Release 0.9.4

- Improve QEMU driver performance by integrating io_uring in it (up to 1.5x total iops improvement)
- Fix QEMU driver deadlocks which started to reproduce in qemu-img after iothread fixes
- Fix `vitastor-cli status` reporting more etcds than actually exists (fix etcd address duplication in config on reload)
- Fix `vitastor-cli ls` crashing on inodes in non-existing pools
- Delete old garbage /pool/stats/ keys for non-existing (deleted) pools
- Reduce memory usage of etcds initialized by make-etcd script
- Fix OSDs almost always crashing on etcd restart due to "revisions were compacted" (support reloading state from etcd)
- Fix a crash and a stall possible mostly in HDD setups with small journal and big (512k, 900k) random writes
- Add notes about HDDs to documentation. You are officially allowed to use HDD-only Vitastor with HGST/Toshiba/EXOS :)

.gitea/workflows/test.Dockerfile

@@ -10,6 +10,9 @@ RUN set -e -x; \
     ln -s /root/fio-build/fio-*/ ./fio; \
     ln -s /root/qemu-build/qemu-*/ ./qemu; \
     ls /usr/include/linux/raw.h || cp ./debian/raw.h /usr/include/linux/raw.h; \
+    cd mon; \
+    npm install; \
+    cd ..; \
     mkdir build; \
     cd build; \
     cmake .. -DWITH_ASAN=yes -DWITH_QEMU=yes; \

.gitea/workflows/test.yml

@@ -190,24 +190,6 @@ jobs:
           echo ""
         done
 
-  test_failure_domain:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_failure_domain.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
   test_interrupted_rebalance:
     runs-on: ubuntu-latest
     needs: build
@@ -280,6 +262,60 @@ jobs:
           echo ""
         done
 
+  test_failure_domain:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: /root/vitastor/tests/test_failure_domain.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
+  test_snapshot:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: /root/vitastor/tests/test_snapshot.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
+  test_snapshot_ec:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: SCHEME=ec /root/vitastor/tests/test_snapshot.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
   test_minsize_1:
     runs-on: ubuntu-latest
     needs: build
@@ -316,6 +352,114 @@ jobs:
           echo ""
         done
 
+  test_rm:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: /root/vitastor/tests/test_rm.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
+  test_snapshot_chain:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: /root/vitastor/tests/test_snapshot_chain.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
+  test_snapshot_chain_ec:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: SCHEME=ec /root/vitastor/tests/test_snapshot_chain.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
+  test_snapshot_down:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: /root/vitastor/tests/test_snapshot_down.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
+  test_snapshot_down_ec:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: SCHEME=ec /root/vitastor/tests/test_snapshot_down.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
+  test_splitbrain:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: /root/vitastor/tests/test_splitbrain.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+
   test_rebalance_verify:
     runs-on: ubuntu-latest
     needs: build
@@ -388,78 +532,6 @@ jobs:
           echo ""
         done
 
-  test_rm:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_rm.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
-  test_snapshot:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_snapshot.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
-  test_snapshot_ec:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: SCHEME=ec /root/vitastor/tests/test_snapshot.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
-  test_splitbrain:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_splitbrain.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
   test_write:
     runs-on: ubuntu-latest
     needs: build

CMakeLists.txt

@@ -2,6 +2,6 @@ cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 
-set(VERSION "0.9.2")
+set(VERSION "0.9.4")
 
 add_subdirectory(src)

README-ru.md

@@ -15,7 +15,7 @@ Vitastor архитектурно похож на Ceph, что означает
 и автоматическое распределение данных по любому числу дисков любого размера с настраиваемыми схемами
 избыточности - репликацией или с произвольными кодами коррекции ошибок.
 
-Vitastor нацелен на SSD и SSD+HDD кластеры с как минимум 10 Гбит/с сетью, поддерживает
+Vitastor нацелен в первую очередь на SSD и SSD+HDD кластеры с как минимум 10 Гбит/с сетью, поддерживает
 TCP и RDMA и на хорошем железе может достигать задержки 4 КБ чтения и записи на уровне ~0.1 мс,
 что примерно в 10 раз быстрее, чем Ceph и другие популярные программные СХД.
 

README.md

@@ -14,8 +14,8 @@ Vitastor is 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).
 
-Vitastor targets SSD and SSD+HDD clusters with at least 10 Gbit/s network, supports
-TCP and RDMA and may achieve 4 KB read and write latency as low as ~0.1 ms
+Vitastor targets primarily SSD and SSD+HDD clusters with at least 10 Gbit/s network,
+supports TCP and RDMA and may achieve 4 KB read and write latency as low as ~0.1 ms
 with proper hardware which is ~10 times faster than other popular SDS's like Ceph
 or internal systems of public clouds.
 

csi/Makefile

@@ -1,4 +1,4 @@
-VERSION ?= v0.9.2
+VERSION ?= v0.9.4
 
 all: build push
 

csi/deploy/004-csi-nodeplugin.yaml

@@ -49,7 +49,7 @@ spec:
             capabilities:
               add: ["SYS_ADMIN"]
             allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v0.9.2
+          image: vitalif/vitastor-csi:v0.9.4
           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:v0.9.2
+          image: vitalif/vitastor-csi:v0.9.4
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/src/config.go

@@ -5,7 +5,7 @@ package vitastor
 
 const (
     vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "0.9.2"
+    vitastorCSIDriverVersion = "0.9.4"
 )
 
 // Config struct fills the parameters of request or user input

debian/build-pve-qemu.sh

@@ -0,0 +1,58 @@
+exit
+
+git clone https://git.yourcmc.ru/vitalif/pve-qemu .
+
+# bookworm
+
+docker run -it -v `pwd`/pve-qemu:/root/pve-qemu --name pve-qemu-bullseye debian:bullseye bash
+
+perl -i -pe 's/Types: deb$/Types: deb deb-src/' /etc/apt/sources.list.d/debian.sources
+echo 'deb [arch=amd64] http://download.proxmox.com/debian/pve bookworm pve-no-subscription' >> /etc/apt/sources.list
+echo 'deb https://vitastor.io/debian bookworm main' >> /etc/apt/sources.list
+echo 'APT::Install-Recommends false;' >> /etc/apt/apt.conf
+echo 'ru_RU UTF-8' >> /etc/locale.gen
+echo 'en_US UTF-8' >> /etc/locale.gen
+apt-get update
+apt-get install wget ca-certificates
+wget https://enterprise.proxmox.com/debian/proxmox-release-bookworm.gpg -O /etc/apt/trusted.gpg.d/proxmox-release-bookworm.gpg
+wget https://vitastor.io/debian/pubkey.gpg -O /etc/apt/trusted.gpg.d/vitastor.gpg
+apt-get update
+apt-get install git devscripts equivs wget mc libjemalloc-dev vitastor-client-dev lintian locales
+mk-build-deps --install ./control
+
+# bullseye
+
+docker run -it -v `pwd`/pve-qemu:/root/pve-qemu --name pve-qemu-bullseye debian:bullseye bash
+
+grep '^deb ' /etc/apt/sources.list | perl -pe 's/^deb /deb-src /' >> /etc/apt/sources.list
+echo 'deb [arch=amd64] http://download.proxmox.com/debian/pve bullseye pve-no-subscription' >> /etc/apt/sources.list
+echo 'deb https://vitastor.io/debian bullseye main' >> /etc/apt/sources.list
+echo 'APT::Install-Recommends false;' >> /etc/apt/apt.conf
+echo 'ru_RU UTF-8' >> /etc/locale.gen
+echo 'en_US UTF-8' >> /etc/locale.gen
+apt-get update
+apt-get install wget
+wget https://enterprise.proxmox.com/debian/proxmox-release-bullseye.gpg -O /etc/apt/trusted.gpg.d/proxmox-release-bullseye.gpg
+wget https://vitastor.io/debian/pubkey.gpg -O /etc/apt/trusted.gpg.d/vitastor.gpg
+apt-get update
+apt-get install git devscripts equivs wget mc libjemalloc-dev vitastor-client-dev lintian locales
+mk-build-deps --install ./control
+
+# buster
+
+docker run -it -v `pwd`/pve-qemu:/root/pve-qemu --name pve-qemu-buster debian:buster bash
+
+grep '^deb ' /etc/apt/sources.list | perl -pe 's/^deb /deb-src /' >> /etc/apt/sources.list
+echo 'deb [arch=amd64] http://download.proxmox.com/debian/pve buster pve-no-subscription' >> /etc/apt/sources.list
+echo 'deb https://vitastor.io/debian buster main' >> /etc/apt/sources.list
+echo 'deb http://deb.debian.org/debian buster-backports main' >> /etc/apt/sources.list
+echo 'APT::Install-Recommends false;' >> /etc/apt/apt.conf
+echo 'ru_RU UTF-8' >> /etc/locale.gen
+echo 'en_US UTF-8' >> /etc/locale.gen
+apt-get update
+apt-get install wget ca-certificates
+wget http://download.proxmox.com/debian/proxmox-ve-release-6.x.gpg -O /etc/apt/trusted.gpg.d/proxmox-ve-release-6.x.gpg
+wget https://vitastor.io/debian/pubkey.gpg -O /etc/apt/trusted.gpg.d/vitastor.gpg
+apt-get update
+apt-get install git devscripts equivs wget mc libjemalloc-dev vitastor-client-dev lintian locales
+mk-build-deps --install ./control

debian/changelog

@@ -1,10 +1,10 @@
-vitastor (0.9.2-1) unstable; urgency=medium
+vitastor (0.9.4-1) unstable; urgency=medium
 
   * Bugfixes
 
  -- Vitaliy Filippov <vitalif@yourcmc.ru>  Fri, 03 Jun 2022 02:09:44 +0300
 
-vitastor (0.9.2-1) unstable; urgency=medium
+vitastor (0.9.4-1) unstable; urgency=medium
 
   * Implement NFS proxy
   * Add documentation

debian/patched-qemu.Dockerfile

@@ -28,13 +28,19 @@ RUN apt-get --download-only source qemu
 
 ADD patches /root/vitastor/patches
 ADD src/qemu_driver.c /root/vitastor/src/qemu_driver.c
+
+#RUN set -e; \
+#    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 $REL main > /etc/apt/sources.list.d/vitastor.list); \
+#    (echo "APT::Install-Recommends false;" > /etc/apt/apt.conf) && \
+#    apt-get update; \
+#    apt-get install -y vitastor-client vitastor-client-dev quilt
+
 RUN set -e; \
-    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 $REL main > /etc/apt/sources.list.d/vitastor.list); \
-    (echo "APT::Install-Recommends false;" > /etc/apt/apt.conf) && \
+    dpkg -i /root/packages/vitastor-$REL/vitastor-client_*.deb /root/packages/vitastor-$REL/vitastor-client-dev_*.deb; \
     apt-get update; \
-    apt-get install -y vitastor-client vitastor-client-dev quilt; \
+    apt-get install -y quilt; \
     mkdir -p /root/packages/qemu-$REL; \
     rm -rf /root/packages/qemu-$REL/*; \
     cd /root/packages/qemu-$REL; \
@@ -48,7 +54,7 @@ RUN set -e; \
     quilt add block/vitastor.c; \
     cp /root/vitastor/src/qemu_driver.c block/vitastor.c; \
     quilt refresh; \
-    V=$(head -n1 debian/changelog | perl -pe 's/^.*\((.*?)(~bpo[\d\+]*)?\).*$/$1/')+vitastor1; \
+    V=$(head -n1 debian/changelog | perl -pe 's/^.*\((.*?)(~bpo[\d\+]*)?\).*$/$1/')+vitastor3; \
     DEBEMAIL="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/vitastor.Dockerfile

@@ -35,8 +35,8 @@ RUN set -e -x; \
     mkdir -p /root/packages/vitastor-$REL; \
     rm -rf /root/packages/vitastor-$REL/*; \
     cd /root/packages/vitastor-$REL; \
-    cp -r /root/vitastor vitastor-0.9.2; \
-    cd vitastor-0.9.2; \
+    cp -r /root/vitastor vitastor-0.9.4; \
+    cd vitastor-0.9.4; \
     ln -s /root/fio-build/fio-*/ ./fio; \
     FIO=$(head -n1 fio/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     ls /usr/include/linux/raw.h || cp ./debian/raw.h /usr/include/linux/raw.h; \
@@ -49,8 +49,8 @@ RUN set -e -x; \
     rm -rf a b; \
     echo "dep:fio=$FIO" > debian/fio_version; \
     cd /root/packages/vitastor-$REL; \
-    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.9.2.orig.tar.xz vitastor-0.9.2; \
-    cd vitastor-0.9.2; \
+    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.9.4.orig.tar.xz vitastor-0.9.4; \
+    cd vitastor-0.9.4; \
     V=$(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; \

docs/config.en.md

@@ -21,7 +21,7 @@ Configuration parameters can be set in 3 places:
   mon, fio and QEMU options, OpenStack/Proxmox/etc configuration. The latter
   doesn't allow to set all variables directly, but it allows to override the
   configuration file and set everything you need inside it.
-- OSD superblocks created by [vitastor-disk](../usage/disk.en.md) contain
+- OSD superblocks created by [vitastor-disk](usage/disk.en.md) contain
   primarily disk layout parameters of specific OSDs. In fact, these parameters
   are automatically passed into the command line of vitastor-osd process, so
   they have the same "status" as command-line parameters.

docs/config.ru.md

@@ -23,7 +23,7 @@
   монитора, опциях fio и QEMU, настроек OpenStack, Proxmox и т.п. Последние,
   как правило, не включают полный набор параметров напрямую, но позволяют
   определить путь к файлу конфигурации и задать любые параметры в нём.
-- В суперблоке OSD, записываемом [vitastor-disk](../usage/disk.ru.md) - параметры,
+- В суперблоке OSD, записываемом [vitastor-disk](usage/disk.ru.md) - параметры,
   связанные с дисковым форматом и с этим конкретным OSD. На самом деле,
   при запуске OSD эти параметры автоматически передаются в командную строку
   процесса vitastor-osd, то есть по "статусу" они эквивалентны параметрам

docs/config/layout-cluster.en.md

@@ -33,12 +33,13 @@ Size of objects (data blocks) into which all physical and virtual drives
 in Vitastor, affects memory usage, write amplification and I/O load
 distribution effectiveness.
 
-Recommended default block size is 128 KB for SSD and 4 MB for HDD. In fact,
-it's possible to use 4 MB for SSD too - it will lower memory usage, but
+Recommended default block size is 128 KB for SSD and 1 MB for HDD. In fact,
+it's possible to use 1 MB for SSD too - it will lower memory usage, but
 may increase average WA and reduce linear performance.
 
 OSD memory usage is roughly (SIZE / BLOCK * 68 bytes) which is roughly
 544 MB per 1 TB of used disk space with the default 128 KB block size.
+With 1 MB it's 8 times lower.
 
 ## bitmap_granularity
 

docs/config/layout-cluster.ru.md

@@ -33,14 +33,14 @@ OSD) могут сосуществовать в одном кластере Vita
 настроек, влияет на потребление памяти, объём избыточной записи (write
 amplification) и эффективность распределения нагрузки по OSD.
 
-Рекомендуемые по умолчанию размеры блока - 128 килобайт для SSD и 4
-мегабайта для HDD. В принципе, для SSD можно тоже использовать 4 мегабайта,
+Рекомендуемые по умолчанию размеры блока - 128 килобайт для SSD и 1 мегабайт
+для HDD. В принципе, для SSD можно тоже использовать блок размером 1 мегабайт,
 это понизит использование памяти, но ухудшит распределение нагрузки и в
 среднем увеличит WA.
 
 Потребление памяти OSD составляет примерно (РАЗМЕР / БЛОК * 68 байт),
 т.е. примерно 544 МБ памяти на 1 ТБ занятого места на диске при
-стандартном 128 КБ блоке.
+стандартном 128 КБ блоке. При 1 МБ блоке памяти нужно в 8 раз меньше.
 
 ## bitmap_granularity
 

docs/config/src/include.js

@@ -0,0 +1,145 @@
+#!/usr/bin/nodejs
+
+const fsp = require('fs').promises;
+
+run(process.argv).catch(console.error);
+
+async function run(argv)
+{
+    if (argv.length < 3)
+    {
+        console.log('Markdown preprocessor\nUSAGE: ./include.js file.md');
+        return;
+    }
+    const index_file = await fsp.realpath(argv[2]);
+    const re = /(\{\{[\s\S]*?\}\}|\[[^\]]+\]\([^\)]+\)|(?:^|\n)#[^\n]+)/;
+    let text = await fsp.readFile(index_file, { encoding: 'utf-8' });
+    text = text.split(re);
+    let included = {};
+    let heading = 0, heading_name = '', m;
+    for (let i = 0; i < text.length; i++)
+    {
+        if (text[i].substr(0, 2) == '{{')
+        {
+            // Inclusion
+            let incfile = text[i].substr(2, text[i].length-4);
+            let section = null;
+            let indent = heading;
+            incfile = incfile.replace(/\s*\|\s*indent\s*=\s*(-?\d+)\s*$/, (m, m1) => { indent = parseInt(m1); return ''; });
+            incfile = incfile.replace(/\s*#\s*([^#]+)$/, (m, m1) => { section = m1; return ''; });
+            let inc_heading = section;
+            incfile = rel2abs(index_file, incfile);
+            let inc = await fsp.readFile(incfile, { encoding: 'utf-8' });
+            inc = inc.trim().replace(/^[\s\S]+?\n#/, '#'); // remove until the first header
+            inc = inc.split(re);
+            const indent_str = new Array(indent+1).join('#');
+            let section_start = -1, section_end = -1;
+            for (let j = 0; j < inc.length; j++)
+            {
+                if ((m = /^(\n?)(#+\s*)([\s\S]+)$/.exec(inc[j])))
+                {
+                    if (!inc_heading)
+                    {
+                        inc_heading = m[3].trim();
+                    }
+                    if (section)
+                    {
+                        if (m[3].trim() == section)
+                            section_start = j;
+                        else if (section_start >= 0)
+                        {
+                            section_end = j;
+                            break;
+                        }
+                    }
+                    inc[j] = m[1] + indent_str + m[2] + m[3];
+                }
+                else if ((m = /^(\[[^\]]+\]\()([^\)]+)(\))$/.exec(inc[j])) && !/^https?:(\/\/)|^#/.exec(m[2]))
+                {
+                    const abs_m2 = rel2abs(incfile, m[2]);
+                    const rel_m = abs2rel(__filename, abs_m2);
+                    if (rel_m.substr(0, 9) == '../../../') // outside docs
+                        inc[j] = m[1] + 'https://git.yourcmc.ru/vitalif/vitastor/src/branch/master/'+rel2abs('docs/config/src/include.js', rel_m) + m[3];
+                    else
+                        inc[j] = m[1] + abs_m2 + m[3];
+                }
+            }
+            if (section)
+            {
+                inc = section_start >= 0 ? inc.slice(section_start, section_end < 0 ? inc.length : section_end) : [];
+            }
+            if (inc.length)
+            {
+                if (!inc_heading)
+                    inc_heading = heading_name||'';
+                included[incfile+(section ? '#'+section : '')] = '#'+inc_heading.toLowerCase().replace(/\P{L}+/ug, '-').replace(/^-|-$/g, '');
+                inc[0] = inc[0].replace(/^\s+/, '');
+                inc[inc.length-1] = inc[inc.length-1].replace(/\s+$/, '');
+            }
+            text.splice(i, 1, ...inc);
+            i = i + inc.length - 1;
+        }
+        else if ((m = /^\n?(#+)\s*([\s\S]+)$/.exec(text[i])))
+        {
+            // Heading
+            heading = m[1].length;
+            heading_name = m[2].trim();
+        }
+    }
+    for (let i = 0; i < text.length; i++)
+    {
+        if ((m = /^(\[[^\]]+\]\()([^\)]+)(\))$/.exec(text[i])) && !/^https?:(\/\/)|^#/.exec(m[2]))
+        {
+            const p = m[2].indexOf('#');
+            if (included[m[2]])
+            {
+                text[i] = m[1]+included[m[2]]+m[3];
+            }
+            else if (p >= 0 && included[m[2].substr(0, p)])
+            {
+                text[i] = m[1]+m[2].substr(p)+m[3];
+            }
+        }
+    }
+    console.log(text.join(''));
+}
+
+function rel2abs(ref, rel)
+{
+    rel = [ ...ref.replace(/^(.*)\/[^\/]+$/, '$1').split(/\/+/), ...rel.split(/\/+/) ];
+    return killdots(rel).join('/');
+}
+
+function abs2rel(ref, abs)
+{
+    ref = ref.split(/\/+/);
+    abs = abs.split(/\/+/);
+    while (ref.length > 1 && ref[0] == abs[0])
+    {
+        ref.shift();
+        abs.shift();
+    }
+    for (let i = 1; i < ref.length; i++)
+    {
+        abs.unshift('..');
+    }
+    return killdots(abs).join('/');
+}
+
+function killdots(rel)
+{
+    for (let i = 0; i < rel.length; i++)
+    {
+        if (rel[i] == '.')
+        {
+            rel.splice(i, 1);
+            i--;
+        }
+        else if (i >= 1 && rel[i] == '..' && rel[i-1] != '..')
+        {
+            rel.splice(i-1, 2);
+            i -= 2;
+        }
+    }
+    return rel;
+}

