                kk = 'etcd_host'
+            if v:
+                args += ':'+kk+'='+v.replace(':', '\\:')
+        return args
+
+    def delete_volume(self, volume):
+        """Deletes a logical volume."""
+
+        vol_name = utils.convert_str(volume.name)
+
+        # Find the volume and all its snapshots
+        range_end = b'index/image/' + vol_name.encode('utf-8')
+        range_end = range_end[0 : len(range_end)-1] + six.int2byte(range_end[len(range_end)-1] + 1)
+        resp = self._etcd_txn({ 'success': [
+            { 'request_range': { 'key': 'index/image/'+vol_name, 'range_end': range_end } },
+        ] })
+        if len(resp['responses'][0]['kvs']) == 0:
+            # already deleted
+            LOG.info("volume %s no longer exists in backend", vol_name)
+            return
+        layers = resp['responses'][0]['kvs']
+        layer_ids = {}
+        for kv in layers:
+            inode_id = int(kv['value']['id'])
+            pool_id = int(kv['value']['pool_id'])
+            inode_pool_id = (pool_id << 48) | (inode_id & 0xffffffffffff)
+            layer_ids[inode_pool_id] = True
+
+        # Check if the volume has clones and raise 'busy' if so
+        children = self._child_count(layer_ids)
+        if children > 0:
+            raise exception.VolumeIsBusy(volume_name = vol_name)
+
+        # Clear data
+        for kv in layers:
+            args = [
+                'vitastor-rm', '--pool', str(kv['value']['pool_id']),
+                '--inode', str(kv['value']['id']), '--progress', '0',
+                *(self._vitastor_args())
+            ]
+            try:
+                self._execute(*args)
+            except processutils.ProcessExecutionError as exc:
+                LOG.error("Failed to remove layer "+kv['key']+": "+exc)
+                raise exception.VolumeBackendAPIException(data = exc.stderr)
+
+        # Delete all layers from etcd
+        requests = []
+        for kv in layers:
+            requests.append({ 'request_delete_range': { 'key': kv['key'] } })
+            requests.append({ 'request_delete_range': { 'key': 'config/inode/'+str(kv['value']['pool_id'])+'/'+str(kv['value']['id']) } })
+        self._etcd_txn({ 'success': requests })
+
+    def retype(self, context, volume, new_type, diff, host):
+        """Change extra type specifications for a volume."""
+
+        # FIXME Maybe (in the future) support multiple pools as different types
+        return True, {}
+
+    def ensure_export(self, context, volume):
+        """Synchronously recreates an export for a logical volume."""
+        pass
+
+    def create_export(self, context, volume, connector):
+        """Exports the volume."""
+        pass
+
+    def remove_export(self, context, volume):
+        """Removes an export for a logical volume."""
+        pass
+
+    def _create_image(self, vol_name, cfg):
+        pool_s = str(self.cfg['pool_id'])
+        image_id = 0
+        while image_id == 0:
+            # check if the image already exists and find a free ID
+            resp = self._etcd_txn({ 'success': [
+                { 'request_range': { 'key': 'index/image/'+vol_name } },
+                { 'request_range': { 'key': 'index/maxid/'+pool_s } },
+            ] })
+            if len(resp['responses'][0]['kvs']) > 0:
+                # already exists
+                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' already exists')
+            image_id, id_mod = self._next_id(resp['responses'][1])
+            # try to create the image
+            resp = self._etcd_txn({ 'compare': [
+                { 'target': 'MOD', 'mod_revision': id_mod, 'key': 'index/maxid/'+pool_s },
+                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+vol_name },
+                { 'target': 'VERSION', 'version': 0, 'key': 'config/inode/'+pool_s+'/'+str(image_id) },
+            ], 'success': [
+                { 'request_put': { 'key': 'index/maxid/'+pool_s, 'value': image_id } },
+                { 'request_put': { 'key': 'index/image/'+vol_name, 'value': json.dumps({
+                    'id': image_id, 'pool_id': self.cfg['pool_id']
+                }) } },
+                { 'request_put': { 'key': 'config/inode/'+pool_s+'/'+str(image_id), 'value': json.dumps({
+                    **cfg, 'name': vol_name,
+                }) } },
+            ] })
+            if not resp.get('succeeded'):
+                # repeat
+                image_id = 0
+
+    def _create_snapshot(self, vol_name, snap_vol_name, allow_existing = False):
+        while True:
+            # check if the image already exists and snapshot doesn't
+            resp = self._etcd_txn({ 'success': [
+                { 'request_range': { 'key': 'index/image/'+vol_name } },
+                { 'request_range': { 'key': 'index/image/'+snap_vol_name } },
+            ] })
+            if len(resp['responses'][0]['kvs']) == 0:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
+            if len(resp['responses'][1]['kvs']) > 0:
+                if allow_existing:
+                    snap_idx = resp['responses'][1]['kvs'][0]['value']
+                    resp = self._etcd_txn({ 'success': [
+                        { 'request_range': { 'key': 'config/inode/'+str(snap_idx['pool_id'])+'/'+str(snap_idx['id']) } },
+                    ] })
+                    if len(resp['responses'][0]['kvs']) == 0:
+                        raise exception.VolumeBackendAPIException(data =
+                            'Volume '+snap_vol_name+' is already indexed, but does not exist'
+                        )
+                    return resp['responses'][0]['kvs'][0]['value']
+                raise exception.VolumeBackendAPIException(
+                    data = 'Volume '+snap_vol_name+' already exists'
+                )
+            vol_idx = resp['responses'][0]['kvs'][0]['value']
+            vol_idx_mod = resp['responses'][0]['kvs'][0]['mod_revision']
+            # get image inode config and find a new ID
+            resp = self._etcd_txn({ 'success': [
+                { 'request_range': { 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']) } },
+                { 'request_range': { 'key': 'index/maxid/'+str(self.cfg['pool_id']) } },
+            ] })
+            if len(resp['responses'][0]['kvs']) == 0:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
+            vol_cfg = resp['responses'][0]['kvs'][0]['value']
+            vol_mod = resp['responses'][0]['kvs'][0]['mod_revision']
+            new_id, id_mod = self._next_id(resp['responses'][1])
+            # try to redirect image to the new inode
+            new_cfg = {
+                **vol_cfg, 'name': vol_name, 'parent_id': vol_idx['id'], 'parent_pool_id': vol_idx['pool_id']
+            }
+            resp = self._etcd_txn({ 'compare': [
+                { 'target': 'MOD', 'mod_revision': vol_idx_mod, 'key': 'index/image/'+vol_name },
+                { 'target': 'MOD', 'mod_revision': vol_mod, 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']) },
+                { 'target': 'MOD', 'mod_revision': id_mod, 'key': 'index/maxid/'+str(self.cfg['pool_id']) },
+                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+snap_vol_name },
+                { 'target': 'VERSION', 'version': 0, 'key': 'config/inode/'+str(self.cfg['pool_id'])+'/'+str(new_id) },
+            ], 'success': [
+                { 'request_put': { 'key': 'index/maxid/'+str(self.cfg['pool_id']), 'value': new_id } },
+                { 'request_put': { 'key': 'index/image/'+vol_name, 'value': json.dumps({
+                    'id': new_id, 'pool_id': self.cfg['pool_id']
+                }) } },
+                { 'request_put': { 'key': 'config/inode/'+str(self.cfg['pool_id'])+'/'+str(new_id), 'value': json.dumps(new_cfg) } },
+                { 'request_put': { 'key': 'index/image/'+snap_vol_name, 'value': json.dumps({
+                    'id': vol_idx['id'], 'pool_id': vol_idx['pool_id']
+                }) } },
+                { 'request_put': { 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']), 'value': json.dumps({
+                    **vol_cfg, 'name': snap_vol_name, 'readonly': True
+                }) } }
+            ] })
+            if resp.get('succeeded'):
+                return new_cfg
+
+    def initialize_connection(self, volume, connector):
+        data = {
+            'driver_volume_type': 'vitastor',
+            'data': {
+                'config_path': self.configuration.vitastor_config_path,
+                'etcd_address': self.configuration.vitastor_etcd_address,
+                'etcd_prefix': self.configuration.vitastor_etcd_prefix,
+                'name': volume.name,
+                'logical_block_size': 512,
+                'physical_block_size': 4096,
+            }
+        }
+        LOG.debug('connection data: %s', data)
+        return data
+
+    def terminate_connection(self, volume, connector, **kwargs):
+        pass
+
+    def clone_image(self, context, volume, image_location, image_meta, image_service):
+        if image_location:
+            # Note: image_location[0] is glance image direct_url.
+            # image_location[1] contains the list of all locations (including
+            # direct_url) or None if show_multiple_locations is False in
+            # glance configuration.
+            if image_location[1]:
+                url_locations = [location['url'] for location in image_location[1]]
+            else:
+                url_locations = [image_location[0]]
+            # iterate all locations to look for a cloneable one.
+            for url_location in url_locations:
+                if url_location and url_location.startswith('cinder://'):
+                    # The idea is to use cinder://<volume-id> Glance volumes as base images
+                    base_vol = self.db.volume_get(context, url_location[len('cinder://') : ])
+                    if not base_vol or base_vol.volume_type_id != volume.volume_type_id:
+                        continue
+                    size = int(volume.size) * units.Gi
+                    dest_name = utils.convert_str(volume.name)
+                    # Find or create the base snapshot
+                    snap_cfg = self._create_snapshot(base_vol.name, base_vol.name+'@.clone_snap', True)
+                    # Then create a clone from it
+                    new_cfg = self._create_image(dest_name, {
+                        'size': size,
+                        'parent_id': snap_cfg['parent_id'],
+                        'parent_pool_id': snap_cfg['parent_pool_id'],
+                    })
+                    return ({}, True)
+        return ({}, False)
+
+    def copy_image_to_encrypted_volume(self, context, volume, image_service, image_id):
+        self.copy_image_to_volume(context, volume, image_service, image_id, encrypted = True)
+
+    def copy_image_to_volume(self, context, volume, image_service, image_id, encrypted = False):
+        tmp_dir = volume_utils.image_conversion_dir()
+        with tempfile.NamedTemporaryFile(dir = tmp_dir) as tmp:
+            image_utils.fetch_to_raw(
+                context, image_service, image_id, tmp.name,
+                self.configuration.volume_dd_blocksize, size = volume.size
+            )
+            out_format = [ '-O', 'raw' ]
+            if encrypted:
+                key_file, opts = self._encrypt_opts(volume, context)
+                out_format = [ '-O', 'luks', *opts ]
+            dest_name = utils.convert_str(volume.name)
+            self._try_execute(
+                'qemu-img', 'convert', '-f', 'raw', tmp.name, *out_format,
+                'vitastor:image='+dest_name.replace(':', '\\:')+self._qemu_args()
+            )
+            if encrypted:
+                key_file.close()
+
+    def copy_volume_to_image(self, context, volume, image_service, image_meta):
+        tmp_dir = volume_utils.image_conversion_dir()
+        tmp_file = os.path.join(tmp_dir, volume.name + '-' + image_meta['id'])
+        with fileutils.remove_path_on_error(tmp_file):
+            vol_name = utils.convert_str(volume.name)
+            self._try_execute(
+                'qemu-img', 'convert', '-f', 'raw',
+                'vitastor:image='+vol_name.replace(':', '\\:')+self._qemu_args(),
+                '-O', 'raw', tmp_file
+            )
+            # FIXME: Copy directly if the destination image is also in Vitastor
+            volume_utils.upload_volume(context, image_service, image_meta, tmp_file, volume)
+        os.unlink(tmp_file)
+
+    def _get_image(self, vol_name):
+        # find the image
+        resp = self._etcd_txn({ 'success': [
+            { 'request_range': { 'key': 'index/image/'+vol_name } },
+        ] })
+        if len(resp['responses'][0]['kvs']) == 0:
+            return None
+        vol_idx = resp['responses'][0]['kvs'][0]['value']
+        vol_idx_mod = resp['responses'][0]['kvs'][0]['mod_revision']
+        # get image inode config
+        resp = self._etcd_txn({ 'success': [
+            { 'request_range': { 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']) } },
+        ] })
+        if len(resp['responses'][0]['kvs']) == 0:
+            return None
+        vol_cfg = resp['responses'][0]['kvs'][0]['value']
+        vol_cfg_mod = resp['responses'][0]['kvs'][0]['mod_revision']
+        return {
+            'cfg': vol_cfg,
+            'cfg_mod': vol_cfg_mod,
+            'idx': vol_idx,
+            'idx_mod': vol_idx_mod,
+        }
+
+    def extend_volume(self, volume, new_size):
+        """Extend an existing volume."""
+        vol_name = utils.convert_str(volume.name)
+        while True:
+            vol = self._get_image(vol_name)
+            if not vol:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
+            # change size
+            size = int(new_size) * units.Gi
+            if size == vol['cfg']['size']:
+                break
+            resp = self._etcd_txn({ 'compare': [ {
+                'target': 'MOD',
+                'mod_revision': vol['cfg_mod'],
+                'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
+            } ], 'success': [
+                { 'request_put': {
+                    'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
+                    'value': json.dumps({ **vol['cfg'], 'size': size }),
+                } },
+            ] })
+            if resp.get('succeeded'):
+                break
+        LOG.debug(
+            "Extend volume from %(old_size)s GB to %(new_size)s GB.",
+            {'old_size': volume.size, 'new_size': new_size}
+        )
+
+    def _add_manageable_volume(self, kv, manageable_volumes, cinder_ids):
+        cfg = kv['value']
+        if kv['key'].find('@') >= 0:
+            # snapshot
+            return
+        image_id = volume_utils.extract_id_from_volume_name(cfg['name'])
+        image_info = {
+            'reference': {'source-name': image_name},
+            'size': int(math.ceil(float(cfg['size']) / units.Gi)),
+            'cinder_id': None,
+            'extra_info': None,
+        }
+        if image_id in cinder_ids:
+            image_info['cinder_id'] = image_id
+            image_info['safe_to_manage'] = False
+            image_info['reason_not_safe'] = 'already managed'
+        else:
+            image_info['safe_to_manage'] = True
+            image_info['reason_not_safe'] = None
+        manageable_volumes.append(image_info)
+
+    def get_manageable_volumes(self, cinder_volumes, marker, limit, offset, sort_keys, sort_dirs):
+        manageable_volumes = []
+        cinder_ids = [resource['id'] for resource in cinder_volumes]
+
+        # List all volumes
+        # FIXME: It's possible to use pagination in our case, but.. do we want it?
+        self._etcd_foreach('config/inode/'+str(self.cfg['pool_id']),
+            lambda kv: self._add_manageable_volume(kv, manageable_volumes, cinder_ids))
+
+        return volume_utils.paginate_entries_list(
+            manageable_volumes, marker, limit, offset, sort_keys, sort_dirs)
+
+    def _get_existing_name(existing_ref):
+        if not isinstance(existing_ref, dict):
+            existing_ref = {"source-name": existing_ref}
+        if 'source-name' not in existing_ref:
+            reason = _('Reference must contain source-name element.')
+            raise exception.ManageExistingInvalidReference(existing_ref=existing_ref, reason=reason)
+        src_name = utils.convert_str(existing_ref['source-name'])
+        if not src_name:
+            reason = _('Reference must contain source-name element.')
+            raise exception.ManageExistingInvalidReference(existing_ref=existing_ref, reason=reason)
+        return src_name
+
+    def manage_existing_get_size(self, volume, existing_ref):
+        """Return size of an existing image for manage_existing.
+
+        :param volume: volume ref info to be set
+        :param existing_ref: {'source-name': <image name>}
+        """
+        src_name = self._get_existing_name(existing_ref)
+        vol = self._get_image(src_name)
+        if not vol:
+            raise exception.VolumeBackendAPIException(data = 'Volume '+src_name+' does not exist')
+        return int(math.ceil(float(vol['cfg']['size']) / units.Gi))
+
+    def manage_existing(self, volume, existing_ref):
+        """Manages an existing image.
+
+        Renames the image name to match the expected name for the volume.
+
+        :param volume: volume ref info to be set
+        :param existing_ref: {'source-name': <image name>}
+        """
+        from_name = self._get_existing_name(existing_ref)
+        to_name = utils.convert_str(volume.name)
+        self._rename(from_name, to_name)
+
+    def _rename(self, from_name, to_name):
+        while True:
+            vol = self._get_image(from_name)
+            if not vol:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+from_name+' does not exist')
+            to = self._get_image(to_name)
+            if to:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+to_name+' already exists')
+            resp = self._etcd_txn({ 'compare': [
+                { 'target': 'MOD', 'mod_revision': vol['idx_mod'], 'key': 'index/image/'+vol['cfg']['name'] },
+                { 'target': 'MOD', 'mod_revision': vol['cfg_mod'], 'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']) },
+                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+to_name },
+            ], 'success': [
+                { 'request_delete_range': { 'key': 'index/image/'+vol['cfg']['name'] } },
+                { 'request_put': { 'key': 'index/image/'+to_name, 'value': json.dumps(vol['idx']) } },
+                { 'request_put': { 'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
+                    'value': json.dumps({ **vol['cfg'], 'name': to_name }) } },
+            ] })
+            if resp.get('succeeded'):
+                break
+
+    def unmanage(self, volume):
+        pass
+
+    def _add_manageable_snapshot(self, kv, manageable_snapshots, cinder_ids):
+        cfg = kv['value']
+        dog = kv['key'].find('@')
+        if dog < 0:
+            # snapshot
+            return
+        image_name = kv['key'][0 : dog]
+        snap_name = kv['key'][dog+1 : ]
+        snapshot_id = volume_utils.extract_id_from_snapshot_name(snap_name)
+        snapshot_info = {
+            'reference': {'source-name': snap_name},
+            'size': int(math.ceil(float(cfg['size']) / units.Gi)),
+            'cinder_id': None,
+            'extra_info': None,
+            'safe_to_manage': False,
+            'reason_not_safe': None,
+            'source_reference': {'source-name': image_name}
+        }
+        if snapshot_id in cinder_ids:
+            # Exclude snapshots already managed.
+            snapshot_info['reason_not_safe'] = ('already managed')
+            snapshot_info['cinder_id'] = snapshot_id
+        elif snap_name.endswith('.clone_snap'):
+            # Exclude clone snapshot.
+            snapshot_info['reason_not_safe'] = ('used for clone snap')
+        else:
+            snapshot_info['safe_to_manage'] = True
+        manageable_snapshots.append(snapshot_info)
+
+    def get_manageable_snapshots(self, cinder_snapshots, marker, limit, offset, sort_keys, sort_dirs):
+        """List manageable snapshots in Vitastor."""
+        manageable_snapshots = []
+        cinder_snapshot_ids = [resource['id'] for resource in cinder_snapshots]
+        # List all volumes
+        # FIXME: It's possible to use pagination in our case, but.. do we want it?
+        self._etcd_foreach('config/inode/'+str(self.cfg['pool_id']),
+            lambda kv: self._add_manageable_volume(kv, manageable_snapshots, cinder_snapshot_ids))
+        return volume_utils.paginate_entries_list(
+            manageable_snapshots, marker, limit, offset, sort_keys, sort_dirs)
+
+    def manage_existing_snapshot_get_size(self, snapshot, existing_ref):
+        """Return size of an existing image for manage_existing.
+
+        :param snapshot: snapshot ref info to be set
+        :param existing_ref: {'source-name': <name of snapshot>}
+        """
+        vol_name = utils.convert_str(snapshot.volume_name)
+        snap_name = self._get_existing_name(existing_ref)
+        vol = self._get_image(vol_name+'@'+snap_name)
+        if not vol:
+            raise exception.ManageExistingInvalidReference(
+                existing_ref=snapshot_name, reason='Specified snapshot does not exist.'
+            )
+        return int(math.ceil(float(vol['cfg']['size']) / units.Gi))
+
+    def manage_existing_snapshot(self, snapshot, existing_ref):
+        """Manages an existing snapshot.
+
+        Renames the snapshot name to match the expected name for the snapshot.
+        Error checking done by manage_existing_get_size is not repeated.
+
+        :param snapshot: snapshot ref info to be set
+        :param existing_ref: {'source-name': <name of snapshot>}
+        """
+        vol_name = utils.convert_str(snapshot.volume_name)
+        snap_name = self._get_existing_name(existing_ref)
+        from_name = vol_name+'@'+snap_name
+        to_name = vol_name+'@'+utils.convert_str(snapshot.name)
+        self._rename(from_name, to_name)
+
+    def unmanage_snapshot(self, snapshot):
+        """Removes the specified snapshot from Cinder management."""
+        pass
+
+    def _dumps(self, obj):
+        return json.dumps(obj, separators=(',', ':'), sort_keys=True)

