Update stock/hsx.vn.md

- QEMU virtual disk migration with enabled iothread is finally fixed correctly.
- Fixed operation of Proxmox VMs with swTPM without enabling NBD for all disks.
- Debian packages are now again built with stable Antietcd released instead of
  the unstable master branch.
- Antietcd cluster mode previously broken in that master branch is fixed. The
  symptom was Antietcd being unable to elect the leader in a cluster.
- Fixed monitor startup with embedded Antietcd when using IPv6.
- Fixed space statistics calculation for FS and S3 pools in the new storage ([PR #127](https://github.com/vitalif/vitastor/pull/127)).

.gitea/workflows/test.yml

@@ -234,60 +234,6 @@ jobs:
           echo ""
         done
 
-  test_etcd_fail_https:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 10
-      run: ETCD_SCHEME=https /root/vitastor/tests/test_etcd_fail.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_etcd_fail_https_antietcd:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 10
-      run: ETCD_SCHEME=https ANTIETCD=1 /root/vitastor/tests/test_etcd_fail.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_https:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: ETCD_SCHEME=https /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_interrupted_rebalance:
     runs-on: ubuntu-latest
     needs: build
@@ -360,6 +306,78 @@ jobs:
           echo ""
         done
 
+  test_dump_load:
+    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_dump_load.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_dump_load_32k:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: TEST_NAME=32k OSD_ARGS="--data_csum_type crc32c --csum_block_size 32k" OFFSET_ARGS="$OSD_ARGS" /root/vitastor/tests/test_dump_load.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_old_dump_load:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: OLD=1 /root/vitastor/tests/test_dump_load.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_dump_load_old_32k:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: TEST_NAME=old_32k OLD=1 OSD_ARGS="--data_csum_type crc32c --csum_block_size 32k" OFFSET_ARGS="$OSD_ARGS" /root/vitastor/tests/test_dump_load.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_old_interrupted_rebalance:
     runs-on: ubuntu-latest
     needs: build
@@ -702,24 +720,6 @@ jobs:
           echo ""
         done
 
-  test_snapshot_chain_encrypted:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: ENCRYPTED=1 /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_old_snapshot_chain:
     runs-on: ubuntu-latest
     needs: build
@@ -1296,24 +1296,6 @@ jobs:
           echo ""
         done
 
-  test_checksum_xxhash:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: TEST_NAME=xxhash OSD_ARGS="--data_csum_type xxh3_32" /root/vitastor/tests/test_checksum.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_old_checksum:
     runs-on: ubuntu-latest
     needs: build
@@ -1548,6 +1530,24 @@ jobs:
           echo ""
         done
 
+  test_resize_last:
+    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_resize_last.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_resize_auto:
     runs-on: ubuntu-latest
     needs: build
@@ -1584,6 +1584,24 @@ jobs:
           echo ""
         done
 
+  test_old_resize_last:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: OLD=1 /root/vitastor/tests/test_resize_last.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_old_resize_auto:
     runs-on: ubuntu-latest
     needs: build
@@ -2142,39 +2160,3 @@ jobs:
           echo ""
         done
 
-  test_write_encrypted:
-    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_write_encrypted.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_encrypted_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_write_encrypted.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
-

.gitea/workflows/tests-to-yaml.pl

@@ -38,14 +38,6 @@ for my $line (<>)
             {
                 $test_name .= '_antietcd';
             }
-            elsif ($1 eq 'ETCD_SCHEME' && $2 eq 'https')
-            {
-                $test_name .= '_https';
-            }
-            elsif ($1 eq 'ENCRYPTED')
-            {
-                $test_name .= '_encrypted';
-            }
             elsif ($1 eq 'OLD')
             {
                 $test_name =~ s/^test_/test_old_/s;

.gitignore

@@ -3,4 +3,3 @@
 package-lock.json
 fio
 qemu
-node_modules

CMakeLists.txt

@@ -2,7 +2,7 @@ cmake_minimum_required(VERSION 2.8...3.30)
 
 project(vitastor)
 
-set(VITASTOR_VERSION "3.0.8")
+set(VITASTOR_VERSION "3.0.15")
 
 include(CTest)
 

README-ru.md

@@ -62,7 +62,6 @@ Vitastor поддерживает QEMU-драйвер, протоколы UBLK,
     - [Дисковые параметры OSD](docs/config/layout-osd.ru.md)
     - [Прочие параметры OSD](docs/config/osd.ru.md)
     - [Параметры мониторов](docs/config/monitor.ru.md)
-    - [Безопасность](docs/config/security.ru.md)
   - [Настройки пулов](docs/config/pool.ru.md)
   - [Метаданные образов в etcd](docs/config/inode.ru.md)
 - Использование

README.md

@@ -1,113 +1,183 @@
-# Vitastor
+# tromcho.net
 
-[Читать на русском](README-ru.md)
+Repository này chứa toàn bộ source code của website **tromcho.net**.
 
-## The Idea
+## Giới thiệu
 
-Make Clustered Block Storage Fast Again.
+`tromcho.net` là mã nguồn website được tổ chức để phục vụ phát triển, triển khai và vận hành theo quy trình chuẩn trên GitHub.
+README này đóng vai trò tài liệu khởi đầu cho lập trình viên, DevOps engineer và cộng tác viên khi tiếp cận repository.
 
-Vitastor is a distributed block, file and object SDS, direct replacement of Ceph RBD, CephFS and RGW,
-and also internal SDS's of public clouds. However, in contrast to them, Vitastor is fast
-and simple at the same time. The only thing is it's slightly young :-).
+## Mục tiêu repository
 
-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).
+- Quản lý tập trung toàn bộ source code của website.
+- Chuẩn hóa quy trình phát triển, review và triển khai.
+- Tạo nền tảng rõ ràng cho việc CI/CD, kiểm thử và vận hành production.
+- Hỗ trợ onboarding nhanh cho thành viên mới.
 
