vmsplice+splice experiment in stub_osd to test it too

40% CPU according to perf is lost inside do_splice() -> unix_stream_sendpage()
without io_uring and in various exc_page_fault() with io_uring

CMakeLists.txt

@@ -2,6 +2,6 @@ cmake_minimum_required(VERSION 2.8)
 
 project(vitastor)
 
-set(VERSION "0.6.7")
+set(VERSION "0.6.8")
 
 add_subdirectory(src)

csi/Makefile

@@ -1,4 +1,4 @@
-VERSION ?= v0.6.7
+VERSION ?= v0.6.8
 
 all: build push
 

csi/deploy/004-csi-nodeplugin.yaml

@@ -49,7 +49,7 @@ spec:
             capabilities:
               add: ["SYS_ADMIN"]
             allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v0.6.7
+          image: vitalif/vitastor-csi:v0.6.8
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/deploy/007-csi-provisioner.yaml

@@ -116,7 +116,7 @@ spec:
             privileged: true
             capabilities:
               add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v0.6.7
+          image: vitalif/vitastor-csi:v0.6.8
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/src/config.go

@@ -5,7 +5,7 @@ package vitastor
 
 const (
     vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "0.6.7"
+    vitastorCSIDriverVersion = "0.6.8"
 )
 
 // Config struct fills the parameters of request or user input

debian/changelog

@@ -1,4 +1,4 @@
-vitastor (0.6.7-1) unstable; urgency=medium
+vitastor (0.6.8-1) unstable; urgency=medium
 
   * RDMA support
   * Bugfixes

debian/control

@@ -9,9 +9,47 @@ 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
+Depends: vitastor-osd, vitastor-mon, vitastor-client, vitastor-client-dev, vitastor-fio, vitastor-qemu
 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).
+
+Package: vitastor-osd
+Architecture: amd64
+Depends: ${shlibs:Depends}, ${misc:Depends}, vitastor-client (= ${binary:Version})
+Description: Vitastor, a fast software-defined clustered block storage - object storage daemon
+ Vitastor object storage daemon, i.e. server program that stores data.
+
+Package: vitastor-mon
+Architecture: amd64
+Depends: ${misc:Depends}, nodejs (>= 10), node-sprintf-js, node-ws (>= 7), lp-solve
+Description: Vitastor, a fast software-defined clustered block storage - monitor
+ Vitastor monitor, i.e. server program responsible for watching cluster state and
+ scheduling cluster-level operations.
+
+Package: vitastor-client
+Architecture: amd64
+Depends: ${shlibs:Depends}, ${misc:Depends}
+Description: Vitastor, a fast software-defined clustered block storage - client
+ Vitastor client library and command-line interface.
+
+Package: vitastor-client-dev
+Section: devel
+Architecture: amd64
+Depends: ${misc:Depends}, vitastor-client (= ${binary:Version})
+Description: Vitastor, a fast software-defined clustered block storage - development files
+ Vitastor library headers for development.
+
+Package: vitastor-fio
+Architecture: amd64
+Depends: ${shlibs:Depends}, ${misc:Depends}, vitastor-client (= ${binary:Version}), fio (= ${dep:fio})
+Description: Vitastor, a fast software-defined clustered block storage - fio drivers
+ Vitastor fio drivers for benchmarking.
+
+Package: vitastor-qemu
+Architecture: amd64
+Depends: ${shlibs:Depends}, ${misc:Depends}, vitastor-client (= ${binary:Version}), qemu (= ${dep:qemu})
+Description: Vitastor, a fast software-defined clustered block storage - QEMU driver
+ Vitastor QEMU block device driver.

debian/fio_version

@@ -0,0 +1 @@
+dep:fio=3.16-1

debian/install

@@ -1,3 +1,4 @@
 VNPL-1.1.txt usr/share/doc/vitastor
 GPL-2.0.txt usr/share/doc/vitastor
-mon usr/lib/vitastor
+README.md usr/share/doc/vitastor
+README-ru.md usr/share/doc/vitastor

debian/patched-qemu.Dockerfile

@@ -7,11 +7,11 @@ ARG REL=
 
 WORKDIR /root
 
-RUN if [ "$REL" = "buster" ]; then \
+RUN if [ "$REL" = "buster" -o "$REL" = "bullseye" ]; then \
-        echo 'deb http://deb.debian.org/debian buster-backports main' >> /etc/apt/sources.list; \
+        echo "deb http://deb.debian.org/debian $REL-backports main" >> /etc/apt/sources.list; \
         echo >> /etc/apt/preferences; \
         echo 'Package: *' >> /etc/apt/preferences; \
-        echo 'Pin: release a=buster-backports' >> /etc/apt/preferences; \
+        echo "Pin: release a=$REL-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; \
@@ -27,15 +27,20 @@ RUN apt-get -y build-dep fio
 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/
+ADD patches/qemu-5.0-vitastor.patch patches/qemu-5.1-vitastor.patch patches/qemu-6.1-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 \
+    if ls -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; \
+        D=$(ls -d /root/packages/qemu-$REL/qemu-5.0*); \
-        echo qemu-5.0-vitastor.patch >> /root/packages/qemu-$REL/qemu-5.0/debian/patches/series; \
+        cp /root/vitastor/patches/qemu-5.0-vitastor.patch $D/debian/patches; \
+        echo qemu-5.0-vitastor.patch >> $D/debian/patches/series; \
+    elif ls /root/packages/qemu-$REL/qemu-6.1*; then \
+        D=$(ls -d /root/packages/qemu-$REL/qemu-6.1*); \
+        cp /root/vitastor/patches/qemu-6.1-vitastor.patch $D/debian/patches; \
+        echo qemu-6.1-vitastor.patch >> $D/debian/patches/series; \
     else \
         cp /root/vitastor/patches/qemu-5.1-vitastor.patch /root/packages/qemu-$REL/qemu-*/debian/patches; \
         P=`ls -d /root/packages/qemu-$REL/qemu-*/debian/patches`; \

debian/qemu_version

@@ -0,0 +1 @@
+dep:qemu=1:5.2+dfsg-10+vitastor1

debian/rules

@@ -5,5 +5,6 @@ export DH_VERBOSE = 1
 	dh $@
 
 override_dh_installdeb:
-	cat debian/substvars >> debian/vitastor.substvars
+	cat debian/fio_version >> debian/vitastor-fio.substvars
+	cat debian/qemu_version >> debian/vitastor-qemu.substvars
 	dh_installdeb

debian/substvars

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

debian/vitastor-client-dev.install

@@ -0,0 +1,2 @@
+usr/include
+usr/lib/*/pkgconfig

debian/vitastor-client.install

@@ -0,0 +1,5 @@
+usr/bin/vita
+usr/bin/vitastor-cli
+usr/bin/vitastor-rm
+usr/bin/vitastor-nbd
+usr/lib/*/libvitastor*.so*

debian/vitastor-fio.install

@@ -0,0 +1 @@
+usr/lib/*/libfio*.so*

debian/vitastor-mon.install

@@ -0,0 +1 @@
+mon usr/lib/vitastor

debian/vitastor-osd.install

@@ -0,0 +1,3 @@
+usr/bin/vitastor-osd
+usr/bin/vitastor-dump-journal
+mon/make-osd.sh /usr/lib/vitastor

debian/vitastor-qemu.install

@@ -0,0 +1 @@
+usr/lib/*/qemu/*

debian/vitastor.Dockerfile

@@ -42,10 +42,10 @@ RUN set -e -x; \
     mkdir -p /root/packages/vitastor-$REL; \
     rm -rf /root/packages/vitastor-$REL/*; \
     cd /root/packages/vitastor-$REL; \
-    cp -r /root/vitastor vitastor-0.6.7; \
+    cp -r /root/vitastor vitastor-0.6.8; \
-    ln -s /root/packages/qemu-$REL/qemu-*/ vitastor-0.6.7/qemu; \
+    ln -s /root/packages/qemu-$REL/qemu-*/ vitastor-0.6.8/qemu; \
-    ln -s /root/fio-build/fio-*/ vitastor-0.6.7/fio; \
+    ln -s /root/fio-build/fio-*/ vitastor-0.6.8/fio; \
-    cd vitastor-0.6.7; \
+    cd vitastor-0.6.8; \
     FIO=$(head -n1 fio/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     QEMU=$(head -n1 qemu/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     sh copy-qemu-includes.sh; \
@@ -58,11 +58,11 @@ RUN set -e -x; \
     echo qemu-fio-headers.patch >> debian/patches/series; \
     rm -rf a b; \
     rm -rf /root/packages/qemu-$REL/qemu*/; \
-    echo "dep:fio=$FIO" > debian/substvars; \
+    echo "dep:fio=$FIO" > debian/fio_version; \
-    echo "dep:qemu=$QEMU" >> debian/substvars; \
+    echo "dep:qemu=$QEMU" > debian/qemu_version; \
     cd /root/packages/vitastor-$REL; \
-    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.6.7.orig.tar.xz vitastor-0.6.7; \
+    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.6.8.orig.tar.xz vitastor-0.6.8; \
-    cd vitastor-0.6.7; \
+    cd vitastor-0.6.8; \
     V=$(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; \

mon/lp-optimizer.js

@@ -50,7 +50,7 @@ async function lp_solve(text)
     return { score, vars };
 }
 
-async function optimize_initial({ osd_tree, pg_count, pg_size = 3, pg_minsize = 2, max_combinations = 10000, parity_space = 1 })
+async function optimize_initial({ osd_tree, pg_count, pg_size = 3, pg_minsize = 2, max_combinations = 10000, parity_space = 1, round_robin = false })
 {
     if (!pg_count || !osd_tree)
     {
@@ -92,7 +92,7 @@ async function optimize_initial({ osd_tree, pg_count, pg_size = 3, pg_minsize =
         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 int_pgs = make_int_pgs(lp_result.vars, pg_count, round_robin);
     const eff = pg_list_space_efficiency(int_pgs, all_weights, pg_minsize, parity_space);
     const res = {
         score: lp_result.score,
@@ -115,7 +115,7 @@ function shuffle(array)
     }
 }
 
-function make_int_pgs(weights, pg_count)
+function make_int_pgs(weights, pg_count, round_robin)
 {
     const total_weight = Object.values(weights).reduce((a, c) => Number(a) + Number(c), 0);
     let int_pgs = [];
@@ -123,10 +123,15 @@ function make_int_pgs(weights, pg_count)
     let weight_left = total_weight;
     for (const pg_name in weights)
     {
+        let cur_pg = pg_name.substr(3).split('_');
         let n = Math.round(weights[pg_name] / weight_left * pg_left);
         for (let i = 0; i < n; i++)
         {
-            int_pgs.push(pg_name.substr(3).split('_'));
+            int_pgs.push([ ...cur_pg ]);
+            if (round_robin)
+            {
+                cur_pg.push(cur_pg.shift());
+            }
         }
         weight_left -= weights[pg_name];
         pg_left -= n;

mon/make-osd.sh

@@ -17,23 +17,14 @@ 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
+OSD_NUM=$(vitastor-cli alloc-osd)
-while true; do
-    ST=$(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 $DEV --format options | tr '\n' ' '`
+OPT=$(vitastor-cli simple-offsets --format options $DEV | tr '\n' ' ')
-META=`echo $OPT | grep -Po '(?<=data_offset )\d+'`
+META=$(vitastor-cli simple-offsets --format json $DEV | jq .data_offset)
 dd if=/dev/zero of=$DEV bs=1048576 count=$(((META+1048575)/1048576)) oflag=direct
 
 cat >/etc/systemd/system/vitastor-osd$OSD_NUM.service <<EOF

mon/mon.js

@@ -1022,6 +1022,7 @@ class Mon
                     pg_size: pool_cfg.pg_size,
                     pg_minsize: pool_cfg.pg_minsize,
                     max_combinations: pool_cfg.max_osd_combinations,
+                    round_robin: pool_cfg.scheme != 'replicated',
                 };
                 let optimize_result;
                 if (old_pg_count > 0)

patches/cinder-vitastor.py

@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils
 
-VERSION = '0.6.7'
+VERSION = '0.6.8'
 