patches/devstack-local.conf

@@ -0,0 +1,23 @@
+# Devstack configuration for bridged networking
+
+[[local|localrc]]
+ADMIN_PASSWORD=secret
+DATABASE_PASSWORD=$ADMIN_PASSWORD
+RABBIT_PASSWORD=$ADMIN_PASSWORD
+SERVICE_PASSWORD=$ADMIN_PASSWORD
+HOST_IP=10.0.2.15
+Q_USE_SECGROUP=True
+FLOATING_RANGE="10.0.2.0/24"
+IPV4_ADDRS_SAFE_TO_USE="10.0.5.0/24"
+Q_FLOATING_ALLOCATION_POOL=start=10.0.2.50,end=10.0.2.100
+PUBLIC_NETWORK_GATEWAY=10.0.2.2
+PUBLIC_INTERFACE=ens3
+Q_USE_PROVIDERNET_FOR_PUBLIC=True
+Q_AGENT=linuxbridge
+Q_ML2_PLUGIN_MECHANISM_DRIVERS=linuxbridge
+LB_PHYSICAL_INTERFACE=ens3
+PUBLIC_PHYSICAL_NETWORK=default
+LB_INTERFACE_MAPPINGS=default:ens3
+Q_SERVICE_PLUGIN_CLASSES=
+Q_ML2_PLUGIN_TYPE_DRIVERS=flat
+Q_ML2_PLUGIN_EXT_DRIVERS=