-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.
+## Cấu trúc thư mục đề xuất
 
-Vitastor supports QEMU, UBLK, NBD, NFS protocols, OpenStack, OpenNebula, Proxmox, Kubernetes drivers.
-More drivers may be created easily.
+```text
+.
+├── app/                # Source code ứng dụng chính
+├── public/             # Static files, images, favicon, robots.txt
+├── config/             # Cấu hình môi trường, app, service integration
+├── database/           # Migration, seed, schema
+├── tests/              # Unit test, integration test, e2e test
+├── scripts/            # Script hỗ trợ build, deploy, backup, maintenance
+├── docs/               # Tài liệu kỹ thuật, kiến trúc, quy trình
+├── .github/            # GitHub Actions, issue template, PR template
+├── Dockerfile          # Build image ứng dụng
+├── docker-compose.yml  # Chạy local/dev bằng container
+├── .env.example        # Biến môi trường mẫu
+└── README.md
+```
 
-Read more details in the documentation. You can start from here: [Quick Start](docs/intro/quickstart.en.md).
+> Cấu trúc thực tế có thể thay đổi theo framework đang sử dụng.
 
-## Talks and presentations
+## Yêu cầu môi trường
 
-- KuberConf'2025: [video](https://vitastor.io/presentation/kuberconf.webm)
-- Highload'2025: [video](https://vitastor.io/presentation/hl2025/hl2025.webm),
-  [youtube](https://www.youtube.com/watch?v=0R8MLjFtz7g), presentation
-  ([in Russian](https://vitastor.io/presentation/hl2025/), [in English](https://vitastor.io/presentation/hl2025/en.html))
-- Highload'2022: presentation ([in Russian](https://vitastor.io/presentation/highload/highload.html)),
-  [video](https://vitastor.io/presentation/highload/talk.webm)
-- DevOpsConf'2021: presentation ([in Russian](https://vitastor.io/presentation/devopsconf/devopsconf.html),
-  [in English](https://vitastor.io/presentation/devopsconf/devopsconf_en.html)),
-  [video](https://vitastor.io/presentation/devopsconf/talk.webm)
+Tùy theo stack công nghệ của website, môi trường phát triển nên có:
 
-## Documentation
+- Git
+- Docker và Docker Compose
+- Node.js / PHP / Python / runtime phù hợp với dự án
+- Make (khuyến nghị)
+- Truy cập vào file cấu hình môi trường `.env`
 
-- Introduction
-  - [Quick Start](docs/intro/quickstart.en.md)
-  - [Features](docs/intro/features.en.md)
-  - [Architecture](docs/intro/architecture.en.md)
-  - [Author and license](docs/intro/author.en.md)
-- Installation
-  - [Packages](docs/installation/packages.en.md)
-  - [Docker](docs/installation/docker.en.md)
-  - [Proxmox](docs/installation/proxmox.en.md)
-  - [OpenNebula](docs/installation/opennebula.en.md)
-  - [OpenStack](docs/installation/openstack.en.md)
-  - [Kubernetes CSI](docs/installation/kubernetes.en.md)
-  - [S3](docs/installation/s3.en.md)
-  - [Building from Source](docs/installation/source.en.md)
-- Configuration
-  - [Overview](docs/config.en.md)
-  - Parameter Reference
-    - [Common](docs/config/common.en.md)
-    - [Network](docs/config/network.en.md)
-    - [Client](docs/config/client.en.md)
-    - [Global Disk Layout](docs/config/layout-cluster.en.md)
-    - [OSD Disk Layout](docs/config/layout-osd.en.md)
-    - [OSD Runtime Parameters](docs/config/osd.en.md)
-    - [Monitor](docs/config/monitor.en.md)
-    - [Security](docs/config/security.en.md)
-  - [Pool configuration](docs/config/pool.en.md)
-  - [Image metadata in etcd](docs/config/inode.en.md)
-- Usage
-  - [vitastor-cli](docs/usage/cli.en.md) (command-line interface)
-  - [vitastor-disk](docs/usage/disk.en.md) (disk management tool)
-  - [fio](docs/usage/fio.en.md) for benchmarks
-  - [UBLK](docs/usage/ublk.en.md) for kernel mounts
-  - [NBD](docs/usage/nbd.en.md) - old interface for kernel mounts
-  - [QEMU, qemu-img and VDUSE](docs/usage/qemu.en.md)
-  - [NFS](docs/usage/nfs.en.md) clustered file system and pseudo-FS proxy
-  - [Administration](docs/usage/admin.en.md)
-- Performance
-  - [Understanding storage performance](docs/performance/understanding.en.md)
-  - [Theoretical performance](docs/performance/theoretical.en.md)
-  - [Example comparison with Ceph](docs/performance/comparison1.en.md)
-  - [Newer benchmark of Vitastor 1.3.1](docs/performance/bench2.en.md)
+## Bắt đầu nhanh
 
-## Author and License
+### 1. Clone repository
 
-Copyright (c) Vitaliy Filippov (vitalif [at] yourcmc.ru), 2019+
+```bash
+git clone https://github.com/<your-org>/tromcho.net.git
+cd tromcho.net
+```
 
-Join Vitastor Telegram Chat: https://t.me/vitastor
+### 2. Tạo file môi trường
 
-All server-side code (OSD, Monitor and so on) is licensed under the terms of
-Vitastor Network Public License 1.1 (VNPL 1.1), a copyleft license based on
-GNU GPLv3.0 with the additional "Network Interaction" clause which requires
-opensourcing all programs directly or indirectly interacting with Vitastor
-through a computer network and expressly designed to be used in conjunction
-with it ("Proxy Programs"). Proxy Programs may be made public not only under
-the terms of the same license, but also under the terms of any GPL-Compatible
-Free Software License, as listed by the Free Software Foundation.
-This is a stricter copyleft license than the Affero GPL.
+```bash
+cp .env.example .env
+```
 
-Please note that VNPL doesn't require you to open the code of proprietary
-software running inside a VM if it's not specially designed to be used with
-Vitastor.
+Sau đó cập nhật các biến cấu hình cần thiết trong file `.env`.
 