 LOG = logging.getLogger(__name__)
 
@@ -489,7 +489,7 @@ class VitastorDriver(driver.CloneableImageVD,
                 # FIXME use etcd_address in qemu driver
                 kk = 'etcd_host'
             if v:
-                args += ':'+kk+'='+v.replace(':', '\\:')
+                args += ':'+kk.replace('_', '-')+'='+v.replace(':', '\\:')
         return args
 
     def delete_volume(self, volume):

patches/libvirt-5.0-vitastor.diff

@@ -245,9 +245,9 @@ index cbf0aa4..096700d 100644
 +
 +    if (virJSONValueObjectCreate(&ret,
 +                                 "s:driver", "vitastor",
-+                                 "S:etcd_host", etcd,
++                                 "S:etcd-host", etcd,
-+                                 "S:etcd_prefix", src->relPath,
++                                 "S:etcd-prefix", src->relPath,
-+                                 "S:config_path", src->configFile,
++                                 "S:config-path", src->configFile,
 +                                 "s:image", src->path,
 +                                 NULL) < 0)
 +        goto cleanup;
@@ -293,7 +293,7 @@ index 822d5f8..e375cef 100644
 +            virBufferStrcat(&buf, "vitastor:image=", src->path, NULL);
 +
 +            if (src->nhosts > 0) {
-+                virBufferAddLit(&buf, ":etcd_host=");
++                virBufferAddLit(&buf, ":etcd-host=");
 +                for (i = 0; i < src->nhosts; i++) {
 +                    if (i)
 +                        virBufferAddLit(&buf, ",");
@@ -311,10 +311,10 @@ index 822d5f8..e375cef 100644
 +            }
 +
 +            if (src->configFile)
-+                virBufferEscape(&buf, '\\', ":", ":config_path=%s", src->configFile);
++                virBufferEscape(&buf, '\\', ":", ":config-path=%s", src->configFile);
 +
 +            if (src->relPath)
-+                virBufferEscape(&buf, '\\', ":", ":etcd_prefix=%s", src->relPath);
++                virBufferEscape(&buf, '\\', ":", ":etcd-prefix=%s", src->relPath);
 +
 +            ret = virBufferContentAndReset(&buf);
 +            break;
@@ -438,16 +438,16 @@ index bd4b027..b323cd6 100644
 +        if (STRPREFIX(p, "image=")) {
 +            if (VIR_STRDUP(src->path, p + strlen("image=")) < 0)
 +                return -1;
-+        } else if (STRPREFIX(p, "etcd_prefix=")) {
++        } else if (STRPREFIX(p, "etcd-prefix=")) {
-+            if (VIR_STRDUP(src->relPath, p + strlen("etcd_prefix=")) < 0)
++            if (VIR_STRDUP(src->relPath, p + strlen("etcd-prefix=")) < 0)
 +                return -1;
-+        } else if (STRPREFIX(p, "config_file=")) {
++        } else if (STRPREFIX(p, "config-path=")) {
-+            if (VIR_STRDUP(src->configFile, p + strlen("config_file=")) < 0)
++            if (VIR_STRDUP(src->configFile, p + strlen("config-path=")) < 0)
 +                return -1;
-+        } else if (STRPREFIX(p, "etcd_host=")) {
++        } else if (STRPREFIX(p, "etcd-host=")) {
 +            char *h, *sep;
 +
-+            h = p + strlen("etcd_host=");
++            h = p + strlen("etcd-host=");
 +            while (h < e) {
 +                for (sep = h; sep < e; ++sep) {
 +                    if (*sep == '\\' && (sep[1] == ',' ||
@@ -507,8 +507,8 @@ index bd4b027..b323cd6 100644
 +{
 +    const char *filename;
 +    const char *image = virJSONValueObjectGetString(json, "image");
-+    const char *conf = virJSONValueObjectGetString(json, "config_path");
++    const char *conf = virJSONValueObjectGetString(json, "config-path");
-+    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd_prefix");
++    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd-prefix");
 +    virJSONValuePtr servers = virJSONValueObjectGetArray(json, "server");
 +    size_t nservers;
 +    size_t i;

patches/libvirt-7.0-vitastor.diff

@@ -244,9 +244,9 @@ index f9c6da2..922dde5 100644
 +    }
 +
 +    if (virJSONValueObjectCreate(&ret,
-+                                 "S:etcd_host", etcd,
++                                 "S:etcd-host", etcd,
-+                                 "S:etcd_prefix", src->query,
++                                 "S:etcd-prefix", src->query,
-+                                 "S:config_path", src->configFile,
++                                 "S:config-path", src->configFile,
 +                                 "s:image", src->path,
 +                                 NULL) < 0)
 +        return NULL;
@@ -311,7 +311,7 @@ index 6f970a3..10b39ca 100644
 +            virBufferStrcat(&buf, "vitastor:image=", src->path, NULL);
 +
 +            if (src->nhosts > 0) {
-+                virBufferAddLit(&buf, ":etcd_host=");
++                virBufferAddLit(&buf, ":etcd-host=");
 +                for (i = 0; i < src->nhosts; i++) {
 +                    if (i)
 +                        virBufferAddLit(&buf, ",");
@@ -329,10 +329,10 @@ index 6f970a3..10b39ca 100644
 +            }
 +
 +            if (src->configFile)
-+                virBufferEscape(&buf, '\\', ":", ":config_path=%s", src->configFile);
++                virBufferEscape(&buf, '\\', ":", ":config-path=%s", src->configFile);
 +
 +            if (src->query)
-+                virBufferEscape(&buf, '\\', ":", ":etcd_prefix=%s", src->query);
++                virBufferEscape(&buf, '\\', ":", ":etcd-prefix=%s", src->query);
 +
 +            ret = virBufferContentAndReset(&buf);
 +            break;
@@ -462,14 +462,14 @@ index 0d3c2af..36e3afc 100644
 +
 +        if (STRPREFIX(p, "image=")) {
 +            src->path = g_strdup(p + strlen("image="));
-+        } else if (STRPREFIX(p, "etcd_prefix=")) {
++        } else if (STRPREFIX(p, "etcd-prefix=")) {
-+            src->query = g_strdup(p + strlen("etcd_prefix="));
++            src->query = g_strdup(p + strlen("etcd-prefix="));
-+        } else if (STRPREFIX(p, "config_file=")) {
++        } else if (STRPREFIX(p, "config-path=")) {
-+            src->configFile = g_strdup(p + strlen("config_file="));
++            src->configFile = g_strdup(p + strlen("config-path="));
-+        } else if (STRPREFIX(p, "etcd_host=")) {
++        } else if (STRPREFIX(p, "etcd-host=")) {
 +            char *h, *sep;
 +
-+            h = p + strlen("etcd_host=");
++            h = p + strlen("etcd-host=");
 +            while (h < e) {
 +                for (sep = h; sep < e; ++sep) {
 +                    if (*sep == '\\' && (sep[1] == ',' ||
@@ -526,8 +526,8 @@ index 0d3c2af..36e3afc 100644
 +{
 +    const char *filename;
 +    const char *image = virJSONValueObjectGetString(json, "image");
-+    const char *conf = virJSONValueObjectGetString(json, "config_path");
++    const char *conf = virJSONValueObjectGetString(json, "config-path");
-+    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd_prefix");
++    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd-prefix");
 +    virJSONValuePtr servers = virJSONValueObjectGetArray(json, "server");
 +    size_t nservers;
 +    size_t i;

patches/libvirt-7.5-vitastor.diff

@@ -276,9 +276,9 @@ index 6627d04..c33f428 100644
 +    }
 +
 +    if (virJSONValueObjectCreate(&ret,
-+                                 "S:etcd_host", etcd,
++                                 "S:etcd-host", etcd,
-+                                 "S:etcd_prefix", src->query,
++                                 "S:etcd-prefix", src->query,
-+                                 "S:config_path", src->configFile,
++                                 "S:config-path", src->configFile,
 +                                 "s:image", src->path,
 +                                 NULL) < 0)
 +        return NULL;
@@ -343,7 +343,7 @@ index ea51369..8258632 100644
 +            virBufferStrcat(&buf, "vitastor:image=", src->path, NULL);
 +
 +            if (src->nhosts > 0) {
-+                virBufferAddLit(&buf, ":etcd_host=");
++                virBufferAddLit(&buf, ":etcd-host=");
 +                for (i = 0; i < src->nhosts; i++) {
 +                    if (i)
 +                        virBufferAddLit(&buf, ",");
@@ -361,10 +361,10 @@ index ea51369..8258632 100644
 +            }
 +
 +            if (src->configFile)
-+                virBufferEscape(&buf, '\\', ":", ":config_path=%s", src->configFile);
++                virBufferEscape(&buf, '\\', ":", ":config-path=%s", src->configFile);
 +
 +            if (src->query)
-+                virBufferEscape(&buf, '\\', ":", ":etcd_prefix=%s", src->query);
++                virBufferEscape(&buf, '\\', ":", ":etcd-prefix=%s", src->query);
 +
 +            ret = virBufferContentAndReset(&buf);
 +            break;
@@ -474,14 +474,14 @@ index e48ae72..d7a9b72 100644
 +
 +        if (STRPREFIX(p, "image=")) {
 +            src->path = g_strdup(p + strlen("image="));
-+        } else if (STRPREFIX(p, "etcd_prefix=")) {
++        } else if (STRPREFIX(p, "etcd-prefix=")) {
-+            src->query = g_strdup(p + strlen("etcd_prefix="));
++            src->query = g_strdup(p + strlen("etcd-prefix="));
-+        } else if (STRPREFIX(p, "config_file=")) {
++        } else if (STRPREFIX(p, "config-path=")) {
-+            src->configFile = g_strdup(p + strlen("config_file="));
++            src->configFile = g_strdup(p + strlen("config-path="));
-+        } else if (STRPREFIX(p, "etcd_host=")) {
++        } else if (STRPREFIX(p, "etcd-host=")) {
 +            char *h, *sep;
 +
-+            h = p + strlen("etcd_host=");
++            h = p + strlen("etcd-host=");
 +            while (h < e) {
 +                for (sep = h; sep < e; ++sep) {
 +                    if (*sep == '\\' && (sep[1] == ',' ||
@@ -538,8 +538,8 @@ index e48ae72..d7a9b72 100644
 +{
 +    const char *filename;
 +    const char *image = virJSONValueObjectGetString(json, "image");
-+    const char *conf = virJSONValueObjectGetString(json, "config_path");
++    const char *conf = virJSONValueObjectGetString(json, "config-path");
-+    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd_prefix");
++    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd-prefix");
 +    virJSONValue *servers = virJSONValueObjectGetArray(json, "server");
 +    size_t nservers;
 +    size_t i;