patches/libvirt-5.0-vitastor.diff

@@ -0,0 +1,609 @@
+commit bd283191b3e7a4c6d1c100d3d96e348a1ebffe55
+Author: Vitaliy Filippov <vitalif@yourcmc.ru>
+Date:   Sun Jun 27 12:52:40 2021 +0300
+
+    Add Vitastor support
+
+diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng
+index aa50eac..082b4f8 100644
+--- a/docs/schemas/domaincommon.rng
++++ b/docs/schemas/domaincommon.rng
+@@ -1728,6 +1728,35 @@
+     </element>
+   </define>
+ 
++  <define name="diskSourceNetworkProtocolVitastor">
++    <element name="source">
++      <interleave>
++        <attribute name="protocol">
++          <value>vitastor</value>
++        </attribute>
++        <ref name="diskSourceCommon"/>
++        <optional>
++          <attribute name="name"/>
++        </optional>
++        <optional>
++          <attribute name="query"/>
++        </optional>
++        <zeroOrMore>
++          <ref name="diskSourceNetworkHost"/>
++        </zeroOrMore>
++        <optional>
++          <element name="config">
++            <attribute name="file">
++              <ref name="absFilePath"/>
++            </attribute>
++            <empty/>
++          </element>
++        </optional>
++        <empty/>
++      </interleave>
++    </element>
++  </define>
++
+   <define name="diskSourceNetworkProtocolISCSI">
+     <element name="source">
+       <attribute name="protocol">
+@@ -1851,6 +1880,7 @@
+       <ref name="diskSourceNetworkProtocolHTTP"/>
+       <ref name="diskSourceNetworkProtocolSimple"/>
+       <ref name="diskSourceNetworkProtocolVxHS"/>
++      <ref name="diskSourceNetworkProtocolVitastor"/>
+     </choice>
+   </define>
+ 
+diff --git a/include/libvirt/libvirt-storage.h b/include/libvirt/libvirt-storage.h
+index 4bf2b5f..dbc011b 100644
+--- a/include/libvirt/libvirt-storage.h
++++ b/include/libvirt/libvirt-storage.h
+@@ -240,6 +240,7 @@ typedef enum {
+     VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER       = 1 << 16,
+     VIR_CONNECT_LIST_STORAGE_POOLS_ZFS           = 1 << 17,
+     VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE      = 1 << 18,
++    VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR      = 1 << 20,
+ } virConnectListAllStoragePoolsFlags;
+ 
+ int                     virConnectListAllStoragePools(virConnectPtr conn,
+diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
+index 222bb8c..685d255 100644
+--- a/src/conf/domain_conf.c
++++ b/src/conf/domain_conf.c
+@@ -8653,6 +8653,10 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
+         goto cleanup;
+     }
+ 
++    if (src->protocol == VIR_STORAGE_NET_PROTOCOL_VITASTOR) {
++        src->relPath = virXMLPropString(node, "query");
++    }
++
+     if ((haveTLS = virXMLPropString(node, "tls")) &&
+         (src->haveTLS = virTristateBoolTypeFromString(haveTLS)) <= 0) {
+         virReportError(VIR_ERR_XML_ERROR,
+@@ -23849,6 +23853,10 @@ virDomainDiskSourceFormatNetwork(virBufferPtr attrBuf,
+ 
+     virBufferEscapeString(attrBuf, " name='%s'", path ? path : src->path);
+ 
++    if (src->protocol == VIR_STORAGE_NET_PROTOCOL_VITASTOR && src->relPath != NULL) {
++        virBufferEscapeString(attrBuf, " query='%s'", src->relPath);
++    }
++
+     VIR_FREE(path);
+ 
+     if (src->haveTLS != VIR_TRISTATE_BOOL_ABSENT &&
+@@ -30930,6 +30938,7 @@ virDomainDiskTranslateSourcePool(virDomainDiskDefPtr def)
+ 
+     case VIR_STORAGE_POOL_MPATH:
+     case VIR_STORAGE_POOL_RBD:
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_SHEEPDOG:
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_LAST:
+diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c
+index 55db7a9..7cbe937 100644
+--- a/src/conf/storage_conf.c
++++ b/src/conf/storage_conf.c
+@@ -58,7 +58,7 @@ VIR_ENUM_IMPL(virStoragePool,
+               "logical", "disk", "iscsi",
+               "iscsi-direct", "scsi", "mpath",
+               "rbd", "sheepdog", "gluster",
+-              "zfs", "vstorage")
++              "zfs", "vstorage", "vitastor")
+ 
+ VIR_ENUM_IMPL(virStoragePoolFormatFileSystem,
+               VIR_STORAGE_POOL_FS_LAST,
+@@ -232,6 +232,18 @@ static virStoragePoolTypeInfo poolTypeInfo[] = {
+           .formatToString = virStorageFileFormatTypeToString,
+       }
+     },
++    {.poolType = VIR_STORAGE_POOL_VITASTOR,
++     .poolOptions = {
++         .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
++                   VIR_STORAGE_POOL_SOURCE_NETWORK |
++                   VIR_STORAGE_POOL_SOURCE_NAME),
++      },
++      .volOptions = {
++          .defaultFormat = VIR_STORAGE_FILE_RAW,
++          .formatFromString = virStorageVolumeFormatFromString,
++          .formatToString = virStorageFileFormatTypeToString,
++      }
++    },
+     {.poolType = VIR_STORAGE_POOL_SHEEPDOG,
+      .poolOptions = {
+          .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
+@@ -434,6 +446,11 @@ virStoragePoolDefParseSource(xmlXPathContextPtr ctxt,
+                        _("element 'name' is mandatory for RBD pool"));
+         goto cleanup;
+     }
++    if (pool_type == VIR_STORAGE_POOL_VITASTOR && source->name == NULL) {
++        virReportError(VIR_ERR_XML_ERROR, "%s",
++                       _("element 'name' is mandatory for Vitastor pool"));
++        return -1;
++    }
+ 
+     if (options->formatFromString) {
+         char *format = virXPathString("string(./format/@type)", ctxt);
+@@ -1009,6 +1026,7 @@ virStoragePoolDefFormatBuf(virBufferPtr buf,
+     /* RBD, Sheepdog, Gluster and Iscsi-direct devices are not local block devs nor
+      * files, so they don't have a target */
+     if (def->type != VIR_STORAGE_POOL_RBD &&
++        def->type != VIR_STORAGE_POOL_VITASTOR &&
+         def->type != VIR_STORAGE_POOL_SHEEPDOG &&
+         def->type != VIR_STORAGE_POOL_GLUSTER &&
+         def->type != VIR_STORAGE_POOL_ISCSI_DIRECT) {
+diff --git a/src/conf/storage_conf.h b/src/conf/storage_conf.h
+index dc0aa2a..ed4983d 100644
+--- a/src/conf/storage_conf.h
++++ b/src/conf/storage_conf.h
+@@ -91,6 +91,7 @@ typedef enum {
+     VIR_STORAGE_POOL_GLUSTER,  /* Gluster device */
+     VIR_STORAGE_POOL_ZFS,      /* ZFS */
+     VIR_STORAGE_POOL_VSTORAGE, /* Virtuozzo Storage */
++    VIR_STORAGE_POOL_VITASTOR, /* Vitastor */
+ 
+     VIR_STORAGE_POOL_LAST,
+ } virStoragePoolType;
+@@ -422,6 +423,7 @@ VIR_ENUM_DECL(virStoragePartedFs)
+                  VIR_CONNECT_LIST_STORAGE_POOLS_SCSI     | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_MPATH    | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_RBD      | \
++                 VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER  | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_ZFS      | \
+diff --git a/src/conf/virstorageobj.c b/src/conf/virstorageobj.c
+index 6ea6a97..3ba45b9 100644
+--- a/src/conf/virstorageobj.c
++++ b/src/conf/virstorageobj.c
+@@ -1478,6 +1478,7 @@ virStoragePoolObjSourceFindDuplicateCb(const void *payload,
+             return 1;
+         break;
+ 
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_RBD:
+     case VIR_STORAGE_POOL_LAST:
+         break;
+@@ -1971,6 +1972,8 @@ virStoragePoolObjMatch(virStoragePoolObjPtr obj,
+                (obj->def->type == VIR_STORAGE_POOL_MPATH))   ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_RBD) &&
+                (obj->def->type == VIR_STORAGE_POOL_RBD))     ||
++              (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR) &&
++               (obj->def->type == VIR_STORAGE_POOL_VITASTOR)) ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG) &&
+                (obj->def->type == VIR_STORAGE_POOL_SHEEPDOG)) ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER) &&
+diff --git a/src/libvirt-storage.c b/src/libvirt-storage.c
+index 2ea3e94..d5d2273 100644
+--- a/src/libvirt-storage.c
++++ b/src/libvirt-storage.c
+@@ -92,6 +92,7 @@ virStoragePoolGetConnect(virStoragePoolPtr pool)
+  * VIR_CONNECT_LIST_STORAGE_POOLS_SCSI
+  * VIR_CONNECT_LIST_STORAGE_POOLS_MPATH
+  * VIR_CONNECT_LIST_STORAGE_POOLS_RBD
++ * VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR
+  * VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG
+  *
+  * Returns the number of storage pools found or -1 and sets @pools to
+diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
+index 73e988a..ab7bb81 100644
+--- a/src/libxl/libxl_conf.c
++++ b/src/libxl/libxl_conf.c
+@@ -905,6 +905,7 @@ libxlMakeNetworkDiskSrcStr(virStorageSourcePtr src,
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_SSH:
+     case VIR_STORAGE_NET_PROTOCOL_VXHS:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+         virReportError(VIR_ERR_NO_SUPPORT,
+diff --git a/src/qemu/qemu_block.c b/src/qemu/qemu_block.c
+index cbf0aa4..096700d 100644
+--- a/src/qemu/qemu_block.c
++++ b/src/qemu/qemu_block.c
+@@ -959,6 +959,42 @@ qemuBlockStorageSourceGetRBDProps(virStorageSourcePtr src)
+ }
+ 
+ 
++static virJSONValuePtr
++qemuBlockStorageSourceGetVitastorProps(virStorageSource *src)
++{
++    virJSONValuePtr ret = NULL;
++    virStorageNetHostDefPtr host;
++    size_t i;
++    virBuffer buf = VIR_BUFFER_INITIALIZER;
++    char *etcd = NULL;
++
++    for (i = 0; i < src->nhosts; i++) {
++        host = src->hosts + i;
++        if ((virStorageNetHostTransport)host->transport != VIR_STORAGE_NET_HOST_TRANS_TCP) {
++            goto cleanup;
++        }
++        virBufferAsprintf(&buf, i > 0 ? ",%s:%u" : "%s:%u", host->name, host->port);
++    }
++    if (src->nhosts > 0) {
++        etcd = virBufferContentAndReset(&buf);
++    }
++
++    if (virJSONValueObjectCreate(&ret,
++                                 "s:driver", "vitastor",
++                                 "S:etcd_host", etcd,
++                                 "S:etcd_prefix", src->relPath,
++                                 "S:config_path", src->configFile,
++                                 "s:image", src->path,
++                                 NULL) < 0)
++        goto cleanup;
++
++cleanup:
++    VIR_FREE(etcd);
++    virBufferFreeAndReset(&buf);
++    return ret;
++}
++
++
+ static virJSONValuePtr
+ qemuBlockStorageSourceGetSheepdogProps(virStorageSourcePtr src)
+ {
+@@ -1174,6 +1210,11 @@ qemuBlockStorageSourceGetBackendProps(virStorageSourcePtr src,
+                 return NULL;
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            if (!(fileprops = qemuBlockStorageSourceGetVitastorProps(src)))
++                return NULL;
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             if (!(fileprops = qemuBlockStorageSourceGetSheepdogProps(src)))
+                 return NULL;
+diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
+index 822d5f8..e375cef 100644
+--- a/src/qemu/qemu_command.c
++++ b/src/qemu/qemu_command.c
+@@ -975,6 +975,43 @@ qemuBuildNetworkDriveStr(virStorageSourcePtr src,
+             ret = virBufferContentAndReset(&buf);
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            if (strchr(src->path, ':')) {
++                virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
++                               _("':' not allowed in Vitastor source volume name '%s'"),
++                               src->path);
++                return NULL;
++            }
++
++            virBufferStrcat(&buf, "vitastor:image=", src->path, NULL);
++
++            if (src->nhosts > 0) {
++                virBufferAddLit(&buf, ":etcd_host=");
++                for (i = 0; i < src->nhosts; i++) {
++                    if (i)
++                        virBufferAddLit(&buf, ",");
++
++                    /* assume host containing : is ipv6 */
++                    if (strchr(src->hosts[i].name, ':'))
++                        virBufferEscape(&buf, '\\', ":", "[%s]",
++                                        src->hosts[i].name);
++                    else
++                        virBufferAsprintf(&buf, "%s", src->hosts[i].name);
++
++                    if (src->hosts[i].port)
++                        virBufferAsprintf(&buf, "\\:%u", src->hosts[i].port);
++                }
++            }
++
++            if (src->configFile)
++                virBufferEscape(&buf, '\\', ":", ":config_path=%s", src->configFile);
++
++            if (src->relPath)
++                virBufferEscape(&buf, '\\', ":", ":etcd_prefix=%s", src->relPath);
++
++            ret = virBufferContentAndReset(&buf);
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_VXHS:
+             virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+                            _("VxHS protocol does not support URI syntax"));
+diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
+index ec6b340..f399efa 100644
+--- a/src/qemu/qemu_domain.c
++++ b/src/qemu/qemu_domain.c
+@@ -10881,6 +10881,7 @@ qemuDomainPrepareStorageSourceTLS(virStorageSourcePtr src,
+         break;
+ 
+     case VIR_STORAGE_NET_PROTOCOL_RBD:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+     case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
+index 1d96170..2d24396 100644
+--- a/src/qemu/qemu_driver.c
++++ b/src/qemu/qemu_driver.c
+@@ -14687,6 +14687,7 @@ qemuDomainSnapshotPrepareDiskExternalInactive(virDomainSnapshotDiskDefPtr snapdi
+         case VIR_STORAGE_NET_PROTOCOL_TFTP:
+         case VIR_STORAGE_NET_PROTOCOL_SSH:
+         case VIR_STORAGE_NET_PROTOCOL_VXHS:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_LAST:
+             virReportError(VIR_ERR_INTERNAL_ERROR,
+                            _("external inactive snapshots are not supported on "
+@@ -14764,6 +14765,7 @@ qemuDomainSnapshotPrepareDiskExternalActive(virDomainSnapshotDiskDefPtr snapdisk
+         case VIR_STORAGE_NET_PROTOCOL_TFTP:
+         case VIR_STORAGE_NET_PROTOCOL_SSH:
+         case VIR_STORAGE_NET_PROTOCOL_VXHS:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_LAST:
+             virReportError(VIR_ERR_INTERNAL_ERROR,
+                            _("external active snapshots are not supported on "
+@@ -14887,6 +14889,7 @@ qemuDomainSnapshotPrepareDiskInternal(virDomainDiskDefPtr disk,
+         case VIR_STORAGE_NET_PROTOCOL_TFTP:
+         case VIR_STORAGE_NET_PROTOCOL_SSH:
+         case VIR_STORAGE_NET_PROTOCOL_VXHS:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_LAST:
+             virReportError(VIR_ERR_INTERNAL_ERROR,
+                            _("internal inactive snapshots are not supported on "
+diff --git a/src/qemu/qemu_parse_command.c b/src/qemu/qemu_parse_command.c
+index c4650f0..551da41 100644
+--- a/src/qemu/qemu_parse_command.c
++++ b/src/qemu/qemu_parse_command.c
+@@ -2184,6 +2184,7 @@ qemuParseCommandLine(virFileCachePtr capsCache,
+                 case VIR_STORAGE_NET_PROTOCOL_TFTP:
+                 case VIR_STORAGE_NET_PROTOCOL_SSH:
+                 case VIR_STORAGE_NET_PROTOCOL_LAST:
++                case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+                 case VIR_STORAGE_NET_PROTOCOL_NONE:
+                     /* ignored for now */
+                     break;
+diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c
+index 4a13e90..33301c7 100644
+--- a/src/storage/storage_driver.c
++++ b/src/storage/storage_driver.c
+@@ -1568,6 +1568,7 @@ storageVolLookupByPathCallback(virStoragePoolObjPtr obj,
+         case VIR_STORAGE_POOL_RBD:
+         case VIR_STORAGE_POOL_SHEEPDOG:
+         case VIR_STORAGE_POOL_ZFS:
++        case VIR_STORAGE_POOL_VITASTOR:
+         case VIR_STORAGE_POOL_LAST:
+             ignore_value(VIR_STRDUP(stable_path, data->path));
+             break;
+diff --git a/src/util/virstoragefile.c b/src/util/virstoragefile.c
+index bd4b027..b323cd6 100644
+--- a/src/util/virstoragefile.c
++++ b/src/util/virstoragefile.c
+@@ -84,7 +84,8 @@ VIR_ENUM_IMPL(virStorageNetProtocol, VIR_STORAGE_NET_PROTOCOL_LAST,
+               "ftps",
+               "tftp",
+               "ssh",
+-              "vxhs")
++              "vxhs",
++              "vitastor")
+ 
+ VIR_ENUM_IMPL(virStorageNetHostTransport, VIR_STORAGE_NET_HOST_TRANS_LAST,
+               "tcp",
+@@ -2839,6 +2840,83 @@ virStorageSourceParseRBDColonString(const char *rbdstr,
+ }
+ 
+ 
++static int
++virStorageSourceParseVitastorColonString(const char *colonstr,
++                                         virStorageSourcePtr src)
++{
++    char *p, *e, *next;
++    char *options = NULL;
++
++    /* optionally skip the "vitastor:" prefix if provided */
++    if (STRPREFIX(colonstr, "vitastor:"))
++        colonstr += strlen("vitastor:");
++
++    if (VIR_STRDUP(options, colonstr) < 0)
++        return -1;
++
++    p = options;
++    while (*p) {
++        /* find : delimiter or end of string */
++        for (e = p; *e && *e != ':'; ++e) {
++            if (*e == '\\') {
++                e++;
++                if (*e == '\0')
++                    break;
++            }
++        }
++        if (*e == '\0') {
++            next = e;    /* last kv pair */
++        } else {
++            next = e + 1;
++            *e = '\0';
++        }
++
++        if (STRPREFIX(p, "image=")) {
++            if (VIR_STRDUP(src->path, p + strlen("image=")) < 0)
++                return -1;
++        } else if (STRPREFIX(p, "etcd_prefix=")) {
++            if (VIR_STRDUP(src->relPath, p + strlen("etcd_prefix=")) < 0)
++                return -1;
++        } else if (STRPREFIX(p, "config_file=")) {
++            if (VIR_STRDUP(src->configFile, p + strlen("config_file=")) < 0)
++                return -1;
++        } else if (STRPREFIX(p, "etcd_host=")) {
++            char *h, *sep;
++
++            h = p + strlen("etcd_host=");
++            while (h < e) {
++                for (sep = h; sep < e; ++sep) {
++                    if (*sep == '\\' && (sep[1] == ',' ||
++                                         sep[1] == ';' ||
++                                         sep[1] == ' ')) {
++                        *sep = '\0';
++                        sep += 2;
++                        break;
++                    }
++                }
++
++                if (virStorageSourceRBDAddHost(src, h) < 0)
++                    goto error;
++
++                h = sep;
++            }
++        }
++
++        p = next;
++    }
++
++    if (!src->path) {
++        goto error;
++    }
++
++    return 0;
++
++error:
++    VIR_FREE(options);
++    return -1;
++}
++
++
+ static int
+ virStorageSourceParseNBDColonString(const char *nbdstr,
+                                     virStorageSourcePtr src)
+@@ -2942,6 +3020,11 @@ virStorageSourceParseBackingColon(virStorageSourcePtr src,
+             goto cleanup;
+         break;
+ 
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++        if (virStorageSourceParseVitastorColonString(path, src) < 0)
++            return -1;
++        break;
++
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+@@ -3441,6 +3524,56 @@ virStorageSourceParseBackingJSONRBD(virStorageSourcePtr src,
+     return ret;
+ }
+ 
++static int
++virStorageSourceParseBackingJSONVitastor(virStorageSourcePtr src,
++                                         virJSONValuePtr json,
++                                         int opaque ATTRIBUTE_UNUSED)
++{
++    const char *filename;
++    const char *image = virJSONValueObjectGetString(json, "image");
++    const char *conf = virJSONValueObjectGetString(json, "config_path");
++    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd_prefix");
++    virJSONValuePtr servers = virJSONValueObjectGetArray(json, "server");
++    size_t nservers;
++    size_t i;
++
++    src->type = VIR_STORAGE_TYPE_NETWORK;
++    src->protocol = VIR_STORAGE_NET_PROTOCOL_VITASTOR;
++
++    /* legacy syntax passed via 'filename' option */
++    if ((filename = virJSONValueObjectGetString(json, "filename")))
++        return virStorageSourceParseVitastorColonString(filename, src);
++
++    if (!image) {
++        virReportError(VIR_ERR_INVALID_ARG, "%s",
++                       _("missing image name in Vitastor backing volume "
++                         "JSON specification"));
++        return -1;
++    }
++
++    if (VIR_STRDUP(src->path, image) < 0 ||
++        VIR_STRDUP(src->configFile, conf) < 0 ||
++        VIR_STRDUP(src->relPath, etcd_prefix) < 0)
++        return -1;
++
++    if (servers) {
++        nservers = virJSONValueArraySize(servers);
++
++        if (VIR_ALLOC_N(src->hosts, nservers) < 0)
++            return -1;
++
++        src->nhosts = nservers;
++
++        for (i = 0; i < nservers; i++) {
++            if (virStorageSourceParseBackingJSONInetSocketAddress(src->hosts + i,
++                                                                  virJSONValueArrayGet(servers, i)) < 0)
++                return -1;
++        }
++    }
++
++    return 0;
++}
++
+ static int
+ virStorageSourceParseBackingJSONRaw(virStorageSourcePtr src,
+                                     virJSONValuePtr json,
+@@ -3507,6 +3640,7 @@ static const struct virStorageSourceJSONDriverParser jsonParsers[] = {
+     {"sheepdog", virStorageSourceParseBackingJSONSheepdog, 0},
+     {"ssh", virStorageSourceParseBackingJSONSSH, 0},
+     {"rbd", virStorageSourceParseBackingJSONRBD, 0},
++    {"vitastor", virStorageSourceParseBackingJSONVitastor, 0},
+     {"raw", virStorageSourceParseBackingJSONRaw, 0},
+     {"vxhs", virStorageSourceParseBackingJSONVxHS, 0},
+ };
+@@ -4276,6 +4410,7 @@ virStorageSourceNetworkDefaultPort(virStorageNetProtocol protocol)
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+             return 24007;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
+             /* we don't provide a default for RBD */
+             return 0;
+diff --git a/src/util/virstoragefile.h b/src/util/virstoragefile.h
+index 1d6161a..8d83bf3 100644
+--- a/src/util/virstoragefile.h
++++ b/src/util/virstoragefile.h
+@@ -134,6 +134,7 @@ typedef enum {
+     VIR_STORAGE_NET_PROTOCOL_TFTP,
+     VIR_STORAGE_NET_PROTOCOL_SSH,
+     VIR_STORAGE_NET_PROTOCOL_VXHS,
++    VIR_STORAGE_NET_PROTOCOL_VITASTOR,
+ 
+     VIR_STORAGE_NET_PROTOCOL_LAST
+ } virStorageNetProtocol;
+diff --git a/src/xenconfig/xen_xl.c b/src/xenconfig/xen_xl.c
+index accfc3a..a18f9c3 100644
+--- a/src/xenconfig/xen_xl.c
++++ b/src/xenconfig/xen_xl.c
+@@ -1535,6 +1535,7 @@ xenFormatXLDiskSrcNet(virStorageSourcePtr src)
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_SSH:
+     case VIR_STORAGE_NET_PROTOCOL_VXHS:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+         virReportError(VIR_ERR_NO_SUPPORT,
+diff --git a/tools/virsh-pool.c b/tools/virsh-pool.c
+index 70ca39b..9caef51 100644
+--- a/tools/virsh-pool.c
++++ b/tools/virsh-pool.c
+@@ -1219,6 +1219,9 @@ cmdPoolList(vshControl *ctl, const vshCmd *cmd ATTRIBUTE_UNUSED)
+             case VIR_STORAGE_POOL_VSTORAGE:
+                 flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE;
+                 break;
++            case VIR_STORAGE_POOL_VITASTOR:
++                flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR;
++                break;
+             case VIR_STORAGE_POOL_LAST:
+                 break;
+             }