-Basically, you can't use the software in a proprietary environment to provide
-its functionality to users without opensourcing all intermediary components
-standing between the user and Vitastor or purchasing a commercial license
-from the author 😀.
+### 3. Chạy môi trường local
 
-Client libraries (cluster_client and so on) are dual-licensed under the same
-VNPL 1.1 and also GNU GPL 2.0 or later to allow for compatibility with GPLed
-software like QEMU and fio.
+Nếu dự án dùng Docker:
 
-You can find the full text of VNPL-1.1 in the file [VNPL-1.1.txt](VNPL-1.1.txt).
-GPL 2.0 is also included in this repository as [GPL-2.0.txt](GPL-2.0.txt).
+```bash
+docker compose up -d --build
+```
+
+Nếu dự án chạy trực tiếp theo framework, sử dụng lệnh tương ứng của stack hiện tại.
+
+## Quy trình phát triển
+
+- Tạo branch mới từ `main` hoặc `develop`.
+- Đặt tên branch rõ ràng, ví dụ: `feature/homepage-banner`, `fix/login-timeout`.
+- Commit ngắn gọn, đúng ngữ cảnh thay đổi.
+- Tạo Pull Request để review trước khi merge.
+- Không commit file bí mật như `.env`, private key hoặc credential.
+
+## Quy ước commit
+
+Khuyến nghị dùng convention sau:
+
+```text
+feat: thêm chức năng mới
+fix: sửa lỗi
+refactor: tái cấu trúc mã nguồn
+chore: cập nhật tác vụ phụ trợ
+ci: thay đổi pipeline CI/CD
+docs: cập nhật tài liệu
+test: bổ sung hoặc cập nhật kiểm thử
+```
+
+## CI/CD
+
+Repository nên tích hợp các bước tự động sau:
+
+- Lint source code
+- Chạy unit test / integration test
+- Build artifact hoặc Docker image
+- Scan bảo mật dependency/container
+- Deploy tới staging hoặc production theo rule xác định
+
+Ví dụ vị trí cấu hình pipeline:
+
+```text
+.github/workflows/
+```
+
+## Biến môi trường
+
+Không commit file `.env` thật lên GitHub.
+Nên cung cấp `.env.example` với:
+
+- Danh sách biến bắt buộc
+- Giá trị mẫu an toàn
+- Ghi chú ngắn cho từng biến quan trọng
+
+Ví dụ:
+
+```env
+APP_ENV=local
+APP_DEBUG=true
+APP_URL=http://localhost
+DB_HOST=127.0.0.1
+DB_PORT=3306
+DB_NAME=tromcho
+DB_USER=user
+DB_PASSWORD=change_me
+```
+
+## Triển khai
+
+Khuyến nghị tách rõ các môi trường:
+
+- local
+- development
+- staging
+- production
+
+Các thành phần nên được chuẩn hóa khi triển khai:
+
+- Biến môi trường
+- Reverse proxy / web server
+- TLS certificate
+- Database migration
+- Backup strategy
+- Log rotation và monitoring
+
+## Bảo mật
+
+- Không đưa secrets vào source code.
+- Bật branch protection cho nhánh quan trọng.
+- Review dependency định kỳ.
+- Áp dụng nguyên tắc least privilege cho tài khoản deploy.
+- Theo dõi log, audit và cảnh báo bất thường.
+
+## Đóng góp
+
+Khi đóng góp vào repository:
+
+1. Fork hoặc tạo branch làm việc.
+2. Cập nhật mã nguồn theo phạm vi thay đổi.
+3. Kiểm tra local trước khi tạo Pull Request.
+4. Viết mô tả PR rõ ràng: mục tiêu, phạm vi ảnh hưởng, cách kiểm thử.
+
+## Tài liệu nên bổ sung
+
+Repository này nên có thêm các tài liệu sau trong thư mục `docs/`:
+
+- Kiến trúc hệ thống
+- Sơ đồ database
+- Luồng deploy
+- Quy trình backup/restore
+- Hướng dẫn xử lý sự cố
+- Checklist release
+
+## License
+
+Copyright 2026 Trộm Chó chấm Nét
+---

csi/Makefile

@@ -1,4 +1,4 @@
-VITASTOR_VERSION ?= v3.0.8
+VITASTOR_VERSION ?= v3.0.15
 
 all: build push
 

csi/deploy/004-csi-nodeplugin.yaml

@@ -49,7 +49,7 @@ spec:
             capabilities:
               add: ["SYS_ADMIN"]
             allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v3.0.8
+          image: vitalif/vitastor-csi:v3.0.15
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/deploy/007-csi-provisioner.yaml

@@ -121,7 +121,7 @@ spec:
             privileged: true
             capabilities:
               add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v3.0.8
+          image: vitalif/vitastor-csi:v3.0.15
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/src/config.go

@@ -5,7 +5,7 @@ package vitastor
 
 const (
     vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "3.0.8"
+    vitastorCSIDriverVersion = "3.0.15"
 )
 
 // Config struct fills the parameters of request or user input

debian/changelog

@@ -1,4 +1,4 @@
-vitastor (3.0.8-1) unstable; urgency=medium
+vitastor (3.0.15-1) unstable; urgency=medium
 