patches/nova-20.diff

@@ -124,7 +124,7 @@ index 391231c527..34dc60dcdd 100644
 +                            if kk == 'etcd_address':
 +                                # FIXME use etcd_address in qemu driver
 +                                kk = 'etcd_host'
-+                            path += ":"+kk+"="+connection_info['data'][k].replace(':', '\\:')
++                            path += ":"+kk.replace('_', '-')+"="+connection_info['data'][k].replace(':', '\\:')
                  else:
                      path = 'unknown'
                      raise exception.DiskNotFound(location='unknown')

patches/qemu-3.1-vitastor.patch

@@ -23,18 +23,18 @@ Index: qemu-3.1+dfsg/qapi/block-core.json
 +# @inode:       Inode number
 +# @pool:        Pool ID
 +# @size:        Desired image size in bytes
-+# @config_path: Path to Vitastor configuration
++# @config-path: Path to Vitastor configuration
-+# @etcd_host:   etcd connection address(es)
++# @etcd-host:   etcd connection address(es)
-+# @etcd_prefix: etcd key/value prefix
++# @etcd-prefix: etcd key/value prefix
 +##
 +{ 'struct': 'BlockdevOptionsVitastor',
 +  'data': { '*inode': 'uint64',
 +            '*pool': 'uint64',
 +            '*size': 'uint64',
 +            '*image': 'str',
-+            '*config_path': 'str',
++            '*config-path': 'str',
-+            '*etcd_host': 'str',
++            '*etcd-host': 'str',
-+            '*etcd_prefix': 'str' } }
++            '*etcd-prefix': 'str' } }
 +
 +##
  # @ReplicationMode:

patches/qemu-4.2-vitastor.patch

@@ -23,18 +23,18 @@ Index: qemu/qapi/block-core.json
 +# @inode:       Inode number
 +# @pool:        Pool ID
 +# @size:        Desired image size in bytes
-+# @config_path: Path to Vitastor configuration
++# @config-path: Path to Vitastor configuration
-+# @etcd_host:   etcd connection address(es)
++# @etcd-host:   etcd connection address(es)
-+# @etcd_prefix: etcd key/value prefix
++# @etcd-prefix: etcd key/value prefix
 +##
 +{ 'struct': 'BlockdevOptionsVitastor',
 +  'data': { '*inode': 'uint64',
 +            '*pool': 'uint64',
 +            '*size': 'uint64',
 +            '*image': 'str',
-+            '*config_path': 'str',
++            '*config-path': 'str',
-+            '*etcd_host': 'str',
++            '*etcd-host': 'str',
-+            '*etcd_prefix': 'str' } }
++            '*etcd-prefix': 'str' } }
 +
 +##
  # @ReplicationMode:

patches/qemu-5.0-vitastor.patch

@@ -23,18 +23,18 @@ Index: qemu/qapi/block-core.json
 +# @inode:       Inode number
 +# @pool:        Pool ID
 +# @size:        Desired image size in bytes
-+# @config_path: Path to Vitastor configuration
++# @config-path: Path to Vitastor configuration
-+# @etcd_host:   etcd connection address(es)
++# @etcd-host:   etcd connection address(es)
-+# @etcd_prefix: etcd key/value prefix
++# @etcd-prefix: etcd key/value prefix
 +##
 +{ 'struct': 'BlockdevOptionsVitastor',
 +  'data': { '*inode': 'uint64',
 +            '*pool': 'uint64',
 +            '*size': 'uint64',
 +            '*image': 'str',
-+            '*config_path': 'str',
++            '*config-path': 'str',
-+            '*etcd_host': 'str',
++            '*etcd-host': 'str',
-+            '*etcd_prefix': 'str' } }
++            '*etcd-prefix': 'str' } }
 +
 +##
  # @ReplicationMode:

patches/qemu-5.1-vitastor.patch

@@ -23,18 +23,18 @@ Index: qemu-5.1+dfsg/qapi/block-core.json
 +# @inode:       Inode number
 +# @pool:        Pool ID
 +# @size:        Desired image size in bytes
-+# @config_path: Path to Vitastor configuration
++# @config-path: Path to Vitastor configuration
-+# @etcd_host:   etcd connection address(es)
++# @etcd-host:   etcd connection address(es)
-+# @etcd_prefix: etcd key/value prefix
++# @etcd-prefix: etcd key/value prefix
 +##
 +{ 'struct': 'BlockdevOptionsVitastor',
 +  'data': { '*inode': 'uint64',
 +            '*pool': 'uint64',
 +            '*size': 'uint64',
 +            '*image': 'str',
-+            '*config_path': 'str',
++            '*config-path': 'str',
-+            '*etcd_host': 'str',
++            '*etcd-host': 'str',
-+            '*etcd_prefix': 'str' } }
++            '*etcd-prefix': 'str' } }
 +
 +##
  # @ReplicationMode:

patches/qemu-6.1-vitastor.patch

@@ -0,0 +1,88 @@
+Index: qemu-6.1+dfsg/qapi/block-core.json
+===================================================================
+--- qemu-6.1+dfsg.orig/qapi/block-core.json
++++ qemu-6.1+dfsg/qapi/block-core.json
+@@ -2838,7 +2838,7 @@
+             'luks', 'nbd', 'nfs', 'null-aio', 'null-co', 'nvme', 'parallels',
+             'preallocate', 'qcow', 'qcow2', 'qed', 'quorum', 'raw', 'rbd',
+             { 'name': 'replication', 'if': 'defined(CONFIG_REPLICATION)' },
+-            'ssh', 'throttle', 'vdi', 'vhdx', 'vmdk', 'vpc', 'vvfat' ] }
++            'ssh', 'throttle', 'vdi', 'vhdx', 'vitastor', 'vmdk', 'vpc', 'vvfat' ] }
+ 
+ ##
+ # @BlockdevOptionsFile:
+@@ -3763,6 +3763,28 @@
+             '*server': ['InetSocketAddressBase'] } }
+ 
+ ##
++# @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.
+@@ -4134,6 +4156,7 @@
+       'throttle':   'BlockdevOptionsThrottle',
+       'vdi':        'BlockdevOptionsGenericFormat',
+       'vhdx':       'BlockdevOptionsGenericFormat',
++      'vitastor':   'BlockdevOptionsVitastor',
+       'vmdk':       'BlockdevOptionsGenericCOWFormat',
+       'vpc':        'BlockdevOptionsGenericFormat',
+       'vvfat':      'BlockdevOptionsVVFAT'
+@@ -4523,6 +4546,17 @@
+             '*encrypt' :        'RbdEncryptionCreateOptions' } }
+ 
+ ##
++# @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
+@@ -4718,6 +4752,7 @@
+       'ssh':            'BlockdevCreateOptionsSsh',
+       'vdi':            'BlockdevCreateOptionsVdi',
+       'vhdx':           'BlockdevCreateOptionsVhdx',
++      'vitastor':       'BlockdevCreateOptionsVitastor',
+       'vmdk':           'BlockdevCreateOptionsVmdk',
+       'vpc':            'BlockdevCreateOptionsVpc'
+   } }
+Index: qemu-6.1+dfsg/scripts/modules/module_block.py
+===================================================================
+--- qemu-6.1+dfsg.orig/scripts/modules/module_block.py
++++ qemu-6.1+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

@@ -48,4 +48,4 @@ FIO=`rpm -qi fio | perl -e 'while(<>) { /^Epoch[\s:]+(\S+)/ && print "$1:"; /^Ve
 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-0.6.7/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.6.7$(rpm --eval '%dist').tar.gz *
+tar --transform 's#^#vitastor-0.6.8/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.6.8$(rpm --eval '%dist').tar.gz *

rpm/vitastor-el7.Dockerfile

@@ -38,7 +38,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.6.7.el7.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.6.8.el7.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el7.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el7.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.6.7
+Version:        0.6.8
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-0.6.7.el7.tar.gz
+Source0:        vitastor-0.6.8.el7.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
@@ -16,13 +16,12 @@ BuildRequires:  jerasure-devel
 BuildRequires:  gf-complete-devel
 BuildRequires:  libibverbs-devel
 BuildRequires:  cmake
-Requires:       fio = 3.7-1.el7
+Requires:       vitastor-osd = %{version}-%{release}
-Requires:       qemu-kvm = 2.0.0-1.el7.6
+Requires:       vitastor-mon = %{version}-%{release}
-Requires:       rh-nodejs12
+Requires:       vitastor-client = %{version}-%{release}
-Requires:       rh-nodejs12-npm
+Requires:       vitastor-client-devel = %{version}-%{release}
-Requires:       liburing >= 0.6
+Requires:       vitastor-fio = %{version}-%{release}
-Requires:       libJerasure2
+Requires:       vitastor-qemu = %{version}-%{release}
-Requires:       lpsolve
 
 %description
 Vitastor is a small, simple and fast clustered block storage (storage for VM drives),
@@ -31,6 +30,70 @@ symmetric clustering and automatic data distribution over any number of drives o
 size with configurable redundancy (replication or erasure codes/XOR).
 
 
+%package -n vitastor-osd
+Summary:        Vitastor - OSD
+Requires:       libJerasure2
+Requires:       liburing >= 0.6
+Requires:       vitastor-client = %{version}-%{release}
+
+
+%description -n vitastor-osd
+Vitastor object storage daemon, i.e. server program that stores data.
+
+
+%package -n vitastor-mon
+Summary:        Vitastor - monitor
+Requires:       rh-nodejs12
+Requires:       rh-nodejs12-npm
+Requires:       lpsolve
+
+
+%description -n vitastor-mon
+Vitastor monitor, i.e. server program responsible for watching cluster state and
+scheduling cluster-level operations.
+
+
+%package -n vitastor-client
+Summary:        Vitastor - client
+Requires:       liburing >= 0.6
+
+
+%description -n vitastor-client
+Vitastor client library and command-line interface.
+
+
+%package -n vitastor-client-devel
+Summary:        Vitastor - development files
+Group:          Development/Libraries
+Requires:       vitastor-client = %{version}-%{release}
+
+
+%description -n vitastor-client-devel
+Vitastor library headers for development.
+
+
+%package -n vitastor-fio
+Summary:        Vitastor - fio drivers
+Group:          Development/Libraries
+Requires:       vitastor-client = %{version}-%{release}
+Requires:       fio = 3.7-1.el7
+
+
+%description -n vitastor-fio
+Vitastor fio drivers for benchmarking.
+
+
+%package -n vitastor-qemu
+Summary:        Vitastor - QEMU driver
+Group:          Development/Libraries
+Requires:       vitastor-client = %{version}-%{release}
+Requires:       qemu-kvm = 2.0.0-1.el7.6
+
+
+%description -n vitastor-qemu
+Vitastor QEMU block device driver.
+
+
 %prep
 %setup -q
 
@@ -49,25 +112,46 @@ cd mon
 npm install
 cd ..
 mkdir -p %buildroot/usr/lib/vitastor
-cp -r mon %buildroot/usr/lib/vitastor/mon
+cp mon/make-osd.sh %buildroot/usr/lib/vitastor
+cp -r mon %buildroot/usr/lib/vitastor
 