patches/libvirt-7.0-vitastor.diff

@@ -0,0 +1,657 @@
+commit 41cdfe8317d98f70aadedfdbb381effed2641bdd
+Author: Vitaliy Filippov <vitalif@yourcmc.ru>
+Date:   Fri Jul 9 01:31:57 2021 +0300
+
+    Add Vitastor support
+
+diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng
+index 7dc419b..875433b 100644
+--- a/docs/schemas/domaincommon.rng
++++ b/docs/schemas/domaincommon.rng
+@@ -1827,6 +1827,35 @@
+     </element>
+   </define>
+ 
++  <define name="diskSourceNetworkProtocolVitastor">
++    <element name="source">
++      <interleave>
++        <attribute name="protocol">
++          <value>vitastor</value>
++        </attribute>
++        <ref name="diskSourceCommon"/>
++        <optional>
++          <attribute name="name"/>
++        </optional>
++        <optional>
++          <attribute name="query"/>
++        </optional>
++        <zeroOrMore>
++          <ref name="diskSourceNetworkHost"/>
++        </zeroOrMore>
++        <optional>
++          <element name="config">
++            <attribute name="file">
++              <ref name="absFilePath"/>
++            </attribute>
++            <empty/>
++          </element>
++        </optional>
++        <empty/>
++      </interleave>
++    </element>
++  </define>
++
+   <define name="diskSourceNetworkProtocolISCSI">
+     <element name="source">
+       <attribute name="protocol">
+@@ -2083,6 +2112,7 @@
+       <ref name="diskSourceNetworkProtocolSimple"/>
+       <ref name="diskSourceNetworkProtocolVxHS"/>
+       <ref name="diskSourceNetworkProtocolNFS"/>
++      <ref name="diskSourceNetworkProtocolVitastor"/>
+     </choice>
+   </define>
+ 
+diff --git a/include/libvirt/libvirt-storage.h b/include/libvirt/libvirt-storage.h
+index 089e1e0..d7e7ef4 100644
+--- a/include/libvirt/libvirt-storage.h
++++ b/include/libvirt/libvirt-storage.h
+@@ -245,6 +245,7 @@ typedef enum {
+     VIR_CONNECT_LIST_STORAGE_POOLS_ZFS           = 1 << 17,
+     VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE      = 1 << 18,
+     VIR_CONNECT_LIST_STORAGE_POOLS_ISCSI_DIRECT  = 1 << 19,
++    VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR      = 1 << 20,
+ } virConnectListAllStoragePoolsFlags;
+ 
+ int                     virConnectListAllStoragePools(virConnectPtr conn,
+diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
+index 01b7187..c6e9702 100644
+--- a/src/conf/domain_conf.c
++++ b/src/conf/domain_conf.c
+@@ -8261,7 +8261,8 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
+     src->configFile = virXPathString("string(./config/@file)", ctxt);
+ 
+     if (src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTP ||
+-        src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTPS)
++        src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTPS ||
++        src->protocol == VIR_STORAGE_NET_PROTOCOL_VITASTOR)
+         src->query = virXMLPropString(node, "query");
+ 
+     if (virDomainStorageNetworkParseHosts(node, ctxt, &src->hosts, &src->nhosts) < 0)
+@@ -31392,6 +31393,7 @@ virDomainStorageSourceTranslateSourcePool(virStorageSourcePtr src,
+ 
+     case VIR_STORAGE_POOL_MPATH:
+     case VIR_STORAGE_POOL_RBD:
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_SHEEPDOG:
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_LAST:
+diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c
+index 0c50529..fe97574 100644
+--- a/src/conf/storage_conf.c
++++ b/src/conf/storage_conf.c
+@@ -60,7 +60,7 @@ VIR_ENUM_IMPL(virStoragePool,
+               "logical", "disk", "iscsi",
+               "iscsi-direct", "scsi", "mpath",
+               "rbd", "sheepdog", "gluster",
+-              "zfs", "vstorage",
++              "zfs", "vstorage", "vitastor",
+ );
+ 
+ VIR_ENUM_IMPL(virStoragePoolFormatFileSystem,
+@@ -249,6 +249,18 @@ static virStoragePoolTypeInfo poolTypeInfo[] = {
+           .formatToString = virStorageFileFormatTypeToString,
+       }
+     },
++    {.poolType = VIR_STORAGE_POOL_VITASTOR,
++     .poolOptions = {
++         .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
++                   VIR_STORAGE_POOL_SOURCE_NETWORK |
++                   VIR_STORAGE_POOL_SOURCE_NAME),
++      },
++      .volOptions = {
++          .defaultFormat = VIR_STORAGE_FILE_RAW,
++          .formatFromString = virStorageVolumeFormatFromString,
++          .formatToString = virStorageFileFormatTypeToString,
++      }
++    },
+     {.poolType = VIR_STORAGE_POOL_SHEEPDOG,
+      .poolOptions = {
+          .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
+@@ -551,6 +563,11 @@ virStoragePoolDefParseSource(xmlXPathContextPtr ctxt,
+                        _("element 'name' is mandatory for RBD pool"));
+         goto cleanup;
+     }
++    if (pool_type == VIR_STORAGE_POOL_VITASTOR && source->name == NULL) {
++        virReportError(VIR_ERR_XML_ERROR, "%s",
++                       _("element 'name' is mandatory for Vitastor pool"));
++        return -1;
++    }
+ 
+     if (options->formatFromString) {
+         g_autofree char *format = NULL;
+@@ -1217,6 +1234,7 @@ virStoragePoolDefFormatBuf(virBufferPtr buf,
+     /* RBD, Sheepdog, Gluster and Iscsi-direct devices are not local block devs nor
+      * files, so they don't have a target */
+     if (def->type != VIR_STORAGE_POOL_RBD &&
++        def->type != VIR_STORAGE_POOL_VITASTOR &&
+         def->type != VIR_STORAGE_POOL_SHEEPDOG &&
+         def->type != VIR_STORAGE_POOL_GLUSTER &&
+         def->type != VIR_STORAGE_POOL_ISCSI_DIRECT) {
+diff --git a/src/conf/storage_conf.h b/src/conf/storage_conf.h
+index ffd406e..8868a05 100644
+--- a/src/conf/storage_conf.h
++++ b/src/conf/storage_conf.h
+@@ -110,6 +110,7 @@ typedef enum {
+     VIR_STORAGE_POOL_GLUSTER,  /* Gluster device */
+     VIR_STORAGE_POOL_ZFS,      /* ZFS */
+     VIR_STORAGE_POOL_VSTORAGE, /* Virtuozzo Storage */
++    VIR_STORAGE_POOL_VITASTOR, /* Vitastor */
+ 
+     VIR_STORAGE_POOL_LAST,
+ } virStoragePoolType;
+@@ -474,6 +475,7 @@ VIR_ENUM_DECL(virStoragePartedFs);
+                  VIR_CONNECT_LIST_STORAGE_POOLS_SCSI     | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_MPATH    | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_RBD      | \
++                 VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER  | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_ZFS      | \
+diff --git a/src/conf/virstorageobj.c b/src/conf/virstorageobj.c
+index 9fe8b3f..bf595b0 100644
+--- a/src/conf/virstorageobj.c
++++ b/src/conf/virstorageobj.c
+@@ -1491,6 +1491,7 @@ virStoragePoolObjSourceFindDuplicateCb(const void *payload,
+             return 1;
+         break;
+ 
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_RBD:
+     case VIR_STORAGE_POOL_LAST:
+         break;
+@@ -1990,6 +1991,8 @@ virStoragePoolObjMatch(virStoragePoolObjPtr obj,
+                (obj->def->type == VIR_STORAGE_POOL_MPATH))   ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_RBD) &&
+                (obj->def->type == VIR_STORAGE_POOL_RBD))     ||
++              (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR) &&
++               (obj->def->type == VIR_STORAGE_POOL_VITASTOR)) ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG) &&
+                (obj->def->type == VIR_STORAGE_POOL_SHEEPDOG)) ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER) &&
+diff --git a/src/libvirt-storage.c b/src/libvirt-storage.c
+index 2a7cdca..f756be1 100644
+--- a/src/libvirt-storage.c
++++ b/src/libvirt-storage.c
+@@ -92,6 +92,7 @@ virStoragePoolGetConnect(virStoragePoolPtr pool)
+  * VIR_CONNECT_LIST_STORAGE_POOLS_SCSI
+  * VIR_CONNECT_LIST_STORAGE_POOLS_MPATH
+  * VIR_CONNECT_LIST_STORAGE_POOLS_RBD
++ * VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR
+  * VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG
+  * VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER
+  * VIR_CONNECT_LIST_STORAGE_POOLS_ZFS
+diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
+index 6a8ae27..a735bc6 100644
+--- a/src/libxl/libxl_conf.c
++++ b/src/libxl/libxl_conf.c
+@@ -942,6 +942,7 @@ libxlMakeNetworkDiskSrcStr(virStorageSourcePtr src,
+     case VIR_STORAGE_NET_PROTOCOL_SSH:
+     case VIR_STORAGE_NET_PROTOCOL_VXHS:
+     case VIR_STORAGE_NET_PROTOCOL_NFS:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+         virReportError(VIR_ERR_NO_SUPPORT,
+diff --git a/src/libxl/xen_xl.c b/src/libxl/xen_xl.c
+index 17b93d0..c5a0084 100644
+--- a/src/libxl/xen_xl.c
++++ b/src/libxl/xen_xl.c
+@@ -1601,6 +1601,7 @@ xenFormatXLDiskSrcNet(virStorageSourcePtr src)
+     case VIR_STORAGE_NET_PROTOCOL_SSH:
+     case VIR_STORAGE_NET_PROTOCOL_VXHS:
+     case VIR_STORAGE_NET_PROTOCOL_NFS:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+         virReportError(VIR_ERR_NO_SUPPORT,
+diff --git a/src/qemu/qemu_block.c b/src/qemu/qemu_block.c
+index f9c6da2..922dde5 100644
+--- a/src/qemu/qemu_block.c
++++ b/src/qemu/qemu_block.c
+@@ -938,6 +938,38 @@ qemuBlockStorageSourceGetRBDProps(virStorageSourcePtr src,
+ }
+ 
+ 
++static virJSONValuePtr
++qemuBlockStorageSourceGetVitastorProps(virStorageSource *src)
++{
++    virJSONValuePtr ret = NULL;
++    virStorageNetHostDefPtr host;
++    size_t i;
++    g_auto(virBuffer) buf = VIR_BUFFER_INITIALIZER;
++    g_autofree char *etcd = NULL;
++
++    for (i = 0; i < src->nhosts; i++) {
++        host = src->hosts + i;
++        if ((virStorageNetHostTransport)host->transport != VIR_STORAGE_NET_HOST_TRANS_TCP) {
++            return NULL;
++        }
++        virBufferAsprintf(&buf, i > 0 ? ",%s:%u" : "%s:%u", host->name, host->port);
++    }
++    if (src->nhosts > 0) {
++        etcd = virBufferContentAndReset(&buf);
++    }
++
++    if (virJSONValueObjectCreate(&ret,
++                                 "S:etcd_host", etcd,
++                                 "S:etcd_prefix", src->query,
++                                 "S:config_path", src->configFile,
++                                 "s:image", src->path,
++                                 NULL) < 0)
++        return NULL;
++
++    return ret;
++}
++
++
+ static virJSONValuePtr
+ qemuBlockStorageSourceGetSheepdogProps(virStorageSourcePtr src)
+ {
+@@ -1224,6 +1256,12 @@ qemuBlockStorageSourceGetBackendProps(virStorageSourcePtr src,
+                 return NULL;
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            driver = "vitastor";
++            if (!(fileprops = qemuBlockStorageSourceGetVitastorProps(src)))
++                return NULL;
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             driver = "sheepdog";
+             if (!(fileprops = qemuBlockStorageSourceGetSheepdogProps(src)))
+@@ -2183,6 +2221,7 @@ qemuBlockGetBackingStoreString(virStorageSourcePtr src,
+ 
+             case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             case VIR_STORAGE_NET_PROTOCOL_RBD:
++            case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+             case VIR_STORAGE_NET_PROTOCOL_VXHS:
+             case VIR_STORAGE_NET_PROTOCOL_NFS:
+             case VIR_STORAGE_NET_PROTOCOL_SSH:
+@@ -2560,6 +2599,12 @@ qemuBlockStorageSourceCreateGetStorageProps(virStorageSourcePtr src,
+                 return -1;
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            driver = "vitastor";
++            if (!(location = qemuBlockStorageSourceGetVitastorProps(src)))
++                return -1;
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             driver = "sheepdog";
+             if (!(location = qemuBlockStorageSourceGetSheepdogProps(src)))
+diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
+index 6f970a3..10b39ca 100644
+--- a/src/qemu/qemu_command.c
++++ b/src/qemu/qemu_command.c
+@@ -1034,6 +1034,43 @@ qemuBuildNetworkDriveStr(virStorageSourcePtr src,
+             ret = virBufferContentAndReset(&buf);
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            if (strchr(src->path, ':')) {
++                virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
++                               _("':' not allowed in Vitastor source volume name '%s'"),
++                               src->path);
++                return NULL;
++            }
++
++            virBufferStrcat(&buf, "vitastor:image=", src->path, NULL);
++
++            if (src->nhosts > 0) {
++                virBufferAddLit(&buf, ":etcd_host=");
++                for (i = 0; i < src->nhosts; i++) {
++                    if (i)
++                        virBufferAddLit(&buf, ",");
++
++                    /* assume host containing : is ipv6 */
++                    if (strchr(src->hosts[i].name, ':'))
++                        virBufferEscape(&buf, '\\', ":", "[%s]",
++                                        src->hosts[i].name);
++                    else
++                        virBufferAsprintf(&buf, "%s", src->hosts[i].name);
++
++                    if (src->hosts[i].port)
++                        virBufferAsprintf(&buf, "\\:%u", src->hosts[i].port);
++                }
++            }
++
++            if (src->configFile)
++                virBufferEscape(&buf, '\\', ":", ":config_path=%s", src->configFile);
++
++            if (src->query)
++                virBufferEscape(&buf, '\\', ":", ":etcd_prefix=%s", src->query);
++
++            ret = virBufferContentAndReset(&buf);
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_VXHS:
+             virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+                            _("VxHS protocol does not support URI syntax"));
+diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
+index 0765dc7..4cff344 100644
+--- a/src/qemu/qemu_domain.c
++++ b/src/qemu/qemu_domain.c
+@@ -4610,7 +4610,8 @@ qemuDomainValidateStorageSource(virStorageSourcePtr src,
+     if (src->query &&
+         (actualType != VIR_STORAGE_TYPE_NETWORK ||
+          (src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTPS &&
+-          src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTP))) {
++          src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTP &&
++          src->protocol != VIR_STORAGE_NET_PROTOCOL_VITASTOR))) {
+         virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+                        _("query is supported only with HTTP(S) protocols"));
+         return -1;
+@@ -9704,6 +9705,7 @@ qemuDomainPrepareStorageSourceTLS(virStorageSourcePtr src,
+         break;
+ 
+     case VIR_STORAGE_NET_PROTOCOL_RBD:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+     case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+diff --git a/src/qemu/qemu_snapshot.c b/src/qemu/qemu_snapshot.c
+index ee333c3..674aa58 100644
+--- a/src/qemu/qemu_snapshot.c
++++ b/src/qemu/qemu_snapshot.c
+@@ -403,6 +403,7 @@ qemuSnapshotPrepareDiskExternalInactive(virDomainSnapshotDiskDefPtr snapdisk,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+@@ -493,6 +494,7 @@ qemuSnapshotPrepareDiskExternalActive(virDomainObjPtr vm,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+         case VIR_STORAGE_NET_PROTOCOL_HTTP:
+@@ -623,6 +625,7 @@ qemuSnapshotPrepareDiskInternal(virDomainDiskDefPtr disk,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c
+index 16bc53a..1e5d820 100644
+--- a/src/storage/storage_driver.c
++++ b/src/storage/storage_driver.c
+@@ -1645,6 +1645,7 @@ storageVolLookupByPathCallback(virStoragePoolObjPtr obj,
+ 
+         case VIR_STORAGE_POOL_GLUSTER:
+         case VIR_STORAGE_POOL_RBD:
++        case VIR_STORAGE_POOL_VITASTOR:
+         case VIR_STORAGE_POOL_SHEEPDOG:
+         case VIR_STORAGE_POOL_ZFS:
+         case VIR_STORAGE_POOL_LAST:
+diff --git a/src/test/test_driver.c b/src/test/test_driver.c
+index 29c4c86..a27ad94 100644
+--- a/src/test/test_driver.c
++++ b/src/test/test_driver.c
+@@ -7096,6 +7096,7 @@ testStorageVolumeTypeForPool(int pooltype)
+     case VIR_STORAGE_POOL_ISCSI_DIRECT:
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_RBD:
++    case VIR_STORAGE_POOL_VITASTOR:
+         return VIR_STORAGE_VOL_NETWORK;
+     case VIR_STORAGE_POOL_LOGICAL:
+     case VIR_STORAGE_POOL_DISK:
+diff --git a/src/util/virstoragefile.c b/src/util/virstoragefile.c
+index 0d3c2af..36e3afc 100644
+--- a/src/util/virstoragefile.c
++++ b/src/util/virstoragefile.c
+@@ -91,6 +91,7 @@ VIR_ENUM_IMPL(virStorageNetProtocol,
+               "ssh",
+               "vxhs",
+               "nfs",
++              "vitastor",
+ );
+ 
+ VIR_ENUM_IMPL(virStorageNetHostTransport,
+@@ -2880,6 +2881,75 @@ virStorageSourceParseRBDColonString(const char *rbdstr,
+ }
+ 
+ 
++static int
++virStorageSourceParseVitastorColonString(const char *colonstr,
++                                         virStorageSourcePtr src)
++{
++    char *p, *e, *next;
++    g_autofree char *options = NULL;
++
++    /* optionally skip the "vitastor:" prefix if provided */
++    if (STRPREFIX(colonstr, "vitastor:"))
++        colonstr += strlen("vitastor:");
++
++    options = g_strdup(colonstr);
++
++    p = options;
++    while (*p) {
++        /* find : delimiter or end of string */
++        for (e = p; *e && *e != ':'; ++e) {
++            if (*e == '\\') {
++                e++;
++                if (*e == '\0')
++                    break;
++            }
++        }
++        if (*e == '\0') {
++            next = e;    /* last kv pair */
++        } else {
++            next = e + 1;
++            *e = '\0';
++        }
++
++        if (STRPREFIX(p, "image=")) {
++            src->path = g_strdup(p + strlen("image="));
++        } else if (STRPREFIX(p, "etcd_prefix=")) {
++            src->query = g_strdup(p + strlen("etcd_prefix="));
++        } else if (STRPREFIX(p, "config_file=")) {
++            src->configFile = g_strdup(p + strlen("config_file="));
++        } else if (STRPREFIX(p, "etcd_host=")) {
++            char *h, *sep;
++
++            h = p + strlen("etcd_host=");
++            while (h < e) {
++                for (sep = h; sep < e; ++sep) {
++                    if (*sep == '\\' && (sep[1] == ',' ||
++                                         sep[1] == ';' ||
++                                         sep[1] == ' ')) {
++                        *sep = '\0';
++                        sep += 2;
++                        break;
++                    }
++                }
++
++                if (virStorageSourceRBDAddHost(src, h) < 0)
++                    return -1;
++
++                h = sep;
++            }
++        }
++
++        p = next;
++    }
++
++    if (!src->path) {
++        return -1;
++    }
++
++    return 0;
++}
++
++
+ static int
+ virStorageSourceParseNBDColonString(const char *nbdstr,
+                                     virStorageSourcePtr src)
+@@ -2992,6 +3062,11 @@ virStorageSourceParseBackingColon(virStorageSourcePtr src,
+             return -1;
+         break;
+ 
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++        if (virStorageSourceParseVitastorColonString(path, src) < 0)
++            return -1;
++        break;
++
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+@@ -3581,6 +3656,54 @@ virStorageSourceParseBackingJSONRBD(virStorageSourcePtr src,
+     return 0;
+ }
+ 
++static int
++virStorageSourceParseBackingJSONVitastor(virStorageSourcePtr src,
++                                         virJSONValuePtr json,
++                                         const char *jsonstr G_GNUC_UNUSED,
++                                         int opaque G_GNUC_UNUSED)
++{
++    const char *filename;
++    const char *image = virJSONValueObjectGetString(json, "image");
++    const char *conf = virJSONValueObjectGetString(json, "config_path");
++    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd_prefix");
++    virJSONValuePtr servers = virJSONValueObjectGetArray(json, "server");
++    size_t nservers;
++    size_t i;
++
++    src->type = VIR_STORAGE_TYPE_NETWORK;
++    src->protocol = VIR_STORAGE_NET_PROTOCOL_VITASTOR;
++
++    /* legacy syntax passed via 'filename' option */
++    if ((filename = virJSONValueObjectGetString(json, "filename")))
++        return virStorageSourceParseVitastorColonString(filename, src);
++
++    if (!image) {
++        virReportError(VIR_ERR_INVALID_ARG, "%s",
++                       _("missing image name in Vitastor backing volume "
++                         "JSON specification"));
++        return -1;
++    }
++
++    src->path = g_strdup(image);
++    src->configFile = g_strdup(conf);
++    src->query = g_strdup(etcd_prefix);
++
++    if (servers) {
++        nservers = virJSONValueArraySize(servers);
++
++        src->hosts = g_new0(virStorageNetHostDef, nservers);
++        src->nhosts = nservers;
++
++        for (i = 0; i < nservers; i++) {
++            if (virStorageSourceParseBackingJSONInetSocketAddress(src->hosts + i,
++                                                                  virJSONValueArrayGet(servers, i)) < 0)
++                return -1;
++        }
++    }
++
++    return 0;
++}
++
+ static int
+ virStorageSourceParseBackingJSONRaw(virStorageSourcePtr src,
+                                     virJSONValuePtr json,
+@@ -3759,6 +3882,7 @@ static const struct virStorageSourceJSONDriverParser jsonParsers[] = {
+     {"sheepdog", false, virStorageSourceParseBackingJSONSheepdog, 0},
+     {"ssh", false, virStorageSourceParseBackingJSONSSH, 0},
+     {"rbd", false, virStorageSourceParseBackingJSONRBD, 0},
++    {"vitastor", false, virStorageSourceParseBackingJSONVitastor, 0},
+     {"raw", true, virStorageSourceParseBackingJSONRaw, 0},
+     {"nfs", false, virStorageSourceParseBackingJSONNFS, 0},
+     {"vxhs", false, virStorageSourceParseBackingJSONVxHS, 0},
+@@ -4503,6 +4627,7 @@ virStorageSourceNetworkDefaultPort(virStorageNetProtocol protocol)
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+             return 24007;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
+             /* we don't provide a default for RBD */
+             return 0;
+diff --git a/src/util/virstoragefile.h b/src/util/virstoragefile.h
+index 5689c39..3eb4e3c 100644
+--- a/src/util/virstoragefile.h
++++ b/src/util/virstoragefile.h
+@@ -136,6 +136,7 @@ typedef enum {
+     VIR_STORAGE_NET_PROTOCOL_SSH,
+     VIR_STORAGE_NET_PROTOCOL_VXHS,
+     VIR_STORAGE_NET_PROTOCOL_NFS,
++    VIR_STORAGE_NET_PROTOCOL_VITASTOR,
+ 
+     VIR_STORAGE_NET_PROTOCOL_LAST
+ } virStorageNetProtocol;
+diff --git a/tests/storagepoolcapsschemadata/poolcaps-fs.xml b/tests/storagepoolcapsschemadata/poolcaps-fs.xml
+index eee75af..8bd0a57 100644
+--- a/tests/storagepoolcapsschemadata/poolcaps-fs.xml
++++ b/tests/storagepoolcapsschemadata/poolcaps-fs.xml
+@@ -204,4 +204,11 @@
+       </enum>
+     </volOptions>
+   </pool>
++  <pool type='vitastor' supported='no'>
++    <volOptions>
++      <defaultFormat type='raw'/>
++      <enum name='targetFormatType'>
++      </enum>
++    </volOptions>
++  </pool>
+ </storagepoolCapabilities>
+diff --git a/tests/storagepoolcapsschemadata/poolcaps-full.xml b/tests/storagepoolcapsschemadata/poolcaps-full.xml
+index 805950a..852df0d 100644
+--- a/tests/storagepoolcapsschemadata/poolcaps-full.xml
++++ b/tests/storagepoolcapsschemadata/poolcaps-full.xml
+@@ -204,4 +204,11 @@
+       </enum>
+     </volOptions>
+   </pool>
++  <pool type='vitastor' supported='yes'>
++    <volOptions>
++      <defaultFormat type='raw'/>
++      <enum name='targetFormatType'>
++      </enum>
++    </volOptions>
++  </pool>
+ </storagepoolCapabilities>
+diff --git a/tests/storagepoolxml2argvtest.c b/tests/storagepoolxml2argvtest.c
+index 967d1f2..1e8ff7a 100644
+--- a/tests/storagepoolxml2argvtest.c
++++ b/tests/storagepoolxml2argvtest.c
+@@ -68,6 +68,7 @@ testCompareXMLToArgvFiles(bool shouldFail,
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_ZFS:
+     case VIR_STORAGE_POOL_VSTORAGE:
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_LAST:
+     default:
+         VIR_TEST_DEBUG("pool type '%s' has no xml2argv test", defTypeStr);
+diff --git a/tools/virsh-pool.c b/tools/virsh-pool.c
+index 7835fa6..8841fcf 100644
+--- a/tools/virsh-pool.c
++++ b/tools/virsh-pool.c
+@@ -1237,6 +1237,9 @@ cmdPoolList(vshControl *ctl, const vshCmd *cmd G_GNUC_UNUSED)
+             case VIR_STORAGE_POOL_VSTORAGE:
+                 flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE;
+                 break;
++            case VIR_STORAGE_POOL_VITASTOR:
++                flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR;
++                break;
+             case VIR_STORAGE_POOL_LAST:
+                 break;
+             }