   * Bugfixes
 

debian/control

@@ -3,7 +3,7 @@ Section: admin
 Priority: optional
 Maintainer: Vitaliy Filippov <vitalif@yourcmc.ru>
 Build-Depends: debhelper, g++ (>= 8), libstdc++6 (>= 8),
-  linux-libc-dev, libgoogle-perftools-dev, libjerasure-dev, libgf-complete-dev, libc-ares-dev,
+  linux-libc-dev, libgoogle-perftools-dev, libjerasure-dev, libgf-complete-dev,
   libibverbs-dev, librdmacm-dev, libisal-dev, cmake, pkg-config, libnl-3-dev, libnl-genl-3-dev,
   node-bindings <!nocheck>, node-gyp, node-nan
 Standards-Version: 4.5.0

debian/rules

@@ -11,6 +11,7 @@ override_dh_install:
 	cp -v node-binding/package.json node-binding/index.js node-binding/addon.cc node-binding/addon.h node-binding/client.cc node-binding/client.h debian/tmp/usr/lib/x86_64-linux-gnu/nodejs/vitastor
 	cp -v node-binding/build/Release/addon.node debian/tmp/usr/lib/x86_64-linux-gnu/nodejs/vitastor/build/Release
 	dh_install
+	cd debian/vitastor-mon/usr/lib/vitastor/mon && npm install --production
 
 override_dh_installdeb:
 	cat debian/fio_version >> debian/vitastor-fio.substvars

debian/vitastor-build.sh

@@ -37,12 +37,6 @@ rm -rf a b
 
 echo "dep:fio=$FIO" > debian/fio_version
 
-cd /root/vitastor/packages/vitastor-$REL/vitastor-$VER
-mkdir mon/node_modules
-cd mon/node_modules
-curl -s https://git.yourcmc.ru/vitalif/antietcd/archive/master.tar.gz | tar -zx
-curl -s https://git.yourcmc.ru/vitalif/tinyraft/archive/master.tar.gz | tar -zx
-
 cd /root/vitastor/packages/vitastor-$REL
 if [[ ( "$REL" = "trixie" || "$REL" = "resolute" ) && -e ../vitastor-bookworm/vitastor_$VER.orig.tar.xz ]]; then
     # Fucking shit, archives differ between bookworm (xz 5.4.1) and trixie (xz 5.8.1)

debian/vitastor-buildenv.Dockerfile

@@ -25,7 +25,7 @@ RUN set -e -x; \
     echo 'APT::Install-Suggests false;' >> /etc/apt/apt.conf
 
 RUN apt-get update && \
-    apt-get -y install fio libgoogle-perftools-dev devscripts libjerasure-dev cmake libc-ares-dev \
+    apt-get -y install fio libgoogle-perftools-dev devscripts libjerasure-dev cmake \
         libibverbs-dev librdmacm-dev libisal-dev libnl-3-dev libnl-genl-3-dev curl nodejs npm node-nan node-bindings && \
     apt-get -y build-dep fio && \
     apt-get --download-only source fio

docker/Makefile

@@ -1,4 +1,4 @@
-VITASTOR_VERSION ?= v3.0.8
+VITASTOR_VERSION ?= v3.0.15
 
 all: build push
 

docker/etc/vitastor/docker.conf

@@ -4,7 +4,7 @@
 #
 
 # Desired Vitastor version
-VITASTOR_VERSION=v3.0.8
+VITASTOR_VERSION=v3.0.15
 
 # Additional arguments for all containers
 # For example, you may want to specify a custom logging driver here

docs/config.en.md

@@ -38,4 +38,3 @@ In the future, additional configuration methods may be added:
 - [OSD Disk Layout](config/layout-osd.en.md)
 - [OSD Runtime Parameters](config/osd.en.md)
 - [Monitor](config/monitor.en.md)
-- [Security Parameters](config/security.en.md)

docs/config.ru.md

@@ -41,4 +41,3 @@
 - [Дисковые параметры OSD](config/layout-osd.ru.md)
 - [Прочие параметры OSD](config/osd.ru.md)
 - [Параметры мониторов](config/monitor.ru.md)
-- [Параметры безопасности](config/security.ru.md)

docs/config/layout-osd.en.md

@@ -198,14 +198,8 @@ put a modified value into etcd key /vitastor/config/global.
 - Type: string
 - Default: none
 
-Data and metadata checksum type to use. May be "crc32c", "xxh3_32" or "none".
-Select crc32c or xxh3_32 and set csum_block_size to enable data checksums.
-
-Both crc32c and xxh3_32 are almost equally fast, xxh3_32 is safer. xxh3_32 is
-the xxhash3 algorithm truncated from 64 to 32 bits (which is still a good hash).
-
-Note that enabled data checksums either increase memory usage or reduce
-performance. Check details in [csum_block_size](#csum_block_size) description.
+Data checksum type to use. May be "crc32c" or "none". Set to "crc32c" to
+enable data checksums.
 
 ## csum_block_size
 

docs/config/layout-osd.ru.md

@@ -209,12 +209,8 @@ journal_block_size и meta_block_size. Однако на данный момен
 - Тип: строка
 - Значение по умолчанию: none
 
-Тип используемых OSD контрольных сумм данных и метаданных. Может быть "crc32c",
-"xxh3_32" или "none". Выберите crc32c или xxh3_32 и установите csum_block_size,
-чтобы включить контрольные суммы данных.
-
-И crc32c, и xxh3_32 примерно одинаково быстры, xxh3_32 надёжней. xxh3_32 - это
-алгоритм xxhash3, обрезанный с 64 до 32 бит (это всё равно хороший хеш).
+Тип используемых OSD контрольных сумм данных. Может быть "crc32c" или "none".
+Установите в "crc32c", чтобы включить расчёт и проверку контрольных сумм данных.
 