 
 %files
-%doc
+%doc GPL-2.0.txt VNPL-1.1.txt README.md README-ru.md
-%_bindir/vitastor-dump-journal
+
-%_bindir/vitastor-nbd
+
+%files -n vitastor-osd
 %_bindir/vitastor-osd
+%_bindir/vitastor-dump-journal
+/usr/lib/vitastor/make-osd.sh
+
+
+%files -n vitastor-mon
+/usr/lib/vitastor/mon
+
+
+%files -n vitastor-client
+%_bindir/vitastor-nbd
 %_bindir/vitastor-cli
 %_bindir/vitastor-rm
 %_bindir/vita
-%_libdir/qemu-kvm/block-vitastor.so
+%_libdir/libvitastor_blk.so*
+%_libdir/libvitastor_client.so*
+
+
+%files -n vitastor-client-devel
+%_includedir/vitastor_c.h
+%_libdir/pkgconfig
+
+
+%files -n vitastor-fio
 %_libdir/libfio_vitastor.so
 %_libdir/libfio_vitastor_blk.so
 %_libdir/libfio_vitastor_sec.so
-%_libdir/libvitastor_blk.so*
+
-%_libdir/libvitastor_client.so*
+
-%_includedir/vitastor_c.h
+%files -n vitastor-qemu
-/usr/lib/vitastor
+%_libdir/qemu-kvm/block-vitastor.so
 
 
 %changelog

rpm/vitastor-el8.Dockerfile

@@ -36,7 +36,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.6.7.el8.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.6.8.el8.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el8.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el8.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.6.7
+Version:        0.6.8
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-0.6.7.el8.tar.gz
+Source0:        vitastor-0.6.8.el8.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
@@ -15,12 +15,12 @@ BuildRequires:  jerasure-devel
 BuildRequires:  gf-complete-devel
 BuildRequires:  libibverbs-devel
 BuildRequires:  cmake
-Requires:       fio = 3.7-3.el8
+Requires:       vitastor-osd = %{version}-%{release}
-Requires:       qemu-kvm = 4.2.0-29.el8.6
+Requires:       vitastor-mon = %{version}-%{release}
-Requires:       nodejs >= 10
+Requires:       vitastor-client = %{version}-%{release}
-Requires:       liburing >= 0.6
+Requires:       vitastor-client-devel = %{version}-%{release}
-Requires:       libJerasure2
+Requires:       vitastor-fio = %{version}-%{release}
-Requires:       lpsolve
+Requires:       vitastor-qemu = %{version}-%{release}
 
 %description
 Vitastor is a small, simple and fast clustered block storage (storage for VM drives),
@@ -29,6 +29,69 @@ symmetric clustering and automatic data distribution over any number of drives o
 size with configurable redundancy (replication or erasure codes/XOR).
 
 
+%package -n vitastor-osd
+Summary:        Vitastor - OSD
+Requires:       libJerasure2
+Requires:       liburing >= 0.6
+Requires:       vitastor-client = %{version}-%{release}
+
+
+%description -n vitastor-osd
+Vitastor object storage daemon, i.e. server program that stores data.
+
+
+%package -n vitastor-mon
+Summary:        Vitastor - monitor
+Requires:       nodejs >= 10
+Requires:       lpsolve
+
+
+%description -n vitastor-mon
+Vitastor monitor, i.e. server program responsible for watching cluster state and
+scheduling cluster-level operations.
+
+
+%package -n vitastor-client
+Summary:        Vitastor - client
+Requires:       liburing >= 0.6
+
+
+%description -n vitastor-client
+Vitastor client library and command-line interface.
+
+
+%package -n vitastor-client-devel
+Summary:        Vitastor - development files
+Group:          Development/Libraries
+Requires:       vitastor-client = %{version}-%{release}
+
+
+%description -n vitastor-client-devel
+Vitastor library headers for development.
+
+
+%package -n vitastor-fio
+Summary:        Vitastor - fio drivers
+Group:          Development/Libraries
+Requires:       vitastor-client = %{version}-%{release}
+Requires:       fio = 3.7-3.el8
+
+
+%description -n vitastor-fio
+Vitastor fio drivers for benchmarking.
+
+
+%package -n vitastor-qemu
+Summary:        Vitastor - QEMU driver
+Group:          Development/Libraries
+Requires:       vitastor-client = %{version}-%{release}
+Requires:       qemu-kvm = 4.2.0-29.el8.6
+
+
+%description -n vitastor-qemu
+Vitastor QEMU block device driver.
+
+
 %prep
 %setup -q
 
@@ -46,25 +109,46 @@ cd mon
 npm install
 cd ..
 mkdir -p %buildroot/usr/lib/vitastor
+cp mon/make-osd.sh %buildroot/usr/lib/vitastor
 cp -r mon %buildroot/usr/lib/vitastor
 
 
 %files
-%doc
+%doc GPL-2.0.txt VNPL-1.1.txt README.md README-ru.md
-%_bindir/vitastor-dump-journal
+
-%_bindir/vitastor-nbd
+
+%files -n vitastor-osd
 %_bindir/vitastor-osd
+%_bindir/vitastor-dump-journal
+/usr/lib/vitastor/make-osd.sh
+
+
+%files -n vitastor-mon
+/usr/lib/vitastor/mon
+
+
+%files -n vitastor-client
+%_bindir/vitastor-nbd
 %_bindir/vitastor-cli
 %_bindir/vitastor-rm
 %_bindir/vita
-%_libdir/qemu-kvm/block-vitastor.so
+%_libdir/libvitastor_blk.so*
+%_libdir/libvitastor_client.so*
+
+
+%files -n vitastor-client-devel
+%_includedir/vitastor_c.h
+%_libdir/pkgconfig
+
+
+%files -n vitastor-fio
 %_libdir/libfio_vitastor.so
 %_libdir/libfio_vitastor_blk.so
 %_libdir/libfio_vitastor_sec.so
-%_libdir/libvitastor_blk.so*
+
-%_libdir/libvitastor_client.so*
+
-%_includedir/vitastor_c.h
+%files -n vitastor-qemu
-/usr/lib/vitastor
+%_libdir/qemu-kvm/block-vitastor.so
 
 
 %changelog

src/CMakeLists.txt

@@ -15,7 +15,7 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 	set(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR}")
 endif()
 
-add_definitions(-DVERSION="0.6.7")
+add_definitions(-DVERSION="0.6.8")
 add_definitions(-Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fdiagnostics-color=always -I ${CMAKE_SOURCE_DIR}/src)
 if (${WITH_ASAN})
 	add_definitions(-fsanitize=address -fno-omit-frame-pointer)
@@ -145,7 +145,7 @@ endif (${WITH_FIO})
 
 # vitastor-nbd
 add_executable(vitastor-nbd
-	nbd_proxy.cpp
+	nbd_proxy.cpp mmap_manager.cpp
 )
 target_link_libraries(vitastor-nbd
 	vitastor_client
@@ -153,11 +153,13 @@ target_link_libraries(vitastor-nbd
 
 # vitastor-cli
 add_executable(vitastor-cli
-	cli.cpp cli_ls.cpp cli_create.cpp cli_modify.cpp cli_flatten.cpp cli_merge.cpp cli_rm.cpp cli_snap_rm.cpp
+	cli.cpp cli_alloc_osd.cpp cli_simple_offsets.cpp
+	cli_ls.cpp cli_create.cpp cli_modify.cpp cli_flatten.cpp cli_merge.cpp cli_rm.cpp cli_snap_rm.cpp
 )
 target_link_libraries(vitastor-cli
 	vitastor_client
 )
+configure_file(vitastor.pc.in vitastor.pc @ONLY)
 
 # vitastor-dump-journal
 add_executable(vitastor-dump-journal
@@ -186,7 +188,7 @@ endif (${WITH_QEMU})
 ### Test stubs
 
 # stub_osd, stub_bench, osd_test
-add_executable(stub_osd stub_osd.cpp rw_blocking.cpp)
+add_executable(stub_osd stub_osd.cpp rw_blocking.cpp mmap_manager.cpp)
 target_link_libraries(stub_osd tcmalloc_minimal)
 add_executable(stub_bench stub_bench.cpp rw_blocking.cpp)
 target_link_libraries(stub_bench tcmalloc_minimal)
@@ -248,6 +250,7 @@ install(
 	LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR}
 	PUBLIC_HEADER DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}
 )
+install(FILES ${CMAKE_CURRENT_BINARY_DIR}/vitastor.pc DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
 if (${WITH_FIO})
 	install(TARGETS fio_vitastor fio_vitastor_blk fio_vitastor_sec LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR})
 endif (${WITH_FIO})

src/cli.cpp

@@ -130,6 +130,18 @@ void cli_tool_t::help()
         "  <to> must be a child of <from> and <target> may be one of the layers between\n"
         "  <from> and <to>, including <from> and <to>.\n"
         "\n"
+        "%s alloc-osd\n"
+        "  Allocate a new OSD number and reserve it by creating empty /osd/stats/<n> key.\n"
+        "%s simple-offsets <device>\n"
+        "  Calculate offsets for simple&stupid (no superblock) OSD deployment. Options:\n"
+        "  --object_size 128k       Set blockstore block size\n"
+        "  --bitmap_granularity 4k  Set bitmap granularity\n"
+        "  --journal_size 16M       Set journal size\n"
+        "  --device_block_size 4k   Set device block size\n"
+        "  --journal_offset 0       Set journal offset\n"
+        "  --device_size 0          Set device size\n"
+        "  --format text            Result format: json, options, env, or text\n"
+        "\n"
         "GLOBAL OPTIONS:\n"
         "  --etcd_address <etcd_address>\n"
         "  --iodepth N         Send N operations in parallel to each OSD when possible (default 32)\n"
@@ -139,7 +151,7 @@ void cli_tool_t::help()
         "  --no-color          Disable colored output\n"
         "  --json              JSON output\n"
         ,
-        exe_name, exe_name, exe_name, exe_name, exe_name, exe_name, exe_name, exe_name, exe_name
+        exe_name, exe_name, exe_name, exe_name, exe_name, exe_name, exe_name, exe_name, exe_name, exe_name, exe_name
     );
     exit(0);
 }
@@ -268,6 +280,16 @@ void cli_tool_t::run(json11::Json cfg)
         // Remove multiple snapshots and rebase their children
         action_cb = start_snap_rm(cfg);
     }
+    else if (cmd[0] == "alloc-osd")
+    {
+        // Allocate a new OSD number
+        action_cb = start_alloc_osd(cfg);
+    }
+    else if (cmd[0] == "simple-offsets")
+    {
+        // Calculate offsets for simple & stupid OSD deployment without superblock
+        action_cb = simple_offsets(cfg);
+    }
     else
     {
         fprintf(stderr, "unknown command: %s\n", cmd[0].string_value().c_str());

src/cli.h

@@ -50,11 +50,16 @@ public:
     friend struct snap_flattener_t;
     friend struct snap_remover_t;
 