docs/config/src/included.en.md

@@ -0,0 +1,65 @@
+# Vitastor
+
+{{../../../README.md#The Idea}}
+
+{{../../../README.md#Talks and presentations}}
+
+{{../../intro/features.en.md}}
+
+{{../../intro/quickstart.en.md}}
+
+{{../../intro/architecture.en.md}}
+
+## Installation
+
+{{../../installation/packages.en.md}}
+
+{{../../installation/proxmox.en.md}}
+
+{{../../installation/openstack.en.md}}
+
+{{../../installation/kubernetes.en.md}}
+
+{{../../installation/source.en.md}}
+
+{{../../config.en.md|indent=1}}
+
+{{../../config/common.en.md|indent=2}}
+
+{{../../config/network.en.md|indent=2}}
+
+{{../../config/layout-cluster.en.md|indent=2}}
+
+{{../../config/layout-osd.en.md|indent=2}}
+
+{{../../config/osd.en.md|indent=2}}
+
+{{../../config/monitor.en.md|indent=2}}
+
+{{../../config/pool.en.md|indent=2}}
+
+{{../../config/inode.en.md|indent=2}}
+
+## Usage
+
+{{../../usage/cli.en.md}}
+
+{{../../usage/disk.en.md}}
+
+{{../../usage/fio.en.md}}
+
+{{../../usage/nbd.en.md}}
+
+{{../../usage/qemu.en.md}}
+
+{{../../usage/nfs.en.md}}
+
+## Performance
+
+{{../../performance/understanding.en.md}}
+
+{{../../performance/theoretical.en.md}}
+
+{{../../performance/comparison1.en.md}}
+
+{{../../intro/author.en.md|indent=1}}

docs/config/src/included.ru.md

@@ -0,0 +1,65 @@
+# Vitastor
+
+{{../../../README-ru.md#Идея|indent=0}}
+
+{{../../../README-ru.md#Презентации и записи докладов|indent=0}}
+
+{{../../intro/features.ru.md}}
+
+{{../../intro/quickstart.ru.md}}
+
+{{../../intro/architecture.ru.md}}
+
+## Установка
+
+{{../../installation/packages.ru.md}}
+
+{{../../installation/proxmox.ru.md}}
+
+{{../../installation/openstack.ru.md}}
+
+{{../../installation/kubernetes.ru.md}}
+
+{{../../installation/source.ru.md}}
+
+{{../../config.ru.md|indent=1}}
+
+{{../../config/common.ru.md|indent=2}}
+
+{{../../config/network.ru.md|indent=2}}
+
+{{../../config/layout-cluster.ru.md|indent=2}}
+
+{{../../config/layout-osd.ru.md|indent=2}}
+
+{{../../config/osd.ru.md|indent=2}}
+
+{{../../config/monitor.ru.md|indent=2}}
+
+{{../../config/pool.ru.md|indent=2}}
+
+{{../../config/inode.ru.md|indent=2}}
+
+## Использование
+
+{{../../usage/cli.ru.md}}
+
+{{../../usage/disk.ru.md}}
+
+{{../../usage/fio.ru.md}}
+
+{{../../usage/nbd.ru.md}}
+
+{{../../usage/qemu.ru.md}}
+
+{{../../usage/nfs.ru.md}}
+
+## Производительность
+
+{{../../performance/understanding.ru.md}}
+
+{{../../performance/theoretical.ru.md}}
+
+{{../../performance/comparison1.ru.md}}
+
+{{../../intro/author.ru.md|indent=1}}

docs/config/src/layout-cluster.yml

@@ -7,26 +7,27 @@
     in Vitastor, affects memory usage, write amplification and I/O load
     distribution effectiveness.
 
-    Recommended default block size is 128 KB for SSD and 4 MB for HDD. In fact,
-    it's possible to use 4 MB for SSD too - it will lower memory usage, but
+    Recommended default block size is 128 KB for SSD and 1 MB for HDD. In fact,
+    it's possible to use 1 MB for SSD too - it will lower memory usage, but
     may increase average WA and reduce linear performance.
 
     OSD memory usage is roughly (SIZE / BLOCK * 68 bytes) which is roughly
     544 MB per 1 TB of used disk space with the default 128 KB block size.
+    With 1 MB it's 8 times lower.
   info_ru: |
     Размер объектов (блоков данных), на которые делятся физические и виртуальные
     диски в Vitastor (в рамках каждого пула). Одна из ключевых на данный момент
     настроек, влияет на потребление памяти, объём избыточной записи (write
     amplification) и эффективность распределения нагрузки по OSD.
 
-    Рекомендуемые по умолчанию размеры блока - 128 килобайт для SSD и 4
-    мегабайта для HDD. В принципе, для SSD можно тоже использовать 4 мегабайта,
+    Рекомендуемые по умолчанию размеры блока - 128 килобайт для SSD и 1 мегабайт
+    для HDD. В принципе, для SSD можно тоже использовать блок размером 1 мегабайт,
     это понизит использование памяти, но ухудшит распределение нагрузки и в
     среднем увеличит WA.
 
     Потребление памяти OSD составляет примерно (РАЗМЕР / БЛОК * 68 байт),
     т.е. примерно 544 МБ памяти на 1 ТБ занятого места на диске при
-    стандартном 128 КБ блоке.
+    стандартном 128 КБ блоке. При 1 МБ блоке памяти нужно в 8 раз меньше.
 - name: bitmap_granularity
   type: int
   default: 4096

docs/installation/kubernetes.ru.md

@@ -8,13 +8,13 @@
 
 У Vitastor есть CSI-плагин для Kubernetes, поддерживающий RWO, а также блочные RWX, 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](../csi/deploy/001-csi-config-map.yaml),
-настройте StorageClass в [csi/deploy/009-storage-class.yaml](../csi/deploy/009-storage-class.yaml)
+Для установки возьмите манифесты из директории [csi/deploy/](../../csi/deploy/), поместите
+вашу конфигурацию подключения к Vitastor в [csi/deploy/001-csi-config-map.yaml](../../csi/deploy/001-csi-config-map.yaml),
+настройте StorageClass в [csi/deploy/009-storage-class.yaml](../../csi/deploy/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).
+После этого вы сможете создавать PersistentVolume. Пример смотрите в файле [csi/deploy/example-pvc.yaml](../../csi/deploy/example-pvc.yaml).

docs/installation/openstack.en.md

@@ -36,5 +36,5 @@ vitastor_pool_id = 1
 image_upload_use_cinder_backend = True
 ```
 
-To put Glance images in Vitastor, use [https://docs.openstack.org/cinder/pike/admin/blockstorage-volume-backed-image.html](volume-backed images),
+To put Glance images in Vitastor, use [volume-backed images](https://docs.openstack.org/cinder/pike/admin/blockstorage-volume-backed-image.html),
 although the support has not been verified yet.

docs/installation/openstack.ru.md

@@ -36,5 +36,5 @@ image_upload_use_cinder_backend = True
 ```
 
 Чтобы помещать в Vitastor Glance-образы, нужно использовать
-[https://docs.openstack.org/cinder/pike/admin/blockstorage-volume-backed-image.html](образы на основе томов Cinder),
+[образы на основе томов Cinder](https://docs.openstack.org/cinder/pike/admin/blockstorage-volume-backed-image.html),
 однако, поддержка этой функции ещё не проверялась.

docs/installation/proxmox.en.md

@@ -6,10 +6,10 @@
 
 # Proxmox VE
 
-To enable Vitastor support in Proxmox Virtual Environment (6.4-7.4 are supported):
+To enable Vitastor support in Proxmox Virtual Environment (6.4-8.0 are supported):
 
 - Add the corresponding Vitastor Debian repository into sources.list on Proxmox hosts:
-  buster for 6.4, bullseye for 7.4, pve7.1 for 7.1, pve7.2 for 7.2, pve7.3 for 7.3
+  bookworm for 8.0, bullseye for 7.4, pve7.3 for 7.3, pve7.2 for 7.2, pve7.1 for 7.1, buster for 6.4
 - Install vitastor-client, pve-qemu-kvm, pve-storage-vitastor (* or see note) packages from Vitastor repository
 - Define storage in `/etc/pve/storage.cfg` (see below)
 - Block network access from VMs to Vitastor network (to OSDs and etcd),
@@ -35,5 +35,5 @@ vitastor: vitastor
     vitastor_nbd 0
 ```
 
-\* Note: you can also manually copy [patches/VitastorPlugin.pm](patches/VitastorPlugin.pm) to Proxmox hosts
+\* Note: you can also manually copy [patches/VitastorPlugin.pm](../../patches/VitastorPlugin.pm) to Proxmox hosts
 as `/usr/share/perl5/PVE/Storage/Custom/VitastorPlugin.pm` instead of installing pve-storage-vitastor.

docs/installation/proxmox.ru.md

@@ -1,15 +1,15 @@
-[Документация](../../README-ru.md#документация) → Установка → Proxmox
+[Документация](../../README-ru.md#документация) → Установка → Proxmox VE
 
 -----
 
 [Read in English](proxmox.en.md)
 
-# Proxmox
+# Proxmox VE
 
-Чтобы подключить Vitastor к Proxmox Virtual Environment (поддерживаются версии 6.4-7.4):
+Чтобы подключить Vitastor к Proxmox Virtual Environment (поддерживаются версии 6.4-8.0):
 
 - Добавьте соответствующий Debian-репозиторий Vitastor в sources.list на хостах Proxmox:
-  buster для 6.4, bullseye для 7.4, pve7.1 для 7.1, pve7.2 для 7.2, pve7.3 для 7.3
+  bookworm для 8.0, bullseye для 7.4, pve7.3 для 7.3, pve7.2 для 7.2, pve7.1 для 7.1, buster для 6.4
 - Установите пакеты vitastor-client, pve-qemu-kvm, pve-storage-vitastor (* или см. сноску) из репозитория Vitastor
 - Определите тип хранилища в `/etc/pve/storage.cfg` (см. ниже)
 - Обязательно заблокируйте доступ от виртуальных машин к сети Vitastor (OSD и etcd), т.к. Vitastor (пока) не поддерживает аутентификацию
@@ -35,5 +35,5 @@ vitastor: vitastor
 ```
 
 \* Примечание: вместо установки пакета pve-storage-vitastor вы можете вручную скопировать файл
-[patches/VitastorPlugin.pm](patches/VitastorPlugin.pm) на хосты Proxmox как
+[patches/VitastorPlugin.pm](../../patches/VitastorPlugin.pm) на хосты Proxmox как
 `/usr/share/perl5/PVE/Storage/Custom/VitastorPlugin.pm`.

docs/installation/source.en.md

@@ -21,7 +21,7 @@
 
 ## Basic instructions
 
-Download source, for example using git: `git clone --recurse-submodules https://yourcmc.ru/git/vitalif/vitastor/`
+Download source, for example using git: `git clone --recurse-submodules https://git.yourcmc.ru/vitalif/vitastor/`
 
 Get `fio` source and symlink it into `<vitastor>/fio`. If you don't want to build fio engine,
 you can disable it by passing `-DWITH_FIO=no` to cmake.
@@ -41,7 +41,7 @@ It's recommended to build the QEMU driver (qemu_driver.c) in-tree, as a part of
 QEMU build process. To do that:
 - Install vitastor client library headers (from source or from vitastor-client-dev package)
 - Take a corresponding patch from `patches/qemu-*-vitastor.patch` and apply it to QEMU source
-- Copy `src/qemu_driver.c` to QEMU source directory as `block/block-vitastor.c`
+- Copy `src/qemu_driver.c` to QEMU source directory as `block/vitastor.c`
 - Build QEMU as usual
 
 But it is also possible to build it out-of-tree. To do that:

docs/installation/source.ru.md

@@ -21,7 +21,7 @@
 
 ## Базовая инструкция
 
-Скачайте исходные коды, например, из git: `git clone --recurse-submodules https://yourcmc.ru/git/vitalif/vitastor/`
+Скачайте исходные коды, например, из git: `git clone --recurse-submodules https://git.yourcmc.ru/vitalif/vitastor/`
 
 Скачайте исходные коды пакета `fio`, распакуйте их и создайте символическую ссылку на них
 в директории исходников Vitastor: `<vitastor>/fio`. Либо, если вы не хотите собирать плагин fio,
@@ -41,7 +41,7 @@ cmake .. && make -j8 install
 Драйвер QEMU (qemu_driver.c) рекомендуется собирать вместе с самим QEMU. Для этого:
 - Установите заголовки клиентской библиотеки Vitastor (из исходников или из пакета vitastor-client-dev)
 - Возьмите соответствующий патч из `patches/qemu-*-vitastor.patch` и примените его к исходникам QEMU
-- Скопируйте [src/qemu_driver.c](../../src/qemu_driver.c) в директорию исходников QEMU как `block/block-vitastor.c`
+- Скопируйте [src/qemu_driver.c](../../src/qemu_driver.c) в директорию исходников QEMU как `block/vitastor.c`
 - Соберите QEMU как обычно
 
 Однако в целях отладки драйвер также можно собирать отдельно от QEMU. Для этого:
@@ -60,7 +60,7 @@ cmake .. && make -j8 install
       * 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
 - Сконфигурируйте cmake Vitastor с `WITH_QEMU=yes` (`cmake .. -DWITH_QEMU=yes`) и, если вы
-  используете RHEL-подобый дистрибутив, также с `QEMU_PLUGINDIR=qemu-kvm`.
+  используете RHEL-подобный дистрибутив, также с `QEMU_PLUGINDIR=qemu-kvm`.
 - После этого в процессе сборки Vitastor также будет собираться подходящий для вашей
   версии QEMU `block-vitastor.so`.
 - Таким образом можно использовать драйвер даже с немодифицированным QEMU, но в этом случае

docs/intro/architecture.en.md

@@ -44,7 +44,7 @@
   depends linearly on drive capacity and data store block size which is 128 KB by default.
   With 128 KB blocks metadata takes around 512 MB per 1 TB (which is still less than Ceph wants).
   Journal is also kept in memory by default, but in SSD-only clusters it's only 32 MB, and in SSD+HDD
-  clusters, where it's beneficial to increase it, [inmemory_journal](docs/config/osd.en.md#inmemory_journal) can be disabled.
+  clusters, where it's beneficial to increase it, [inmemory_journal](../config/osd.en.md#inmemory_journal) can be disabled.
 - 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.

docs/intro/architecture.ru.md

@@ -156,7 +156,7 @@
   блока хранилища (block_size, по умолчанию 128 КБ). С 128 КБ блоком потребление памяти
   составляет примерно 512 МБ на 1 ТБ данных. Журналы по умолчанию тоже хранятся в памяти,
   но в SSD-кластерах нужный размер журнала составляет всего 32 МБ, а в гибридных (SSD+HDD)
-  кластерах, в которых есть смысл делать журналы больше, можно отключить [inmemory_journal](../docs/config/osd.ru.md#inmemory_journal).
+  кластерах, в которых есть смысл делать журналы больше, можно отключить [inmemory_journal](../config/osd.ru.md#inmemory_journal).
 - 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.

docs/intro/features.en.md

@@ -35,7 +35,7 @@
 
 - [Debian and CentOS packages](../installation/packages.en.md)
 - [Image management CLI (vitastor-cli)](../usage/cli.en.md)
-- [Disk management CLI (vitastor-disk)](docs/usage/disk.en.md)
+- [Disk management CLI (vitastor-disk)](../usage/disk.en.md)
 - Generic user-space client library
 - [Native QEMU driver](../usage/qemu.en.md)
 - [Loadable fio engine for benchmarks](../usage/fio.en.md)

docs/intro/features.ru.md

@@ -13,7 +13,7 @@
 ## Серверные функции
 
 - Базовая часть - надёжное кластерное блочное хранилище без единой точки отказа
-- [Производительность](../comparison1.ru.md) ;-D
+- [Производительность](../performance/comparison1.ru.md) ;-D
 - [Несколько схем отказоустойчивости](../config/pool.ru.md#scheme): репликация, XOR n+1 (1 диск чётности), коды коррекции ошибок
   Рида-Соломона на основе библиотек jerasure и ISA-L с любым числом дисков данных и чётности в группе
 - Конфигурация через простые человекочитаемые JSON-структуры в etcd
@@ -37,7 +37,7 @@
 
 - [Пакеты для Debian и CentOS](../installation/packages.ru.md)
 - [Консольный интерфейс управления образами (vitastor-cli)](../usage/cli.ru.md)
-- [Инструмент управления дисками (vitastor-disk)](docs/usage/disk.ru.md)
+- [Инструмент управления дисками (vitastor-disk)](../usage/disk.ru.md)
 - Общая пользовательская клиентская библиотека для работы с кластером
 - [Драйвер диска для QEMU](../usage/qemu.ru.md)
 - [Драйвер диска для утилиты тестирования производительности fio](../usage/fio.ru.md)

docs/intro/quickstart.en.md

@@ -7,6 +7,7 @@
 # Quick Start
 
 - [Preparation](#preparation)
+- [Recommended drives](#recommended-drives)
 - [Configure monitors](#configure-monitors)
 - [Configure OSDs](#configure-osds)
 - [Create a pool](#create-a-pool)
@@ -19,10 +20,20 @@
 - 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. Read more about capacitors
   [here](../config/layout-cluster.en.md#immediate_commit).
+- If you want to use HDDs, get modern HDDs with Media Cache or SSD Cache: HGST Ultrastar,
+  Toshiba MG08, Seagate EXOS or something similar. If your drives don't have such cache then
+  you also need small SSDs for journal and metadata (even 2 GB per 1 TB of HDD space is enough).
 - Get a fast network (at least 10 Gbit/s). Something like Mellanox ConnectX-4 with RoCEv2 is ideal.
 - Disable CPU powersaving: `cpupower idle-set -D 0 && cpupower frequency-set -g performance`.
 - [Install Vitastor packages](../installation/packages.en.md).
 