patches/libvirt-7.5-vitastor.diff

@@ -0,0 +1,661 @@
+commit c6e1958a1b4974828e8e5852beb252ce6594e670
+Author: Vitaliy Filippov <vitalif@yourcmc.ru>
+Date:   Mon Jun 28 01:20:19 2021 +0300
+
+    Add Vitastor support
+
+diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng
+index 5ea14b6..a9df168 100644
+--- a/docs/schemas/domaincommon.rng
++++ b/docs/schemas/domaincommon.rng
+@@ -1859,6 +1859,35 @@
+     </element>
+   </define>
+ 
++  <define name="diskSourceNetworkProtocolVitastor">
++    <element name="source">
++      <interleave>
++        <attribute name="protocol">
++          <value>vitastor</value>
++        </attribute>
++        <ref name="diskSourceCommon"/>
++        <optional>
++          <attribute name="name"/>
++        </optional>
++        <optional>
++          <attribute name="query"/>
++        </optional>
++        <zeroOrMore>
++          <ref name="diskSourceNetworkHost"/>
++        </zeroOrMore>
++        <optional>
++          <element name="config">
++            <attribute name="file">
++              <ref name="absFilePath"/>
++            </attribute>
++            <empty/>
++          </element>
++        </optional>
++        <empty/>
++      </interleave>
++    </element>
++  </define>
++
+   <define name="diskSourceNetworkProtocolISCSI">
+     <element name="source">
+       <attribute name="protocol">
+@@ -2115,6 +2144,7 @@
+       <ref name="diskSourceNetworkProtocolSimple"/>
+       <ref name="diskSourceNetworkProtocolVxHS"/>
+       <ref name="diskSourceNetworkProtocolNFS"/>
++      <ref name="diskSourceNetworkProtocolVitastor"/>
+     </choice>
+   </define>
+ 
+diff --git a/include/libvirt/libvirt-storage.h b/include/libvirt/libvirt-storage.h
+index 089e1e0..d7e7ef4 100644
+--- a/include/libvirt/libvirt-storage.h
++++ b/include/libvirt/libvirt-storage.h
+@@ -245,6 +245,7 @@ typedef enum {
+     VIR_CONNECT_LIST_STORAGE_POOLS_ZFS           = 1 << 17,
+     VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE      = 1 << 18,
+     VIR_CONNECT_LIST_STORAGE_POOLS_ISCSI_DIRECT  = 1 << 19,
++    VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR      = 1 << 20,
+ } virConnectListAllStoragePoolsFlags;
+ 
+ int                     virConnectListAllStoragePools(virConnectPtr conn,
+diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
+index d78f846..f7222e3 100644
+--- a/src/conf/domain_conf.c
++++ b/src/conf/domain_conf.c
+@@ -8251,7 +8251,8 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
+     src->configFile = virXPathString("string(./config/@file)", ctxt);
+ 
+     if (src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTP ||
+-        src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTPS)
++        src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTPS ||
++        src->protocol == VIR_STORAGE_NET_PROTOCOL_VITASTOR)
+         src->query = virXMLPropString(node, "query");
+ 
+     if (virDomainStorageNetworkParseHosts(node, ctxt, &src->hosts, &src->nhosts) < 0)
+@@ -30775,6 +30776,7 @@ virDomainStorageSourceTranslateSourcePool(virStorageSource *src,
+ 
+     case VIR_STORAGE_POOL_MPATH:
+     case VIR_STORAGE_POOL_RBD:
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_SHEEPDOG:
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_LAST:
+diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c
+index 2aa9a3d..166ca1f 100644
+--- a/src/conf/storage_conf.c
++++ b/src/conf/storage_conf.c
+@@ -60,7 +60,7 @@ VIR_ENUM_IMPL(virStoragePool,
+               "logical", "disk", "iscsi",
+               "iscsi-direct", "scsi", "mpath",
+               "rbd", "sheepdog", "gluster",
+-              "zfs", "vstorage",
++              "zfs", "vstorage", "vitastor",
+ );
+ 
+ VIR_ENUM_IMPL(virStoragePoolFormatFileSystem,
+@@ -246,6 +246,18 @@ static virStoragePoolTypeInfo poolTypeInfo[] = {
+           .formatToString = virStorageFileFormatTypeToString,
+       }
+     },
++    {.poolType = VIR_STORAGE_POOL_VITASTOR,
++     .poolOptions = {
++         .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
++                   VIR_STORAGE_POOL_SOURCE_NETWORK |
++                   VIR_STORAGE_POOL_SOURCE_NAME),
++      },
++      .volOptions = {
++          .defaultFormat = VIR_STORAGE_FILE_RAW,
++          .formatFromString = virStorageVolumeFormatFromString,
++          .formatToString = virStorageFileFormatTypeToString,
++      }
++    },
+     {.poolType = VIR_STORAGE_POOL_SHEEPDOG,
+      .poolOptions = {
+          .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
+@@ -546,6 +558,11 @@ virStoragePoolDefParseSource(xmlXPathContextPtr ctxt,
+                        _("element 'name' is mandatory for RBD pool"));
+         return -1;
+     }
++    if (pool_type == VIR_STORAGE_POOL_VITASTOR && source->name == NULL) {
++        virReportError(VIR_ERR_XML_ERROR, "%s",
++                       _("element 'name' is mandatory for Vitastor pool"));
++        return -1;
++    }
+ 
+     if (options->formatFromString) {
+         g_autofree char *format = NULL;
+@@ -1182,6 +1199,7 @@ virStoragePoolDefFormatBuf(virBuffer *buf,
+     /* RBD, Sheepdog, Gluster and Iscsi-direct devices are not local block devs nor
+      * files, so they don't have a target */
+     if (def->type != VIR_STORAGE_POOL_RBD &&
++        def->type != VIR_STORAGE_POOL_VITASTOR &&
+         def->type != VIR_STORAGE_POOL_SHEEPDOG &&
+         def->type != VIR_STORAGE_POOL_GLUSTER &&
+         def->type != VIR_STORAGE_POOL_ISCSI_DIRECT) {
+diff --git a/src/conf/storage_conf.h b/src/conf/storage_conf.h
+index 76efaac..928149a 100644
+--- a/src/conf/storage_conf.h
++++ b/src/conf/storage_conf.h
+@@ -106,6 +106,7 @@ typedef enum {
+     VIR_STORAGE_POOL_GLUSTER,  /* Gluster device */
+     VIR_STORAGE_POOL_ZFS,      /* ZFS */
+     VIR_STORAGE_POOL_VSTORAGE, /* Virtuozzo Storage */
++    VIR_STORAGE_POOL_VITASTOR, /* Vitastor */
+ 
+     VIR_STORAGE_POOL_LAST,
+ } virStoragePoolType;
+@@ -465,6 +466,7 @@ VIR_ENUM_DECL(virStoragePartedFs);
+                  VIR_CONNECT_LIST_STORAGE_POOLS_SCSI     | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_MPATH    | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_RBD      | \
++                 VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER  | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_ZFS      | \
+diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
+index 5ca06fa..05ded49 100644
+--- a/src/conf/storage_source_conf.c
++++ b/src/conf/storage_source_conf.c
+@@ -85,6 +85,7 @@ VIR_ENUM_IMPL(virStorageNetProtocol,
+               "ssh",
+               "vxhs",
+               "nfs",
++              "vitastor",
+ );
+ 
+ 
+@@ -1262,6 +1263,7 @@ virStorageSourceNetworkDefaultPort(virStorageNetProtocol protocol)
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+             return 24007;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
+             /* we don't provide a default for RBD */
+             return 0;
+diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
+index 389c7b5..dbf02e3 100644
+--- a/src/conf/storage_source_conf.h
++++ b/src/conf/storage_source_conf.h
+@@ -127,6 +127,7 @@ typedef enum {
+     VIR_STORAGE_NET_PROTOCOL_SSH,
+     VIR_STORAGE_NET_PROTOCOL_VXHS,
+     VIR_STORAGE_NET_PROTOCOL_NFS,
++    VIR_STORAGE_NET_PROTOCOL_VITASTOR,
+ 
+     VIR_STORAGE_NET_PROTOCOL_LAST
+ } virStorageNetProtocol;
+diff --git a/src/conf/virstorageobj.c b/src/conf/virstorageobj.c
+index 24957d6..4520a73 100644
+--- a/src/conf/virstorageobj.c
++++ b/src/conf/virstorageobj.c
+@@ -1487,6 +1487,7 @@ virStoragePoolObjSourceFindDuplicateCb(const void *payload,
+             return 1;
+         break;
+ 
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_RBD:
+     case VIR_STORAGE_POOL_LAST:
+         break;
+@@ -1986,6 +1987,8 @@ virStoragePoolObjMatch(virStoragePoolObj *obj,
+                (obj->def->type == VIR_STORAGE_POOL_MPATH))   ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_RBD) &&
+                (obj->def->type == VIR_STORAGE_POOL_RBD))     ||
++              (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR) &&
++               (obj->def->type == VIR_STORAGE_POOL_VITASTOR)) ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG) &&
+                (obj->def->type == VIR_STORAGE_POOL_SHEEPDOG)) ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER) &&
+diff --git a/src/libvirt-storage.c b/src/libvirt-storage.c
+index 2a7cdca..f756be1 100644
+--- a/src/libvirt-storage.c
++++ b/src/libvirt-storage.c
+@@ -92,6 +92,7 @@ virStoragePoolGetConnect(virStoragePoolPtr pool)
+  * VIR_CONNECT_LIST_STORAGE_POOLS_SCSI
+  * VIR_CONNECT_LIST_STORAGE_POOLS_MPATH
+  * VIR_CONNECT_LIST_STORAGE_POOLS_RBD
++ * VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR
+  * VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG
+  * VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER
+  * VIR_CONNECT_LIST_STORAGE_POOLS_ZFS
+diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
+index 56cb9ab..dfb31b9 100644
+--- a/src/libxl/libxl_conf.c
++++ b/src/libxl/libxl_conf.c
+@@ -972,6 +972,7 @@ libxlMakeNetworkDiskSrcStr(virStorageSource *src,
+     case VIR_STORAGE_NET_PROTOCOL_SSH:
+     case VIR_STORAGE_NET_PROTOCOL_VXHS:
+     case VIR_STORAGE_NET_PROTOCOL_NFS:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+         virReportError(VIR_ERR_NO_SUPPORT,
+diff --git a/src/libxl/xen_xl.c b/src/libxl/xen_xl.c
+index c0905b0..c172378 100644
+--- a/src/libxl/xen_xl.c
++++ b/src/libxl/xen_xl.c
+@@ -1540,6 +1540,7 @@ xenFormatXLDiskSrcNet(virStorageSource *src)
+     case VIR_STORAGE_NET_PROTOCOL_SSH:
+     case VIR_STORAGE_NET_PROTOCOL_VXHS:
+     case VIR_STORAGE_NET_PROTOCOL_NFS:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+         virReportError(VIR_ERR_NO_SUPPORT,
+diff --git a/src/qemu/qemu_block.c b/src/qemu/qemu_block.c
+index 6627d04..c33f428 100644
+--- a/src/qemu/qemu_block.c
++++ b/src/qemu/qemu_block.c
+@@ -928,6 +928,38 @@ qemuBlockStorageSourceGetRBDProps(virStorageSource *src,
+ }
+ 
+ 
++static virJSONValue *
++qemuBlockStorageSourceGetVitastorProps(virStorageSource *src)
++{
++    virJSONValuePtr ret = NULL;
++    virStorageNetHostDefPtr host;
++    size_t i;
++    g_auto(virBuffer) buf = VIR_BUFFER_INITIALIZER;
++    g_autofree char *etcd = NULL;
++
++    for (i = 0; i < src->nhosts; i++) {
++        host = src->hosts + i;
++        if ((virStorageNetHostTransport)host->transport != VIR_STORAGE_NET_HOST_TRANS_TCP) {
++            return NULL;
++        }
++        virBufferAsprintf(&buf, i > 0 ? ",%s:%u" : "%s:%u", host->name, host->port);
++    }
++    if (src->nhosts > 0) {
++        etcd = virBufferContentAndReset(&buf);
++    }
++
++    if (virJSONValueObjectCreate(&ret,
++                                 "S:etcd_host", etcd,
++                                 "S:etcd_prefix", src->query,
++                                 "S:config_path", src->configFile,
++                                 "s:image", src->path,
++                                 NULL) < 0)
++        return NULL;
++
++    return ret;
++}
++
++
+ static virJSONValue *
+ qemuBlockStorageSourceGetSheepdogProps(virStorageSource *src)
+ {
+@@ -1218,6 +1250,12 @@ qemuBlockStorageSourceGetBackendProps(virStorageSource *src,
+                 return NULL;
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            driver = "vitastor";
++            if (!(fileprops = qemuBlockStorageSourceGetVitastorProps(src)))
++                return NULL;
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             driver = "sheepdog";
+             if (!(fileprops = qemuBlockStorageSourceGetSheepdogProps(src)))
+@@ -2231,6 +2269,7 @@ qemuBlockGetBackingStoreString(virStorageSource *src,
+ 
+             case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             case VIR_STORAGE_NET_PROTOCOL_RBD:
++            case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+             case VIR_STORAGE_NET_PROTOCOL_VXHS:
+             case VIR_STORAGE_NET_PROTOCOL_NFS:
+             case VIR_STORAGE_NET_PROTOCOL_SSH:
+@@ -2608,6 +2647,12 @@ qemuBlockStorageSourceCreateGetStorageProps(virStorageSource *src,
+                 return -1;
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            driver = "vitastor";
++            if (!(location = qemuBlockStorageSourceGetVitastorProps(src)))
++                return -1;
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             driver = "sheepdog";
+             if (!(location = qemuBlockStorageSourceGetSheepdogProps(src)))
+diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
+index ea51369..8258632 100644
+--- a/src/qemu/qemu_command.c
++++ b/src/qemu/qemu_command.c
+@@ -1074,6 +1074,43 @@ qemuBuildNetworkDriveStr(virStorageSource *src,
+             ret = virBufferContentAndReset(&buf);
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            if (strchr(src->path, ':')) {
++                virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