-    std::function<bool(void)> start_ls(json11::Json cfg);
+    std::function<bool(void)> start_ls(json11::Json);
-    std::function<bool(void)> start_create(json11::Json cfg);
+    std::function<bool(void)> start_create(json11::Json);
-    std::function<bool(void)> start_modify(json11::Json cfg);
+    std::function<bool(void)> start_modify(json11::Json);
     std::function<bool(void)> start_rm(json11::Json);
     std::function<bool(void)> start_merge(json11::Json);
     std::function<bool(void)> start_flatten(json11::Json);
     std::function<bool(void)> start_snap_rm(json11::Json);
+    std::function<bool(void)> start_alloc_osd(json11::Json cfg, uint64_t *out = NULL);
+    std::function<bool(void)> simple_offsets(json11::Json cfg);
 };
+
+std::string format_size(uint64_t size);
+uint64_t parse_size(std::string size_str);

src/cli_alloc_osd.cpp

@@ -0,0 +1,141 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
+#include <ctype.h>
+#include "cli.h"
+#include "cluster_client.h"
+#include "base64.h"
+
+#include <algorithm>
+
+// Safely allocate an OSD number
+struct alloc_osd_t
+{
+    cli_tool_t *parent;
+
+    json11::Json result;
+    uint64_t new_id = 1;
+
+    int state = 0;
+
+    bool is_done()
+    {
+        return state == 100;
+    }
+
+    void loop()
+    {
+        if (state == 1)
+            goto resume_1;
+        do
+        {
+            etcd_txn(json11::Json::object {
+                { "compare", json11::Json::array {
+                    json11::Json::object {
+                        { "target", "VERSION" },
+                        { "version", 0 },
+                        { "key", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/osd/stats/"+std::to_string(new_id)
+                        ) },
+                    },
+                } },
+                { "success", json11::Json::array {
+                    json11::Json::object {
+                        { "request_put", json11::Json::object {
+                            { "key", base64_encode(
+                                parent->cli->st_cli.etcd_prefix+"/osd/stats/"+std::to_string(new_id)
+                            ) },
+                            { "value", base64_encode("{}") },
+                        } },
+                    },
+                } },
+                { "failure", json11::Json::array {
+                    json11::Json::object {
+                        { "request_range", json11::Json::object {
+                            { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/osd/stats/") },
+                            { "range_end", base64_encode(parent->cli->st_cli.etcd_prefix+"/osd/stats0") },
+                            { "keys_only", true },
+                        } },
+                    },
+                } },
+            });
+        resume_1:
+            state = 1;
+            if (parent->waiting > 0)
+                return;
+            if (!result["succeeded"].bool_value())
+            {
+                std::vector<osd_num_t> used;
+                for (auto kv: result["responses"][0]["response_range"]["kvs"].array_items())
+                {
+                    std::string key = base64_decode(kv["key"].string_value());
+                    osd_num_t cur_osd;
+                    char null_byte = 0;
+                    sscanf(key.c_str() + parent->cli->st_cli.etcd_prefix.length(), "/osd/stats/%lu%c", &cur_osd, &null_byte);
+                    if (!cur_osd || null_byte != 0)
+                    {
+                        fprintf(stderr, "Invalid key in etcd: %s\n", key.c_str());
+                        continue;
+                    }
+                    used.push_back(cur_osd);
+                }
+                std::sort(used.begin(), used.end());
+                if (used[used.size()-1] == used.size())
+                {
+                    new_id = used.size()+1;
+                }
+                else
+                {
+                    int s = 0, e = used.size();
+                    while (e > s+1)
+                    {
+                        int c = (s+e)/2;
+                        if (used[c] == c+1)
+                            s = c;
+                        else
+                            e = c;
+                    }
+                    new_id = used[e-1]+1;
+                }
+            }
+        } while (!result["succeeded"].bool_value());
+        state = 100;
+    }
+
+    void etcd_txn(json11::Json txn)
+    {
+        parent->waiting++;
+        parent->cli->st_cli.etcd_txn(txn, ETCD_SLOW_TIMEOUT, [this](std::string err, json11::Json res)
+        {
+            parent->waiting--;
+            if (err != "")
+            {
+                fprintf(stderr, "Error reading from etcd: %s\n", err.c_str());
+                exit(1);
+            }
+            this->result = res;
+            parent->ringloop->wakeup();
+        });
+    }
+};
+
+std::function<bool(void)> cli_tool_t::start_alloc_osd(json11::Json cfg, uint64_t *out)
+{
+    json11::Json::array cmd = cfg["command"].array_items();
+    auto alloc_osd = new alloc_osd_t();
+    alloc_osd->parent = this;
+    return [alloc_osd, &out]()
+    {
+        alloc_osd->loop();
+        if (alloc_osd->is_done())
+        {
+            if (out)
+                *out = alloc_osd->new_id;
+            else if (alloc_osd->new_id)
+                printf("%lu\n", alloc_osd->new_id);
+            delete alloc_osd;
+            return true;
+        }
+        return false;
+    };
+}

src/cli_create.cpp

@@ -94,6 +94,11 @@ struct image_creator_t
             goto resume_2;
         else if (state == 3)
             goto resume_3;
+        if (!size)
+        {
+            fprintf(stderr, "Image size is missing\n");
+            exit(1);
+        }
         for (auto & ic: parent->cli->st_cli.inode_config)
         {
             if (ic.second.name == image_name)
@@ -432,6 +437,31 @@ resume_3:
     }
 };
 
+uint64_t parse_size(std::string size_str)
+{
+    uint64_t mul = 1;
+    char type_char = tolower(size_str[size_str.length()-1]);
+    if (type_char == 'k' || type_char == 'm' || type_char == 'g' || type_char == 't')
+    {
+        if (type_char == 'k')
+            mul = 1l<<10;
+        else if (type_char == 'm')
+            mul = 1l<<20;
+        else if (type_char == 'g')
+            mul = 1l<<30;
+        else /*if (type_char == 't')*/
+            mul = 1l<<40;
+        size_str = size_str.substr(0, size_str.length()-1);
+    }
+    uint64_t size = json11::Json(size_str).uint64_value() * mul;
+    if (size == 0 && size_str != "0" && (size_str != "" || mul != 1))
+    {
+        fprintf(stderr, "Invalid syntax for size: %s\n", size_str.c_str());
+        exit(1);
+    }
+    return size;
+}
+
 std::function<bool(void)> cli_tool_t::start_create(json11::Json cfg)
 {
     json11::Json::array cmd = cfg["command"].array_items();
@@ -458,33 +488,12 @@ std::function<bool(void)> cli_tool_t::start_create(json11::Json cfg)
     image_creator->new_parent = cfg["parent"].string_value();
     if (cfg["size"].string_value() != "")
     {
-        std::string size_str = cfg["size"].string_value();
+        image_creator->size = parse_size(cfg["size"].string_value());
-        uint64_t mul = 1;
+        if (image_creator->size % 4096)
-        char type_char = tolower(size_str[size_str.length()-1]);
-        if (type_char == 'k' || type_char == 'm' || type_char == 'g' || type_char == 't')
         {
-            if (type_char == 'k')
+            fprintf(stderr, "Size should be a multiple of 4096\n");
-                mul = 1l<<10;
-            else if (type_char == 'm')
-                mul = 1l<<20;
-            else if (type_char == 'g')
-                mul = 1l<<30;
-            else /*if (type_char == 't')*/
-                mul = 1l<<40;
-            size_str = size_str.substr(0, size_str.length()-1);
-        }
-        uint64_t size = json11::Json(size_str).uint64_value() * mul;
-        if (size == 0)
-        {
-            fprintf(stderr, "Invalid syntax for size: %s\n", cfg["size"].string_value().c_str());
             exit(1);
         }
-        if (size % 4096)
-        {
-            fprintf(stderr, "Image size should be a multiple of 4096\n");
-            exit(1);
-        }
-        image_creator->size = size;
         if (image_creator->new_snap != "")
         {
             fprintf(stderr, "--size can't be specified for snapshots\n");

src/cli_ls.cpp

@@ -505,7 +505,7 @@ std::function<bool(void)> cli_tool_t::start_ls(json11::Json cfg)
     lister->sort_field = cfg["sort"].string_value();
     lister->reverse = cfg["reverse"].bool_value();
     lister->max_count = cfg["count"].uint64_value();
-    for (int i = 0; i < cmd.size(); i++)
+    for (int i = 1; i < cmd.size(); i++)
     {
         lister->only_names.insert(cmd[i].string_value());
     }

src/cli_simple_offsets.cpp

@@ -0,0 +1,143 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
+#include <fcntl.h>
+#include <sys/ioctl.h>
+#include <ctype.h>
+#include <unistd.h>
+#include "cli.h"
+#include "cluster_client.h"
+#include "base64.h"
+
+// Calculate offsets for a block device and print OSD command line parameters
+std::function<bool(void)> cli_tool_t::simple_offsets(json11::Json cfg)
+{
+    std::string device = cfg["command"][1].string_value();
+    uint64_t object_size = parse_size(cfg["object_size"].string_value());
+    uint64_t bitmap_granularity = parse_size(cfg["bitmap_granularity"].string_value());
+    uint64_t journal_size = parse_size(cfg["journal_size"].string_value());
+    uint64_t device_block_size = parse_size(cfg["device_block_size"].string_value());
+    uint64_t journal_offset = parse_size(cfg["journal_offset"].string_value());
+    uint64_t device_size = parse_size(cfg["device_size"].string_value());
+    std::string format = cfg["format"].string_value();
+    if (json_output)
+        format = "json";
+    if (!object_size)
+        object_size = DEFAULT_BLOCK_SIZE;
+    if (!bitmap_granularity)
+        bitmap_granularity = DEFAULT_BITMAP_GRANULARITY;
+    if (!journal_size)
+        journal_size = 16*1024*1024;
+    if (!device_block_size)
+        device_block_size = 4096;
+    uint64_t orig_device_size = device_size;
+    if (!device_size)
+    {
+        struct stat st;
+        if (stat(device.c_str(), &st) < 0)
+        {
+            fprintf(stderr, "Can't stat %s: %s\n", device.c_str(), strerror(errno));
+            exit(1);
+        }
+        if (S_ISBLK(st.st_mode))
+        {
+            int fd = open(device.c_str(), O_DIRECT|O_RDONLY);
+            if (fd < 0 || ioctl(fd, BLKGETSIZE64, &device_size) < 0)
+            {
+                fprintf(stderr, "Failed to get device size for %s: %s\n", device.c_str(), strerror(errno));
+                exit(1);
+            }
+            close(fd);
+            if (st.st_blksize < device_block_size)
+            {
+                fprintf(
+                    stderr, "Warning: %s reports %lu byte blocks, but we use %lu."
+                    " Set --device_block_size=%lu if you're sure it works well with %lu byte blocks.\n",
+                    device.c_str(), st.st_blksize, device_block_size, st.st_blksize, st.st_blksize
+                );
+            }
+        }
+        else if (S_ISREG(st.st_mode))
+        {
+            device_size = st.st_size;
+        }
+        else
+        {
+            fprintf(stderr, "%s is neither a block device nor a regular file\n", device.c_str());
+            exit(1);
+        }
+    }
+    if (!device_size)
+    {
+        fprintf(stderr, "Failed to get device size for %s\n", device.c_str());
+        exit(1);
+    }
+    if (device_block_size < 512 || device_block_size > 1048576 ||
+        device_block_size & (device_block_size-1) != 0)
+    {
+        fprintf(stderr, "Invalid device block size specified: %lu\n", device_block_size);
+        exit(1);
+    }
+    if (object_size < device_block_size || object_size > MAX_BLOCK_SIZE ||
+        object_size & (object_size-1) != 0)
+    {
+        fprintf(stderr, "Invalid object size specified: %lu\n", object_size);
+        exit(1);
+    }
+    if (bitmap_granularity < device_block_size || bitmap_granularity > object_size ||
+        bitmap_granularity & (bitmap_granularity-1) != 0)
+    {
+        fprintf(stderr, "Invalid bitmap granularity specified: %lu\n", bitmap_granularity);
+        exit(1);
+    }
+    journal_offset = ((journal_offset+device_block_size-1)/device_block_size)*device_block_size;
+    uint64_t meta_offset = journal_offset + ((journal_size+device_block_size-1)/device_block_size)*device_block_size;
+    uint64_t entries_per_block = (device_block_size / (24 + 2*object_size/bitmap_granularity/8));
+    uint64_t object_count = ((device_size-meta_offset)/object_size);
+    uint64_t meta_size = (1 + (object_count+entries_per_block-1)/entries_per_block) * device_block_size;
+    uint64_t data_offset = meta_offset + meta_size;
+    if (format == "json")
+    {
+        // JSON
+        printf("%s\n", json11::Json(json11::Json::object {
+            { "meta_block_size", device_block_size },
+            { "journal_block_size", device_block_size },
+            { "data_size", device_size-data_offset },
+            { "data_device", device },
+            { "journal_offset", journal_offset },
+            { "meta_offset", meta_offset },
+            { "data_offset", data_offset },
+        }).dump().c_str());
+    }
+    else if (format == "env")
+    {
+        // Env
+        printf(
+            "meta_block_size=%lu\njournal_block_size=%lu\ndata_size=%lu\n"
+            "data_device=%s\njournal_offset=%lu\nmeta_offset=%lu\ndata_offset=%lu\n",
+            device_block_size, device_block_size, device_size-data_offset,
+            device.c_str(), journal_offset, meta_offset, data_offset
+        );
+    }
+    else
+    {
+        // OSD command-line options
+        if (format != "options")
+        {
+            fprintf(stderr, "Metadata size: %s\nOptions for the OSD:\n", format_size(meta_size).c_str());
+        }
+        if (device_block_size != 4096)
+        {
+            printf("--meta_block_size %lu\n--journal_block_size %lu\n", device_block_size, device_block_size);
+        }
+        if (orig_device_size)
+        {
+            printf("--data_size %lu\n", device_size-data_offset);
+        }
+        printf(
+            "--data_device %s\n--journal_offset %lu\n--meta_offset %lu\n--data_offset %lu\n",
+            device.c_str(), journal_offset, meta_offset, data_offset
+        );
+    }
+    return NULL;
+}