+## Recommended drives
+
+- SATA SSD: Micron 5100/5200/5300/5400, Samsung PM863/PM883/PM893, Intel D3-S4510/4520/4610/4620, Kingston DC500M
+- NVMe: Micron 9100/9200/9300/9400, Micron 7300/7450, Samsung PM983/PM9A3, Samsung PM1723/1735/1743,
+  Intel DC-P3700/P4500/P4600, Intel D7-P5500/P5600, Intel Optane, Kingston DC1000B/DC1500M
+- HDD: HGST Ultrastar, Toshiba MG06/MG07/MG08, Seagate EXOS
+
 ## Configure monitors
 
 On the monitor hosts:
@@ -45,9 +56,10 @@ On the monitor hosts:
   }
   ```
 - Initialize OSDs:
-  - SSD-only: `vitastor-disk prepare /dev/sdXXX [/dev/sdYYY ...]`. You can add
-    `--disable_data_fsync off` to leave disk cache enabled if you use desktop
-    SSDs without capacitors.
+  - SSD-only or HDD-only: `vitastor-disk prepare /dev/sdXXX [/dev/sdYYY ...]`.
+    Add `--disable_data_fsync off` to leave disk write cache enabled if you use
+    desktop SSDs without capacitors. Do NOT add `--disable_data_fsync off` if you
+    use HDDs or SSD+HDD.
   - Hybrid, SSD+HDD: `vitastor-disk prepare --hybrid /dev/sdXXX [/dev/sdYYY ...]`.
     Pass all your devices (HDD and SSD) to this script — it will partition disks and initialize journals on its own.
     This script skips HDDs which are already partitioned so if you want to use non-empty disks for

docs/intro/quickstart.ru.md

@@ -7,6 +7,7 @@
 # Быстрый старт
 
 - [Подготовка](#подготовка)
+- [Рекомендуемые диски](#рекомендуемые-диски)
 - [Настройте мониторы](#настройте-мониторы)
 - [Настройте OSD](#настройте-osd)
 - [Создайте пул](#создайте-пул)
@@ -19,10 +20,20 @@
 - Возьмите серверы с SSD (SATA или NVMe), желательно с конденсаторами (серверные SSD). Можно
   использовать и десктопные SSD, включив режим отложенного fsync, но производительность будет хуже.
   О конденсаторах читайте [здесь](../config/layout-cluster.ru.md#immediate_commit).
+- Если хотите использовать HDD, берите современные модели с Media или SSD кэшем - HGST Ultrastar,
+  Toshiba MG08, Seagate EXOS или что-то похожее. Если такого кэша у ваших дисков нет,
+  обязательно возьмите SSD под метаданные и журнал (маленькие, буквально 2 ГБ на 1 ТБ HDD-места).
 - Возьмите быструю сеть, минимум 10 гбит/с. Идеал - что-то вроде Mellanox ConnectX-4 с RoCEv2.
 - Для лучшей производительности отключите энергосбережение CPU: `cpupower idle-set -D 0 && cpupower frequency-set -g performance`.
 - [Установите пакеты Vitastor](../installation/packages.ru.md).
 
+## Рекомендуемые диски
+
+- SATA SSD: Micron 5100/5200/5300/5400, Samsung PM863/PM883/PM893, Intel D3-S4510/4520/4610/4620, Kingston DC500M
+- NVMe: Micron 9100/9200/9300/9400, Micron 7300/7450, Samsung PM983/PM9A3, Samsung PM1723/1735/1743,
+  Intel DC-P3700/P4500/P4600, Intel D7-P5500/P5600, Intel Optane, Kingston DC1000B/DC1500M
+- HDD: HGST Ultrastar, Toshiba MG06/MG07/MG08, Seagate EXOS
+
 ## Настройте мониторы
 
 На хостах, выделенных под мониторы:
@@ -45,9 +56,10 @@
   }
   ```
 - Инициализуйте OSD:
-  - SSD: `vitastor-disk prepare /dev/sdXXX [/dev/sdYYY ...]`. Если вы используете
-    десктопные SSD без конденсаторов, можете оставить кэш включённым, добавив
-    опцию `--disable_data_fsync off`.
+  - Только SSD или только HDD: `vitastor-disk prepare /dev/sdXXX [/dev/sdYYY ...]`.
+    Если вы используете десктопные SSD без конденсаторов, добавьте опцию `--disable_data_fsync off`,
+    чтобы оставить кэш записи диска включённым. НЕ добавляйте эту опцию, если используете
+    жёсткие диски (HDD).
   - Гибридные, SSD+HDD: `vitastor-disk prepare --hybrid /dev/sdXXX [/dev/sdYYY ...]`.
     Передайте все ваши SSD и HDD скрипту в командной строке подряд, скрипт автоматически выделит
     разделы под журналы на SSD и данные на HDD. Скрипт пропускает HDD, на которых уже есть разделы

docs/usage/nbd.en.md

@@ -13,6 +13,8 @@ remains decent (see an example [here](../performance/comparison1.en.md#vitastor-
 
 Vitastor Kubernetes CSI driver is based on NBD.
 
+See also [VDUSE](qemu.en.md#vduse).
+
 ## Map image
 
 To create a local block device for a Vitastor image run:

docs/usage/nbd.ru.md

@@ -16,6 +16,8 @@ NBD немного снижает производительность из-за
 
 CSI-драйвер Kubernetes Vitastor основан на NBD.
 
+Смотрите также [VDUSE](qemu.ru.md#vduse).
+
 ## Подключить устройство
 
 Чтобы создать локальное блочное устройство для образа, выполните команду:

docs/usage/nfs.ru.md

@@ -29,7 +29,7 @@ vitastor-nfs [--etcd_address ADDR] [ДРУГИЕ ОПЦИИ]
 --bind <IP>       принимать соединения по адресу <IP> (по умолчанию 0.0.0.0 - на всех)
 --nfspath <PATH>  установить путь NFS-экспорта в <PATH> (по умолчанию /)
 --port <PORT>     использовать порт <PORT> для NFS-сервисов (по умолчанию 2049)
---pool <POOL>     исползовать пул <POOL> для новых образов (обязательно, если пул в кластере не один)
+--pool <POOL>     использовать пул <POOL> для новых образов (обязательно, если пул в кластере не один)
 --foreground 1    не уходить в фон после запуска
 ```
 

docs/usage/qemu.en.md

@@ -83,3 +83,43 @@ qemu-img rebase -u -b '' testimg.qcow2
 This can be used for backups. Just note that exporting an image that is currently being written to
 is of course unsafe and doesn't produce a consistent result, so only export snapshots if you do this
 on a live VM.
+
+## VDUSE
+
+Linux kernel, starting with version 5.15, supports a new interface for attaching virtual disks
+to the host - VDUSE (vDPA Device in Userspace). QEMU, starting with 7.2, has support for
+exporting QEMU block devices over this protocol using qemu-storage-daemon.
+
+VDUSE has the same problem as other FUSE-like interfaces in Linux: if a userspace process hangs,
+for example, if it loses connectivity with Vitastor cluster - active processes doing I/O may
+hang in the D state (uninterruptible sleep) and you won't be able to kill them even with kill -9.
+In this case reboot will be the only way to remove VDUSE devices from system.
+
+On the other hand, VDUSE is faster than [NBD](nbd.en.md), so you may prefer to use it if
+performance is important for you. Approximate performance numbers:
+direct fio benchmark - 115000 iops, NBD - 60000 iops, VDUSE - 90000 iops.
+
+To try VDUSE you need at least Linux 5.15, built with VDUSE support
+(CONFIG_VIRTIO_VDPA=m and CONFIG_VDPA_USER=m). Debian Linux kernels have these options
+disabled by now, so if you want to try it on Debian, use a kernel from Ubuntu
+[kernel-ppa/mainline](https://kernel.ubuntu.com/~kernel-ppa/mainline/) or Proxmox.
+
+Commands to attach Vitastor image as a VDUSE device:
+
+```
+modprobe vduse virtio-vdpa
+qemu-storage-daemon --daemonize --blockdev '{"node-name":"test1","driver":"vitastor",\
+  "etcd-host":"192.168.7.2:2379/v3","image":"testosd1","cache":{"direct":true,"no-flush":false},"discard":"unmap"}' \
+  --export vduse-blk,id=test1,node-name=test1,name=test1,num-queues=16,queue-size=128,writable=true
+vdpa dev add name test1 mgmtdev vduse
+```
+
+After running these commands /dev/vda device will appear in the system and you'll be able to
+use it as a normal disk.
+
+To remove the device:
+
+```
+vdpa dev del test1
+kill <qemu-storage-daemon_process_PID>
+```

docs/usage/qemu.ru.md

@@ -87,3 +87,43 @@ qemu-img rebase -u -b '' testimg.qcow2
 Это можно использовать для резервного копирования. Только помните, что экспортировать образ, в который
 в то же время идёт запись, небезопасно - результат чтения не будет целостным. Так что если вы работаете
 с активными виртуальными машинами, экспортируйте только их снимки, но не сам образ.
+
+## VDUSE
+
+В Linux, начиная с версии ядра 5.15, доступен новый интерфейс для подключения виртуальных дисков
+к системе - VDUSE (vDPA Device in Userspace), а в QEMU, начиная с версии 7.2, есть поддержка
+экспорта блочных устройств QEMU по этому протоколу через qemu-storage-daemon.
+
+VDUSE страдает общей проблемой FUSE-подобных интерфейсов в Linux: если пользовательский процесс
+подвиснет, например, если будет потеряна связь с кластером Vitastor - читающие/пишущие в кластер
+процессы могут "залипнуть" в состоянии D (непрерываемый сон) и их будет невозможно убить даже
+через kill -9. В этом случае удалить из системы устройство можно только перезагрузившись.
+
+С другой стороны, VDUSE быстрее по сравнению с [NBD](nbd.ru.md), поэтому его может
+быть предпочтительно использовать там, где производительность важнее. Порядок показателей:
+прямое тестирование через fio - 115000 iops, NBD - 60000 iops, VDUSE - 90000 iops.
+
+Чтобы использовать VDUSE, вам нужно ядро Linux версии хотя бы 5.15, собранное с поддержкой
+VDUSE (CONFIG_VIRTIO_VDPA=m и CONFIG_VDPA_USER=m). В ядрах в Debian Linux поддержка пока
+отключена - если хотите попробовать эту функцию на Debian, поставьте ядро из Ubuntu
+[kernel-ppa/mainline](https://kernel.ubuntu.com/~kernel-ppa/mainline/) или из Proxmox.
+
+Команды для подключения виртуального диска через VDUSE:
+
+```
+modprobe vduse virtio-vdpa
+qemu-storage-daemon --daemonize --blockdev '{"node-name":"test1","driver":"vitastor",\
+  "etcd-host":"192.168.7.2:2379/v3","image":"testosd1","cache":{"direct":true,"no-flush":false},"discard":"unmap"}' \
+  --export vduse-blk,id=test1,node-name=test1,name=test1,num-queues=16,queue-size=128,writable=true
+vdpa dev add name test1 mgmtdev vduse
+```
+
+После этого в системе появится устройство /dev/vda, которое можно будет использовать как
+обычный диск.
+
+Для удаления устройства из системы:
+
+```
+vdpa dev del test1
+kill <PID_процесса_qemu-storage-daemon>
+```

mon/make-etcd

@@ -63,8 +63,9 @@ Wants=network-online.target local-fs.target time-sync.target
 
 [Service]
 Restart=always
-ExecStart=/usr/local/bin/etcd -name etcd${num} --data-dir /var/lib/etcd${num}.etcd \\
-    --advertise-client-urls http://${etcds[num]}:2379 --listen-client-urls http://${etcds[num]}:2379 \\
+Environment=GOGC=50
+ExecStart=etcd -name etcd${num} --data-dir /var/lib/etcd${num}.etcd \\
+    --snapshot-count 10000 --advertise-client-urls http://${etcds[num]}:2379 --listen-client-urls http://${etcds[num]}:2379 \\
     --initial-advertise-peer-urls http://${etcds[num]}:2380 --listen-peer-urls http://${etcds[num]}:2380 \\
     --initial-cluster-token vitastor-etcd-1 --initial-cluster ${etcd_cluster} \\
     --initial-cluster-state new --max-txn-ops=100000 --max-request-bytes=104857600 \\

mon/mon.js

@@ -1608,7 +1608,7 @@ class Mon
                 }
             }
         }
-        return inode_stats;
+        return { inode_stats, seen_pools };
     }
 
     serialize_bigints(obj)
@@ -1634,7 +1634,7 @@ class Mon
         const timestamp = Date.now();
         const { object_counts, object_bytes } = this.sum_object_counts();
         let stats = this.sum_op_stats(timestamp, this.prev_stats);
-        let inode_stats = this.sum_inode_stats(
+        let { inode_stats, seen_pools } = this.sum_inode_stats(
             this.prev_stats ? this.prev_stats.inode_stats : null,
             timestamp, this.prev_stats ? this.prev_stats.timestamp : null
         );
@@ -1669,12 +1669,22 @@ class Mon
         }
         for (const pool_id in this.state.pool.stats)
         {
-            const pool_stats = { ...this.state.pool.stats[pool_id] };
-            this.serialize_bigints(pool_stats);
-            txn.push({ requestPut: {
-                key: b64(this.etcd_prefix+'/pool/stats/'+pool_id),
-                value: b64(JSON.stringify(pool_stats)),
-            } });
+            if (!seen_pools[pool_id])
+            {
+                txn.push({ requestDeleteRange: {
+                    key: b64(this.etcd_prefix+'/pool/stats/'+pool_id),
+                } });
+                delete this.state.pool.stats[pool_id];
+            }
+            else
+            {
+                const pool_stats = { ...this.state.pool.stats[pool_id] };
+                this.serialize_bigints(pool_stats);
+                txn.push({ requestPut: {
+                    key: b64(this.etcd_prefix+'/pool/stats/'+pool_id),
+                    value: b64(JSON.stringify(pool_stats)),
+                } });
+            }
         }
         if (txn.length)
         {

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 = '0.9.2'
+VERSION = '0.9.4'
 
 LOG = logging.getLogger(__name__)
 

patches/pve-qemu-8.0-vitastor.patch

@@ -0,0 +1,190 @@
+diff --git a/block/meson.build b/block/meson.build
+index 382bec0e7d..af6207dbce 100644
+--- a/block/meson.build
++++ b/block/meson.build
+@@ -114,6 +114,7 @@ foreach m : [
+   [libnfs, 'nfs', files('nfs.c')],
+   [libssh, 'ssh', files('ssh.c')],
+   [rbd, 'rbd', files('rbd.c')],
++  [vitastor, 'vitastor', files('vitastor.c')],
+ ]
+   if m[0].found()
+     module_ss = ss.source_set()
+diff --git a/meson.build b/meson.build
+index c44d05a13f..ebedb42843 100644
+--- a/meson.build
++++ b/meson.build
+@@ -1028,6 +1028,26 @@ if not get_option('rbd').auto() or have_block
+   endif
+ endif
+ 
++vitastor = not_found
++if not get_option('vitastor').auto() or have_block
++  libvitastor_client = cc.find_library('vitastor_client', has_headers: ['vitastor_c.h'],
++    required: get_option('vitastor'), kwargs: static_kwargs)
++  if libvitastor_client.found()
++    if cc.links('''
++      #include <vitastor_c.h>
++      int main(void) {
++        vitastor_c_create_qemu(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
++        return 0;
++      }''', dependencies: libvitastor_client)
++      vitastor = declare_dependency(dependencies: libvitastor_client)
++    elif get_option('vitastor').enabled()
++      error('could not link libvitastor_client')
++    else
++      warning('could not link libvitastor_client, disabling')
++    endif
++  endif
++endif
++
+ glusterfs = not_found
+ glusterfs_ftruncate_has_stat = false
+ glusterfs_iocb_has_stat = false
+@@ -1882,6 +1902,7 @@ endif
+ config_host_data.set('CONFIG_OPENGL', opengl.found())
+ config_host_data.set('CONFIG_PROFILER', get_option('profiler'))
+ config_host_data.set('CONFIG_RBD', rbd.found())
++config_host_data.set('CONFIG_VITASTOR', vitastor.found())
+ config_host_data.set('CONFIG_RDMA', rdma.found())
+ config_host_data.set('CONFIG_SDL', sdl.found())
+ config_host_data.set('CONFIG_SDL_IMAGE', sdl_image.found())
+@@ -4020,6 +4041,7 @@ if spice_protocol.found()
+   summary_info += {'  spice server support': spice}
+ endif
+ summary_info += {'rbd support':       rbd}
++summary_info += {'vitastor support':  vitastor}
+ summary_info += {'smartcard support': cacard}
+ summary_info += {'U2F support':       u2f}
+ summary_info += {'libusb':            libusb}
+diff --git a/meson_options.txt b/meson_options.txt
+index fc9447d267..c4ac55c283 100644
+--- a/meson_options.txt
++++ b/meson_options.txt
+@@ -173,6 +173,8 @@ option('lzo', type : 'feature', value : 'auto',
+        description: 'lzo compression support')
+ option('rbd', type : 'feature', value : 'auto',
+        description: 'Ceph block device driver')
++option('vitastor', type : 'feature', value : 'auto',
++       description: 'Vitastor block device driver')
+ option('opengl', type : 'feature', value : 'auto',
+        description: 'OpenGL support')
+ option('rdma', type : 'feature', value : 'auto',
+diff --git a/qapi/block-core.json b/qapi/block-core.json
+index c05ad0c07e..f5eb701604 100644
+--- a/qapi/block-core.json
++++ b/qapi/block-core.json
+@@ -3308,7 +3308,7 @@
+             'raw', 'rbd',
+             { 'name': 'replication', 'if': 'CONFIG_REPLICATION' },
+             'pbs',
+-            'ssh', 'throttle', 'vdi', 'vhdx',
++            'ssh', 'throttle', 'vdi', 'vhdx', 'vitastor',
+             { 'name': 'virtio-blk-vfio-pci', 'if': 'CONFIG_BLKIO' },
+             { 'name': 'virtio-blk-vhost-user', 'if': 'CONFIG_BLKIO' },
+             { 'name': 'virtio-blk-vhost-vdpa', 'if': 'CONFIG_BLKIO' },
+@@ -4338,6 +4338,28 @@
+             '*key-secret': 'str',
+             '*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:
+ #
+@@ -4787,6 +4809,7 @@
+       'throttle':   'BlockdevOptionsThrottle',
+       'vdi':        'BlockdevOptionsGenericFormat',
+       'vhdx':       'BlockdevOptionsGenericFormat',
++      'vitastor':   'BlockdevOptionsVitastor',
+       'virtio-blk-vfio-pci':
+                     { 'type': 'BlockdevOptionsVirtioBlkVfioPci',
+                       'if': 'CONFIG_BLKIO' },
+@@ -5187,6 +5210,17 @@
+             '*cluster-size' :   'size',
+             '*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:
+ #
+@@ -5385,6 +5419,7 @@
+       'ssh':            'BlockdevCreateOptionsSsh',
+       'vdi':            'BlockdevCreateOptionsVdi',
+       'vhdx':           'BlockdevCreateOptionsVhdx',
++      'vitastor':       'BlockdevCreateOptionsVitastor',
+       'vmdk':           'BlockdevCreateOptionsVmdk',
+       'vpc':            'BlockdevCreateOptionsVpc'
+   } }
+diff --git a/scripts/ci/org.centos/stream/8/x86_64/configure b/scripts/ci/org.centos/stream/8/x86_64/configure
+index 6e8983f39c..1b0b9fcf3e 100755
+--- a/scripts/ci/org.centos/stream/8/x86_64/configure
++++ b/scripts/ci/org.centos/stream/8/x86_64/configure
+@@ -32,7 +32,7 @@
+ --with-git=meson \
+ --with-git-submodules=update \
+ --target-list="x86_64-softmmu" \
+---block-drv-rw-whitelist="qcow2,raw,file,host_device,nbd,iscsi,rbd,blkdebug,luks,null-co,nvme,copy-on-read,throttle,gluster" \
++--block-drv-rw-whitelist="qcow2,raw,file,host_device,nbd,iscsi,rbd,vitastor,blkdebug,luks,null-co,nvme,copy-on-read,throttle,gluster" \
+ --audio-drv-list="" \
+ --block-drv-ro-whitelist="vmdk,vhdx,vpc,https,ssh" \
+ --with-coroutine=ucontext \
+@@ -179,6 +179,7 @@
+ --enable-opengl \
+ --enable-pie \
+ --enable-rbd \
++--enable-vitastor \
+ --enable-rdma \
+ --enable-seccomp \
+ --enable-snappy \
+diff --git a/scripts/meson-buildoptions.sh b/scripts/meson-buildoptions.sh
+index 009fab1515..95914e6ebc 100644
+--- a/scripts/meson-buildoptions.sh
++++ b/scripts/meson-buildoptions.sh
+@@ -144,6 +144,7 @@ meson_options_help() {
+   printf "%s\n" '  qed             qed image format support'
+   printf "%s\n" '  qga-vss         build QGA VSS support (broken with MinGW)'
+   printf "%s\n" '  rbd             Ceph block device driver'
++  printf "%s\n" '  vitastor        Vitastor block device driver'
+   printf "%s\n" '  rdma            Enable RDMA-based migration'
+   printf "%s\n" '  replication     replication support'
+   printf "%s\n" '  sdl             SDL user interface'
+@@ -392,6 +393,8 @@ _meson_option_parse() {
+     --disable-qom-cast-debug) printf "%s" -Dqom_cast_debug=false ;;
+     --enable-rbd) printf "%s" -Drbd=enabled ;;
+     --disable-rbd) printf "%s" -Drbd=disabled ;;
++    --enable-vitastor) printf "%s" -Dvitastor=enabled ;;
++    --disable-vitastor) printf "%s" -Dvitastor=disabled ;;
+     --enable-rdma) printf "%s" -Drdma=enabled ;;
+     --disable-rdma) printf "%s" -Drdma=disabled ;;
+     --enable-replication) printf "%s" -Dreplication=enabled ;;

patches/qemu-2.12-vitastor.patch

@@ -0,0 +1,176 @@
+diff --git a/block/Makefile.objs b/block/Makefile.objs
+index d644bac60a..e404236291 100644
+--- a/block/Makefile.objs
++++ b/block/Makefile.objs
+@@ -19,6 +19,7 @@ block-obj-$(if $(CONFIG_LIBISCSI),y,n) += iscsi-opts.o
+ block-obj-$(CONFIG_LIBNFS) += nfs.o
+ block-obj-$(CONFIG_CURL) += curl.o
+ block-obj-$(CONFIG_RBD) += rbd.o
++block-obj-$(CONFIG_VITASTOR) += vitastor.o
+ block-obj-$(CONFIG_GLUSTERFS) += gluster.o
+ block-obj-$(CONFIG_VXHS) += vxhs.o
+ block-obj-$(CONFIG_LIBSSH2) += ssh.o
+@@ -39,6 +40,8 @@ curl.o-cflags      := $(CURL_CFLAGS)
+ curl.o-libs        := $(CURL_LIBS)
+ rbd.o-cflags       := $(RBD_CFLAGS)
+ rbd.o-libs         := $(RBD_LIBS)
++vitastor.o-cflags  := $(VITASTOR_CFLAGS)
++vitastor.o-libs    := $(VITASTOR_LIBS)
+ gluster.o-cflags   := $(GLUSTERFS_CFLAGS)
+ gluster.o-libs     := $(GLUSTERFS_LIBS)
+ vxhs.o-libs        := $(VXHS_LIBS)
+diff --git a/configure b/configure
+index 0a19b033bc..58b7fbf24c 100755
+--- a/configure
++++ b/configure
+@@ -398,6 +398,7 @@ trace_backends="log"
+ trace_file="trace"
+ spice=""
+ rbd=""
++vitastor=""
+ smartcard=""
+ libusb=""
+ usb_redir=""
+@@ -1213,6 +1214,10 @@ for opt do
+   ;;
+   --enable-rbd) rbd="yes"
+   ;;
++  --disable-vitastor) vitastor="no"
++  ;;
++  --enable-vitastor) vitastor="yes"
++  ;;
+   --disable-xfsctl) xfs="no"
+   ;;
+   --enable-xfsctl) xfs="yes"
+@@ -1601,6 +1606,7 @@ disabled with --disable-FEATURE, default is enabled if available:
+   vhost-crypto    vhost-crypto acceleration support
+   spice           spice
+   rbd             rados block device (rbd)
++  vitastor        vitastor block device
+   libiscsi        iscsi support
+   libnfs          nfs support
+   smartcard       smartcard support (libcacard)
+@@ -3594,6 +3600,27 @@ EOF
+   fi
+ fi
+ 
++##########################################
++# vitastor probe
++if test "$vitastor" != "no" ; then
++  cat > $TMPC <<EOF
++#include <vitastor_c.h>
++int main(void) {
++  vitastor_c_create_qemu(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
++  return 0;
++}
++EOF
++  vitastor_libs="-lvitastor_client"
++  if compile_prog "" "$vitastor_libs" ; then
++    vitastor=yes
++  else
++    if test "$vitastor" = "yes" ; then
++      feature_not_found "vitastor block device" "Install vitastor-client-dev"
++    fi
++    vitastor=no
++  fi
++fi
++
+ ##########################################
+ # libssh2 probe
+ min_libssh2_version=1.2.8
+@@ -5837,6 +5864,7 @@ echo "Trace output file $trace_file-<pid>"
+ fi
+ echo "spice support     $spice $(echo_version $spice $spice_protocol_version/$spice_server_version)"
+ echo "rbd support       $rbd"
++echo "vitastor support  $vitastor"
+ echo "xfsctl support    $xfs"
+ echo "smartcard support $smartcard"
+ echo "libusb            $libusb"
+@@ -6416,6 +6444,11 @@ if test "$rbd" = "yes" ; then
+   echo "RBD_CFLAGS=$rbd_cflags" >> $config_host_mak
+   echo "RBD_LIBS=$rbd_libs" >> $config_host_mak
+ fi
++if test "$vitastor" = "yes" ; then
++  echo "CONFIG_VITASTOR=m" >> $config_host_mak
++  echo "VITASTOR_CFLAGS=$vitastor_cflags" >> $config_host_mak
++  echo "VITASTOR_LIBS=$vitastor_libs" >> $config_host_mak
++fi
+ 
+ echo "CONFIG_COROUTINE_BACKEND=$coroutine" >> $config_host_mak
+ if test "$coroutine_pool" = "yes" ; then
+diff --git a/qapi/block-core.json b/qapi/block-core.json
+index c50517bff3..c780bb2c1c 100644
+--- a/qapi/block-core.json
++++ b/qapi/block-core.json
+@@ -2514,7 +2514,7 @@
+             'dmg', 'file', 'ftp', 'ftps', 'gluster', 'host_cdrom',
+             'host_device', 'http', 'https', 'iscsi', 'luks', 'nbd', 'nfs',
+             'null-aio', 'null-co', 'nvme', 'parallels', 'qcow', 'qcow2', 'qed',
+-            'quorum', 'raw', 'rbd', 'replication', 'sheepdog', 'ssh',
++            'quorum', 'raw', 'rbd', 'vitastor', 'replication', 'sheepdog', 'ssh',
+             'throttle', 'vdi', 'vhdx', 'vmdk', 'vpc', 'vvfat', 'vxhs' ] }
+ 
+ ##
+@@ -3217,6 +3217,28 @@
+             '*snap-id': 'uint32',
+             '*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:
+ #
+@@ -3547,6 +3569,7 @@
+       'rbd':        'BlockdevOptionsRbd',
+       'replication':'BlockdevOptionsReplication',
+       'sheepdog':   'BlockdevOptionsSheepdog',
++      'vitastor':   'BlockdevOptionsVitastor',
+       'ssh':        'BlockdevOptionsSsh',
+       'throttle':   'BlockdevOptionsThrottle',
+       'vdi':        'BlockdevOptionsGenericFormat',
+@@ -3991,6 +4014,17 @@
+             '*subformat':           'BlockdevVhdxSubformat',
+             '*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:
+ #
+@@ -4074,6 +4108,7 @@
+       'rbd':            'BlockdevCreateOptionsRbd',
+       'replication':    'BlockdevCreateNotSupported',
+       'sheepdog':       'BlockdevCreateOptionsSheepdog',
++      'vitastor':       'BlockdevCreateOptionsVitastor',
+       'ssh':            'BlockdevCreateOptionsSsh',
+       'throttle':       'BlockdevCreateNotSupported',
+       'vdi':            'BlockdevCreateOptionsVdi',

patches/qemu-5.2-vitastor.patch

@@ -0,0 +1,181 @@
+Index: qemu-5.2+dfsg/qapi/block-core.json
+===================================================================
+--- qemu-5.2+dfsg.orig/qapi/block-core.json
++++ qemu-5.2+dfsg/qapi/block-core.json
+@@ -2831,7 +2831,7 @@
+             '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' ] }
+ 
+ ##
+@@ -3668,6 +3668,28 @@
+             '*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.
+@@ -4015,6 +4037,7 @@
+       'replication': { 'type': 'BlockdevOptionsReplication',
+                        'if': 'defined(CONFIG_REPLICATION)' },
+       'sheepdog':   'BlockdevOptionsSheepdog',
++      'vitastor':   'BlockdevOptionsVitastor',
+       'ssh':        'BlockdevOptionsSsh',
+       'throttle':   'BlockdevOptionsThrottle',
+       'vdi':        'BlockdevOptionsGenericFormat',
+@@ -4404,6 +4427,17 @@
+             '*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
+@@ -4665,6 +4699,7 @@
+       'qed':            'BlockdevCreateOptionsQed',
+       'rbd':            'BlockdevCreateOptionsRbd',
+       'sheepdog':       'BlockdevCreateOptionsSheepdog',
++      'vitastor':       'BlockdevCreateOptionsVitastor',
+       'ssh':            'BlockdevCreateOptionsSsh',
+       'vdi':            'BlockdevCreateOptionsVdi',
+       'vhdx':           'BlockdevCreateOptionsVhdx',
+Index: qemu-5.2+dfsg/block/meson.build
+===================================================================
+--- qemu-5.2+dfsg.orig/block/meson.build
++++ qemu-5.2+dfsg/block/meson.build
+@@ -76,6 +76,7 @@ foreach m : [
+   ['CONFIG_LIBNFS', 'nfs', libnfs, 'nfs.c'],
+   ['CONFIG_LIBSSH', 'ssh', libssh, 'ssh.c'],
+   ['CONFIG_RBD', 'rbd', rbd, 'rbd.c'],
++  ['CONFIG_VITASTOR', 'vitastor', vitastor, 'vitastor.c'],