 Следует понимать, что контрольные суммы в зависимости от размера блока их
 расчёта либо увеличивают потребление памяти, либо снижают производительность.

docs/config/osd.en.md

@@ -70,6 +70,7 @@ with an OSD restart or, for some of them, even without restarting by updating co
 - [use_atomic_flag](#use_atomic_flag)
 - [pg_reshard_chunk_size](#pg_reshard_chunk_size)
 - [pg_reshard_chunk_pause_ms](#pg_reshard_chunk_pause_ms)
+- [gc_on_start](#gc_on_start)
 
 ## bind_address
 
@@ -753,3 +754,9 @@ This option sets the maximum number of object is a chunk. Moving 100k objects us
 - Default: 100
 
 This option sets the interval between handling two PG count change chunks.
+
+## gc_on_start
+
+- Type: boolean
+
+Forcibly clean all garbage entries in the new store on every OSD restart.

docs/config/osd.ru.md

@@ -71,6 +71,7 @@
 - [use_atomic_flag](#use_atomic_flag)
 - [pg_reshard_chunk_size](#pg_reshard_chunk_size)
 - [pg_reshard_chunk_pause_ms](#pg_reshard_chunk_pause_ms)
+- [gc_on_start](#gc_on_start)
 
 ## bind_address
 
@@ -793,3 +794,9 @@ pg_minsize OSD во время переключений, что может по
 - Значение по умолчанию: 100
 
 Данная опция задаёт интервал между обработкой двух порций изменения числа PG пулов.
+
+## gc_on_start
+
+- Тип: булево (да/нет)
+
+Принудительно очищать все мусорные записи в новом хранилище при каждом запуске OSD.

docs/config/security.en.md

@@ -1,150 +0,0 @@
-[Documentation](../../README.md#documentation) → [Configuration](../config.en.md) → Security Parameters
-
------
-
-[Читать на русском](security.ru.md)
-
-# Security Parameters
-
-These parameters affect your Vitastor installation security and apply to OSDs, monitors and clients.
-
-Most of them can be set in /etc/vitastor/vitastor.conf and in etcd, but don't support online modification.
-
-- [etcd_client_cert](#etcd_client_cert)
-- [etcd_client_key](#etcd_client_key)
-- [etcd_ca](#etcd_ca)
-- [osd_etcd_client_cert](#osd_etcd_client_cert)
-- [osd_etcd_client_key](#osd_etcd_client_key)
-- [mon_etcd_client_cert](#mon_etcd_client_cert)
-- [mon_etcd_client_key](#mon_etcd_client_key)
-- [vault_url](#vault_url)
-- [vault_secret_api_path](#vault_secret_api_path)
-- [vault_client_cert](#vault_client_cert)
-- [vault_client_key](#vault_client_key)
-- [vault_ca](#vault_ca)
-- [vault_timeout_ms](#vault_timeout_ms)
-- [vault_error_timeout_sec](#vault_error_timeout_sec)
-- [vault_refresh_leeway_sec](#vault_refresh_leeway_sec)
-- [max_aes_xts_pool_size](#max_aes_xts_pool_size)
-
-## etcd_client_cert
-
-- Type: string
-
-Client TLS certificate to use for Vitastor client (not OSD and not monitor)
-etcd https connections. May be path to a file or just a PEM string with certificate.
-In the latter case, string must begin with "-----BEGIN CERTIFICATE-----".
-
-## etcd_client_key
-
-- Type: string
-
-Private key for etcd_client_cert (also a file or a PEM string).
-
-## etcd_ca
-
-- Type: string
-
-Trusted TLS CA to verify etcd server certificate. May be path to a file,
-directory or just a PEM string with certificate.
-
-## osd_etcd_client_cert
-
-- Type: string
-
-Same as [etcd_client_cert](#etcd_client_cert), but only for OSDs.
-OSDs, clients and monitors should have different permissions, so they should
-use different certificates.
-
-## osd_etcd_client_key
-
-- Type: string
-
-Same as [etcd_client_key](#etcd_client_key), but only for OSDs.
-
-## mon_etcd_client_cert
-
-- Type: string
-
-Same as [etcd_client_cert](#etcd_client_cert), but only for Vitastor monitors.
-
-## mon_etcd_client_key
-
-- Type: string
-
-Same as [etcd_client_key](#etcd_client_key), but only for Vitastor monitors.
-
-## vault_url
-
-- Type: string
-
-Vault base URL.
-
-Vitastor clients support AES-256-XTS image data encryption with different per-image keys.
-Encryption is performed by the client, OSDs don't have access to decrypted data.
-
-Encryption keys may be stored in etcd or, for the increased security level, in an external
-[HashiCorp Vault](https://developer.hashicorp.com/vault/) or [OpenBao](https://openbao.org/)
-instance.
-
-Vitastor clients use [v1 k/v secrets engine](https://openbao.org/api-docs/secret/kv/kv-v1/)
-and [TLS authentication engine](https://openbao.org/api-docs/auth/cert/) in Vault.
-
-In that case, only key IDs are stored in etcd.
-
-## vault_secret_api_path
-
-- Type: string
-- Default: /v1/secret/
-
-Vault v1 secret API mount path to use.
-
-## vault_client_cert
-
-- Type: string
-
-Client TLS certificate to use for Vault connections. Just like [etcd_client_cert](#etcd_client_cert),
-may be path to a file or just a certificate in PEM string.
-
-## vault_client_key
-
-- Type: string
-
-Private key for vault_client_cert (also a file or a PEM string).
-
-## vault_ca
-
-- Type: string
-
-Trusted TLS CA to verify Vault server certificate. May be path to a file,
-directory or just a PEM string with certificate.
-
-## vault_timeout_ms
-
-- Type: integer
-- Default: 5000
-
-Timeout for Vault requests in milliseconds.
-
-## vault_error_timeout_sec
-
-- Type: integer
-- Default: 60
-
-Time (in seconds) to wait before retrying after receiving an error from Vault.
-
-## vault_refresh_leeway_sec
-
-- Type: integer
-- Default: 60
-
-Extra time (in seconds) before real Vault token lease_timeout to refresh it, just
-in case of system clock drift.
-
-## max_aes_xts_pool_size
-
-- Type: integer
-- Default: 256
-
-Maximum number of OpenSSL encryption contexts cached in OSD memory. Probably
-doesn't require modification.