src/cli_snap_rm.cpp

@@ -1,6 +1,7 @@
 // Copyright (c) Vitaliy Filippov, 2019+
 // License: VNPL-1.1 (see README.md for details)
 
+#include <fcntl.h>
 #include "cli.h"
 #include "cluster_client.h"
 #include "base64.h"

src/messenger.cpp

@@ -147,13 +147,13 @@ void osd_messenger_t::parse_config(const json11::Json & config)
         this->rdma_max_sge = 128;
     this->rdma_max_send = config["rdma_max_send"].uint64_value();
     if (!this->rdma_max_send)
-        this->rdma_max_send = 32;
+        this->rdma_max_send = 1;
     this->rdma_max_recv = config["rdma_max_recv"].uint64_value();
     if (!this->rdma_max_recv)
-        this->rdma_max_recv = 8;
+        this->rdma_max_recv = 128;
     this->rdma_max_msg = config["rdma_max_msg"].uint64_value();
     if (!this->rdma_max_msg || this->rdma_max_msg > 128*1024*1024)
-        this->rdma_max_msg = 1024*1024;
+        this->rdma_max_msg = 129*1024;
 #endif
     this->receive_buffer_size = (uint32_t)config["tcp_header_buffer_size"].uint64_value();
     if (!this->receive_buffer_size || this->receive_buffer_size > 1024*1024*1024)

src/messenger.h

@@ -133,7 +133,7 @@ protected:
     std::string rdma_device;
     uint64_t rdma_port_num = 1, rdma_gid_index = 0, rdma_mtu = 0;
     msgr_rdma_context_t *rdma_context = NULL;
-    uint64_t rdma_max_sge = 0, rdma_max_send = 0, rdma_max_recv = 8;
+    uint64_t rdma_max_sge = 0, rdma_max_send = 0, rdma_max_recv = 0;
     uint64_t rdma_max_msg = 0;
 #endif
 

src/mmap_manager.cpp

@@ -0,0 +1,82 @@
+#include <stdexcept>
+#include <cassert>
+#include <sys/mman.h>
+#include "mmap_manager.h"
+
+mmap_manager_t::mmap_manager_t(uint64_t mmap_size)
+{
+    this->mmap_size = mmap_size;
+}
+
+mmap_manager_t::~mmap_manager_t()
+{
+    for (auto & kv: past_buffers)
+    {
+        munmap(kv.second.addr, kv.second.size);
+    }
+    if (active_buffer.addr != NULL)
+    {
+        munmap(active_buffer.addr, active_buffer.size);
+    }
+}
+
+void *mmap_manager_t::alloc(uint64_t size)
+{
+    if (!active_buffer.addr || (active_buffer.pos + size) > active_buffer.size)
+    {
+        if (active_buffer.addr)
+        {
+            if (active_buffer.freed >= active_buffer.pos)
+                munmap(active_buffer.addr, active_buffer.size);
+            else
+                past_buffers[active_buffer.addr] = active_buffer;
+            active_buffer = { 0 };
+        }
+        uint64_t new_size = size < mmap_size ? mmap_size : size;
+        void *buf = mmap(NULL, new_size, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
+        if (!buf)
+            throw std::runtime_error(std::string("can't mmap "+std::to_string(new_size)+" bytes"));
+        active_buffer = {
+            .addr = buf,
+            .size = new_size,
+            .freed = 0,
+            .pos = 0,
+        };
+    }
+    void *res = active_buffer.addr + active_buffer.pos;
+    active_buffer.pos += size;
+    return res;
+}
+
+void mmap_manager_t::free(void *addr, uint64_t size)
+{
+    auto it = past_buffers.upper_bound(addr);
+    if (it != past_buffers.begin())
+    {
+        if (it == past_buffers.end())
+        {
+            it--;
+            if (addr < it->second.addr || addr >= it->second.addr+it->second.size)
+                it = past_buffers.end();
+        }
+        else
+            it--;
+    }
+    else
+        it = past_buffers.end();
+    if (it != past_buffers.end())
+    {
+        assert(addr >= it->second.addr && addr+size <= it->second.addr+it->second.size);
+        it->second.freed += size;
+        if (it->second.freed >= it->second.pos)
+        {
+            munmap(it->second.addr, it->second.size);
+            past_buffers.erase(it);
+        }
+    }
+    else
+    {
+        assert(addr < active_buffer.addr+active_buffer.size);
+        active_buffer.freed += size;
+    }
+}

src/mmap_manager.h

@@ -0,0 +1,26 @@
+#pragma once
+
+#include <stdint.h>
+#include <map>
+
+struct mmap_buffer_t
+{
+    void *addr = NULL;
+    uint64_t size = 0;
+    uint64_t freed = 0;
+    uint64_t pos = 0;
+};
+
+class mmap_manager_t
+{
+protected:
+    uint64_t mmap_size = 32*1024*1024;
+    std::map<void*, mmap_buffer_t> past_buffers;
+    mmap_buffer_t active_buffer;
+
+public:
+    mmap_manager_t(uint64_t mmap_size = 32*1024*1024);
+    ~mmap_manager_t();
+    void *alloc(uint64_t size);
+    void free(void *addr, uint64_t size);
+};

src/nbd_proxy.cpp

@@ -17,6 +17,8 @@
 
 #include "epoll_manager.h"
 #include "cluster_client.h"
+#include "mmap_manager.h"
+#include <stdexcept>
 
 #ifndef MSG_ZEROCOPY
 #define MSG_ZEROCOPY 0
@@ -24,6 +26,24 @@
 
 const char *exe_name = NULL;
 
+static inline void my_uring_prep_splice(struct io_uring_sqe *sqe,
+					int fd_in, int64_t off_in,
+					int fd_out, int64_t off_out,
+					unsigned int nbytes,
+					unsigned int splice_flags)
+{
+	my_uring_prep_rw(IORING_OP_SPLICE, sqe, fd_out, NULL, nbytes, (__u64) off_out);
+	sqe->splice_off_in = (__u64) off_in;
+	sqe->splice_fd_in = fd_in;
+	sqe->splice_flags = splice_flags;
+}
+
+struct buf_to_free_t
+{
+    void *buf = NULL;
+    uint64_t unmap = 0;
+};
+
 class nbd_proxy
 {
 protected:
@@ -38,7 +58,7 @@ protected:
     ring_consumer_t consumer;
 
     std::vector<iovec> send_list, next_send_list;
-    std::vector<void*> to_free;
+    std::vector<buf_to_free_t> to_free;
     int nbd_fd = -1;
     void *recv_buf = NULL;
     int receive_buffer_size = 9000;
@@ -51,6 +71,10 @@ protected:
     msghdr read_msg = { 0 }, send_msg = { 0 };
     iovec read_iov = { 0 };
 
+    mmap_manager_t mm;
+    int pipe_fd[2];
+    int vmspliced = 0;
+
 public:
     static json11::Json::object parse_args(int narg, const char *args[])
     {
@@ -174,6 +198,12 @@ public:
                 exit(1);
             }
         }
+        // Create pipe for splicing
+        if (pipe(pipe_fd) < 0)
+        {
+            fprintf(stderr, "pipe failed: %s\n", strerror(errno));
+            exit(1);
+        }
         // Create client
         ringloop = new ring_loop_t(512);
         epmgr = new epoll_manager_t(ringloop);
@@ -522,16 +552,76 @@ protected:
         {
             return;
         }
-        io_uring_sqe* sqe = ringloop->get_sqe();
+        int i;
+        //uint64_t len = 0;
+        for (i = 0; i < send_list.size(); i++)
+        {
+            if (to_free[i].unmap)
+            {
+                break;
+            }
+            //len += send_list[i].iov_len;
+        }
+        //if (true)
+        if (i > 0)
+        {
+            /*io_uring_sqe* sqe = ringloop->get_sqe();
             if (!sqe)
             {
                 return;
             }
             ring_data_t* data = ((ring_data_t*)sqe->user_data);
-        data->callback = [this](ring_data_t *data) { handle_send(data->res); };
+            data->callback = [this](ring_data_t *data) { handle_send(data->res); };*/
             send_msg.msg_iov = send_list.data();
-        send_msg.msg_iovlen = send_list.size();
+            //send_msg.msg_iovlen = send_list.size();
-        my_uring_prep_sendmsg(sqe, nbd_fd, &send_msg, MSG_ZEROCOPY);
+            send_msg.msg_iovlen = i;
+            //my_uring_prep_sendmsg(sqe, nbd_fd, &send_msg, MSG_ZEROCOPY);
+            int res = sendmsg(nbd_fd, &send_msg, MSG_ZEROCOPY);
+            if (res < 0)
+                res = -errno;
+            handle_send(res);
+            //int r = sendmsg(int sockfd, const struct msghdr *msg, int flags);
+        }
+        else
+        {
+            io_uring_sqe* sqe = ringloop->get_sqe();
+            if (!sqe)
+            {
+                return;
+            }
+            if (vmspliced <= 0)
+            {
+                vmspliced = vmsplice(pipe_fd[1], send_list.data(), 1, SPLICE_F_GIFT);
+                if (vmspliced < 0)
+                {
+                    throw std::runtime_error(std::string("vmsplice: ")+strerror(errno));
+                }
+            }
+            send_msg.msg_iovlen = 1;
+            ring_data_t* data = ((ring_data_t*)sqe->user_data);
+            data->callback = [this](ring_data_t *data)
+            {
+                if (data->res > 0)
+                    vmspliced -= data->res;
+                handle_send(data->res);
+            };
+            my_uring_prep_splice(sqe, pipe_fd[0], -1l, nbd_fd, -1l, vmspliced, SPLICE_F_MOVE);
+            /*int sent = res, spl = res;
+            while (spl > 0)
+            {
+                res = splice(pipe_fd[0], NULL, nbd_fd, NULL, spl, SPLICE_F_MOVE);
+                if (res < 0)
+                {
+                    if (errno != EAGAIN)
+                        throw std::runtime_error(std::string("splice: ")+strerror(errno));
+                }
+                else
+                {
+                    spl -= res;
+                }
+            }
+            handle_send(sent);*/
+        }
     }
 