+ ]
+   if config_host.has_key(m[0])
+     if enable_modules
+Index: qemu-5.2+dfsg/configure
+===================================================================
+--- qemu-5.2+dfsg.orig/configure
++++ qemu-5.2+dfsg/configure
+@@ -372,6 +372,7 @@ trace_backends="log"
+ trace_file="trace"
+ spice=""
+ rbd=""
++vitastor=""
+ smartcard=""
+ u2f="auto"
+ libusb=""
+@@ -1263,6 +1264,10 @@ for opt do
+   ;;
+   --enable-rbd) rbd="yes"
+   ;;
++  --disable-vitastor) vitastor="no"
++  ;;
++  --enable-vitastor) vitastor="yes"
++  ;;
+   --disable-xfsctl) xfs="no"
+   ;;
+   --enable-xfsctl) xfs="yes"
+@@ -1827,6 +1832,7 @@ disabled with --disable-FEATURE, default
+   vhost-vdpa      vhost-vdpa kernel backend support
+   spice           spice
+   rbd             rados block device (rbd)
++  vitastor        vitastor block device
+   libiscsi        iscsi support
+   libnfs          nfs support
+   smartcard       smartcard support (libcacard)
+@@ -3719,6 +3725,27 @@ EOF
+ fi
+ 
+ ##########################################
++# vitastor probe
++if test "$vitastor" != "no" ; then
++  cat > $TMPC <<EOF
++#include <vitastor_c.h>
++int main(void) {
++  vitastor_c_create_qemu(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
++  return 0;
++}
++EOF
++  vitastor_libs="-lvitastor_client"
++  if compile_prog "" "$vitastor_libs" ; then
++    vitastor=yes
++  else
++    if test "$vitastor" = "yes" ; then
++      feature_not_found "vitastor block device" "Install vitastor-client-dev"
++    fi
++    vitastor=no
++  fi
++fi
++
++##########################################
+ # libssh probe
+ if test "$libssh" != "no" ; then
+   if $pkg_config --exists libssh; then
+@@ -6456,6 +6483,10 @@ if test "$rbd" = "yes" ; then
+   echo "CONFIG_RBD=y" >> $config_host_mak
+   echo "RBD_LIBS=$rbd_libs" >> $config_host_mak
+ fi
++if test "$vitastor" = "yes" ; then
++  echo "CONFIG_VITASTOR=y" >> $config_host_mak
++  echo "VITASTOR_LIBS=$vitastor_libs" >> $config_host_mak
++fi
+ 
+ echo "CONFIG_COROUTINE_BACKEND=$coroutine" >> $config_host_mak
+ if test "$coroutine_pool" = "yes" ; then
+Index: qemu-5.2+dfsg/meson.build
+===================================================================
+--- qemu-5.2+dfsg.orig/meson.build
++++ qemu-5.2+dfsg/meson.build
+@@ -596,6 +596,10 @@ rbd = not_found
+ if 'CONFIG_RBD' in config_host
+   rbd = declare_dependency(link_args: config_host['RBD_LIBS'].split())
+ endif
++vitastor = not_found
++if 'CONFIG_VITASTOR' in config_host
++  vitastor = declare_dependency(link_args: config_host['VITASTOR_LIBS'].split())
++endif
+ glusterfs = not_found
+ if 'CONFIG_GLUSTERFS' in config_host
+   glusterfs = declare_dependency(compile_args: config_host['GLUSTERFS_CFLAGS'].split(),
+@@ -2145,6 +2149,7 @@ endif
+ # TODO: add back protocol and server version
+ summary_info += {'spice support':     config_host.has_key('CONFIG_SPICE')}
+ summary_info += {'rbd support':       config_host.has_key('CONFIG_RBD')}
++summary_info += {'vitastor support':  config_host.has_key('CONFIG_VITASTOR')}
+ summary_info += {'xfsctl support':    config_host.has_key('CONFIG_XFS')}
+ summary_info += {'smartcard support': config_host.has_key('CONFIG_SMARTCARD')}
+ summary_info += {'U2F support':       u2f.found()}

rpm/build-tarball.sh

@@ -24,4 +24,4 @@ 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"; }'`
 perl -i -pe 's/(Requires:\s*fio)([^\n]+)?/$1 = '$FIO'/' $VITASTOR/rpm/vitastor-el$EL.spec
-tar --transform 's#^#vitastor-0.9.2/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.9.2$(rpm --eval '%dist').tar.gz *
+tar --transform 's#^#vitastor-0.9.4/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.9.4$(rpm --eval '%dist').tar.gz *

rpm/qemu-kvm-4.2-el7.spec.patch

@@ -22,7 +22,7 @@
  Name: qemu-kvm
  Version: 4.2.0
 -Release: 29.vitastor%{?dist}.6
-+Release: 32.vitastor%{?dist}.6
++Release: 34.vitastor%{?dist}.6
  # Epoch because we pushed a qemu-1.0 package. AIUI this can't ever be dropped
  Epoch: 15
  License: GPLv2 and GPLv2+ and CC-BY

rpm/qemu-kvm-4.2-el8.spec.patch

@@ -13,7 +13,7 @@
  Name: qemu-kvm
  Version: 4.2.0
 -Release: 29%{?dist}.6
-+Release: 32.vitastor%{?dist}.6
++Release: 33.vitastor%{?dist}.6
  # Epoch because we pushed a qemu-1.0 package. AIUI this can't ever be dropped
  Epoch: 15
  License: GPLv2 and GPLv2+ and CC-BY

rpm/qemu-kvm-6.2-el8.spec.patch

@@ -0,0 +1,103 @@
+--- qemu-kvm-6.2.spec.orig	2023-07-18 13:52:57.636625440 +0000
++++ qemu-kvm-6.2.spec	2023-07-18 13:52:19.011683886 +0000
+@@ -73,6 +73,7 @@ Requires: %{name}-hw-usbredir = %{epoch}
+ %endif                                                           \
+ Requires: %{name}-block-iscsi = %{epoch}:%{version}-%{release}   \
+ Requires: %{name}-block-rbd = %{epoch}:%{version}-%{release}     \
++Requires: %{name}-block-vitastor = %{epoch}:%{version}-%{release}\
+ Requires: %{name}-block-ssh = %{epoch}:%{version}-%{release}
+ 
+ # Macro to properly setup RHEL/RHEV conflict handling
+@@ -83,7 +84,7 @@ Obsoletes: %1-rhev <= %{epoch}:%{version
+ Summary: QEMU is a machine emulator and virtualizer
+ Name: qemu-kvm
+ Version: 6.2.0
+-Release: 32%{?rcrel}%{?dist}
++Release: 32.vitastor%{?rcrel}%{?dist}
+ # Epoch because we pushed a qemu-1.0 package. AIUI this can't ever be dropped
+ Epoch: 15
+ License: GPLv2 and GPLv2+ and CC-BY
+@@ -122,6 +123,7 @@ Source37: tests_data_acpi_pc_SSDT.dimmpx
+ Source38: tests_data_acpi_q35_FACP.slic
+ Source39: tests_data_acpi_q35_SSDT.dimmpxm
+ Source40: tests_data_acpi_virt_SSDT.memhp
++Source41: qemu-vitastor.c
+ 
+ Patch0001: 0001-redhat-Adding-slirp-to-the-exploded-tree.patch
+ Patch0005: 0005-Initial-redhat-build.patch
+@@ -652,6 +654,7 @@ Patch255: kvm-scsi-protect-req-aiocb-wit
+ Patch256: kvm-dma-helpers-prevent-dma_blk_cb-vs-dma_aio_cancel-rac.patch
+ # For bz#2090990 - qemu crash with error scsi_req_unref(SCSIRequest *): Assertion `req->refcount > 0' failed or scsi_dma_complete(void *, int): Assertion `r->req.aiocb != NULL' failed [8.7.0]
+ Patch257: kvm-virtio-scsi-reset-SCSI-devices-from-main-loop-thread.patch
++Patch258: qemu-6.2-vitastor.patch
+ 
+ BuildRequires: wget
+ BuildRequires: rpm-build
+@@ -689,6 +692,7 @@ BuildRequires: libcurl-devel
+ BuildRequires: libssh-devel
+ BuildRequires: librados-devel
+ BuildRequires: librbd-devel
++BuildRequires: vitastor-client-devel
+ %if %{have_gluster}
+ # For gluster block driver
+ BuildRequires: glusterfs-api-devel
+@@ -926,6 +930,14 @@ Install this package if you want to acce
+ using the rbd protocol.
+ 
+ 
++%package  block-vitastor
++Summary: QEMU Vitastor block driver
++Requires: %{name}-common%{?_isa} = %{epoch}:%{version}-%{release}
++
++%description block-vitastor
++This package provides the additional Vitastor block driver for QEMU.
++
++
+ %package  block-ssh
+ Summary: QEMU SSH block driver
+ Requires: %{name}-common%{?_isa} = %{epoch}:%{version}-%{release}
+@@ -979,6 +991,7 @@ This package provides usbredir support.
+ rm -fr slirp
+ mkdir slirp
+ %autopatch -p1
++cp %{SOURCE41} ./block/vitastor.c
+ 
+ %global qemu_kvm_build qemu_kvm_build
+ mkdir -p %{qemu_kvm_build}
+@@ -994,7 +1007,7 @@ cp -f %{SOURCE40} tests/data/acpi/virt/S
+ # --build-id option is used for giving info to the debug packages.
+ buildldflags="VL_LDFLAGS=-Wl,--build-id"
+ 
+-%global block_drivers_list qcow2,raw,file,host_device,nbd,iscsi,rbd,blkdebug,luks,null-co,nvme,copy-on-read,throttle
++%global block_drivers_list qcow2,raw,file,host_device,nbd,iscsi,rbd,vitastor,blkdebug,luks,null-co,nvme,copy-on-read,throttle
+ 
+ %if 0%{have_gluster}
+     %global block_drivers_list %{block_drivers_list},gluster
+@@ -1149,9 +1162,7 @@ pushd %{qemu_kvm_build}
+   --firmwarepath=%{_prefix}/share/qemu-firmware \
+   --meson="git" \
+   --target-list="%{buildarch}" \
+-  --block-drv-rw-whitelist=%{block_drivers_list} \
+   --audio-drv-list= \
+-  --block-drv-ro-whitelist=vmdk,vhdx,vpc,https,ssh \
+   --with-coroutine=ucontext \
+   --with-git=git \
+   --tls-priority=@QEMU,SYSTEM \
+@@ -1197,6 +1208,7 @@ pushd %{qemu_kvm_build}
+ %endif
+   --enable-pie \
+   --enable-rbd \
++  --enable-vitastor \
+ %if 0%{have_librdma}
+   --enable-rdma \
+ %endif
+@@ -1794,6 +1806,9 @@ sh %{_sysconfdir}/sysconfig/modules/kvm.
+ %files block-rbd
+ %{_libdir}/qemu-kvm/block-rbd.so
+ 
++%files block-vitastor
++%{_libdir}/qemu-kvm/block-vitastor.so
++
+ %files block-ssh
+ %{_libdir}/qemu-kvm/block-ssh.so
+ 

rpm/qemu-kvm-7.2-el9.spec.patch

@@ -0,0 +1,93 @@
+--- qemu-kvm-7.2.spec.orig	2023-06-22 13:56:19.000000000 +0000
++++ qemu-kvm-7.2.spec	2023-07-18 07:55:22.347090196 +0000
+@@ -100,8 +100,6 @@
+ %endif
+ 
+ %global target_list %{kvm_target}-softmmu
+-%global block_drivers_rw_list qcow2,raw,file,host_device,nbd,iscsi,rbd,blkdebug,luks,null-co,nvme,copy-on-read,throttle,compress
+-%global block_drivers_ro_list vdi,vmdk,vhdx,vpc,https
+ %define qemudocdir %{_docdir}/%{name}
+ %global firmwaredirs "%{_datadir}/qemu-firmware:%{_datadir}/ipxe/qemu:%{_datadir}/seavgabios:%{_datadir}/seabios"
+ 
+@@ -126,6 +124,7 @@ Requires: %{name}-device-usb-host = %{ep
+ Requires: %{name}-device-usb-redirect = %{epoch}:%{version}-%{release}   \
+ %endif                                                           \
+ Requires: %{name}-block-rbd = %{epoch}:%{version}-%{release}     \
++Requires: %{name}-block-vitastor = %{epoch}:%{version}-%{release}\
+ Requires: %{name}-audio-pa = %{epoch}:%{version}-%{release}
+ 
+ # Since SPICE is removed from RHEL-9, the following Obsoletes:
+@@ -148,7 +147,7 @@ Obsoletes: %{name}-block-ssh <= %{epoch}
+ Summary: QEMU is a machine emulator and virtualizer
+ Name: qemu-kvm
+ Version: 7.2.0
+-Release: 14%{?rcrel}%{?dist}%{?cc_suffix}.1
++Release: 14.vitastor%{?rcrel}%{?dist}%{?cc_suffix}.1
+ # Epoch because we pushed a qemu-1.0 package. AIUI this can't ever be dropped
+ # Epoch 15 used for RHEL 8
+ # Epoch 17 used for RHEL 9 (due to release versioning offset in RHEL 8.5)
+@@ -171,6 +170,7 @@ Source28: 95-kvm-memlock.conf
+ Source30: kvm-s390x.conf
+ Source31: kvm-x86.conf
+ Source36: README.tests
++Source37: qemu-vitastor.c
+ 
+ 
+ Patch0004: 0004-Initial-redhat-build.patch
+@@ -418,6 +418,7 @@ Patch134: kvm-target-i386-Fix-BZHI-instr
+ Patch135: kvm-intel-iommu-fail-DEVIOTLB_UNMAP-without-dt-mode.patch
+ # For bz#2203745 - Disk detach is unsuccessful while the guest is still booting [rhel-9.2.0.z]
+ Patch136: kvm-acpi-pcihp-allow-repeating-hot-unplug-requests.patch
++Patch137: qemu-7.2-vitastor.patch
+ 
+ %if %{have_clang}
+ BuildRequires: clang
+@@ -449,6 +450,7 @@ BuildRequires: libcurl-devel
+ %if %{have_block_rbd}
+ BuildRequires: librbd-devel
+ %endif
++BuildRequires: vitastor-client-devel
+ # We need both because the 'stap' binary is probed for by configure
+ BuildRequires: systemtap
+ BuildRequires: systemtap-sdt-devel
+@@ -642,6 +644,14 @@ using the rbd protocol.
+ %endif
+ 
+ 
++%package  block-vitastor
++Summary: QEMU Vitastor block driver
++Requires: %{name}-common%{?_isa} = %{epoch}:%{version}-%{release}
++
++%description block-vitastor
++This package provides the additional Vitastor block driver for QEMU.
++
++
+ %package  audio-pa
+ Summary: QEMU PulseAudio audio driver
+ Requires: %{name}-common%{?_isa} = %{epoch}:%{version}-%{release}
+@@ -719,6 +729,7 @@ This package provides usbredir support.
+ %prep
+ %setup -q -n qemu-%{version}%{?rcstr}
+ %autopatch -p1
++cp %{SOURCE37} ./block/vitastor.c
+ 
+ %global qemu_kvm_build qemu_kvm_build
+ mkdir -p %{qemu_kvm_build}
+@@ -946,6 +957,7 @@ run_configure \
+ %if %{have_block_rbd}
+   --enable-rbd \
+ %endif
++  --enable-vitastor \
+ %if %{have_librdma}
+   --enable-rdma \
+ %endif
+@@ -1426,6 +1438,9 @@ useradd -r -u 107 -g qemu -G kvm -d / -s
+ %files block-rbd
+ %{_libdir}/%{name}/block-rbd.so
+ %endif
++%files block-vitastor
++%{_libdir}/%{name}/block-vitastor.so
++
+ %files audio-pa
+ %{_libdir}/%{name}/audio-pa.so
+ 

rpm/vitastor-el7.Dockerfile

@@ -35,7 +35,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.9.2.el7.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.9.4.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:        0.9.2
+Version:        0.9.4
 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-0.9.2.el7.tar.gz
+Source0:        vitastor-0.9.4.el7.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel

rpm/vitastor-el8.Dockerfile

@@ -35,7 +35,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.9.2.el8.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.9.4.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:        0.9.2
+Version:        0.9.4
 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-0.9.2.el8.tar.gz
+Source0:        vitastor-0.9.4.el8.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel

rpm/vitastor-el9.Dockerfile

@@ -18,7 +18,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.9.2.el9.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.9.4.el9.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el9.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el9.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.9.2
+Version:        0.9.4
 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-0.9.2.el9.tar.gz
+Source0:        vitastor-0.9.4.el9.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel

src/CMakeLists.txt

@@ -16,7 +16,7 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 	set(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR}")
 endif()
 
-add_definitions(-DVERSION="0.9.2")
+add_definitions(-DVERSION="0.9.4")
 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)
@@ -56,11 +56,6 @@ if (ISAL_LIBRARIES)
 	add_definitions(-DWITH_ISAL)
 endif (ISAL_LIBRARIES)
 
-find_package(OpenSSL)
-if (OPENSSL_FOUND)
-	add_definitions(-DWITH_OPENSSL)
-endif (OPENSSL_FOUND)
-
 add_custom_target(build_tests)
 add_custom_target(test
 	COMMAND

src/blockstore_flush.cpp

@@ -714,9 +714,15 @@ resume_1:
                             return false;
                         }
                     }
+                    if (new_trim_pos < bs->journal.used_start
+                        ? (bs->journal.dirty_start >= bs->journal.used_start || bs->journal.dirty_start < new_trim_pos)
+                        : (bs->journal.dirty_start >= bs->journal.used_start && bs->journal.dirty_start < new_trim_pos))
+                    {
+                        bs->journal.dirty_start = new_trim_pos;
+                    }
                     bs->journal.used_start = new_trim_pos;
 #ifdef BLOCKSTORE_DEBUG
-                    printf("Journal trimmed to %08lx (next_free=%08lx)\n", bs->journal.used_start, bs->journal.next_free);
+                    printf("Journal trimmed to %08lx (next_free=%08lx dirty_start=%08lx)\n", bs->journal.used_start, bs->journal.next_free, bs->journal.dirty_start);
 #endif
                     if (bs->journal.flush_journal && !flusher->flush_queue.size())
                     {

src/blockstore_flush.h

@@ -103,6 +103,7 @@ public:
     journal_flusher_t(blockstore_impl_t *bs);
     ~journal_flusher_t();
     void loop();
+    bool is_trim_wanted() { return trim_wanted; }
     bool is_active();
     void mark_trim_possible();
     void request_trim();

src/blockstore_rollback.cpp

@@ -218,7 +218,7 @@ void blockstore_impl_t::erase_dirty(blockstore_dirty_db_t::iterator dirty_start,
         auto used = --journal.used_sectors[dirty_it->second.journal_sector];
 #ifdef BLOCKSTORE_DEBUG
         printf(
-            "remove usage of journal offset %08lx by %lx:%lx v%lu (%d refs)\n", dirty_it->second.journal_sector,
+            "remove usage of journal offset %08lx by %lx:%lx v%lu (%lu refs)\n", dirty_it->second.journal_sector,
             dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version, used
         );
 #endif

src/blockstore_write.cpp

@@ -661,8 +661,13 @@ void blockstore_impl_t::release_journal_sectors(blockstore_op_t *op)
         uint64_t s = PRIV(op)->min_flushed_journal_sector;
         while (1)
         {
-            if (s != (1+journal.cur_sector) && journal.sector_info[s-1].flush_count == 0)
+            if (!journal.sector_info[s-1].dirty && journal.sector_info[s-1].flush_count == 0)
             {
+                if (s == (1+journal.cur_sector))
+                {
+                    // Forcibly move to the next sector and move dirty position
+                    journal.in_sector_pos = journal.block_size;
+                }
                 // 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)

src/cli_ls.cpp

@@ -56,14 +56,15 @@ struct image_lister_t
             {
                 continue;
             }
-            auto & pool_cfg = parent->cli->st_cli.pool_config.at(INODE_POOL(ic.second.num));
+            auto pool_it = parent->cli->st_cli.pool_config.find(INODE_POOL(ic.second.num));
+            bool good_pool = pool_it != parent->cli->st_cli.pool_config.end();
             auto item = json11::Json::object {
                 { "name", ic.second.name },
                 { "size", ic.second.size },
                 { "used_size", 0 },
                 { "readonly", ic.second.readonly },
                 { "pool_id", (uint64_t)INODE_POOL(ic.second.num) },
-                { "pool_name", pool_cfg.name },
+                { "pool_name", good_pool ? pool_it->second.name : "? (ID:"+std::to_string(INODE_POOL(ic.second.num))+")" },
                 { "inode_num", INODE_NO_POOL(ic.second.num) },
                 { "inode_id", ic.second.num },
             };
@@ -247,6 +248,8 @@ resume_1:
         if (state == 1)
             goto resume_1;
         get_list();
+        if (state == 100)
+            return;
         if (show_stats)
         {
 resume_1:
@@ -269,7 +272,7 @@ resume_1:
             { "key", "name" },
             { "title", "NAME" },
         });
-        if (!list_pool_id)
+        if (list_pool_name == "")
         {
             cols.push_back(json11::Json::object{
                 { "key", "pool_name" },

src/cli_merge.cpp

@@ -41,7 +41,7 @@ struct snap_merger_t
     int fsync_interval = 128;
 
     // -- STATE --
-    inode_t target;
+    inode_t target, to_num;
     int target_rank;
     bool inside_continue = false;
     int state = 0;
@@ -98,6 +98,7 @@ struct snap_merger_t
             state = 100;
             return;
         }
+        to_num = to_cfg->num;
         // Check that to_cfg is actually a child of from_cfg and target_cfg is somewhere between them
         std::vector<inode_t> chain_list;
         inode_config_t *cur = to_cfg;
@@ -451,7 +452,7 @@ struct snap_merger_t
     {
         cluster_op_t *op = &rwo->op;
         op->opcode = OSD_OP_READ;
-        op->inode = target;
+        op->inode = to_num;
         op->offset = rwo->offset;
         op->len = target_block_size;
         op->iov.push_back(rwo->buf, target_block_size);
@@ -483,7 +484,7 @@ struct snap_merger_t
                 {
                     // write start->end
                     rwo->todo++;
-                    write_subop(rwo, rwo->start*gran, rwo->end*gran, use_cas ? 1+rwo->op.version : 0);
+                    write_subop(rwo, rwo->start*gran, rwo->end*gran, use_cas && to_num == target ? 1+rwo->op.version : 0);
                     rwo->start = rwo->end;
                     if (use_cas)
                     {
@@ -502,7 +503,7 @@ struct snap_merger_t
         {
             // write start->end
             rwo->todo++;
-            write_subop(rwo, rwo->start*gran, rwo->end*gran, use_cas ? 1+rwo->op.version : 0);
+            write_subop(rwo, rwo->start*gran, rwo->end*gran, use_cas && to_num == target ? 1+rwo->op.version : 0);
             rwo->start = rwo->end;
             if (use_cas)
             {
@@ -532,7 +533,7 @@ struct snap_merger_t
                 if (use_cas && subop->retval == -EINTR)
                 {
                     // CAS failure - reread and repeat optimistically
-                    rwo->start = subop->offset - rwo->offset;
+                    rwo->start = rwo->end = 0;
                     rwo_read(rwo);
                     delete subop;
                     return;
@@ -542,7 +543,7 @@ struct snap_merger_t
                 rwo->error_read = false;
             }
             // Increment CAS version
-            rwo->op.version++;
+            rwo->op.version = subop->version;
             if (use_cas)
                 next_write(rwo);
             else

src/cli_rm.cpp

@@ -65,6 +65,9 @@ struct snap_remover_t
     int current_child = 0;
     std::function<bool(cli_result_t &)> cb;
 
+    std::vector<std::string> rebased_images, deleted_images;
+    std::vector<uint64_t> deleted_ids;
+    std::string inverse_child_name, inverse_parent_name;
     cli_result_t result;
 
     bool is_done()
@@ -122,6 +125,7 @@ resume_1:
         {
             if (merge_children[current_child] == inverse_child)
                 continue;
+            rebased_images.push_back(parent->cli->st_cli.inode_config.at(merge_children[current_child]).name);
             start_merge_child(merge_children[current_child], merge_children[current_child]);
             if (state == 100)
                 return;
@@ -134,9 +138,12 @@ resume_2:
             cb = NULL;
             if (result.err)
             {
+                result.data = my_result(result.data);
                 state = 100;