docs/config/security.ru.md

@@ -1,154 +0,0 @@
-[Документация](../../README-ru.md#документация) → [Конфигурация](../config.ru.md) → Параметры безопасности
-
------
-
-[Read in English](security.en.md)
-
-# Параметры безопасности
-
-Данные параметры затрагивают безопасность инсталляций Vitastor и используются
-OSD, мониторами и клиентами.
-
-Большая их часть может задаваться в /etc/vitastor/vitastor.conf и в etcd, но не
-поддерживает онлайн-изменение.
-
-- [etcd_client_cert](#etcd_client_cert)
-- [etcd_client_key](#etcd_client_key)
-- [etcd_ca](#etcd_ca)
-- [osd_etcd_client_cert](#osd_etcd_client_cert)
-- [osd_etcd_client_key](#osd_etcd_client_key)
-- [mon_etcd_client_cert](#mon_etcd_client_cert)
-- [mon_etcd_client_key](#mon_etcd_client_key)
-- [vault_url](#vault_url)
-- [vault_secret_api_path](#vault_secret_api_path)
-- [vault_client_cert](#vault_client_cert)
-- [vault_client_key](#vault_client_key)
-- [vault_ca](#vault_ca)
-- [vault_timeout_ms](#vault_timeout_ms)
-- [vault_error_timeout_sec](#vault_error_timeout_sec)
-- [vault_refresh_leeway_sec](#vault_refresh_leeway_sec)
-- [max_aes_xts_pool_size](#max_aes_xts_pool_size)
-
-## etcd_client_cert
-
-- Тип: строка
-
-Клиентский TLS сертификат для https-подключений к etcd для клиентов Vitastor
-(не OSD и не мониторов). Может быть путём к файлу или просто строкой с
-сертификатом в формате PEM. В последнем случае строка должна начинаться с
-"-----BEGIN CERTIFICATE-----".
-
-## etcd_client_key
-
-- Тип: строка
-
-Закрытый ключ для сертификата etcd_client_cert (также путь к файлу или PEM строка).
-
-## etcd_ca
-
-- Тип: строка
-
-Доверенный корневой TLS-сертификат для проверки сертификата сервера etcd.
-Может быть путём к файлу, директории или просто строкой с сертификатом в
-формате PEM.
-
-## osd_etcd_client_cert
-
-- Тип: строка
-
-Аналогично [etcd_client_cert](#etcd_client_cert), но только для OSD.
-OSD, клиенты и мониторы должны иметь разные привилегии, поэтому они должны
-использовать разные сертификаты.
-
-## osd_etcd_client_key
-
-- Тип: строка
-
-Аналогично [etcd_client_key](#etcd_client_key), но только для OSD.
-
-## mon_etcd_client_cert
-
-- Тип: строка
-
-Аналогично [etcd_client_cert](#etcd_client_cert), но только для мониторов Vitastor.
-
-## mon_etcd_client_key
-
-- Тип: строка
-
-Аналогично [etcd_client_key](#etcd_client_key), но только для мониторов Vitastor.
-
-## vault_url
-
-- Тип: строка
-
-Базовый адрес Vault.
-
-Клиенты Vitastor поддерживают AES-256-XTS шифрование данных образов с отдельными ключами на
-каждый образ. Данные шифруются клиентами, OSD не имеют доступа к незашифрованным данным.
-
-Ключи шифрования могут храниться в etcd или, для повышенного уровня безопасности, во внешнем
-[HashiCorp Vault](https://developer.hashicorp.com/vault/) или [OpenBao](https://openbao.org/).
-
-Клиенты Vitastor используют [движок секретов v1](https://openbao.org/api-docs/secret/kv/kv-v1/)
-и [TLS-аутентификацию](https://openbao.org/api-docs/auth/cert/) в Vault.
-
-В этом случае, только ID ключей хранятся в etcd.
-
-## vault_secret_api_path
-
-- Тип: строка
-- Значение по умолчанию: /v1/secret/
-
-Путь к API секретов v1 для использования клиентами.
-
-## vault_client_cert
-
-- Тип: строка
-
-Клиентский TLS сертификат для подключений к Vault. Как и [etcd_client_cert](#etcd_client_cert),
-может быть путём к файлу или просто PEM-строкой с сертификатом.
-
-## vault_client_key
-
-- Тип: строка
-
-Закрытый ключ для сертификата vault_client_cert (также путь к файлу или PEM строка).
-
-## vault_ca
-
-- Тип: строка
-
-Доверенный корневой TLS-сертификат для проверки сертификата сервера Vault.
-Может быть путём к файлу, директории или просто строкой с сертификатом в
-формате PEM.
-
-## vault_timeout_ms
-
-- Тип: целое число
-- Значение по умолчанию: 5000
-
-Максимально время выполнения Vault-запросов в миллисекундах.
-
-## vault_error_timeout_sec
-
-- Тип: целое число
-- Значение по умолчанию: 60
-
-Время (в секундах) для ожидания перед повторной попыткой при получении ошибки от Vault.
-
-## vault_refresh_leeway_sec
-
-- Тип: целое число
-- Значение по умолчанию: 60
-
-Зазор времени (в секундах), чтобы обновлять токены Vault чуть раньше их реального
-lease_timeout, на случай "ухода" системных часов.
-
-## max_aes_xts_pool_size
-
-- Тип: целое число
-- Значение по умолчанию: 256
-
-Максимальное количество кэшируемых в памяти OSD контекстов шифрования OpenSSL.
-Вряд ли требует изменения.