++                               _("':' not allowed in Vitastor source volume name '%s'"),
++                               src->path);
++                return NULL;
++            }
++
++            virBufferStrcat(&buf, "vitastor:image=", src->path, NULL);
++
++            if (src->nhosts > 0) {
++                virBufferAddLit(&buf, ":etcd_host=");
++                for (i = 0; i < src->nhosts; i++) {
++                    if (i)
++                        virBufferAddLit(&buf, ",");
++
++                    /* assume host containing : is ipv6 */
++                    if (strchr(src->hosts[i].name, ':'))
++                        virBufferEscape(&buf, '\\', ":", "[%s]",
++                                        src->hosts[i].name);
++                    else
++                        virBufferAsprintf(&buf, "%s", src->hosts[i].name);
++
++                    if (src->hosts[i].port)
++                        virBufferAsprintf(&buf, "\\:%u", src->hosts[i].port);
++                }
++            }
++
++            if (src->configFile)
++                virBufferEscape(&buf, '\\', ":", ":config_path=%s", src->configFile);
++
++            if (src->query)
++                virBufferEscape(&buf, '\\', ":", ":etcd_prefix=%s", src->query);
++
++            ret = virBufferContentAndReset(&buf);
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_VXHS:
+             virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+                            _("VxHS protocol does not support URI syntax"));
+diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
+index fc60e15..5ab410d 100644
+--- a/src/qemu/qemu_domain.c
++++ b/src/qemu/qemu_domain.c
+@@ -4829,7 +4829,8 @@ qemuDomainValidateStorageSource(virStorageSource *src,
+     if (src->query &&
+         (actualType != VIR_STORAGE_TYPE_NETWORK ||
+          (src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTPS &&
+-          src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTP))) {
++          src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTP &&
++          src->protocol != VIR_STORAGE_NET_PROTOCOL_VITASTOR))) {
+         virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+                        _("query is supported only with HTTP(S) protocols"));
+         return -1;
+@@ -10027,6 +10028,7 @@ qemuDomainPrepareStorageSourceTLS(virStorageSource *src,
+         break;
+ 
+     case VIR_STORAGE_NET_PROTOCOL_RBD:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+     case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+diff --git a/src/qemu/qemu_snapshot.c b/src/qemu/qemu_snapshot.c
+index 4e74ddd..14e5f2e 100644
+--- a/src/qemu/qemu_snapshot.c
++++ b/src/qemu/qemu_snapshot.c
+@@ -402,6 +402,7 @@ qemuSnapshotPrepareDiskExternalInactive(virDomainSnapshotDiskDef *snapdisk,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+@@ -494,6 +495,7 @@ qemuSnapshotPrepareDiskExternalActive(virDomainObj *vm,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+         case VIR_STORAGE_NET_PROTOCOL_HTTP:
+@@ -647,6 +649,7 @@ qemuSnapshotPrepareDiskInternal(virDomainDiskDef *disk,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c
+index c2ff4b8..70d0689 100644
+--- a/src/storage/storage_driver.c
++++ b/src/storage/storage_driver.c
+@@ -1644,6 +1644,7 @@ storageVolLookupByPathCallback(virStoragePoolObj *obj,
+ 
+         case VIR_STORAGE_POOL_GLUSTER:
+         case VIR_STORAGE_POOL_RBD:
++        case VIR_STORAGE_POOL_VITASTOR:
+         case VIR_STORAGE_POOL_SHEEPDOG:
+         case VIR_STORAGE_POOL_ZFS:
+         case VIR_STORAGE_POOL_LAST:
+diff --git a/src/storage_file/storage_source_backingstore.c b/src/storage_file/storage_source_backingstore.c
+index e48ae72..d7a9b72 100644
+--- a/src/storage_file/storage_source_backingstore.c
++++ b/src/storage_file/storage_source_backingstore.c
+@@ -284,6 +284,75 @@ virStorageSourceParseRBDColonString(const char *rbdstr,
+ }
+ 
+ 
++static int
++virStorageSourceParseVitastorColonString(const char *colonstr,
++                                         virStorageSource *src)
++{
++    char *p, *e, *next;
++    g_autofree char *options = NULL;
++
++    /* optionally skip the "vitastor:" prefix if provided */
++    if (STRPREFIX(colonstr, "vitastor:"))
++        colonstr += strlen("vitastor:");
++
++    options = g_strdup(colonstr);
++
++    p = options;
++    while (*p) {
++        /* find : delimiter or end of string */
++        for (e = p; *e && *e != ':'; ++e) {
++            if (*e == '\\') {
++                e++;
++                if (*e == '\0')
++                    break;
++            }
++        }
++        if (*e == '\0') {
++            next = e;    /* last kv pair */
++        } else {
++            next = e + 1;
++            *e = '\0';
++        }
++
++        if (STRPREFIX(p, "image=")) {
++            src->path = g_strdup(p + strlen("image="));
++        } else if (STRPREFIX(p, "etcd_prefix=")) {
++            src->query = g_strdup(p + strlen("etcd_prefix="));
++        } else if (STRPREFIX(p, "config_file=")) {
++            src->configFile = g_strdup(p + strlen("config_file="));
++        } else if (STRPREFIX(p, "etcd_host=")) {
++            char *h, *sep;
++
++            h = p + strlen("etcd_host=");
++            while (h < e) {
++                for (sep = h; sep < e; ++sep) {
++                    if (*sep == '\\' && (sep[1] == ',' ||
++                                         sep[1] == ';' ||
++                                         sep[1] == ' ')) {
++                        *sep = '\0';
++                        sep += 2;
++                        break;
++                    }
++                }
++
++                if (virStorageSourceRBDAddHost(src, h) < 0)
++                    return -1;
++
++                h = sep;
++            }
++        }
++
++        p = next;
++    }
++
++    if (!src->path) {
++        return -1;
++    }
++
++    return 0;
++}
++
++
+ static int
+ virStorageSourceParseNBDColonString(const char *nbdstr,
+                                     virStorageSource *src)
+@@ -396,6 +465,11 @@ virStorageSourceParseBackingColon(virStorageSource *src,
+             return -1;
+         break;
+ 
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++        if (virStorageSourceParseVitastorColonString(path, src) < 0)
++            return -1;
++        break;
++
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+@@ -984,6 +1058,54 @@ virStorageSourceParseBackingJSONRBD(virStorageSource *src,
+     return 0;
+ }
+ 
++static int
++virStorageSourceParseBackingJSONVitastor(virStorageSource *src,
++                                         virJSONValue *json,
++                                         const char *jsonstr G_GNUC_UNUSED,
++                                         int opaque G_GNUC_UNUSED)
++{
++    const char *filename;
++    const char *image = virJSONValueObjectGetString(json, "image");
++    const char *conf = virJSONValueObjectGetString(json, "config_path");
++    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd_prefix");
++    virJSONValue *servers = virJSONValueObjectGetArray(json, "server");
++    size_t nservers;
++    size_t i;
++
++    src->type = VIR_STORAGE_TYPE_NETWORK;
++    src->protocol = VIR_STORAGE_NET_PROTOCOL_VITASTOR;
++
++    /* legacy syntax passed via 'filename' option */
++    if ((filename = virJSONValueObjectGetString(json, "filename")))
++        return virStorageSourceParseVitastorColonString(filename, src);
++
++    if (!image) {
++        virReportError(VIR_ERR_INVALID_ARG, "%s",
++                       _("missing image name in Vitastor backing volume "
++                         "JSON specification"));
++        return -1;
++    }
++
++    src->path = g_strdup(image);
++    src->configFile = g_strdup(conf);
++    src->query = g_strdup(etcd_prefix);
++
++    if (servers) {
++        nservers = virJSONValueArraySize(servers);
++
++        src->hosts = g_new0(virStorageNetHostDef, nservers);
++        src->nhosts = nservers;
++
++        for (i = 0; i < nservers; i++) {
++            if (virStorageSourceParseBackingJSONInetSocketAddress(src->hosts + i,
++                                                                  virJSONValueArrayGet(servers, i)) < 0)
++                return -1;
++        }
++    }
++
++    return 0;
++}
++
+ static int
+ virStorageSourceParseBackingJSONRaw(virStorageSource *src,
+                                     virJSONValue *json,
+@@ -1162,6 +1284,7 @@ static const struct virStorageSourceJSONDriverParser jsonParsers[] = {
+     {"sheepdog", false, virStorageSourceParseBackingJSONSheepdog, 0},
+     {"ssh", false, virStorageSourceParseBackingJSONSSH, 0},
+     {"rbd", false, virStorageSourceParseBackingJSONRBD, 0},
++    {"vitastor", false, virStorageSourceParseBackingJSONVitastor, 0},
+     {"raw", true, virStorageSourceParseBackingJSONRaw, 0},
+     {"nfs", false, virStorageSourceParseBackingJSONNFS, 0},
+     {"vxhs", false, virStorageSourceParseBackingJSONVxHS, 0},
+diff --git a/src/test/test_driver.c b/src/test/test_driver.c
+index ef0ddab..2173dc3 100644
+--- a/src/test/test_driver.c
++++ b/src/test/test_driver.c
+@@ -7131,6 +7131,7 @@ testStorageVolumeTypeForPool(int pooltype)
+     case VIR_STORAGE_POOL_ISCSI_DIRECT:
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_RBD:
++    case VIR_STORAGE_POOL_VITASTOR:
+         return VIR_STORAGE_VOL_NETWORK;
+     case VIR_STORAGE_POOL_LOGICAL:
+     case VIR_STORAGE_POOL_DISK:
+diff --git a/tests/storagepoolcapsschemadata/poolcaps-fs.xml b/tests/storagepoolcapsschemadata/poolcaps-fs.xml
+index eee75af..8bd0a57 100644
+--- a/tests/storagepoolcapsschemadata/poolcaps-fs.xml
++++ b/tests/storagepoolcapsschemadata/poolcaps-fs.xml
+@@ -204,4 +204,11 @@
+       </enum>
+     </volOptions>
+   </pool>
++  <pool type='vitastor' supported='no'>
++    <volOptions>
++      <defaultFormat type='raw'/>
++      <enum name='targetFormatType'>
++      </enum>
++    </volOptions>
++  </pool>
+ </storagepoolCapabilities>
+diff --git a/tests/storagepoolcapsschemadata/poolcaps-full.xml b/tests/storagepoolcapsschemadata/poolcaps-full.xml
+index 805950a..852df0d 100644
+--- a/tests/storagepoolcapsschemadata/poolcaps-full.xml
++++ b/tests/storagepoolcapsschemadata/poolcaps-full.xml
+@@ -204,4 +204,11 @@
+       </enum>
+     </volOptions>
+   </pool>
++  <pool type='vitastor' supported='yes'>
++    <volOptions>
++      <defaultFormat type='raw'/>
++      <enum name='targetFormatType'>
++      </enum>
++    </volOptions>
++  </pool>
+ </storagepoolCapabilities>
+diff --git a/tests/storagepoolxml2argvtest.c b/tests/storagepoolxml2argvtest.c
+index 449b745..7f95cc8 100644
+--- a/tests/storagepoolxml2argvtest.c
++++ b/tests/storagepoolxml2argvtest.c
+@@ -68,6 +68,7 @@ testCompareXMLToArgvFiles(bool shouldFail,
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_ZFS:
+     case VIR_STORAGE_POOL_VSTORAGE:
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_LAST:
+     default:
+         VIR_TEST_DEBUG("pool type '%s' has no xml2argv test", defTypeStr);
+diff --git a/tools/virsh-pool.c b/tools/virsh-pool.c
+index 18f3839..c8e1436 100644
+--- a/tools/virsh-pool.c
++++ b/tools/virsh-pool.c
+@@ -1231,6 +1231,9 @@ cmdPoolList(vshControl *ctl, const vshCmd *cmd G_GNUC_UNUSED)
+             case VIR_STORAGE_POOL_VSTORAGE:
+                 flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE;
+                 break;
++            case VIR_STORAGE_POOL_VITASTOR:
++                flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR;
++                break;
+             case VIR_STORAGE_POOL_LAST:
+                 break;
+             }

patches/libvirt-example.xml

@@ -0,0 +1,32 @@
+<!-- Example libvirt VM configuration with Vitastor disk -->
+<domain type='kvm'>
+  <name>debian9</name>
+  <uuid>96f277fb-fd9c-49da-bf21-a5cfd54eb162</uuid>
+  <memory unit="KiB">524288</memory>
+  <currentMemory>524288</currentMemory>
+  <vcpu>1</vcpu>
+  <os>
+    <type arch='x86_64'>hvm</type>
+    <boot dev='hd' />
+  </os>
+  <devices>
+    <emulator>/usr/bin/qemu-system-x86_64</emulator>
+    <disk type='network' device='disk'>
+      <target dev='vda' bus='virtio' />
+      <driver name='qemu' type='raw' />
+      <!-- name is Vitastor image name -->
+      <!-- config (optional) is the path to Vitastor's configuration file -->
+      <!-- query (optional) is Vitastor's etcd_prefix -->
+      <source protocol='vitastor' name='debian9' query='/vitastor' config='/etc/vitastor/vitastor.conf'>
+        <!-- hosts = etcd addresses -->
+        <host name='192.168.7.2' port='2379' />
+      </source>
+      <!-- required because Vitastor only supports 4k physical sectors -->
+      <blockio physical_block_size="4096" logical_block_size="512" />
+    </disk>
+    <interface type='network'>
+      <source network='default' />
+    </interface>
+    <graphics type='vnc' port='-1' />
+  </devices>
+</domain>