                 return;
             }
+            else if (parent->progress)
+                printf("%s\n", result.text.c_str());
             parent->change_parent(merge_children[current_child], new_parent, &result);
             state = 3;
 resume_3:
@@ -144,6 +151,7 @@ resume_3:
                 return;
             if (result.err)
             {
+                result.data = my_result(result.data);
                 state = 100;
                 return;
             }
@@ -165,9 +173,12 @@ resume_4:
             cb = NULL;
             if (result.err)
             {
+                result.data = my_result(result.data);
                 state = 100;
                 return;
             }
+            else if (parent->progress)
+                printf("%s\n", result.text.c_str());
             // Delete "inverse" child data
             start_delete_source(inverse_child);
             if (state == 100)
@@ -181,9 +192,12 @@ resume_5:
             cb = NULL;
             if (result.err)
             {
+                result.data = my_result(result.data);
                 state = 100;
                 return;
             }
+            else if (parent->progress)
+                printf("%s\n", result.text.c_str());
             // Delete "inverse" child metadata, rename parent over it,
             // and also change parent links of the previous "inverse" child
             rename_inverse_parent();
@@ -199,6 +213,12 @@ resume_6:
         {
             if (chain_list[current_child] == inverse_parent)
                 continue;
+            {
+                auto parent_it = parent->cli->st_cli.inode_config.find(chain_list[current_child]);
+                if (parent_it != parent->cli->st_cli.inode_config.end())
+                    deleted_images.push_back(parent_it->second.name);
+                deleted_ids.push_back(chain_list[current_child]);
+            }
             start_delete_source(chain_list[current_child]);
 resume_7:
             while (!cb(result))
@@ -209,9 +229,12 @@ resume_7:
             cb = NULL;
             if (result.err)
             {
+                result.data = my_result(result.data);
                 state = 100;
                 return;
             }
+            else if (parent->progress)
+                printf("%s\n", result.text.c_str());
             delete_inode_config(chain_list[current_child]);
             if (state == 100)
                 return;
@@ -221,11 +244,26 @@ resume_8:
                 return;
         }
         state = 100;
+        result = (cli_result_t){
+            .text = "",
+            .data = my_result(result.data),
+        };
 resume_100:
         // Done
         return;
     }
 
+    json11::Json my_result(json11::Json src)
+    {
+        auto obj = src.object_items();
+        obj["deleted_ids"] = deleted_ids;
+        obj["deleted_images"] = deleted_images;
+        obj["rebased_images"] = rebased_images;
+        obj["renamed_from"] = inverse_parent_name;
+        obj["renamed_to"] = inverse_child_name;
+        return obj;
+    }
+
     void get_merge_children()
     {
         // Get all children of from..to
@@ -338,7 +376,11 @@ resume_100:
             }
             for (auto inode_result: data["responses"].array_items())
             {
-                auto kv = parent->cli->st_cli.parse_etcd_kv(inode_result["kvs"][0]);
+                if (inode_result["response_range"]["kvs"].array_items().size() == 0)
+                {
+                    continue;
+                }
+                auto kv = parent->cli->st_cli.parse_etcd_kv(inode_result["response_range"]["kvs"][0]);
                 pool_id_t pool_id = 0;
                 inode_t inode = 0;
                 char null_byte = 0;
@@ -377,7 +419,7 @@ resume_100:
             inode_t child = cp.first;
             uint64_t child_used = inode_used[child];
             int rank = cp.second;
-            for (int i = chain_list.size()-rank; i < chain_list.size(); i++)
+            for (int i = chain_list.size()-1-rank; i < chain_list.size(); i++)
             {
                 inode_t parent = chain_list[i];
                 uint64_t parent_used = inode_used[parent];
@@ -413,8 +455,8 @@ resume_100:
         }
         inode_config_t *child_cfg = &child_it->second;
         inode_config_t *target_cfg = &target_it->second;
-        std::string child_name = child_cfg->name;
-        std::string target_name = target_cfg->name;
+        inverse_child_name = child_cfg->name;
+        inverse_parent_name = target_cfg->name;
         std::string child_cfg_key = base64_encode(
             parent->cli->st_cli.etcd_prefix+
             "/config/inode/"+std::to_string(INODE_POOL(inverse_child))+
@@ -425,6 +467,9 @@ resume_100:
             "/config/inode/"+std::to_string(INODE_POOL(inverse_parent))+
             "/"+std::to_string(INODE_NO_POOL(inverse_parent))
         );
+        std::string target_idx_key = base64_encode(
+            parent->cli->st_cli.etcd_prefix+"/index/image/"+inverse_parent_name
+        );
         // Fill new configuration
         inode_config_t new_cfg = *child_cfg;
         new_cfg.num = target_cfg->num;
@@ -449,6 +494,11 @@ resume_100:
                     { "key", child_cfg_key },
                 } },
             },
+            json11::Json::object {
+                { "request_delete_range", json11::Json::object {
+                    { "key", target_idx_key },
+                } },
+            },
             json11::Json::object {
                 { "request_put", json11::Json::object {
                     { "key", target_cfg_key },
@@ -495,12 +545,12 @@ resume_100:
         parent->cli->st_cli.etcd_txn_slow(json11::Json::object {
             { "compare", cmp },
             { "success", txn },
-        }, [this, target_name, child_name](std::string err, json11::Json res)
+        }, [this](std::string err, json11::Json res)
         {
             parent->waiting--;
             if (err != "")
             {
-                result = (cli_result_t){ .err = EIO, .text = "Error renaming "+target_name+" to "+child_name+": "+err };
+                result = (cli_result_t){ .err = EIO, .text = "Error renaming "+inverse_parent_name+" to "+inverse_child_name+": "+err };
                 state = 100;
                 return;
             }
@@ -508,14 +558,14 @@ resume_100:
             {
                 result = (cli_result_t){
                     .err = EAGAIN,
-                    .text = "Parent ("+target_name+"), child ("+child_name+"), or one of its children"
+                    .text = "Parent ("+inverse_parent_name+"), child ("+inverse_child_name+"), or one of its children"
                         " configuration was modified during rename",
                 };
                 state = 100;
                 return;
             }
             if (parent->progress)
-                printf("Layer %s renamed to %s\n", target_name.c_str(), child_name.c_str());
+                printf("Layer %s renamed to %s\n", inverse_parent_name.c_str(), inverse_child_name.c_str());
             parent->ringloop->wakeup();
         });
     }

src/cli_rm_data.cpp

@@ -28,6 +28,7 @@ struct rm_inode_t
     cli_tool_t *parent = NULL;
     inode_list_t *lister = NULL;
     std::vector<rm_pg_t*> lists;
+    std::vector<osd_num_t> inactive_osds;
     uint64_t total_count = 0, total_done = 0, total_prev_pct = 0;
     uint64_t pgs_to_list = 0;
     bool lists_done = false;
@@ -86,6 +87,16 @@ struct rm_inode_t
             state = 100;
             return;
         }
+        inactive_osds = parent->cli->list_inode_get_inactive_osds(lister);
+        if (inactive_osds.size() && !parent->json_output)
+        {
+            fprintf(stderr, "Some data may remain after delete on OSDs which are currently down: ");
+            for (int i = 0; i < inactive_osds.size(); i++)
+            {
+                fprintf(stderr, i > 0 ? ", %lu" : "%lu", inactive_osds[i]);
+            }
+            fprintf(stderr, "\n");
+        }
         pgs_to_list = parent->cli->list_pg_count(lister);
         parent->cli->list_inode_next(lister, parent->parallel_osds);
     }
@@ -167,16 +178,33 @@ struct rm_inode_t
         }
         if (parent->progress && total_count > 0 && total_done*1000/total_count != total_prev_pct)
         {
-            printf("\rRemoved %lu/%lu objects, %lu more PGs to list...", total_done, total_count, pgs_to_list);
+            fprintf(stderr, "\rRemoved %lu/%lu objects, %lu more PGs to list...", total_done, total_count, pgs_to_list);
             total_prev_pct = total_done*1000/total_count;
         }
         if (lists_done && !lists.size())
         {
+            if (parent->progress && total_count > 0)
+            {
+                fprintf(stderr, "\n");
+            }
+            if (parent->progress && (total_done < total_count || inactive_osds.size() > 0))
+            {
+                fprintf(
+                    stderr, "Warning: Pool:%u,ID:%lu inode data may not have been fully removed.\n"
+                    " Use `vitastor-cli rm-data --pool %u --inode %lu` if you encounter it in listings.\n",
+                    pool_id, INODE_NO_POOL(inode), pool_id, INODE_NO_POOL(inode)
+                );
+            }
             result = (cli_result_t){
                 .err = error_count > 0 ? EIO : 0,
                 .text = error_count > 0 ? "Some blocks were not removed" : (
                     "Done, inode "+std::to_string(INODE_NO_POOL(inode))+" from pool "+
                     std::to_string(pool_id)+" removed"),
+                .data = json11::Json::object {
+                    { "removed_objects", total_done },
+                    { "total_objects", total_count },
+                    { "inactive_osds", inactive_osds },
+                },
             };
             state = 100;
         }

src/cluster_client.cpp

@@ -1209,6 +1209,10 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
             copy_part_bitmap(op, part);
             op->version = op->parts.size() == 1 ? part->op.reply.rw.version : 0;
         }
+        else if (op->opcode == OSD_OP_WRITE)
+        {
+            op->version = op->parts.size() == 1 ? part->op.reply.rw.version : 0;
+        }
         if (op->inflight_count == 0)
         {
             if (op->opcode == OSD_OP_SYNC)

src/cluster_client.h

@@ -130,6 +130,7 @@ public:
     inode_list_t *list_inode_start(inode_t inode,
         std::function<void(inode_list_t* lst, std::set<object_id>&& objects, pg_num_t pg_num, osd_num_t primary_osd, int status)> callback);
     int list_pg_count(inode_list_t *lst);
+    const std::vector<osd_num_t> & list_inode_get_inactive_osds(inode_list_t *lst);
     void list_inode_next(inode_list_t *lst, int next_pgs);
     //inline uint32_t get_bs_bitmap_granularity() { return st_cli.global_bitmap_granularity; }
     //inline uint64_t get_bs_block_size() { return st_cli.global_block_size; }

src/cluster_client_list.cpp

@@ -36,6 +36,7 @@ struct inode_list_t
     inode_t inode = 0;
     int done_pgs = 0;
     int want = 0;
+    std::vector<osd_num_t> inactive_osds;
     std::vector<inode_list_pg_t*> pgs;
     std::function<void(inode_list_t* lst, std::set<object_id>&& objects, pg_num_t pg_num, osd_num_t primary_osd, int status)> callback;
 };
@@ -60,6 +61,7 @@ inode_list_t* cluster_client_t::list_inode_start(inode_t inode,
     lst->inode = inode;
     lst->callback = callback;
     auto pool_cfg = st_cli.pool_config[pool_id];
+    std::set<osd_num_t> inactive_osd_set;
     for (auto & pg_item: pool_cfg.pg_config)
     {
         auto & pg = pg_item.second;
@@ -106,11 +108,18 @@ inode_list_t* cluster_client_t::list_inode_start(inode_t inode,
             }
             for (osd_num_t peer_osd: all_peers)
             {
-                r->list_osds.push_back((inode_list_osd_t){
-                    .pg = r,
-                    .osd_num = peer_osd,
-                    .sent = false,
-                });
+                if (st_cli.peer_states.find(peer_osd) != st_cli.peer_states.end())
+                {
+                    r->list_osds.push_back((inode_list_osd_t){
+                        .pg = r,
+                        .osd_num = peer_osd,
+                        .sent = false,
+                    });
+                }
+                else
+                {
+                    inactive_osd_set.insert(peer_osd);
+                }
             }
         }
         else
@@ -132,6 +141,7 @@ inode_list_t* cluster_client_t::list_inode_start(inode_t inode,
     {
         lst->pgs[i]->pos = i;
     }
+    lst->inactive_osds.insert(lst->inactive_osds.end(), inactive_osd_set.begin(), inactive_osd_set.end());
     lists.push_back(lst);
     return lst;
 }
@@ -141,6 +151,11 @@ int cluster_client_t::list_pg_count(inode_list_t *lst)
     return lst->pgs.size();
 }
 
+const std::vector<osd_num_t> & cluster_client_t::list_inode_get_inactive_osds(inode_list_t *lst)
+{
+    return lst->inactive_osds;
+}
+
 void cluster_client_t::list_inode_next(inode_list_t *lst, int next_pgs)
 {
     if (next_pgs >= 0)

src/disk_tool_prepare.cpp

@@ -99,15 +99,16 @@ int disk_tool_t::prepare_one(std::map<std::string, std::string> options, int is_
     if (options["journal_size"] == "")
     {
         if (options["journal_device"] == "")
-            options["journal_size"] = "32M";
+            options["journal_size"] = is_hdd ? "128M" : "32M";
         else if (is_hdd)
             options["journal_size"] = DEFAULT_HYBRID_JOURNAL;
     }
+    bool is_hybrid = is_hdd && options["journal_device"] != "" && options["journal_device"] != options["data_device"];
     if (is_hdd)
     {
         if (options["block_size"] == "")
             options["block_size"] = "1M";
-        if (options["throttle_small_writes"] == "")
+        if (is_hybrid && options["throttle_small_writes"] == "")
             options["throttle_small_writes"] = "1";
     }
     json11::Json::object sb;
@@ -134,7 +135,7 @@ int disk_tool_t::prepare_one(std::map<std::string, std::string> options, int is_
             { "meta_offset", 4096 + (dsk.meta_device == dsk.journal_device ? dsk.journal_len : 0) },
             { "data_offset", 4096 + (dsk.data_device == dsk.meta_device ? dsk.meta_len : 0) +
                 (dsk.data_device == dsk.journal_device ? dsk.journal_len : 0) },
-            { "journal_no_same_sector_overwrites", true },
+            { "journal_no_same_sector_overwrites", !is_hdd || is_hybrid },
             { "journal_sector_buffer_count", 1024 },
             { "disable_data_fsync", json_is_true(options["disable_data_fsync"]) },
             { "disable_meta_fsync", json_is_true(options["disable_meta_fsync"]) },

src/etcd_state_client.cpp

@@ -187,22 +187,30 @@ void etcd_state_client_t::add_etcd_url(std::string addr)
             check_addr = addr;
         if (pos == std::string::npos)
             addr += "/v3";
+        bool local = false;
         int i;
         for (i = 0; i < local_ips.size(); i++)
         {
             if (local_ips[i] == check_addr)
             {
-                this->etcd_local.push_back(addr);
+                local = true;
                 break;
             }
         }
-        if (i >= local_ips.size())
-            this->etcd_addresses.push_back(addr);
+        auto & to = local ? this->etcd_local : this->etcd_addresses;
+        for (i = 0; i < to.size(); i++)
+        {
+            if (to[i] == addr)
+                break;
+        }
+        if (i >= to.size())
+            to.push_back(addr);
     }
 }
 
 void etcd_state_client_t::parse_config(const json11::Json & config)
 {
+    this->etcd_local.clear();
     this->etcd_addresses.clear();
     if (config["etcd_address"].is_string())
     {
@@ -349,7 +357,7 @@ void etcd_state_client_t::start_etcd_watcher()
                         watch_id == ETCD_OSD_STATE_WATCH_ID)
                         etcd_watches_initialised++;
                     if (etcd_watches_initialised == 4 && this->log_level > 0)
-                        fprintf(stderr, "Successfully subscribed to etcd at %s\n", selected_etcd_address.c_str());
+                        fprintf(stderr, "Successfully subscribed to etcd at %s\n", cur_addr.c_str());
                 }
                 if (data["result"]["canceled"].bool_value())
                 {
@@ -360,15 +368,17 @@ void etcd_state_client_t::start_etcd_watcher()
                         // so we should restart from the beginning if we can
                         if (on_reload_hook != NULL)
                         {
-                            fprintf(stderr, "Revisions before %lu were compacted by etcd, reloading state\n",
-                                data["result"]["compact_revision"].uint64_value());
-                            if (etcd_watch_ws)
+                            // check to not trigger on_reload_hook multiple times
+                            if (etcd_watch_ws != NULL)
                             {
+                                fprintf(stderr, "Revisions before %lu were compacted by etcd, reloading state\n",
+                                    data["result"]["compact_revision"].uint64_value());
                                 http_close(etcd_watch_ws);
                                 etcd_watch_ws = NULL;
+                                etcd_watch_revision = 0;
+                                on_reload_hook();
                             }
-                            etcd_watch_revision = 0;
-                            on_reload_hook();
+                            return;
                         }
                         else
                         {
@@ -415,13 +425,9 @@ void etcd_state_client_t::start_etcd_watcher()
         }
         if (msg->eof)
         {
+            fprintf(stderr, "Disconnected from etcd %s\n", cur_addr.c_str());
             if (cur_addr == selected_etcd_address)
-            {
-                fprintf(stderr, "Disconnected from etcd %s\n", selected_etcd_address.c_str());
                 selected_etcd_address = "";
-            }
-            else
-                fprintf(stderr, "Disconnected from etcd\n");
             if (etcd_watch_ws)
             {
                 http_close(etcd_watch_ws);
@@ -438,6 +444,7 @@ void etcd_state_client_t::start_etcd_watcher()
             else if (etcd_watches_initialised > 0)
             {
                 // Connection was live, retry immediately
+                etcd_watches_initialised = 0;
                 start_etcd_watcher();
             }
         }

src/http_client.cpp

@@ -27,19 +27,10 @@ static void parse_http_headers(std::string & res, http_response_t *parsed);
 
 struct http_co_t
 {
-#ifdef WITH_OPENSSL
-    static SSL_CTX *ssl_ctx = NULL;
-    SSL *ssl_cli = NULL;
-    BIO *ssl_rbio = NULL;
-    BIO *ssl_wbio = NULL;
-    std::vector<uint8_t> encrypted_out;
-#endif
-
     timerfd_manager_t *tfd;
     std::function<void(const http_response_t*)> response_callback;
 
     int request_timeout = 0;
-    bool ssl = false;
     std::string host;
     std::string request;
     std::string ws_outbox;
@@ -55,7 +46,7 @@ struct http_co_t
     int timeout_id = -1;
     int epoll_events = 0;
     int sent = 0;
-    std::vector<uint8_t> rbuf;
+    std::vector<char> rbuf;
     iovec read_iov, send_iov;
     msghdr read_msg = { 0 }, send_msg = { 0 };
     http_response_t parsed;
@@ -268,12 +259,6 @@ void http_response_t::parse_json_response(std::string & error, json11::Json & r)
 
 http_co_t::~http_co_t()
 {
-#ifdef WITH_OPENSSL
-    if (ssl_cli)
-    {
-        SSL_free(ssl_cli);
-    }
-#endif
     close_connection();
 }
 
@@ -290,16 +275,6 @@ void http_co_t::close_connection()
         close(peer_fd);
         peer_fd = -1;
     }
-#ifdef WITH_OPENSSL
-    if (ssl_ctx)
-    {
-        // Frees context, client and bios at once
-        SSL_free(ssl_ctx);
-        ssl_rbio = NULL;
-        ssl_wbio = NULL;
-        ssl_cli = NULL;
-    }
-#endif
     state = HTTP_CO_CLOSED;
     connected_host = "";
     response = "";
@@ -329,27 +304,6 @@ void http_co_t::start_connection()
     }
     fcntl(peer_fd, F_SETFL, fcntl(peer_fd, F_GETFL, 0) | O_NONBLOCK);
     epoll_events = 0;
-#ifdef WITH_OPENSSL
-    // https://wiki.openssl.org/index.php/Hostname_validation
-    if (ssl)
-    {
-        if (!ssl_ctx)
-            ssl_ctx = SSL_CTX_new(TLS_method());
-        ssl_rbio = BIO_new(BIO_s_mem());
-        ssl_wbio = BIO_new(BIO_s_mem());
-        ssl_cli = SSL_new(ssl_ctx);
-        if (!ssl_ctx || !ssl_cli || !ssl_rbio || !ssl_wbio)
-        {
-            parsed = { .error = std::string("openssl initialization failed: ")+ERR_get_error(NULL) };
-            response_callback(&parsed);
-            response_callback = NULL;
-            stackout();
-            return;
-        }
-        SSL_set_connect_state(ssl_cli);
-        SSL_set_bio(ssl_cli, ssl_rbio, ssl_wbio);
-    }
-#endif
     // Finally call connect
     int r = ::connect(peer_fd, (sockaddr*)&addr, sizeof(addr));
     if (r < 0 && errno != EINPROGRESS)
@@ -478,11 +432,11 @@ void http_co_t::submit_read(bool check_timeout)
     stackin();
     int res;
 again:
-    if (rbuf.capacity()-rbuf.size() < READ_BUFFER_SIZE)