docs/config/src/included.en.md

@@ -44,8 +44,6 @@
 
 {{../../config/monitor.en.md|indent=2}}
 
-{{../../config/security.en.md|indent=2}}
-
 {{../../config/pool.en.md|indent=2}}
 
 {{../../config/inode.en.md|indent=2}}

docs/config/src/included.ru.md

@@ -44,8 +44,6 @@
 
 {{../../config/monitor.ru.md|indent=2}}
 
-{{../../config/security.ru.md|indent=2}}
-
 {{../../config/pool.ru.md|indent=2}}
 
 {{../../config/inode.ru.md|indent=2}}

docs/config/src/layout-osd.yml

@@ -233,21 +233,11 @@
   type: string
   default: none
   info: |
-    Data and metadata checksum type to use. May be "crc32c", "xxh3_32" or "none".
-    Select crc32c or xxh3_32 and set csum_block_size to enable data checksums.
-
-    Both crc32c and xxh3_32 are almost equally fast, xxh3_32 is safer. xxh3_32 is
-    the xxhash3 algorithm truncated from 64 to 32 bits (which is still a good hash).
-
-    Note that enabled data checksums either increase memory usage or reduce
-    performance. Check details in [csum_block_size](#csum_block_size) description.
+    Data checksum type to use. May be "crc32c" or "none". Set to "crc32c" to
+    enable data checksums.
   info_ru: |
-    Тип используемых OSD контрольных сумм данных и метаданных. Может быть "crc32c",
-    "xxh3_32" или "none". Выберите crc32c или xxh3_32 и установите csum_block_size,
-    чтобы включить контрольные суммы данных.
-
-    И crc32c, и xxh3_32 примерно одинаково быстры, xxh3_32 надёжней. xxh3_32 - это
-    алгоритм xxhash3, обрезанный с 64 до 32 бит (это всё равно хороший хеш).
+    Тип используемых OSD контрольных сумм данных. Может быть "crc32c" или "none".
+    Установите в "crc32c", чтобы включить расчёт и проверку контрольных сумм данных.
 
     Следует понимать, что контрольные суммы в зависимости от размера блока их
     расчёта либо увеличивают потребление памяти, либо снижают производительность.

docs/config/src/osd.yml

@@ -938,3 +938,7 @@
     This option sets the interval between handling two PG count change chunks.
   info_ru: |
     Данная опция задаёт интервал между обработкой двух порций изменения числа PG пулов.
+- name: gc_on_start
+  type: bool
+  info: Forcibly clean all garbage entries in the new store on every OSD restart.
+  info_ru: Принудительно очищать все мусорные записи в новом хранилище при каждом запуске OSD.

docs/config/src/package.json

@@ -1,5 +0,0 @@
-{
-  "dependencies": {
-    "yaml": "^2.8.2"
-  }
-}

docs/config/src/security.en.md

@@ -1,5 +0,0 @@
-# Security Parameters
-
-These parameters affect your Vitastor installation security and apply to OSDs, monitors and clients.
-
-Most of them can be set in /etc/vitastor/vitastor.conf and in etcd, but don't support online modification.

docs/config/src/security.ru.md

@@ -1,7 +0,0 @@
-# Параметры безопасности
-
-Данные параметры затрагивают безопасность инсталляций Vitastor и используются
-OSD, мониторами и клиентами.
-
-Большая их часть может задаваться в /etc/vitastor/vitastor.conf и в etcd, но не
-поддерживает онлайн-изменение.