patches/nova-20.diff

@@ -0,0 +1,287 @@
+diff --git a/nova/virt/image/model.py b/nova/virt/image/model.py
+index 971f7e9c07..70ed70d5e2 100644
+--- a/nova/virt/image/model.py
++++ b/nova/virt/image/model.py
+@@ -129,3 +129,22 @@ class RBDImage(Image):
+         self.user = user
+         self.password = password
+         self.servers = servers
++
++
++class VitastorImage(Image):
++    """Class for images in a remote Vitastor cluster"""
++
++    def __init__(self, name, etcd_address = None, etcd_prefix = None, config_path = None):
++        """Create a new Vitastor image object
++
++        :param name: name of the image
++        :param etcd_address: etcd URL(s) (optional)
++        :param etcd_prefix: etcd prefix (optional)
++        :param config_path: path to the configuration (optional)
++        """
++        super(RBDImage, self).__init__(FORMAT_RAW)
++
++        self.name = name
++        self.etcd_address = etcd_address
++        self.etcd_prefix = etcd_prefix
++        self.config_path = config_path
+diff --git a/nova/virt/images.py b/nova/virt/images.py
+index 5358f3766a..ebe3d6effb 100644
+--- a/nova/virt/images.py
++++ b/nova/virt/images.py
+@@ -41,7 +41,7 @@ IMAGE_API = glance.API()
+ 
+ def qemu_img_info(path, format=None):
+     """Return an object containing the parsed output from qemu-img info."""
+-    if not os.path.exists(path) and not path.startswith('rbd:'):
++    if not os.path.exists(path) and not path.startswith('rbd:') and not path.startswith('vitastor:'):
+         raise exception.DiskNotFound(location=path)
+ 
+     info = nova.privsep.qemu.unprivileged_qemu_img_info(path, format=format)
+@@ -50,7 +50,7 @@ def qemu_img_info(path, format=None):
+ 
+ def privileged_qemu_img_info(path, format=None, output_format='json'):
+     """Return an object containing the parsed output from qemu-img info."""
+-    if not os.path.exists(path) and not path.startswith('rbd:'):
++    if not os.path.exists(path) and not path.startswith('rbd:') and not path.startswith('vitastor:'):
+         raise exception.DiskNotFound(location=path)
+ 
+     info = nova.privsep.qemu.privileged_qemu_img_info(path, format=format)
+diff --git a/nova/virt/libvirt/config.py b/nova/virt/libvirt/config.py
+index f9475776b3..51573fe41d 100644
+--- a/nova/virt/libvirt/config.py
++++ b/nova/virt/libvirt/config.py
+@@ -1060,6 +1060,8 @@ class LibvirtConfigGuestDisk(LibvirtConfigGuestDevice):
+         self.driver_iommu = False
+         self.source_path = None
+         self.source_protocol = None
++        self.source_query = None
++        self.source_config = None
+         self.source_name = None
+         self.source_hosts = []
+         self.source_ports = []
+@@ -1186,7 +1188,8 @@ class LibvirtConfigGuestDisk(LibvirtConfigGuestDevice):
+         elif self.source_type == "mount":
+             dev.append(etree.Element("source", dir=self.source_path))
+         elif self.source_type == "network" and self.source_protocol:
+-            source = etree.Element("source", protocol=self.source_protocol)
++            source = etree.Element("source", protocol=self.source_protocol,
++                query=self.source_query, config=self.source_config)
+             if self.source_name is not None:
+                 source.set('name', self.source_name)
+             hosts_info = zip(self.source_hosts, self.source_ports)
+diff --git a/nova/virt/libvirt/driver.py b/nova/virt/libvirt/driver.py
+index 391231c527..34dc60dcdd 100644
+--- a/nova/virt/libvirt/driver.py
++++ b/nova/virt/libvirt/driver.py
+@@ -179,6 +179,7 @@ VOLUME_DRIVERS = {
+     'local': 'nova.virt.libvirt.volume.volume.LibvirtVolumeDriver',
+     'fake': 'nova.virt.libvirt.volume.volume.LibvirtFakeVolumeDriver',
+     'rbd': 'nova.virt.libvirt.volume.net.LibvirtNetVolumeDriver',
++    'vitastor': 'nova.virt.libvirt.volume.vitastor.LibvirtVitastorVolumeDriver',
+     'nfs': 'nova.virt.libvirt.volume.nfs.LibvirtNFSVolumeDriver',
+     'smbfs': 'nova.virt.libvirt.volume.smbfs.LibvirtSMBFSVolumeDriver',
+     'fibre_channel': 'nova.virt.libvirt.volume.fibrechannel.LibvirtFibreChannelVolumeDriver',  # noqa:E501
+@@ -385,10 +386,10 @@ class LibvirtDriver(driver.ComputeDriver):
+         # This prevents the risk of one test setting a capability
+         # which bleeds over into other tests.
+ 
+-        # LVM and RBD require raw images. If we are not configured to
++        # LVM, RBD, Vitastor require raw images. If we are not configured to
+         # force convert images into raw format, then we _require_ raw
+         # images only.
+-        raw_only = ('rbd', 'lvm')
++        raw_only = ('rbd', 'lvm', 'vitastor')
+         requires_raw_image = (CONF.libvirt.images_type in raw_only and
+                               not CONF.force_raw_images)
+         requires_ploop_image = CONF.libvirt.virt_type == 'parallels'
+@@ -775,12 +776,12 @@ class LibvirtDriver(driver.ComputeDriver):
+         # Some imagebackends are only able to import raw disk images,
+         # and will fail if given any other format. See the bug
+         # https://bugs.launchpad.net/nova/+bug/1816686 for more details.
+-        if CONF.libvirt.images_type in ('rbd',):
++        if CONF.libvirt.images_type in ('rbd', 'vitastor'):
+             if not CONF.force_raw_images:
+                 msg = _("'[DEFAULT]/force_raw_images = False' is not "
+-                        "allowed with '[libvirt]/images_type = rbd'. "
++                        "allowed with '[libvirt]/images_type = rbd' or 'vitastor'. "
+                         "Please check the two configs and if you really "
+-                        "do want to use rbd as images_type, set "
++                        "do want to use rbd or vitastor as images_type, set "
+                         "force_raw_images to True.")
+                 raise exception.InvalidConfiguration(msg)
+ 
+@@ -2603,6 +2604,16 @@ class LibvirtDriver(driver.ComputeDriver):
+                     if connection_info['data'].get('auth_enabled'):
+                         username = connection_info['data']['auth_username']
+                         path = f"rbd:{volume_name}:id={username}"
++                elif connection_info['driver_volume_type'] == 'vitastor':
++                    volume_name = connection_info['data']['name']
++                    path = 'vitastor:image='+volume_name.replace(':', '\\:')
++                    for k in [ 'config_path', 'etcd_address', 'etcd_prefix' ]:
++                        if k in connection_info['data']:
++                            kk = k
++                            if kk == 'etcd_address':
++                                # FIXME use etcd_address in qemu driver
++                                kk = 'etcd_host'
++                            path += ":"+kk+"="+connection_info['data'][k].replace(':', '\\:')
+                 else:
+                     path = 'unknown'
+                     raise exception.DiskNotFound(location='unknown')
+@@ -2827,8 +2838,8 @@ class LibvirtDriver(driver.ComputeDriver):
+ 
+         image_format = CONF.libvirt.snapshot_image_format or source_type
+ 
+-        # NOTE(bfilippov): save lvm and rbd as raw
+-        if image_format == 'lvm' or image_format == 'rbd':
++        # NOTE(bfilippov): save lvm and rbd and vitastor as raw
++        if image_format == 'lvm' or image_format == 'rbd' or image_format == 'vitastor':
+             image_format = 'raw'
+ 
+         metadata = self._create_snapshot_metadata(instance.image_meta,
+@@ -2899,7 +2910,7 @@ class LibvirtDriver(driver.ComputeDriver):
+                               expected_state=task_states.IMAGE_UPLOADING)
+ 
+             # TODO(nic): possibly abstract this out to the root_disk
+-            if source_type == 'rbd' and live_snapshot:
++            if (source_type == 'rbd' or source_type == 'vitastor') and live_snapshot:
+                 # Standard snapshot uses qemu-img convert from RBD which is
+                 # not safe to run with live_snapshot.
+                 live_snapshot = False
+@@ -4099,7 +4110,7 @@ class LibvirtDriver(driver.ComputeDriver):
+         # cleanup rescue volume
+         lvm.remove_volumes([lvmdisk for lvmdisk in self._lvm_disks(instance)
+                                 if lvmdisk.endswith('.rescue')])
+-        if CONF.libvirt.images_type == 'rbd':
++        if CONF.libvirt.images_type == 'rbd' or CONF.libvirt.images_type == 'vitastor':
+             filter_fn = lambda disk: (disk.startswith(instance.uuid) and
+                                       disk.endswith('.rescue'))
+             rbd_utils.RBDDriver().cleanup_volumes(filter_fn)
+@@ -4356,6 +4367,8 @@ class LibvirtDriver(driver.ComputeDriver):
+         # TODO(mikal): there is a bug here if images_type has
+         # changed since creation of the instance, but I am pretty
+         # sure that this bug already exists.
++        if CONF.libvirt.images_type == 'vitastor':
++            return 'vitastor'
+         return 'rbd' if CONF.libvirt.images_type == 'rbd' else 'raw'
+ 
+     @staticmethod
+@@ -4764,10 +4777,10 @@ class LibvirtDriver(driver.ComputeDriver):
+                 finally:
+                     # NOTE(mikal): if the config drive was imported into RBD,
+                     # then we no longer need the local copy
+-                    if CONF.libvirt.images_type == 'rbd':
++                    if CONF.libvirt.images_type == 'rbd' or CONF.libvirt.images_type == 'vitastor':
+                         LOG.info('Deleting local config drive %(path)s '
+-                                 'because it was imported into RBD.',
+-                                 {'path': config_disk_local_path},
++                                 'because it was imported into %(type).',
++                                 {'path': config_disk_local_path, 'type': CONF.libvirt.images_type},
+                                  instance=instance)
+                         os.unlink(config_disk_local_path)
+ 
+diff --git a/nova/virt/libvirt/utils.py b/nova/virt/libvirt/utils.py
+index da2a6e8b8a..52c02e72f1 100644
+--- a/nova/virt/libvirt/utils.py
++++ b/nova/virt/libvirt/utils.py
+@@ -340,6 +340,10 @@ def find_disk(guest: libvirt_guest.Guest) -> ty.Tuple[str, ty.Optional[str]]:
+             disk_path = disk.source_name
+             if disk_path:
+                 disk_path = 'rbd:' + disk_path
++        elif not disk_path and disk.source_protocol == 'vitastor':
++            disk_path = disk.source_name
++            if disk_path:
++                disk_path = 'vitastor:' + disk_path
+ 
+     if not disk_path:
+         raise RuntimeError(_("Can't retrieve root device path "
+@@ -354,6 +358,8 @@ def get_disk_type_from_path(path: str) -> ty.Optional[str]:
+         return 'lvm'
+     elif path.startswith('rbd:'):
+         return 'rbd'
++    elif path.startswith('vitastor:'):
++        return 'vitastor'
+     elif (os.path.isdir(path) and
+           os.path.exists(os.path.join(path, "DiskDescriptor.xml"))):
+         return 'ploop'
+diff --git a/nova/virt/libvirt/volume/vitastor.py b/nova/virt/libvirt/volume/vitastor.py
+new file mode 100644
+index 0000000000..0256df62c1
+--- /dev/null
++++ b/nova/virt/libvirt/volume/vitastor.py
+@@ -0,0 +1,75 @@
++# Copyright (c) 2021+, Vitaliy Filippov <vitalif@yourcmc.ru>
++#
++#    Licensed under the Apache License, Version 2.0 (the "License"); you may
++#    not use this file except in compliance with the License. You may obtain
++#    a copy of the License at
++#
++#         http://www.apache.org/licenses/LICENSE-2.0
++#
++#    Unless required by applicable law or agreed to in writing, software
++#    distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
++#    WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
++#    License for the specific language governing permissions and limitations
++#    under the License.
++
++from os_brick import exception as os_brick_exception
++from os_brick import initiator
++from os_brick.initiator import connector
++from oslo_log import log as logging
++
++import nova.conf
++from nova import utils
++from nova.virt.libvirt.volume import volume as libvirt_volume
++
++
++CONF = nova.conf.CONF
++LOG = logging.getLogger(__name__)
++
++
++class LibvirtVitastorVolumeDriver(libvirt_volume.LibvirtBaseVolumeDriver):
++    """Driver to attach Vitastor volumes to libvirt."""
++    def __init__(self, host):
++        super(LibvirtVitastorVolumeDriver, self).__init__(host, is_block_dev=False)
++
++    def connect_volume(self, connection_info, instance):
++        pass
++
++    def disconnect_volume(self, connection_info, instance):
++        pass
++
++    def get_config(self, connection_info, disk_info):
++        """Returns xml for libvirt."""
++        conf = super(LibvirtVitastorVolumeDriver, self).get_config(connection_info, disk_info)
++        conf.source_type = 'network'
++        conf.source_protocol = 'vitastor'
++        conf.source_name = connection_info['data'].get('name')
++        conf.source_query = connection_info['data'].get('etcd_prefix') or None
++        conf.source_config = connection_info['data'].get('config_path') or None
++        conf.source_hosts = []
++        conf.source_ports = []
++        addresses = connection_info['data'].get('etcd_address', '')
++        if addresses:
++            if not isinstance(addresses, list):
++                addresses = addresses.split(',')
++            for addr in addresses:
++                if addr.startswith('https://'):
++                    raise NotImplementedError('Vitastor block driver does not support SSL for etcd communication yet')
++                if addr.startswith('http://'):
++                    addr = addr[7:]
++                addr = addr.rstrip('/')
++                if addr.endswith('/v3'):
++                    addr = addr[0:-3]
++                p = addr.find('/')
++                if p > 0:
++                    raise NotImplementedError('libvirt does not support custom URL paths for Vitastor etcd yet. Use /etc/vitastor/vitastor.conf')
++                p = addr.find(':')
++                port = '2379'
++                if p > 0:
++                    port = addr[p+1:]
++                    addr = addr[0:p]
++                conf.source_hosts.append(addr)
++                conf.source_ports.append(port)
++        return conf
++
++    def extend_volume(self, connection_info, instance, requested_size):
++        raise NotImplementedError

patches/qemu-3.1-vitastor.patch

@@ -0,0 +1,88 @@
+Index: qemu-3.1+dfsg/qapi/block-core.json
+===================================================================
+--- qemu-3.1+dfsg.orig/qapi/block-core.json
++++ qemu-3.1+dfsg/qapi/block-core.json
+@@ -2617,7 +2617,7 @@
+ ##
+ { 'enum': 'BlockdevDriver',
+   'data': [ 'blkdebug', 'blklogwrites', 'blkverify', 'bochs', 'cloop',
+-            'copy-on-read', 'dmg', 'file', 'ftp', 'ftps', 'gluster',
++            'copy-on-read', 'dmg', 'file', 'ftp', 'ftps', 'gluster', 'vitastor',
+             'host_cdrom', 'host_device', 'http', 'https', 'iscsi', 'luks',
+             'nbd', 'nfs', 'null-aio', 'null-co', 'nvme', 'parallels', 'qcow',
+             'qcow2', 'qed', 'quorum', 'raw', 'rbd', 'replication', 'sheepdog',
+@@ -3367,6 +3367,28 @@
+             '*tag': 'str' } }
+ 
+ ##
++# @BlockdevOptionsVitastor:
++#
++# Driver specific block device options for vitastor
++#
++# @image:       Image name
++# @inode:       Inode number
++# @pool:        Pool ID
++# @size:        Desired image size in bytes
++# @config_path: Path to Vitastor configuration
++# @etcd_host:   etcd connection address(es)
++# @etcd_prefix: etcd key/value prefix
++##
++{ 'struct': 'BlockdevOptionsVitastor',
++  'data': { '*inode': 'uint64',
++            '*pool': 'uint64',
++            '*size': 'uint64',
++            '*image': 'str',
++            '*config_path': 'str',
++            '*etcd_host': 'str',
++            '*etcd_prefix': 'str' } }
++
++##
+ # @ReplicationMode:
+ #
+ # An enumeration of replication modes.
+@@ -3713,6 +3731,7 @@
+       'rbd':        'BlockdevOptionsRbd',
+       'replication':'BlockdevOptionsReplication',
+       'sheepdog':   'BlockdevOptionsSheepdog',
++      'vitastor':   'BlockdevOptionsVitastor',
+       'ssh':        'BlockdevOptionsSsh',
+       'throttle':   'BlockdevOptionsThrottle',
+       'vdi':        'BlockdevOptionsGenericFormat',
+@@ -4158,6 +4177,17 @@
+             '*block-state-zero':    'bool' } }
+ 
+ ##
++# @BlockdevCreateOptionsVitastor:
++#
++# Driver specific image creation options for Vitastor.
++#
++# @size: Size of the virtual disk in bytes
++##
++{ 'struct': 'BlockdevCreateOptionsVitastor',
++  'data': { 'location':         'BlockdevOptionsVitastor',
++            'size':             'size' } }
++
++##
+ # @BlockdevVpcSubformat:
+ #
+ # @dynamic: Growing image file
+@@ -4212,6 +4242,7 @@
+       'qed':            'BlockdevCreateOptionsQed',
+       'rbd':            'BlockdevCreateOptionsRbd',
+       'sheepdog':       'BlockdevCreateOptionsSheepdog',
++      'vitastor':       'BlockdevCreateOptionsVitastor',
+       'ssh':            'BlockdevCreateOptionsSsh',
+       'vdi':            'BlockdevCreateOptionsVdi',
+       'vhdx':           'BlockdevCreateOptionsVhdx',
+Index: qemu-3.1+dfsg/scripts/modules/module_block.py
+===================================================================
+--- qemu-3.1+dfsg.orig/scripts/modules/module_block.py
++++ qemu-3.1+dfsg/scripts/modules/module_block.py
+@@ -88,6 +88,7 @@ def print_bottom(fheader):
+ output_file = sys.argv[1]
+ with open(output_file, 'w') as fheader:
+     print_top(fheader)
++    add_module(fheader, "vitastor", "vitastor", "vitastor")
+ 
+     for filename in sys.argv[2:]:
+         if os.path.isfile(filename):