     void handle_send(int result)
@@ -547,7 +637,10 @@ protected:
         {
             if (result >= send_list[to_eat].iov_len)
             {
-                free(to_free[to_eat]);
+                if (to_free[to_eat].unmap)
+                    mm.free(to_free[to_eat].buf, to_free[to_eat].unmap);
+                else
+                    free(to_free[to_eat].buf);
                 result -= send_list[to_eat].iov_len;
                 to_eat++;
             }
@@ -659,6 +752,7 @@ protected:
             printf("request %lx +%x %lx\n", be64toh(cur_req.from), be32toh(cur_req.len), handle);
 #endif
             void *buf = NULL;
+            nbd_reply *reply = NULL;
             cluster_op_t *op = new cluster_op_t;
             if (req_type == NBD_CMD_READ || req_type == NBD_CMD_WRITE)
             {
@@ -666,36 +760,51 @@ protected:
                 op->inode = inode ? inode : watch->cfg.num;
                 op->offset = be64toh(cur_req.from);
                 op->len = be32toh(cur_req.len);
+                if (req_type == NBD_CMD_WRITE)
+                {
                     buf = malloc_or_die(sizeof(nbd_reply) + op->len);
+                    reply = (nbd_reply*)buf;
                     op->iov.push_back(buf + sizeof(nbd_reply), op->len);
                 }
+                else
+                {
+                    buf = mm.alloc(op->len);
+                    reply = (nbd_reply*)malloc_or_die(sizeof(nbd_reply));
+                    op->iov.push_back(buf, op->len);
+                }
+            }
             else if (req_type == NBD_CMD_FLUSH)
             {
                 op->opcode = OSD_OP_SYNC;
-                buf = malloc_or_die(sizeof(nbd_reply));
+                reply = (nbd_reply*)malloc_or_die(sizeof(nbd_reply));
             }
-            op->callback = [this, buf, handle](cluster_op_t *op)
+            op->callback = [this, buf, reply, handle](cluster_op_t *op)
             {
 #ifdef DEBUG
                 printf("reply %lx e=%d\n", handle, op->retval);
 #endif
-                nbd_reply *reply = (nbd_reply*)buf;
                 reply->magic = htobe32(NBD_REPLY_MAGIC);
                 memcpy(reply->handle, &handle, 8);
                 reply->error = htobe32(op->retval < 0 ? -op->retval : 0);
                 auto & to_list = send_msg.msg_iovlen > 0 ? next_send_list : send_list;
-                if (op->retval < 0 || op->opcode != OSD_OP_READ)
+                to_list.push_back((iovec){ .iov_base = reply, .iov_len = sizeof(nbd_reply) });
-                    to_list.push_back({ .iov_base = buf, .iov_len = sizeof(nbd_reply) });
+                to_free.push_back((buf_to_free_t){ .buf = reply, .unmap = 0 });
-                else
+                if (op->retval >= 0 && op->opcode == OSD_OP_READ)
-                    to_list.push_back({ .iov_base = buf, .iov_len = sizeof(nbd_reply) + op->len });
+                {
-                to_free.push_back(buf);
+                    to_list.push_back((iovec){ .iov_base = buf, .iov_len = op->len });
+                    to_free.push_back((buf_to_free_t){ .buf = buf, .unmap = op->len });
+                }
+                else if (op->opcode == OSD_OP_READ)
+                {
+                    mm.free(buf, op->len);
+                }
                 delete op;
                 ringloop->wakeup();
             };
             if (req_type == NBD_CMD_WRITE)
             {
                 cur_op = op;
-                cur_buf = buf + sizeof(nbd_reply);
+                cur_buf = buf;
                 cur_left = op->len;
                 read_state = CL_READ_DATA;
             }

src/qemu_driver.c

@@ -132,16 +132,19 @@ static void vitastor_parse_filename(const char *filename, QDict *options, Error
             error_setg(errp, "conf option %s has no value", name);
             break;
         }
+        for (int i = 0; i < strlen(name); i++)
+            if (name[i] == '_')
+                name[i] = '-';
         qemu_vitastor_unescape(name);
         value = qemu_vitastor_next_tok(p, ':', &p);
         qemu_vitastor_unescape(value);
         if (!strcmp(name, "inode") ||
             !strcmp(name, "pool") ||
             !strcmp(name, "size") ||
-            !strcmp(name, "use_rdma") ||
+            !strcmp(name, "use-rdma") ||
-            !strcmp(name, "rdma_port_num") ||
+            !strcmp(name, "rdma-port_num") ||
-            !strcmp(name, "rdma_gid_index") ||
+            !strcmp(name, "rdma-gid-index") ||
-            !strcmp(name, "rdma_mtu"))
+            !strcmp(name, "rdma-mtu"))
         {
             unsigned long long num_val;
             if (parse_uint_full(value, &num_val, 0))
@@ -202,15 +205,15 @@ static int vitastor_file_open(BlockDriverState *bs, QDict *options, int flags, E
     VitastorClient *client = bs->opaque;
     int64_t ret = 0;
     qemu_mutex_init(&client->mutex);
-    client->config_path = g_strdup(qdict_get_try_str(options, "config_path"));
+    client->config_path = g_strdup(qdict_get_try_str(options, "config-path"));
     // FIXME: Rename to etcd_address
-    client->etcd_host = g_strdup(qdict_get_try_str(options, "etcd_host"));
+    client->etcd_host = g_strdup(qdict_get_try_str(options, "etcd-host"));
-    client->etcd_prefix = g_strdup(qdict_get_try_str(options, "etcd_prefix"));
+    client->etcd_prefix = g_strdup(qdict_get_try_str(options, "etcd-prefix"));
-    client->use_rdma = qdict_get_try_int(options, "use_rdma", -1);
+    client->use_rdma = qdict_get_try_int(options, "use-rdma", -1);
-    client->rdma_device = g_strdup(qdict_get_try_str(options, "rdma_device"));
+    client->rdma_device = g_strdup(qdict_get_try_str(options, "rdma-device"));
-    client->rdma_port_num = qdict_get_try_int(options, "rdma_port_num", 0);
+    client->rdma_port_num = qdict_get_try_int(options, "rdma-port-num", 0);
-    client->rdma_gid_index = qdict_get_try_int(options, "rdma_gid_index", 0);
+    client->rdma_gid_index = qdict_get_try_int(options, "rdma-gid-index", 0);
-    client->rdma_mtu = qdict_get_try_int(options, "rdma_mtu", 0);
+    client->rdma_mtu = qdict_get_try_int(options, "rdma-mtu", 0);
     client->proxy = vitastor_c_create_qemu(
         (QEMUSetFDHandler*)aio_set_fd_handler, bdrv_get_aio_context(bs), client->config_path, client->etcd_host, client->etcd_prefix,
         client->use_rdma, client->rdma_device, client->rdma_port_num, client->rdma_gid_index, client->rdma_mtu, 0
@@ -264,14 +267,14 @@ static int vitastor_file_open(BlockDriverState *bs, QDict *options, int flags, E
     }
     bs->total_sectors = client->size / BDRV_SECTOR_SIZE;
     //client->aio_context = bdrv_get_aio_context(bs);
-    qdict_del(options, "use_rdma");
+    qdict_del(options, "use-rdma");
-    qdict_del(options, "rdma_mtu");
+    qdict_del(options, "rdma-mtu");
-    qdict_del(options, "rdma_gid_index");
+    qdict_del(options, "rdma-gid-index");
-    qdict_del(options, "rdma_port_num");
+    qdict_del(options, "rdma-port-num");
-    qdict_del(options, "rdma_device");
+    qdict_del(options, "rdma-device");
-    qdict_del(options, "config_path");
+    qdict_del(options, "config-path");
-    qdict_del(options, "etcd_host");
+    qdict_del(options, "etcd-host");
-    qdict_del(options, "etcd_prefix");
+    qdict_del(options, "etcd-prefix");
     qdict_del(options, "image");
     qdict_del(options, "inode");
     qdict_del(options, "pool");
@@ -515,9 +518,9 @@ static QEMUOptionParameter vitastor_create_opts[] = {
 static const char *vitastor_strong_runtime_opts[] = {
     "inode",
     "pool",
-    "config_path",
+    "config-path",
-    "etcd_host",
+    "etcd-host",
-    "etcd_prefix",
+    "etcd-prefix",
 
     NULL
 };

src/stub_osd.cpp

@@ -37,6 +37,7 @@
 
 #include <stdexcept>
 
+#include "mmap_manager.h"
 #include "rw_blocking.h"
 #include "osd_ops.h"
 
@@ -115,6 +116,12 @@ int bind_stub(const char *bind_address, int bind_port)
 
 void run_stub(int peer_fd)
 {
+    mmap_manager_t mm;
+    int pipe_fd[2];
+    if (pipe(pipe_fd) < 0)
+    {
+        throw std::runtime_error(std::string("pipe: ") + strerror(errno));
+    }
     osd_any_op_t op;
     osd_any_reply_t reply;
     void *buf = NULL;
@@ -136,11 +143,39 @@ void run_stub(int peer_fd)
         if (op.hdr.opcode == OSD_OP_SEC_READ)
         {
             reply.hdr.retval = op.sec_rw.len;
-            buf = malloc(op.sec_rw.len);
+            //buf = malloc(op.sec_rw.len);
+            buf = mm.alloc(op.sec_rw.len);
             r = write_blocking(peer_fd, reply.buf, OSD_PACKET_SIZE);
             if (r == OSD_PACKET_SIZE)
-                r = write_blocking(peer_fd, &buf, op.sec_rw.len);
+            {
-            free(buf);
+                size_t offset = 0;
+                while (offset < op.sec_rw.len)
+                {
+                    iovec iov = { .iov_base = buf+offset, .iov_len = op.sec_rw.len-offset };
+                    int vmspliced = vmsplice(pipe_fd[1], &iov, 1, SPLICE_F_GIFT);
+                    if (vmspliced < 0)
+                    {
+                        throw std::runtime_error(std::string("vmsplice: ")+strerror(errno));
+                    }
+                    int spliced = 0;
+                    while (spliced < vmspliced)
+                    {
+                        int r2 = splice(pipe_fd[0], NULL, peer_fd, NULL, vmspliced-spliced, SPLICE_F_MOVE);
+                        if (r2 < 0)
+                        {
+                            if (errno != EAGAIN)
+                                throw std::runtime_error(std::string("splice: ")+strerror(errno));
+                        }
+                        else
+                            spliced += r2;
+                    }
+                    offset += vmspliced;
+                }
+                r = offset;
+                //r = write_blocking(peer_fd, &buf, op.sec_rw.len);
+            }
+            mm.free(buf, op.sec_rw.len);
+            buf = NULL;
             if (r < op.sec_rw.len)
                 break;
         }
@@ -170,5 +205,6 @@ void run_stub(int peer_fd)
             break;
         }
     }
-    free(buf);
+    close(pipe_fd[0]);
+    close(pipe_fd[1]);
 }