+    if (rbuf.size() != READ_BUFFER_SIZE)
     {
-        rbuf.reserve(rbuf.size() + READ_BUFFER_SIZE);
+        rbuf.resize(READ_BUFFER_SIZE);
     }
-    read_iov = { .iov_base = rbuf.data()+rbuf.size(), .iov_len = READ_BUFFER_SIZE };
+    read_iov = { .iov_base = rbuf.data(), .iov_len = READ_BUFFER_SIZE };
     read_msg.msg_iov = &read_iov;
     read_msg.msg_iovlen = 1;
     res = recvmsg(peer_fd, &read_msg, 0);
@@ -512,177 +466,22 @@ again:
     else if (res <= 0)
     {
         // < 0 means error, 0 means EOF
-        on_read_error(res);
+        epoll_events = epoll_events & ~EPOLLIN;
+        if (state == HTTP_CO_HEADERS_RECEIVED)
+            std::swap(parsed.body, response);
+        close_connection();
+        if (res < 0)
+            parsed = { .error = std::string("recvmsg: ")+strerror(-res) };
+        run_cb_and_clear();
     }
     else
     {
-        if (ssl)
-            handle_ssl_read(rbuf);
-        else
-            response += std::string((char*)rbuf.data(), res);
-        rbuf.resize(0);
+        response += std::string(rbuf.data(), res);
         handle_read();
     }
     stackout();
 }
 
-void http_co_t::on_read_error(int res)
-{
-    epoll_events = epoll_events & ~EPOLLIN;
-    if (state == HTTP_CO_HEADERS_RECEIVED)
-        std::swap(parsed.body, response);
-    close_connection();
-    if (res < 0)
-        parsed = { .error = std::string("recvmsg: ")+strerror(-res) };
-    run_cb_and_clear();
-}
-
-int http_co_t::do_ssl_handshake()
-{
-    stackin();
-    int r;
-    while (1)
-    {
-        r = SSL_do_handshake(ssl_cli);
-        if (r == SSL_ERROR_WANT_WRITE)
-        {
-            r = ssl_encrypt();
-            if (r >= 0)
-                submit_send();
-            else
-            {
-                r = -r;
-                break;
-            }
-        }
-        else
-        {
-            if (r == SSL_ERROR_WANT_READ || r == SSL_ERROR_NONE)
-            {
-                // OK or wait until we have more incoming data
-                r = 0;
-            }
-            break;
-        }
-    }
-    stackout();
-    return r;
-}
-
-// Enqueue outbound encrypted TLS data
-int http_co_t::ssl_encrypt()
-{
-    stackin();
-    int queued = 0;
-    while (true)
-    {
-        if (encrypted_out.size() >= encrypted_out.capacity()/2)
-            encrypted_out.reserve(encrypted_out.size() < READ_BUFFER_SIZE ? encrypted_out.size() + READ_BUFFER_SIZE : 2*encrypted_out.size());
-        int r = BIO_read(ssl_wbio, encrypted_out.data()+encrypted_out.size(), encrypted_out.capacity()-encrypted_out.size());
-        if (r > 0)
-        {
-            queued += r;
-            encrypted_out.resize(encrypted_out.size()+r);
-        }
-        else
-        {
-            if (!BIO_should_retry(ssl_wbio))
-                queued = r;
-            break;
-        }
-    }
-    stackout();
-    return queued;
-}
-
-void http_co_t::handle_ssl_write()
-{
-    stackin();
-    int r = 0;
-    while (sent < request.size())
-    {
-        if (!SSL_is_init_finished(ssl_cli))
-        {
-            if (do_ssl_handshake() != 0)
-            {
-                on_read_error(-EIO);
-                break;
-            }
-            if (!SSL_is_init_finished(ssl_cli))
-                break;
-        }
-        else
-        {
-            int n = SSL_write(ssl_cli, request.data()+sent, request.size()-sent);
-            if (n > 0)
-                sent += n;
-            else if (get_sslstatus(ssl_cli, n) == SSLSTATUS_FAIL)
-            {
-                on_read_error(-EIO);
-                break;
-            }
-            else
-                break;
-        }
-        r = ssl_encrypt();
-        if (r >= 0)
-            submit_send();
-        else
-        {
-            on_read_error(-EIO);
-            break;
-        }
-    }
-    stackout();
-}
-
-// Process incoming encrypted TLS data
-void http_co_t::handle_ssl_read()
-{
-    stackin();
-    int size = rbuf.size();
-    int done = 0;
-    while (done < size)
-    {
-        int n = BIO_write(ssl_rbio, rbuf.data()+done, size-done);
-        if (n > 0)
-        {
-            done += n;
-        }
-        if (n <= 0)
-        {
-            on_read_error(-EIO);
-            break;
-        }
-        if (!SSL_is_init_finished(ssl_cli))
-        {
-            if (do_ssl_handshake() != 0)
-            {
-                on_read_error(-EIO);
-                break;
-            }
-            if (!SSL_is_init_finished(ssl_cli))
-                break;
-        }
-        do
-        {
-            if (response.capacity() - response.size() < READ_BUFFER_SIZE)
-                response.reserve(2*response.size() < response.size() + READ_BUFFER_SIZE ? response.size() + READ_BUFFER_SIZE : 2*response.size());
-            n = SSL_read(ssl_cli, response.data() + response.size(), READ_BUFFER_SIZE);
-            if (n <= 0)
-            {
-                n = SSL_get_error(ssl_cli, n);
-                if (n == SSL_ERROR_WANT_READ)
-                    break;
-            }
-        } while (n > 0);
-    }
-    if (done < size)
-        memmove(rbuf.data(), rbuf.data()+done, size-done);
-    rbuf.resize(size-done);
-    stackout();
-}
-
 bool http_co_t::handle_read()
 {
     stackin();

src/msgr_op.cpp

@@ -9,6 +9,10 @@ osd_op_t::~osd_op_t()
 {
     assert(!bs_op);
     assert(!op_data);
+    if (bitmap_buf)
+    {
+        free(bitmap_buf);
+    }
     if (rmw_buf)
     {
         free(rmw_buf);

src/msgr_op.h

@@ -165,6 +165,7 @@ struct osd_op_t
     void *bitmap = NULL;
     unsigned bitmap_len = 0;
     unsigned bmp_data = 0;
+    void *bitmap_buf = NULL;
     void *rmw_buf = NULL;
     osd_primary_op_data_t* op_data = NULL;
     std::function<void(osd_op_t*)> callback;

src/msgr_receive.cpp

@@ -369,7 +369,7 @@ bool osd_messenger_t::handle_reply_hdr(osd_client_t *cl)
         op->buf = malloc_or_die(op->reply.hdr.retval);
         cl->recv_list.push_back(op->buf, op->reply.hdr.retval);
     }
-    else if (op->reply.hdr.opcode == OSD_OP_DESCRIBE && op->reply.hdr.retval > 0)
+    else if (op->reply.hdr.opcode == OSD_OP_DESCRIBE && op->reply.describe.result_bytes > 0)
     {
         delete cl->read_op;
         cl->read_op = op;

src/msgr_send.cpp

@@ -84,9 +84,12 @@ void osd_messenger_t::outbox_push(osd_op_t *cur_op)
     {
         for (int i = 0; i < cur_op->iov.count; i++)
         {
-            assert(cur_op->iov.buf[i].iov_base);
-            to_send_list.push_back(cur_op->iov.buf[i]);
-            to_outbox.push_back((msgr_sendp_t){ .op = cur_op, .flags = 0 });
+            if (cur_op->iov.buf[i].iov_len > 0)
+            {
+                assert(cur_op->iov.buf[i].iov_base);
+                to_send_list.push_back(cur_op->iov.buf[i]);
+                to_outbox.push_back((msgr_sendp_t){ .op = cur_op, .flags = 0 });
+            }
         }
     }
     if (cur_op->req.hdr.opcode == OSD_OP_SEC_READ_BMP)

src/nbd_proxy.cpp

@@ -149,7 +149,7 @@ public:
             "  --dev_num N\n"
             "    Use the specified device /dev/nbdN instead of automatic selection.\n"
             "  --foreground 1\n"
-            "    Stay in foreground, do not daemonize.n",
+            "    Stay in foreground, do not daemonize.\n",
             exe_name, exe_name, exe_name
         );
         exit(0);

src/osd.h

@@ -198,13 +198,14 @@ class osd_t
     void on_change_pg_history_hook(pool_id_t pool_id, pg_num_t pg_num);
     void on_change_etcd_state_hook(std::map<std::string, etcd_kv_t> & changes);
     void on_load_config_hook(json11::Json::object & changes);
+    void on_reload_config_hook(json11::Json::object & changes);
     json11::Json on_load_pgs_checks_hook();
     void on_load_pgs_hook(bool success);
     void bind_socket();
     void acquire_lease();
     json11::Json get_osd_state();
     void create_osd_state();
-    void renew_lease();
+    void renew_lease(bool reload);
     void print_stats();
     void print_slow();
     void reset_stats();

src/osd_cluster.cpp

@@ -70,6 +70,7 @@ void osd_t::init_cluster()
         st_cli.on_load_config_hook = [this](json11::Json::object & cfg) { on_load_config_hook(cfg); };
         st_cli.load_pgs_checks_hook = [this]() { return on_load_pgs_checks_hook(); };
         st_cli.on_load_pgs_hook = [this](bool success) { on_load_pgs_hook(success); };
+        st_cli.on_reload_hook = [this]() { st_cli.load_global_config(); };
         peering_state = OSD_LOADING_PGS;
         st_cli.load_global_config();
     }
@@ -395,6 +396,14 @@ void osd_t::on_load_config_hook(json11::Json::object & global_config)
     parse_config(true);
     bind_socket();
     acquire_lease();
+    st_cli.on_load_config_hook = [this](json11::Json::object & cfg) { on_reload_config_hook(cfg); };
+}
+
+void osd_t::on_reload_config_hook(json11::Json::object & global_config)
+{
+    etcd_global_config = global_config;
+    parse_config(false);
+    renew_lease(true);
 }
 
 // Acquire lease
@@ -424,7 +433,7 @@ void osd_t::acquire_lease()
     );
     tfd->set_timer(etcd_report_interval*1000, true, [this](int timer_id)
     {
-        renew_lease();
+        renew_lease(false);
     });
 }
 
@@ -499,11 +508,11 @@ void osd_t::create_osd_state()
 }
 
 // Renew lease
-void osd_t::renew_lease()
+void osd_t::renew_lease(bool reload)
 {
     st_cli.etcd_call("/lease/keepalive", json11::Json::object {
         { "ID", etcd_lease_id }
-    }, st_cli.etcd_quick_timeout, 0, 0, [this](std::string err, json11::Json data)
+    }, st_cli.etcd_quick_timeout, 0, 0, [this, reload](std::string err, json11::Json data)
     {
         if (err == "" && data["result"]["TTL"].string_value() == "")
         {
@@ -522,15 +531,20 @@ void osd_t::renew_lease()
                 force_stop(1);
             }
             // Retry
-            tfd->set_timer(st_cli.etcd_quick_timeout, false, [this](int timer_id)
+            tfd->set_timer(st_cli.etcd_quick_timeout, false, [this, reload](int timer_id)
             {
-                renew_lease();
+                renew_lease(reload);
             });
         }
         else
         {
             etcd_failed_attempts = 0;
             report_statistics();
+            // Reload PGs
+            if (reload && run_primary)
+            {
+                st_cli.load_pgs();
+            }
         }
     });
 }
@@ -560,7 +574,6 @@ void osd_t::force_stop(int exitcode)
 
 json11::Json osd_t::on_load_pgs_checks_hook()
 {
-    assert(this->pgs.size() == 0);
     json11::Json::array checks = {
         json11::Json::object {
             { "target", "LEASE" },

src/osd_ops.h

@@ -220,7 +220,7 @@ struct __attribute__((__packed__)) osd_reply_rw_t
     // for reads: bitmap length
     uint32_t bitmap_len;
     uint32_t pad0;
-    // for reads: object version
+    // for reads and writes: object version
     uint64_t version;
 };
 

src/osd_primary.cpp

@@ -87,8 +87,7 @@ bool osd_t::prepare_primary_rw(osd_op_t *cur_op)
         // - op_data
         1, sizeof(osd_primary_op_data_t) +
         // - stripes
-        // - resulting bitmap buffers
-        stripe_count * (clean_entry_bitmap_size + sizeof(osd_rmw_stripe_t)) +
+        stripe_count * sizeof(osd_rmw_stripe_t) +
         chain_size * (
             // - copy of the chain
             sizeof(inode_t) +
@@ -110,11 +109,12 @@ bool osd_t::prepare_primary_rw(osd_op_t *cur_op)
     op_data->pg_size = pg_it->second.pg_size;
     cur_op->op_data = op_data;
     split_stripes(pg_data_size, bs_block_size, (uint32_t)(cur_op->req.rw.offset - oid.stripe), cur_op->req.rw.len, op_data->stripes);
-    // Allocate bitmaps along with stripes to avoid extra allocations and fragmentation
+    // Resulting bitmaps have to survive op_data and be freed with the op itself
+    assert(!cur_op->bitmap_buf);
+    cur_op->bitmap_buf = calloc_or_die(1, clean_entry_bitmap_size * stripe_count);
     for (int i = 0; i < stripe_count; i++)
     {
-        op_data->stripes[i].bmp_buf = data_buf;
-        data_buf = (uint8_t*)data_buf + clean_entry_bitmap_size;
+        op_data->stripes[i].bmp_buf = (uint8_t*)cur_op->bitmap_buf + clean_entry_bitmap_size * i;
     }
     op_data->chain_size = chain_size;
     if (chain_size > 0)
@@ -129,16 +129,19 @@ bool osd_t::prepare_primary_rw(osd_op_t *cur_op)
         data_buf = (uint8_t*)data_buf + chain_size * (pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 0 : pg_it->second.pg_size);
         // Copy chain
         int chain_num = 0;
-        op_data->read_chain[chain_num++] = cur_op->req.rw.inode;
+        op_data->read_chain[chain_num] = cur_op->req.rw.inode;
+        op_data->chain_states[chain_num] = NULL;
+        chain_num++;
         auto inode_it = st_cli.inode_config.find(cur_op->req.rw.inode);
         while (inode_it != st_cli.inode_config.end() && inode_it->second.parent_id &&
             INODE_POOL(inode_it->second.parent_id) == pg_it->second.pool_id &&
             // Check for loops
             inode_it->second.parent_id != cur_op->req.rw.inode)
         {
-            op_data->read_chain[chain_num++] = inode_it->second.parent_id;
-            op_data->chain_states[chain_num++] = NULL;
+            op_data->read_chain[chain_num] = inode_it->second.parent_id;
+            op_data->chain_states[chain_num] = NULL;
             inode_it = st_cli.inode_config.find(inode_it->second.parent_id);
+            chain_num++;
         }
     }
     pg_it->second.inflight++;
@@ -644,12 +647,6 @@ void osd_t::continue_primary_del(osd_op_t *cur_op)
     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;

src/osd_primary_write.cpp

@@ -83,11 +83,13 @@ retry_1:
             // Object is degraded/misplaced and will be moved to <write_osd_set>
             op_data->stripes[0].read_start = 0;
             op_data->stripes[0].read_end = bs_block_size;
+            assert(!cur_op->rmw_buf);
             cur_op->rmw_buf = op_data->stripes[0].read_buf = memalign_or_die(MEM_ALIGNMENT, bs_block_size);
         }
     }
     else
     {
+        assert(!cur_op->rmw_buf);
         cur_op->rmw_buf = calc_rmw(cur_op->buf, op_data->stripes, op_data->prev_set,
             pg.pg_size, op_data->pg_data_size, pg.pg_cursize, pg.cur_set.data(), bs_block_size, clean_entry_bitmap_size);
         if (!cur_op->rmw_buf)

src/qemu_driver.c

@@ -35,6 +35,11 @@
 #define qdict_put_str(options, name, value) qdict_put_obj(options, name, QOBJECT(qstring_from_str(value)))
 #define qobject_unref QDECREF
 #endif
+#if QEMU_VERSION_MAJOR == 4 && QEMU_VERSION_MINOR >= 2 || QEMU_VERSION_MAJOR > 4
+#include "sysemu/replay.h"
+#else
+#include "sysemu/sysemu.h"
+#endif
 
 #include "vitastor_c.h"
 
@@ -48,9 +53,13 @@ void DSO_STAMP_FUN(void)
 }
 #endif
 
+typedef struct VitastorFdData VitastorFdData;
+
 typedef struct VitastorClient
 {
     void *proxy;
+    int uring_eventfd;
+
     void *watch;
     char *config_path;
     char *etcd_host;
@@ -67,12 +76,24 @@ typedef struct VitastorClient
     int rdma_gid_index;
     int rdma_mtu;
     QemuMutex mutex;
+    AioContext *ctx;
+    VitastorFdData **fds;
+    int fd_count, fd_alloc;
+    int bh_uring_scheduled;
 
     uint64_t last_bitmap_inode, last_bitmap_offset, last_bitmap_len;
     uint32_t last_bitmap_granularity;
     uint8_t *last_bitmap;
 } VitastorClient;
 
+typedef struct VitastorFdData
+{
+    VitastorClient *cli;
+    int fd;
+    IOHandler *fd_read, *fd_write;
+    void *opaque;
+} VitastorFdData;
+
 typedef struct VitastorRPC
 {
     BlockDriverState *bs;
@@ -83,10 +104,21 @@ typedef struct VitastorRPC
     uint64_t inode, offset, len;
     uint32_t bitmap_granularity;
     uint8_t *bitmap;
+#if QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR < 8
+    QEMUBH *bh;
+#endif
 } VitastorRPC;
 
+#if QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR < 8
+typedef struct VitastorBH
+{
+    VitastorClient *cli;
+    QEMUBH *bh;
+} VitastorBH;
+#endif
+
 static void vitastor_co_init_task(BlockDriverState *bs, VitastorRPC *task);
-static void vitastor_co_generic_bh_cb(void *opaque, long retval);
+static void vitastor_co_generic_cb(void *opaque, long retval);
 static void vitastor_co_read_cb(void *opaque, long retval, uint64_t version);
 static void vitastor_close(BlockDriverState *bs);
 
@@ -202,6 +234,57 @@ out:
     return;
 }
 
+#if defined VITASTOR_C_API_VERSION && VITASTOR_C_API_VERSION >= 2
+static void vitastor_uring_handler(void *opaque)
+{
+    VitastorClient *client = (VitastorClient*)opaque;
+    qemu_mutex_lock(&client->mutex);
+    client->bh_uring_scheduled = 0;
+    do
+    {
+        vitastor_c_uring_handle_events(client->proxy);
+    } while (vitastor_c_uring_has_work(client->proxy));
+    qemu_mutex_unlock(&client->mutex);
+}
+
+#if QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR < 8
+static void vitastor_bh_uring_handler(void *opaque)
+{
+    VitastorBH *vbh = opaque;
+    vitastor_bh_handler(vbh->cli);
+    qemu_bh_delete(vbh->bh);
+    free(vbh);
+}
+#endif
+
+static void vitastor_schedule_uring_handler(VitastorClient *client)
+{
+    void *opaque = client;
+    if (client->uring_eventfd >= 0 && !client->bh_uring_scheduled)
+    {
+        client->bh_uring_scheduled = 1;
+#if QEMU_VERSION_MAJOR > 4 || QEMU_VERSION_MAJOR == 4 && QEMU_VERSION_MINOR >= 2
+        replay_bh_schedule_oneshot_event(client->ctx, vitastor_uring_handler, opaque);
+#elif QEMU_VERSION_MAJOR >= 3 || QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR >= 8
+        aio_bh_schedule_oneshot(client->ctx, vitastor_uring_handler, opaque);
+#else
+        VitastorBH *vbh = (VitastorBH*)malloc(sizeof(VitastorBH));
+        vbh->cli = client;
+#if QEMU_VERSION_MAJOR >= 2
+        vbh->bh = aio_bh_new(bdrv_get_aio_context(task->bs), vitastor_bh_uring_handler, vbh);
+#else
+        vbh->bh = qemu_bh_new(vitastor_bh_uring_handler, vbh);
+#endif
+        qemu_bh_schedule(vbh->bh);
+#endif
+    }
+}
+#else
+static void vitastor_schedule_uring_handler(VitastorClient *client)
+{
+}
+#endif
+
 static void coroutine_fn vitastor_co_get_metadata(VitastorRPC *task)
 {
     BlockDriverState *bs = task->bs;
@@ -209,7 +292,8 @@ static void coroutine_fn vitastor_co_get_metadata(VitastorRPC *task)
     task->co = qemu_coroutine_self();
 
     qemu_mutex_lock(&client->mutex);
-    vitastor_c_watch_inode(client->proxy, client->image, vitastor_co_generic_bh_cb, task);
+    vitastor_c_watch_inode(client->proxy, client->image, vitastor_co_generic_cb, task);
+    vitastor_schedule_uring_handler(client);
     qemu_mutex_unlock(&client->mutex);
 
     while (!task->complete)
@@ -218,14 +302,32 @@ static void coroutine_fn vitastor_co_get_metadata(VitastorRPC *task)
     }
 }
 
-// FIXME: Fix thread safety of the driver - now it segfaults when iothread is enabled in QEMU
-static void vitastor_aio_set_fd_handler(void *ctx, int fd, int unused1, IOHandler *fd_read, IOHandler *fd_write, void *unused2, void *opaque)
+static void vitastor_aio_fd_read(void *fddv)
+{
+    VitastorFdData *fdd = (VitastorFdData*)fddv;
+    qemu_mutex_lock(&fdd->cli->mutex);
+    fdd->fd_read(fdd->opaque);
+    vitastor_schedule_uring_handler(fdd->cli);
+    qemu_mutex_unlock(&fdd->cli->mutex);
+}
+
+static void vitastor_aio_fd_write(void *fddv)
+{
+    VitastorFdData *fdd = (VitastorFdData*)fddv;
+    qemu_mutex_lock(&fdd->cli->mutex);
+    fdd->fd_write(fdd->opaque);
+    vitastor_schedule_uring_handler(fdd->cli);
+    qemu_mutex_unlock(&fdd->cli->mutex);
+}
+
+static void universal_aio_set_fd_handler(AioContext *ctx, int fd, IOHandler *fd_read, IOHandler *fd_write, void *opaque)
 {
     aio_set_fd_handler(ctx, fd,
 #if QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR >= 5 || QEMU_VERSION_MAJOR >= 3
         0 /*is_external*/,
 #endif
-        fd_read, fd_write,
+        fd_read,
+        fd_write,
 #if QEMU_VERSION_MAJOR == 1 && QEMU_VERSION_MINOR <= 6 || QEMU_VERSION_MAJOR < 1
         NULL /*io_flush*/,
 #endif
@@ -238,6 +340,53 @@ static void vitastor_aio_set_fd_handler(void *ctx, int fd, int unused1, IOHandle
         opaque);
 }
 
+static void vitastor_aio_set_fd_handler(void *vcli, int fd, int unused1, IOHandler *fd_read, IOHandler *fd_write, void *unused2, void *opaque)
+{
+    VitastorClient *client = (VitastorClient*)vcli;
+    VitastorFdData *fdd = NULL;
+    int i;
+    for (i = 0; i < client->fd_count; i++)
+    {
+        if (client->fds[i]->fd == fd)
+        {
+            if (fd_read || fd_write)
+            {
+                fdd = client->fds[i];
+                fdd->opaque = opaque;
+                fdd->fd_read = fd_read;
+                fdd->fd_write = fd_write;
+            }
+            else
+            {
+                for (int j = i+1; j < client->fd_count; j++)
+                    client->fds[j-1] = client->fds[j];
+                client->fd_count--;
+            }
+            break;
+        }
+    }
+    if ((fd_read || fd_write) && !fdd)
+    {
+        fdd = (VitastorFdData*)malloc(sizeof(VitastorFdData));
+        fdd->cli = client;
+        fdd->fd = fd;
+        fdd->fd_read = fd_read;
+        fdd->fd_write = fd_write;
+        fdd->opaque = opaque;
+        if (client->fd_count >= client->fd_alloc)
+        {
+            client->fd_alloc = client->fd_alloc*2;
+            if (client->fd_alloc < 16)