patches/qemu-4.2-vitastor.patch

@@ -0,0 +1,88 @@
+Index: qemu/qapi/block-core.json
+===================================================================
+--- qemu.orig/qapi/block-core.json	2020-11-07 22:57:38.932613674 +0000
++++ qemu.orig/qapi/block-core.json	2020-11-07 22:59:49.890722862 +0000
+@@ -2907,7 +2907,7 @@
+             'nbd', 'nfs', 'null-aio', 'null-co', 'nvme', 'parallels', 'qcow',
+             'qcow2', 'qed', 'quorum', 'raw', 'rbd',
+             { 'name': 'replication', 'if': 'defined(CONFIG_REPLICATION)' },
+-            'sheepdog',
++            'sheepdog', 'vitastor',
+             'ssh', 'throttle', 'vdi', 'vhdx', 'vmdk', 'vpc', 'vvfat', 'vxhs' ] }
+ 
+ ##
+@@ -3725,6 +3725,28 @@
+             '*tag': 'str' } }
+ 
+ ##
++# @BlockdevOptionsVitastor:
++#
++# Driver specific block device options for vitastor
++#
++# @image:       Image name
++# @inode:       Inode number
++# @pool:        Pool ID
++# @size:        Desired image size in bytes
++# @config_path: Path to Vitastor configuration
++# @etcd_host:   etcd connection address(es)
++# @etcd_prefix: etcd key/value prefix
++##
++{ 'struct': 'BlockdevOptionsVitastor',
++  'data': { '*inode': 'uint64',
++            '*pool': 'uint64',
++            '*size': 'uint64',
++            '*image': 'str',
++            '*config_path': 'str',
++            '*etcd_host': 'str',
++            '*etcd_prefix': 'str' } }
++
++##
+ # @ReplicationMode:
+ #
+ # An enumeration of replication modes.
+@@ -4084,6 +4102,7 @@
+       'replication': { 'type': 'BlockdevOptionsReplication',
+                        'if': 'defined(CONFIG_REPLICATION)' },
+       'sheepdog':   'BlockdevOptionsSheepdog',
++      'vitastor':   'BlockdevOptionsVitastor',
+       'ssh':        'BlockdevOptionsSsh',
+       'throttle':   'BlockdevOptionsThrottle',
+       'vdi':        'BlockdevOptionsGenericFormat',
+@@ -4461,6 +4480,17 @@
+             '*cluster-size' :   'size' } }
+ 
+ ##
++# @BlockdevCreateOptionsVitastor:
++#
++# Driver specific image creation options for Vitastor.
++#
++# @size: Size of the virtual disk in bytes
++##
++{ 'struct': 'BlockdevCreateOptionsVitastor',
++  'data': { 'location':         'BlockdevOptionsVitastor',
++            'size':             'size' } }
++
++##
+ # @BlockdevVmdkSubformat:
+ #
+ # Subformat options for VMDK images
+@@ -4722,6 +4752,7 @@
+       'qed':            'BlockdevCreateOptionsQed',
+       'rbd':            'BlockdevCreateOptionsRbd',
+       'sheepdog':       'BlockdevCreateOptionsSheepdog',
++      'vitastor':       'BlockdevCreateOptionsVitastor',
+       'ssh':            'BlockdevCreateOptionsSsh',
+       'vdi':            'BlockdevCreateOptionsVdi',
+       'vhdx':           'BlockdevCreateOptionsVhdx',
+Index: qemu/scripts/modules/module_block.py
+===================================================================
+--- qemu.orig/scripts/modules/module_block.py	2020-11-07 22:57:38.936613739 +0000
++++ qemu/scripts/modules/module_block.py	2020-11-07 22:59:49.890722862 +0000
+@@ -86,6 +86,7 @@ def print_bottom(fheader):
+ output_file = sys.argv[1]
+ with open(output_file, 'w') as fheader:
+     print_top(fheader)
++    add_module(fheader, "vitastor", "vitastor", "vitastor")
+ 
+     for filename in sys.argv[2:]:
+         if os.path.isfile(filename):

patches/qemu-5.0-vitastor.patch

@@ -0,0 +1,88 @@
+Index: qemu/qapi/block-core.json
+===================================================================
+--- qemu.orig/qapi/block-core.json
++++ qemu/qapi/block-core.json
+@@ -2798,7 +2798,7 @@
+             'luks', 'nbd', 'nfs', 'null-aio', 'null-co', 'nvme', 'parallels',
+             'qcow', 'qcow2', 'qed', 'quorum', 'raw', 'rbd',
+             { 'name': 'replication', 'if': 'defined(CONFIG_REPLICATION)' },
+-            'sheepdog',
++            'sheepdog', 'vitastor',
+             'ssh', 'throttle', 'vdi', 'vhdx', 'vmdk', 'vpc', 'vvfat', 'vxhs' ] }
+ 
+ ##
+@@ -3635,6 +3635,28 @@
+             '*tag': 'str' } }
+ 
+ ##
++# @BlockdevOptionsVitastor:
++#
++# Driver specific block device options for vitastor
++#
++# @image:       Image name
++# @inode:       Inode number
++# @pool:        Pool ID
++# @size:        Desired image size in bytes
++# @config_path: Path to Vitastor configuration
++# @etcd_host:   etcd connection address(es)
++# @etcd_prefix: etcd key/value prefix
++##
++{ 'struct': 'BlockdevOptionsVitastor',
++  'data': { '*inode': 'uint64',
++            '*pool': 'uint64',
++            '*size': 'uint64',
++            '*image': 'str',
++            '*config_path': 'str',
++            '*etcd_host': 'str',
++            '*etcd_prefix': 'str' } }
++
++##
+ # @ReplicationMode:
+ #
+ # An enumeration of replication modes.
+@@ -3995,6 +4013,7 @@
+       'replication': { 'type': 'BlockdevOptionsReplication',
+                        'if': 'defined(CONFIG_REPLICATION)' },
+       'sheepdog':   'BlockdevOptionsSheepdog',
++      'vitastor':   'BlockdevOptionsVitastor',
+       'ssh':        'BlockdevOptionsSsh',
+       'throttle':   'BlockdevOptionsThrottle',
+       'vdi':        'BlockdevOptionsGenericFormat',
+@@ -4365,6 +4384,17 @@
+             '*cluster-size' :   'size' } }
+ 
+ ##
++# @BlockdevCreateOptionsVitastor:
++#
++# Driver specific image creation options for Vitastor.
++#
++# @size: Size of the virtual disk in bytes
++##
++{ 'struct': 'BlockdevCreateOptionsVitastor',
++  'data': { 'location':         'BlockdevOptionsVitastor',
++            'size':             'size' } }
++
++##
+ # @BlockdevVmdkSubformat:
+ #
+ # Subformat options for VMDK images
+@@ -4626,6 +4656,7 @@
+       'qed':            'BlockdevCreateOptionsQed',
+       'rbd':            'BlockdevCreateOptionsRbd',
+       'sheepdog':       'BlockdevCreateOptionsSheepdog',
++      'vitastor':       'BlockdevCreateOptionsVitastor',
+       'ssh':            'BlockdevCreateOptionsSsh',
+       'vdi':            'BlockdevCreateOptionsVdi',
+       'vhdx':           'BlockdevCreateOptionsVhdx',
+Index: qemu/scripts/modules/module_block.py
+===================================================================
+--- qemu.orig/scripts/modules/module_block.py
++++ qemu/scripts/modules/module_block.py
+@@ -85,6 +85,7 @@ def print_bottom(fheader):
+ output_file = sys.argv[1]
+ with open(output_file, 'w') as fheader:
+     print_top(fheader)
++    add_module(fheader, "vitastor", "vitastor", "vitastor")
+ 
+     for filename in sys.argv[2:]:
+         if os.path.isfile(filename):

patches/qemu-5.1-vitastor.patch

@@ -0,0 +1,88 @@
+Index: qemu-5.1+dfsg/qapi/block-core.json
+===================================================================
+--- qemu-5.1+dfsg.orig/qapi/block-core.json
++++ qemu-5.1+dfsg/qapi/block-core.json
+@@ -2807,7 +2807,7 @@
+             'luks', 'nbd', 'nfs', 'null-aio', 'null-co', 'nvme', 'parallels',
+             'qcow', 'qcow2', 'qed', 'quorum', 'raw', 'rbd',
+             { 'name': 'replication', 'if': 'defined(CONFIG_REPLICATION)' },
+-            'sheepdog',
++            'sheepdog', 'vitastor',
+             'ssh', 'throttle', 'vdi', 'vhdx', 'vmdk', 'vpc', 'vvfat' ] }
+ 
+ ##
+@@ -3644,6 +3644,28 @@
+             '*tag': 'str' } }
+ 
+ ##
++# @BlockdevOptionsVitastor:
++#
++# Driver specific block device options for vitastor
++#
++# @image:       Image name
++# @inode:       Inode number
++# @pool:        Pool ID
++# @size:        Desired image size in bytes
++# @config_path: Path to Vitastor configuration
++# @etcd_host:   etcd connection address(es)
++# @etcd_prefix: etcd key/value prefix
++##
++{ 'struct': 'BlockdevOptionsVitastor',
++  'data': { '*inode': 'uint64',
++            '*pool': 'uint64',
++            '*size': 'uint64',
++            '*image': 'str',
++            '*config_path': 'str',
++            '*etcd_host': 'str',
++            '*etcd_prefix': 'str' } }
++
++##
+ # @ReplicationMode:
+ #
+ # An enumeration of replication modes.
+@@ -3988,6 +4006,7 @@
+       'replication': { 'type': 'BlockdevOptionsReplication',
+                        'if': 'defined(CONFIG_REPLICATION)' },
+       'sheepdog':   'BlockdevOptionsSheepdog',
++      'vitastor':   'BlockdevOptionsVitastor',
+       'ssh':        'BlockdevOptionsSsh',
+       'throttle':   'BlockdevOptionsThrottle',
+       'vdi':        'BlockdevOptionsGenericFormat',
+@@ -4376,6 +4395,17 @@
+             '*cluster-size' :   'size' } }
+ 
+ ##
++# @BlockdevCreateOptionsVitastor:
++#
++# Driver specific image creation options for Vitastor.
++#
++# @size: Size of the virtual disk in bytes
++##
++{ 'struct': 'BlockdevCreateOptionsVitastor',
++  'data': { 'location':         'BlockdevOptionsVitastor',
++            'size':             'size' } }
++
++##
+ # @BlockdevVmdkSubformat:
+ #
+ # Subformat options for VMDK images
+@@ -4637,6 +4667,7 @@
+       'qed':            'BlockdevCreateOptionsQed',
+       'rbd':            'BlockdevCreateOptionsRbd',
+       'sheepdog':       'BlockdevCreateOptionsSheepdog',
++      'vitastor':       'BlockdevCreateOptionsVitastor',
+       'ssh':            'BlockdevCreateOptionsSsh',
+       'vdi':            'BlockdevCreateOptionsVdi',
+       'vhdx':           'BlockdevCreateOptionsVhdx',
+Index: qemu-5.1+dfsg/scripts/modules/module_block.py
+===================================================================
+--- qemu-5.1+dfsg.orig/scripts/modules/module_block.py
++++ qemu-5.1+dfsg/scripts/modules/module_block.py
+@@ -86,6 +86,7 @@ if __name__ == '__main__':
+     output_file = sys.argv[1]
+     with open(output_file, 'w') as fheader:
+         print_top(fheader)
++        add_module(fheader, "vitastor", "vitastor", "vitastor")
+ 
+         for filename in sys.argv[2:]:
+             if os.path.isfile(filename):

rpm/build-tarball.sh

@@ -0,0 +1,51 @@
+#!/bin/bash
+# Vitastor depends on QEMU and FIO headers, but QEMU and FIO don't have -devel packages
+# So we have to copy their headers into the source tarball
+
+set -e
+
+VITASTOR=$(dirname $0)
+VITASTOR=$(realpath "$VITASTOR/..")
+
+if [ -d /opt/rh/gcc-toolset-9 ]; then
+    # CentOS 8
+    EL=8
+    . /opt/rh/gcc-toolset-9/enable
+else
+    # CentOS 7
+    EL=7
+    . /opt/rh/devtoolset-9/enable
+fi
+cd ~/rpmbuild/SPECS
+rpmbuild -bp fio.spec
+perl -i -pe 's/^make V=1/exit 0; make V=1/' qemu*.spec
+rpmbuild -bc qemu*.spec
+perl -i -pe 's/^exit 0; make V=1/make V=1/' qemu*.spec
+cd ~/rpmbuild/BUILD/qemu*/
+rm -rf $VITASTOR/qemu $VITASTOR/fio
+mkdir -p $VITASTOR/qemu/b/qemu
+make -j8 config-host.h
+cp config-host.h $VITASTOR/qemu/b/qemu
+cp -r include $VITASTOR/qemu
+if [ -f qapi-schema.json ]; then
+    # QEMU 2.0
+    make qapi-types.h
+    cp qapi-types.h $VITASTOR/qemu/b/qemu
+else
+    # QEMU 3.0+
+    make qapi
+    cp -r qapi $VITASTOR/qemu/b/qemu
+fi
+cd $VITASTOR
+sh copy-qemu-includes.sh
+rm -rf qemu
+mv qemu-copy qemu
+ln -s ~/rpmbuild/BUILD/fio*/ fio
+sh copy-fio-includes.sh
+rm fio
+mv fio-copy fio
+FIO=`rpm -qi fio | perl -e 'while(<>) { /^Epoch[\s:]+(\S+)/ && print "$1:"; /^Version[\s:]+(\S+)/ && print $1; /^Release[\s:]+(\S+)/ && print "-$1"; }'`
+QEMU=`rpm -qi qemu qemu-kvm | perl -e 'while(<>) { /^Epoch[\s:]+(\S+)/ && print "$1:"; /^Version[\s:]+(\S+)/ && print $1; /^Release[\s:]+(\S+)/ && print "-$1"; }'`
+perl -i -pe 's/(Requires:\s*fio)([^\n]+)?/$1 = '$FIO'/' $VITASTOR/rpm/vitastor-el$EL.spec
+perl -i -pe 's/(Requires:\s*qemu(?:-kvm)?)([^\n]+)?/$1 = '$QEMU'/' $VITASTOR/rpm/vitastor-el$EL.spec
+tar --transform 's#^#vitastor-0.6.5/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.6.5$(rpm --eval '%dist').tar.gz *

rpm/qemu-el8.Dockerfile

@@ -0,0 +1,31 @@
+# Build packages for CentOS 8 inside a container
+# cd ..; podman build -t qemu-el8 -v `pwd`/packages:/root/packages -f rpm/qemu-el8.Dockerfile .
+
+FROM centos:8
+
+WORKDIR /root
+
+RUN rm -f /etc/yum.repos.d/CentOS-Media.repo
+RUN dnf -y install centos-release-advanced-virtualization epel-release dnf-plugins-core rpm-build
+RUN rm -rf /var/lib/dnf/*; dnf download --disablerepo='*' --enablerepo='centos-advanced-virtualization-source' --source qemu-kvm
+RUN rpm --nomd5 -i qemu*.src.rpm
+RUN cd ~/rpmbuild/SPECS && dnf builddep -y --enablerepo=PowerTools --spec qemu-kvm.spec
+
+ADD patches/qemu-*-vitastor.patch /root/vitastor/patches/
+
+RUN set -e; \
+    mkdir -p /root/packages/qemu-el8; \
+    rm -rf /root/packages/qemu-el8/*; \
+    rpm --nomd5 -i /root/qemu*.src.rpm; \
+    cd ~/rpmbuild/SPECS; \
+    PN=$(grep ^Patch qemu-kvm.spec | tail -n1 | perl -pe 's/Patch(\d+).*/$1/'); \
+    csplit qemu-kvm.spec "/^Patch$PN/"; \
+    cat xx00 > qemu-kvm.spec; \
+    head -n 1 xx01 >> qemu-kvm.spec; \
+    echo "Patch$((PN+1)): qemu-4.2-vitastor.patch" >> qemu-kvm.spec; \
+    tail -n +2 xx01 >> qemu-kvm.spec; \
+    perl -i -pe 's/(^Release:\s*\d+)/$1.vitastor/' qemu-kvm.spec; \
+    cp /root/vitastor/patches/qemu-4.2-vitastor.patch ~/rpmbuild/SOURCES; \
+    rpmbuild --nocheck -ba qemu-kvm.spec; \
+    cp ~/rpmbuild/RPMS/*/*qemu* /root/packages/qemu-el8/; \
+    cp ~/rpmbuild/SRPMS/*qemu* /root/packages/qemu-el8/