src/vitastor.pc.in

@@ -0,0 +1,12 @@
+prefix=@CMAKE_INSTALL_PREFIX@
+exec_prefix=${prefix}
+libdir=${prefix}/@CMAKE_INSTALL_LIBDIR@
+includedir=${prefix}/@CMAKE_INSTALL_INCLUDEDIR@
+
+
+Name: Vitastor
+Description: Vitastor client library
+Version: 0.6.8
+Libs: -L${libdir} -lvitastor_client
+Cflags: -I${includedir}
+

tests/run_3osds.sh

@@ -10,7 +10,7 @@ SCHEME=${SCHEME:-ec}
 
 for i in $(seq 1 $OSD_COUNT); do
     dd if=/dev/zero of=./testdata/test_osd$i.bin bs=1024 count=1 seek=$((OSD_SIZE*1024-1))
-    build/src/vitastor-osd --osd_num $i --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd$i.bin 2>/dev/null) &>./testdata/osd$i.log &
+    build/src/vitastor-osd --osd_num $i --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(build/src/vitastor-cli simple-offsets --format options ./testdata/test_osd$i.bin 2>/dev/null) &>./testdata/osd$i.log &
     eval OSD${i}_PID=$!
 done
 

tests/test_change_pg_count.sh

@@ -10,25 +10,14 @@ else
     NOBJ=1024
 fi
 
-dd if=/dev/zero of=./testdata/test_osd1.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_SIZE=1024
-dd if=/dev/zero of=./testdata/test_osd2.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_COUNT=6
-dd if=/dev/zero of=./testdata/test_osd3.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_ARGS=
-dd if=/dev/zero of=./testdata/test_osd4.bin bs=1024 count=1 seek=$((1024*1024-1))
+for i in $(seq 1 $OSD_COUNT); do
-dd if=/dev/zero of=./testdata/test_osd5.bin bs=1024 count=1 seek=$((1024*1024-1))
+    dd if=/dev/zero of=./testdata/test_osd$i.bin bs=1024 count=1 seek=$((OSD_SIZE*1024-1))
-dd if=/dev/zero of=./testdata/test_osd6.bin bs=1024 count=1 seek=$((1024*1024-1))
+    build/src/vitastor-osd --osd_num $i --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(build/src/vitastor-cli simple-offsets --format options ./testdata/test_osd$i.bin 2>/dev/null) &>./testdata/osd$i.log &
-
+    eval OSD${i}_PID=$!
-build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd1.bin 2>/dev/null) 2>&1 >>./testdata/osd1.log &
+done
-OSD1_PID=$!
-build/src/vitastor-osd --osd_num 2 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd2.bin 2>/dev/null) 2>&1 >>./testdata/osd2.log &
-OSD2_PID=$!
-build/src/vitastor-osd --osd_num 3 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd3.bin 2>/dev/null) 2>&1 >>./testdata/osd3.log &
-OSD3_PID=$!
-build/src/vitastor-osd --osd_num 4 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd4.bin 2>/dev/null) 2>&1 >>./testdata/osd4.log &
-OSD4_PID=$!
-build/src/vitastor-osd --osd_num 5 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd5.bin 2>/dev/null) 2>&1 >>./testdata/osd5.log &
-OSD5_PID=$!
-build/src/vitastor-osd --osd_num 6 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd6.bin 2>/dev/null) 2>&1 >>./testdata/osd6.log &
-OSD6_PID=$!
 
 cd mon
 npm install

tests/test_change_pg_size.sh

@@ -2,16 +2,14 @@
 
 . `dirname $0`/common.sh
 
-dd if=/dev/zero of=./testdata/test_osd1.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_SIZE=1024
-dd if=/dev/zero of=./testdata/test_osd2.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_COUNT=3
-dd if=/dev/zero of=./testdata/test_osd3.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_ARGS=
-
+for i in $(seq 1 $OSD_COUNT); do
-build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd1.bin 2>/dev/null) &>./testdata/osd1.log &
+    dd if=/dev/zero of=./testdata/test_osd$i.bin bs=1024 count=1 seek=$((OSD_SIZE*1024-1))
-OSD1_PID=$!
+    build/src/vitastor-osd --osd_num $i --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(build/src/vitastor-cli simple-offsets --format options ./testdata/test_osd$i.bin 2>/dev/null) &>./testdata/osd$i.log &
-build/src/vitastor-osd --osd_num 2 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd2.bin 2>/dev/null) &>./testdata/osd2.log &
+    eval OSD${i}_PID=$!
-OSD2_PID=$!
+done
-build/src/vitastor-osd --osd_num 3 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd3.bin 2>/dev/null) &>./testdata/osd3.log &
-OSD3_PID=$!
 
 cd mon
 npm install

tests/test_interrupted_rebalance.sh

@@ -10,28 +10,14 @@ else
     $ETCDCTL put /vitastor/config/global '{"recovery_queue_depth":1,"osd_out_time":5}'
 fi
 
-dd if=/dev/zero of=./testdata/test_osd1.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_SIZE=1024
-dd if=/dev/zero of=./testdata/test_osd2.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_COUNT=7
-dd if=/dev/zero of=./testdata/test_osd3.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_ARGS=
-dd if=/dev/zero of=./testdata/test_osd4.bin bs=1024 count=1 seek=$((1024*1024-1))
+for i in $(seq 1 $OSD_COUNT); do
-dd if=/dev/zero of=./testdata/test_osd5.bin bs=1024 count=1 seek=$((1024*1024-1))
+    dd if=/dev/zero of=./testdata/test_osd$i.bin bs=1024 count=1 seek=$((OSD_SIZE*1024-1))
-dd if=/dev/zero of=./testdata/test_osd6.bin bs=1024 count=1 seek=$((1024*1024-1))
+    build/src/vitastor-osd --osd_num $i --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(build/src/vitastor-cli simple-offsets --format options ./testdata/test_osd$i.bin 2>/dev/null) &>./testdata/osd$i.log &
-dd if=/dev/zero of=./testdata/test_osd7.bin bs=1024 count=1 seek=$((1024*1024-1))
+    eval OSD${i}_PID=$!
-
+done
-build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 $NO_SAME --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd1.bin 2>/dev/null) &>./testdata/osd1.log &
-OSD1_PID=$!
-build/src/vitastor-osd --osd_num 2 --bind_address 127.0.0.1 $NO_SAME --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd2.bin 2>/dev/null) &>./testdata/osd2.log &
-OSD2_PID=$!
-build/src/vitastor-osd --osd_num 3 --bind_address 127.0.0.1 $NO_SAME --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd3.bin 2>/dev/null) &>./testdata/osd3.log &
-OSD3_PID=$!
-build/src/vitastor-osd --osd_num 4 --bind_address 127.0.0.1 $NO_SAME --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd4.bin 2>/dev/null) &>./testdata/osd4.log &
-OSD4_PID=$!
-build/src/vitastor-osd --osd_num 5 --bind_address 127.0.0.1 $NO_SAME --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd5.bin 2>/dev/null) &>./testdata/osd5.log &
-OSD5_PID=$!
-build/src/vitastor-osd --osd_num 6 --bind_address 127.0.0.1 $NO_SAME --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd6.bin 2>/dev/null) &>./testdata/osd6.log &
-OSD6_PID=$!
-build/src/vitastor-osd --osd_num 7 --bind_address 127.0.0.1 $NO_SAME --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd7.bin 2>/dev/null) &>./testdata/osd7.log &
-OSD7_PID=$!
 
 cd mon
 npm install

tests/test_move_reappear.sh

@@ -2,22 +2,14 @@
 
 . `dirname $0`/common.sh
 
-dd if=/dev/zero of=./testdata/test_osd1.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_SIZE=1024
-dd if=/dev/zero of=./testdata/test_osd2.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_COUNT=5
-dd if=/dev/zero of=./testdata/test_osd3.bin bs=1024 count=1 seek=$((1024*1024-1))
+OSD_ARGS=
-dd if=/dev/zero of=./testdata/test_osd4.bin bs=1024 count=1 seek=$((1024*1024-1))
+for i in $(seq 1 $OSD_COUNT); do
-dd if=/dev/zero of=./testdata/test_osd5.bin bs=1024 count=1 seek=$((1024*1024-1))
+    dd if=/dev/zero of=./testdata/test_osd$i.bin bs=1024 count=1 seek=$((OSD_SIZE*1024-1))
-
+    build/src/vitastor-osd --osd_num $i --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(build/src/vitastor-cli simple-offsets --format options ./testdata/test_osd$i.bin 2>/dev/null) &>./testdata/osd$i.log &
-build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd1.bin 2>/dev/null) &>./testdata/osd1.log &
+    eval OSD${i}_PID=$!
-OSD1_PID=$!
+done
-build/src/vitastor-osd --osd_num 2 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd2.bin 2>/dev/null) &>./testdata/osd2.log &
-OSD2_PID=$!
-build/src/vitastor-osd --osd_num 3 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd3.bin 2>/dev/null) &>./testdata/osd3.log &
-OSD3_PID=$!
-build/src/vitastor-osd --osd_num 4 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd4.bin 2>/dev/null) &>./testdata/osd4.log &
-OSD4_PID=$!
-build/src/vitastor-osd --osd_num 5 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd5.bin 2>/dev/null) &>./testdata/osd5.log &
-OSD5_PID=$!
 
 $ETCDCTL put /vitastor/config/pools '{"1":{"name":"testpool","scheme":"replicated","pg_size":2,"pg_minsize":1,"pg_count":1,"failure_domain":"osd"}}'
 

tests/test_splitbrain.sh

@@ -20,7 +20,7 @@ LD_PRELOAD=libasan.so.5 \
 # Kill OSD 2, start OSD 1
 
 kill $OSD2_PID
-build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd2.bin 2>/dev/null) >>./testdata/osd2.log 2>&1 &
+build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(build/src/vitastor-cli simple-offsets --format options --device ./testdata/test_osd2.bin 2>/dev/null) >>./testdata/osd2.log 2>&1 &
 sleep 2
 
 # Check PG state - it should NOT become active

tests/test_vm_cont.sh

@@ -7,12 +7,12 @@ etcdctl --endpoints=http://127.0.0.1:12379/v3 del --prefix /vitastor/mon/master
 etcdctl --endpoints=http://127.0.0.1:12379/v3 del --prefix /vitastor/pg/state
 etcdctl --endpoints=http://127.0.0.1:12379/v3 del --prefix /vitastor/osd/state
 
-build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd1.bin 2>/dev/null) &>./testdata/osd1.log &
+OSD_COUNT=3
-OSD1_PID=$!
+OSD_ARGS=
-build/src/vitastor-osd --osd_num 2 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd2.bin 2>/dev/null) &>./testdata/osd2.log &
+for i in $(seq 1 $OSD_COUNT); do
-OSD2_PID=$!
+    build/src/vitastor-osd --osd_num $i --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL $(build/src/vitastor-cli simple-offsets --format options ./testdata/test_osd$i.bin 2>/dev/null) &>./testdata/osd$i.log &
-build/src/vitastor-osd --osd_num 3 --bind_address 127.0.0.1 --etcd_address $ETCD_URL $(node mon/simple-offsets.js --format options --device ./testdata/test_osd3.bin 2>/dev/null) &>./testdata/osd3.log &
+    eval OSD${i}_PID=$!
-OSD3_PID=$!
+done
 
 node mon/mon-main.js --etcd_url http://$ETCD_URL --etcd_prefix "/vitastor" &>./testdata/mon.log &
 MON_PID=$!