+                client->fd_alloc = 16;
+            client->fds = (VitastorFdData**)realloc(client->fds, sizeof(VitastorFdData*) * client->fd_alloc);
+        }
+        client->fds[client->fd_count++] = fdd;
+    }
+    universal_aio_set_fd_handler(
+        client->ctx, fd, fd_read ? vitastor_aio_fd_read : NULL, fd_write ? vitastor_aio_fd_write : NULL, fdd
+    );
+}
+
 static int vitastor_file_open(BlockDriverState *bs, QDict *options, int flags, Error **errp)
 {
     VitastorRPC task;
@@ -255,10 +404,36 @@ static int vitastor_file_open(BlockDriverState *bs, QDict *options, int flags, E
     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_mtu = qdict_get_try_int(options, "rdma-mtu", 0);
-    client->proxy = vitastor_c_create_qemu(
-        vitastor_aio_set_fd_handler, bdrv_get_aio_context(bs), client->config_path, client->etcd_host, client->etcd_prefix,
+    client->ctx = bdrv_get_aio_context(bs);
+#if defined VITASTOR_C_API_VERSION && VITASTOR_C_API_VERSION >= 2
+    client->proxy = vitastor_c_create_qemu_uring(
+        vitastor_aio_set_fd_handler, client, 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
     );
+    if (!client->proxy)
+    {
+        fprintf(stderr, "vitastor: failed to create io_uring: %s - I/O will be slower\n", strerror(errno));
+        client->uring_eventfd = -1;
+#endif
+        client->proxy = vitastor_c_create_qemu(
+            vitastor_aio_set_fd_handler, client, 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
+        );
+#if defined VITASTOR_C_API_VERSION && VITASTOR_C_API_VERSION >= 2
+    }
+    else
+    {
+        client->uring_eventfd = vitastor_c_uring_register_eventfd(client->proxy);
+        if (client->uring_eventfd < 0)
+        {
+            fprintf(stderr, "vitastor: failed to create io_uring eventfd: %s\n", strerror(errno));
+            error_setg(errp, "failed to create io_uring eventfd");
+            vitastor_close(bs);
+            return -1;
+        }
+        universal_aio_set_fd_handler(client->ctx, client->uring_eventfd, vitastor_uring_handler, NULL, client);
+    }
+#endif
     image = client->image = g_strdup(qdict_get_try_str(options, "image"));
     client->readonly = (flags & BDRV_O_RDWR) ? 1 : 0;
     // Get image metadata (size and readonly flag) or just wait until the client is ready
@@ -338,6 +513,12 @@ static void vitastor_close(BlockDriverState *bs)
 {
     VitastorClient *client = bs->opaque;
     vitastor_c_destroy(client->proxy);
+    if (client->fds)
+    {
+        free(client->fds);
+        client->fds = NULL;
+        client->fd_alloc = client->fd_count = 0;
+    }
     qemu_mutex_destroy(&client->mutex);
     if (client->config_path)
         g_free(client->config_path);
@@ -454,25 +635,44 @@ static void vitastor_co_init_task(BlockDriverState *bs, VitastorRPC *task)
     };
 }
 
-static void vitastor_co_generic_bh_cb(void *opaque, long retval)
+static void vitastor_co_generic_bh_cb(void *opaque)
 {
     VitastorRPC *task = opaque;
-    task->ret = retval;
     task->complete = 1;
     if (qemu_coroutine_self() != task->co)
     {
 #if QEMU_VERSION_MAJOR >= 3 || QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR > 8
         aio_co_wake(task->co);
 #else
+#if QEMU_VERSION_MAJOR == 2
+        qemu_bh_delete(task->bh);
+#endif
         qemu_coroutine_enter(task->co, NULL);
         qemu_aio_release(task);
 #endif
     }
 }
 
+static void vitastor_co_generic_cb(void *opaque, long retval)
+{
+    VitastorRPC *task = opaque;
+    task->ret = retval;
+#if QEMU_VERSION_MAJOR > 4 || QEMU_VERSION_MAJOR == 4 && QEMU_VERSION_MINOR >= 2
+    replay_bh_schedule_oneshot_event(bdrv_get_aio_context(task->bs), vitastor_co_generic_bh_cb, opaque);
+#elif QEMU_VERSION_MAJOR >= 3 || QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR >= 8
+    aio_bh_schedule_oneshot(bdrv_get_aio_context(task->bs), vitastor_co_generic_bh_cb, opaque);
+#elif QEMU_VERSION_MAJOR >= 2
+    task->bh = aio_bh_new(bdrv_get_aio_context(task->bs), vitastor_co_generic_bh_cb, opaque);
+    qemu_bh_schedule(task->bh);
+#else
+    task->bh = qemu_bh_new(vitastor_co_generic_bh_cb, opaque);
+    qemu_bh_schedule(task->bh);
+#endif
+}
+
 static void vitastor_co_read_cb(void *opaque, long retval, uint64_t version)
 {
-    vitastor_co_generic_bh_cb(opaque, retval);
+    vitastor_co_generic_cb(opaque, retval);
 }
 
 static int coroutine_fn vitastor_co_preadv(BlockDriverState *bs,
@@ -491,6 +691,7 @@ static int coroutine_fn vitastor_co_preadv(BlockDriverState *bs,
     uint64_t inode = client->watch ? vitastor_c_inode_get_num(client->watch) : client->inode;
     qemu_mutex_lock(&client->mutex);
     vitastor_c_read(client->proxy, inode, offset, bytes, iov->iov, iov->niov, vitastor_co_read_cb, &task);
+    vitastor_schedule_uring_handler(client);
     qemu_mutex_unlock(&client->mutex);
 
     while (!task.complete)
@@ -523,7 +724,8 @@ static int coroutine_fn vitastor_co_pwritev(BlockDriverState *bs,
 
     uint64_t inode = client->watch ? vitastor_c_inode_get_num(client->watch) : client->inode;
     qemu_mutex_lock(&client->mutex);
-    vitastor_c_write(client->proxy, inode, offset, bytes, 0, iov->iov, iov->niov, vitastor_co_generic_bh_cb, &task);
+    vitastor_c_write(client->proxy, inode, offset, bytes, 0, iov->iov, iov->niov, vitastor_co_generic_cb, &task);
+    vitastor_schedule_uring_handler(client);
     qemu_mutex_unlock(&client->mutex);
 
     while (!task.complete)
@@ -541,7 +743,6 @@ static void vitastor_co_read_bitmap_cb(void *opaque, long retval, uint8_t *bitma
     VitastorRPC *task = opaque;
     VitastorClient *client = task->bs->opaque;
     task->ret = retval;
-    task->complete = 1;
     if (retval >= 0)
     {
         task->bitmap = bitmap;
@@ -553,15 +754,17 @@ static void vitastor_co_read_bitmap_cb(void *opaque, long retval, uint8_t *bitma
             client->last_bitmap = bitmap;
         }
     }
-    if (qemu_coroutine_self() != task->co)
-    {
-#if QEMU_VERSION_MAJOR >= 3 || QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR > 8
-        aio_co_wake(task->co);
+#if QEMU_VERSION_MAJOR > 4 || QEMU_VERSION_MAJOR == 4 && QEMU_VERSION_MINOR >= 2
+    replay_bh_schedule_oneshot_event(bdrv_get_aio_context(task->bs), vitastor_co_generic_bh_cb, opaque);
+#elif QEMU_VERSION_MAJOR >= 3 || QEMU_VERSION_MAJOR == 2 && QEMU_VERSION_MINOR >= 8
+    aio_bh_schedule_oneshot(bdrv_get_aio_context(task->bs), vitastor_co_generic_bh_cb, opaque);
+#elif QEMU_VERSION_MAJOR >= 2
+    task->bh = aio_bh_new(bdrv_get_aio_context(task->bs), vitastor_co_generic_bh_cb, opaque);
+    qemu_bh_schedule(task->bh);
 #else
-        qemu_coroutine_enter(task->co, NULL);
-        qemu_aio_release(task);
+    task->bh = qemu_bh_new(vitastor_co_generic_bh_cb, opaque);
+    qemu_bh_schedule(task->bh);
 #endif
-    }
 }
 
 static int coroutine_fn vitastor_co_block_status(
@@ -602,6 +805,7 @@ static int coroutine_fn vitastor_co_block_status(
         task.bitmap = client->last_bitmap = NULL;
         qemu_mutex_lock(&client->mutex);
         vitastor_c_read_bitmap(client->proxy, task.inode, task.offset, task.len, !client->skip_parents, vitastor_co_read_bitmap_cb, &task);
+        vitastor_schedule_uring_handler(client);
         qemu_mutex_unlock(&client->mutex);
         while (!task.complete)
         {
@@ -687,7 +891,8 @@ static int coroutine_fn vitastor_co_flush(BlockDriverState *bs)
     vitastor_co_init_task(bs, &task);
 
     qemu_mutex_lock(&client->mutex);
-    vitastor_c_sync(client->proxy, vitastor_co_generic_bh_cb, &task);
+    vitastor_c_sync(client->proxy, vitastor_co_generic_cb, &task);
+    vitastor_schedule_uring_handler(client);
     qemu_mutex_unlock(&client->mutex);
 
     while (!task.complete)

src/ringloop.cpp

@@ -2,9 +2,12 @@
 // License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
 
 #include <stdlib.h>
+#include <unistd.h>
 
 #include <stdexcept>
 
+#include <sys/eventfd.h>
+
 #include "ringloop.h"
 
 ring_loop_t::ring_loop_t(int qd)
@@ -32,6 +35,10 @@ ring_loop_t::~ring_loop_t()
     free(free_ring_data);
     free(ring_datas);
     io_uring_queue_exit(&ring);
+    if (ring_eventfd)
+    {
+        close(ring_eventfd);
+    }
 }
 
 void ring_loop_t::register_consumer(ring_consumer_t *consumer)
@@ -127,3 +134,24 @@ int ring_loop_t::sqes_left()
     }
     return left;
 }
+
+int ring_loop_t::register_eventfd()
+{
+    if (ring_eventfd >= 0)
+    {
+        return ring_eventfd;
+    }
+    ring_eventfd = eventfd(0, EFD_CLOEXEC|EFD_NONBLOCK);
+    if (ring_eventfd < 0)
+    {
+        return -errno;
+    }
+    int r = io_uring_register_eventfd(&ring, ring_eventfd);
+    if (r < 0)
+    {
+        close(ring_eventfd);
+        ring_eventfd = -1;
+        return r;
+    }
+    return ring_eventfd;
+}

src/ringloop.h

@@ -126,11 +126,13 @@ class ring_loop_t
     unsigned free_ring_data_ptr;
     bool loop_again;
     struct io_uring ring;
+    int ring_eventfd = -1;
 public:
     ring_loop_t(int qd);
     ~ring_loop_t();
     void register_consumer(ring_consumer_t *consumer);
     void unregister_consumer(ring_consumer_t *consumer);
+    int register_eventfd();
 
     inline struct io_uring_sqe* get_sqe()
     {

src/vitastor.pc.in

@@ -6,7 +6,7 @@ includedir=${prefix}/@CMAKE_INSTALL_INCLUDEDIR@
 
 Name: Vitastor
 Description: Vitastor client library
-Version: 0.9.2
+Version: 0.9.4
 Libs: -L${libdir} -lvitastor_client
 Cflags: -I${includedir}
 

src/vitastor_c.cpp

@@ -5,6 +5,7 @@
 // Also acts as a C-C++ proxy for the QEMU driver (QEMU headers don't compile with g++)
 
 #include <sys/epoll.h>
+#include <sys/eventfd.h>
 
 #include "ringloop.h"
 #include "epoll_manager.h"
@@ -25,6 +26,7 @@ struct vitastor_c
     epoll_manager_t *epmgr = NULL;
     timerfd_manager_t *tfd = NULL;
     cluster_client_t *cli = NULL;
+    int uring_eventfd = -1;
 
     QEMUSetFDHandler *aio_set_fd_handler = NULL;
     void *aio_ctx = NULL;
@@ -70,14 +72,8 @@ static void vitastor_c_write_handler(void *opaque)
     data->callback(data->fd, EPOLLOUT);
 }
 
-vitastor_c *vitastor_c_create_qemu(QEMUSetFDHandler *aio_set_fd_handler, void *aio_context,
-    const char *config_path, const char *etcd_host, const char *etcd_prefix,
-    int use_rdma, const char *rdma_device, int rdma_port_num, int rdma_gid_index, int rdma_mtu, int log_level)
+static vitastor_c *vitastor_c_create_qemu_common(QEMUSetFDHandler *aio_set_fd_handler, void *aio_context)
 {
-    json11::Json cfg_json = vitastor_c_common_config(
-        config_path, etcd_host, etcd_prefix, use_rdma,
-        rdma_device, rdma_port_num, rdma_gid_index, rdma_mtu, log_level
-    );
     vitastor_c *self = new vitastor_c;
     self->aio_set_fd_handler = aio_set_fd_handler;
     self->aio_ctx = aio_context;
@@ -95,24 +91,77 @@ vitastor_c *vitastor_c_create_qemu(QEMUSetFDHandler *aio_set_fd_handler, void *a
             self->aio_set_fd_handler(self->aio_ctx, fd, false, NULL, NULL, NULL, NULL);
         }
     });
-    self->cli = new cluster_client_t(NULL, self->tfd, cfg_json);
     return self;
 }
 
-vitastor_c *vitastor_c_create_uring(const char *config_path, const char *etcd_host, const char *etcd_prefix,
+vitastor_c *vitastor_c_create_qemu(QEMUSetFDHandler *aio_set_fd_handler, void *aio_context,
+    const char *config_path, const char *etcd_host, const char *etcd_prefix,
     int use_rdma, const char *rdma_device, int rdma_port_num, int rdma_gid_index, int rdma_mtu, int log_level)
 {
+    json11::Json cfg_json = vitastor_c_common_config(
+        config_path, etcd_host, etcd_prefix, use_rdma,
+        rdma_device, rdma_port_num, rdma_gid_index, rdma_mtu, log_level
+    );
+    auto self = vitastor_c_create_qemu_common(aio_set_fd_handler, aio_context);
+    self->cli = new cluster_client_t(NULL, self->tfd, cfg_json);
+    return self;
+}
+
+vitastor_c *vitastor_c_create_qemu_uring(QEMUSetFDHandler *aio_set_fd_handler, void *aio_context,
+    const char *config_path, const char *etcd_host, const char *etcd_prefix,
+    int use_rdma, const char *rdma_device, int rdma_port_num, int rdma_gid_index, int rdma_mtu, int log_level)
+{
+    ring_loop_t *ringloop = NULL;
+    try
+    {
+        ringloop = new ring_loop_t(512);
+    }
+    catch (std::exception & e)
+    {
+        return NULL;
+    }
+    json11::Json cfg_json = vitastor_c_common_config(
+        config_path, etcd_host, etcd_prefix, use_rdma,
+        rdma_device, rdma_port_num, rdma_gid_index, rdma_mtu, log_level
+    );
+    auto self = vitastor_c_create_qemu_common(aio_set_fd_handler, aio_context);
+    self->ringloop = ringloop;
+    self->cli = new cluster_client_t(self->ringloop, self->tfd, cfg_json);
+    return self;
+}
+
+vitastor_c *vitastor_c_create_uring(const char *config_path, const char *etcd_host, const char *etcd_prefix,
+    int use_rdma, const char *rdma_device, int rdma_port_num, int rdma_gid_index, int rdma_mtu, int log_level)
+{
+    ring_loop_t *ringloop = NULL;
+    try
+    {
+        ringloop = new ring_loop_t(512);
+    }
+    catch (std::exception & e)
+    {
+        return NULL;
+    }
     json11::Json cfg_json = vitastor_c_common_config(
         config_path, etcd_host, etcd_prefix, use_rdma,
         rdma_device, rdma_port_num, rdma_gid_index, rdma_mtu, log_level
     );
     vitastor_c *self = new vitastor_c;
-    self->ringloop = new ring_loop_t(512);
+    self->ringloop = ringloop;
     self->epmgr = new epoll_manager_t(self->ringloop);
     self->cli = new cluster_client_t(self->ringloop, self->epmgr->tfd, cfg_json);
     return self;
 }
 
+int vitastor_c_uring_register_eventfd(vitastor_c *client)
+{
+    if (!client->ringloop)
+    {
+        return -EINVAL;
+    }
+    return client->ringloop->register_eventfd();
+}
+
 vitastor_c *vitastor_c_create_uring_json(const char **options, int options_len)
 {
     json11::Json::object cfg;
@@ -166,6 +215,11 @@ void vitastor_c_uring_wait_events(vitastor_c *client)
     client->ringloop->wait();
 }
 
+int vitastor_c_uring_has_work(vitastor_c *client)
+{
+    return client->ringloop->has_work();
+}
+
 void vitastor_c_read(vitastor_c *client, uint64_t inode, uint64_t offset, uint64_t len,
     struct iovec *iov, int iovcnt, VitastorReadHandler cb, void *opaque)
 {

src/vitastor_c.h

@@ -7,7 +7,7 @@
 #define VITASTOR_QEMU_PROXY_H
 
 // C API wrapper version
-#define VITASTOR_C_API_VERSION 1
+#define VITASTOR_C_API_VERSION 2
 
 #ifndef POOL_ID_BITS
 #define POOL_ID_BITS 16
@@ -34,14 +34,19 @@ typedef void QEMUSetFDHandler(void *ctx, int fd, int is_external, IOHandler *fd_
 vitastor_c *vitastor_c_create_qemu(QEMUSetFDHandler *aio_set_fd_handler, void *aio_context,
     const char *config_path, const char *etcd_host, const char *etcd_prefix,
     int use_rdma, const char *rdma_device, int rdma_port_num, int rdma_gid_index, int rdma_mtu, int log_level);
+vitastor_c *vitastor_c_create_qemu_uring(QEMUSetFDHandler *aio_set_fd_handler, void *aio_context,
+    const char *config_path, const char *etcd_host, const char *etcd_prefix,
+    int use_rdma, const char *rdma_device, int rdma_port_num, int rdma_gid_index, int rdma_mtu, int log_level);
 vitastor_c *vitastor_c_create_uring(const char *config_path, const char *etcd_host, const char *etcd_prefix,
     int use_rdma, const char *rdma_device, int rdma_port_num, int rdma_gid_index, int rdma_mtu, int log_level);
 vitastor_c *vitastor_c_create_uring_json(const char **options, int options_len);
 void vitastor_c_destroy(vitastor_c *client);
 int vitastor_c_is_ready(vitastor_c *client);
+int vitastor_c_uring_register_eventfd(vitastor_c *client);
 void vitastor_c_uring_wait_ready(vitastor_c *client);
 void vitastor_c_uring_handle_events(vitastor_c *client);
 void vitastor_c_uring_wait_events(vitastor_c *client);
+int vitastor_c_uring_has_work(vitastor_c *client);
 void vitastor_c_read(vitastor_c *client, uint64_t inode, uint64_t offset, uint64_t len,
     struct iovec *iov, int iovcnt, VitastorReadHandler cb, void *opaque);
 void vitastor_c_write(vitastor_c *client, uint64_t inode, uint64_t offset, uint64_t len, uint64_t check_version,

tests/run_3osds.sh

@@ -36,9 +36,6 @@ for i in $(seq 1 $OSD_COUNT); do
     start_osd $i
 done
 
-cd mon
-npm install
-cd ..
 (while true; do node mon/mon-main.js --etcd_url $ETCD_URL --etcd_prefix "/vitastor" --verbose 1 || true; done) &>./testdata/mon.log &
 MON_PID=$!
 
@@ -85,7 +82,9 @@ wait_up()
     done
 }
 
-wait_up 60
+if [[ $OSD_COUNT -gt 0 ]]; then
+    wait_up 60
+fi
 
 try_reweight()
 {

tests/run_tests.sh

@@ -16,29 +16,35 @@ SCHEME=ec ./test_change_pg_count.sh
 
 ./test_etcd_fail.sh
 
-./test_failure_domain.sh
-
 ./test_interrupted_rebalance.sh
 IMMEDIATE_COMMIT=1 ./test_interrupted_rebalance.sh
 SCHEME=ec ./test_interrupted_rebalance.sh
 SCHEME=ec IMMEDIATE_COMMIT=1 ./test_interrupted_rebalance.sh
 
+./test_failure_domain.sh
+
+./test_snapshot.sh
+SCHEME=ec ./test_snapshot.sh
+
 ./test_minsize_1.sh
 
 ./test_move_reappear.sh
 
+./test_rm.sh
+
+./test_snapshot_chain.sh
+SCHEME=ec ./test_snapshot_chain.sh
+
+./test_snapshot_down.sh
+SCHEME=ec ./test_snapshot_down.sh
+
+./test_splitbrain.sh
+
 ./test_rebalance_verify.sh
 IMMEDIATE_COMMIT=1 ./test_rebalance_verify.sh
 SCHEME=ec ./test_rebalance_verify.sh
 SCHEME=ec IMMEDIATE_COMMIT=1 ./test_rebalance_verify.sh
 
-./test_rm.sh
-
-./test_snapshot.sh
-SCHEME=ec ./test_snapshot.sh
-
-./test_splitbrain.sh
-
 ./test_write.sh
 SCHEME=xor ./test_write.sh
 

tests/test_failure_domain.sh

@@ -15,9 +15,6 @@ $ETCDCTL put /vitastor/osd/stats/7 '{"host":"host4","size":1073741824,"time":"'$
 $ETCDCTL put /vitastor/osd/stats/8 '{"host":"host4","size":1073741824,"time":"'$TIME'"}'
 $ETCDCTL put /vitastor/config/pools '{"1":{"name":"testpool","scheme":"replicated","pg_size":2,"pg_minsize":1,"pg_count":4,"failure_domain":"rack"}}'
 
-cd mon
-npm install
-cd ..
 node mon/mon-main.js --etcd_url $ETCD_URL --etcd_prefix "/vitastor" &>./testdata/mon.log &
 MON_PID=$!
 

tests/test_snapshot_chain.sh

@@ -0,0 +1,48 @@
+#!/bin/bash -ex
+
+. `dirname $0`/run_3osds.sh
+check_qemu
+
+# Test multiple snapshots
+
+build/src/vitastor-cli --etcd_address $ETCD_URL create -s 32M testchain
+
+LD_PRELOAD="build/src/libfio_vitastor.so" \
+    fio -thread -name=test -ioengine=build/src/libfio_vitastor.so -bs=4M -direct=1 -iodepth=1 -fsync=1 -rw=write \
+        -etcd=$ETCD_URL -image=testchain -mirror_file=./testdata/mirror.bin
+
+for i in {1..10}; do
+    # Create a snapshot
+    build/src/vitastor-cli --etcd_address $ETCD_URL snap-create testchain@$i
+    # Check that the new snapshot is see-through
+    qemu-img convert -p \
+        -f raw "vitastor:etcd_host=127.0.0.1\:$ETCD_PORT/v3:image=testchain" \
+        -O raw ./testdata/check.bin
+    cmp ./testdata/check.bin ./testdata/mirror.bin
+    # Write something to it
+    LD_PRELOAD="build/src/libfio_vitastor.so" \
+    fio -thread -name=test -ioengine=build/src/libfio_vitastor.so -bs=4k -direct=1 -iodepth=1 -fsync=32 -rw=randwrite \
+        -randrepeat=$((i <= 2)) -buffer_pattern=0x$((10+i))$((10+i))$((10+i))$((10+i)) \
+        -etcd=$ETCD_URL -image=testchain -number_ios=1024 -mirror_file=./testdata/mirror.bin
+    # Check the new content
+    qemu-img convert -p \
+        -f raw "vitastor:etcd_host=127.0.0.1\:$ETCD_PORT/v3:image=testchain" \
+        -O raw ./testdata/layer1.bin
+    cmp ./testdata/layer1.bin ./testdata/mirror.bin
+done
+
+build/src/vitastor-cli --etcd_address $ETCD_URL rm testchain@1 testchain@9
+
+# Check the final image
+qemu-img convert -p \
+    -f raw "vitastor:etcd_host=127.0.0.1\:$ETCD_PORT/v3:image=testchain" \
+    -O raw ./testdata/layer1.bin
+cmp ./testdata/layer1.bin ./testdata/mirror.bin
+
+# Check the last remaining snapshot
+qemu-img convert -p \
+    -f raw "vitastor:etcd_host=127.0.0.1\:$ETCD_PORT/v3:image=testchain@10" \
+    -O raw ./testdata/layer0.bin
+cmp ./testdata/layer0.bin ./testdata/check.bin
+
+format_green OK

tests/test_snapshot_down.sh

@@ -0,0 +1,37 @@
+#!/bin/bash -ex
+
+. `dirname $0`/run_3osds.sh
+check_qemu
+
+# Test merge to child (without "inverse rename" optimisation)
+
+build/src/vitastor-cli --etcd_address $ETCD_URL create -s 128M testchain
+
+LD_PRELOAD="build/src/libfio_vitastor.so" \
+    fio -thread -name=test -ioengine=build/src/libfio_vitastor.so -bs=4M -direct=1 -iodepth=1 -fsync=1 -rw=write \
+        -etcd=$ETCD_URL -image=testchain -mirror_file=./testdata/mirror.bin
+
+# Create a snapshot
+build/src/vitastor-cli --etcd_address $ETCD_URL snap-create testchain@0
+
+# Write something to it
+LD_PRELOAD="build/src/libfio_vitastor.so" \
+fio -thread -name=test -ioengine=build/src/libfio_vitastor.so -bs=1M -direct=1 -iodepth=4 -rw=randwrite \
+    -randrepeat=0 -etcd=$ETCD_URL -image=testchain -number_ios=8 -mirror_file=./testdata/mirror.bin
+
+# Check the new content
+qemu-img convert -p \
+    -f raw "vitastor:etcd_host=127.0.0.1\:$ETCD_PORT/v3:image=testchain" \
+    -O raw ./testdata/layer1.bin
+cmp ./testdata/layer1.bin ./testdata/mirror.bin
+
+# Merge
+build/src/vitastor-cli --etcd_address $ETCD_URL rm testchain@0
+
+# Check the final image
+qemu-img convert -p \
+    -f raw "vitastor:etcd_host=127.0.0.1\:$ETCD_PORT/v3:image=testchain" \
+    -O raw ./testdata/layer1.bin
+cmp ./testdata/layer1.bin ./testdata/mirror.bin
+
+format_green OK