docs/config/src/security.yml

@@ -1,131 +0,0 @@
-- name: etcd_client_cert
-  type: string
-  info: |
-    Client TLS certificate to use for Vitastor client (not OSD and not monitor)
-    etcd https connections. May be path to a file or just a PEM string with certificate.
-    In the latter case, string must begin with "-----BEGIN CERTIFICATE-----".
-  info_ru: |
-    Клиентский TLS сертификат для https-подключений к etcd для клиентов Vitastor
-    (не OSD и не мониторов). Может быть путём к файлу или просто строкой с
-    сертификатом в формате PEM. В последнем случае строка должна начинаться с
-    "-----BEGIN CERTIFICATE-----".
-- name: etcd_client_key
-  type: string
-  info: Private key for etcd_client_cert (also a file or a PEM string).
-  info_ru: Закрытый ключ для сертификата etcd_client_cert (также путь к файлу или PEM строка).
-- name: etcd_ca
-  type: string
-  info: |
-    Trusted TLS CA to verify etcd server certificate. May be path to a file,
-    directory or just a PEM string with certificate.
-  info_ru: |
-    Доверенный корневой TLS-сертификат для проверки сертификата сервера etcd.
-    Может быть путём к файлу, директории или просто строкой с сертификатом в
-    формате PEM.
-- name: osd_etcd_client_cert
-  type: string
-  info: |
-    Same as [etcd_client_cert](#etcd_client_cert), but only for OSDs.
-    OSDs, clients and monitors should have different permissions, so they should
-    use different certificates.
-  info_ru: |
-    Аналогично [etcd_client_cert](#etcd_client_cert), но только для OSD.
-    OSD, клиенты и мониторы должны иметь разные привилегии, поэтому они должны
-    использовать разные сертификаты.
-- name: osd_etcd_client_key
-  type: string
-  info: Same as [etcd_client_key](#etcd_client_key), but only for OSDs.
-  info_ru: Аналогично [etcd_client_key](#etcd_client_key), но только для OSD.
-- name: mon_etcd_client_cert
-  type: string
-  info: Same as [etcd_client_cert](#etcd_client_cert), but only for Vitastor monitors.
-  info_ru: Аналогично [etcd_client_cert](#etcd_client_cert), но только для мониторов Vitastor.
-- name: mon_etcd_client_key
-  type: string
-  info: Same as [etcd_client_key](#etcd_client_key), but only for Vitastor monitors.
-  info_ru: Аналогично [etcd_client_key](#etcd_client_key), но только для мониторов Vitastor.
-- name: vault_url
-  type: string
-  info: |
-    Vault base URL.
-
-    Vitastor clients support AES-256-XTS image data encryption with different per-image keys.
-    Encryption is performed by the client, OSDs don't have access to decrypted data.
-
-    Encryption keys may be stored in etcd or, for the increased security level, in an external
-    [HashiCorp Vault](https://developer.hashicorp.com/vault/) or [OpenBao](https://openbao.org/)
-    instance.
-
-    Vitastor clients use [v1 k/v secrets engine](https://openbao.org/api-docs/secret/kv/kv-v1/)
-    and [TLS authentication engine](https://openbao.org/api-docs/auth/cert/) in Vault.
-
-    In that case, only key IDs are stored in etcd.
-  info_ru: |
-    Базовый адрес Vault.
-
-    Клиенты Vitastor поддерживают AES-256-XTS шифрование данных образов с отдельными ключами на
-    каждый образ. Данные шифруются клиентами, OSD не имеют доступа к незашифрованным данным.
-
-    Ключи шифрования могут храниться в etcd или, для повышенного уровня безопасности, во внешнем
-    [HashiCorp Vault](https://developer.hashicorp.com/vault/) или [OpenBao](https://openbao.org/).
-
-    Клиенты Vitastor используют [движок секретов v1](https://openbao.org/api-docs/secret/kv/kv-v1/)
-    и [TLS-аутентификацию](https://openbao.org/api-docs/auth/cert/) в Vault.
-
-    В этом случае, только ID ключей хранятся в etcd.
-- name: vault_secret_api_path
-  type: string
-  default: /v1/secret/
-  info: Vault v1 secret API mount path to use.
-  info_ru: Путь к API секретов v1 для использования клиентами.
-- name: vault_client_cert
-  type: string
-  info: |
-    Client TLS certificate to use for Vault connections. Just like [etcd_client_cert](#etcd_client_cert),
-    may be path to a file or just a certificate in PEM string.
-  info_ru: |
-    Клиентский TLS сертификат для подключений к Vault. Как и [etcd_client_cert](#etcd_client_cert),
-    может быть путём к файлу или просто PEM-строкой с сертификатом.
-- name: vault_client_key
-  type: string
-  info: Private key for vault_client_cert (also a file or a PEM string).
-  info_ru: Закрытый ключ для сертификата vault_client_cert (также путь к файлу или PEM строка).
-- name: vault_ca
-  type: string
-  info: |
-    Trusted TLS CA to verify Vault server certificate. May be path to a file,
-    directory or just a PEM string with certificate.
-  info_ru: |
-    Доверенный корневой TLS-сертификат для проверки сертификата сервера Vault.
-    Может быть путём к файлу, директории или просто строкой с сертификатом в
-    формате PEM.
-- name: vault_timeout_ms
-  type: int
-  default: 5000
-  info: Timeout for Vault requests in milliseconds.
-  info_ru: Максимально время выполнения Vault-запросов в миллисекундах.
-- name: vault_error_timeout_sec
-  type: int
-  default: 60
-  info: |
-    Time (in seconds) to wait before retrying after receiving an error from Vault.
-  info_ru: |
-    Время (в секундах) для ожидания перед повторной попыткой при получении ошибки от Vault.
-- name: vault_refresh_leeway_sec
-  type: int
-  default: 60
-  info: |
-    Extra time (in seconds) before real Vault token lease_timeout to refresh it, just
-    in case of system clock drift.
-  info_ru: |
-    Зазор времени (в секундах), чтобы обновлять токены Vault чуть раньше их реального
-    lease_timeout, на случай "ухода" системных часов.
-- name: max_aes_xts_pool_size
-  type: int
-  default: 256
-  info: |
-    Maximum number of OpenSSL encryption contexts cached in OSD memory. Probably
-    doesn't require modification.
-  info_ru: |
-    Максимальное количество кэшируемых в памяти OSD контекстов шифрования OpenSSL.
-    Вряд ли требует изменения.

docs/installation/docker.en.md

@@ -26,9 +26,9 @@ at Vitastor Kubernetes operator: https://github.com/Antilles7227/vitastor-operat
 The instruction is very simple.
 
 1. Download a Docker image of the desired version: \
-   `docker pull vitalif/vitastor:v3.0.8`
+   `docker pull vitalif/vitastor:v3.0.15`
 2. Install scripts to the host system: \
-   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.8 install.sh`
+   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.15 install.sh`
 3. Reload udev rules: \
    `udevadm control --reload-rules`
 4. Enable the vitastor-host service: \

docs/installation/docker.ru.md

@@ -25,9 +25,9 @@ Vitastor можно установить в Docker/Podman. При этом etcd,
 Инструкция по установке максимально простая.
 
 1. Скачайте Docker-образ желаемой версии: \
-   `docker pull vitalif/vitastor:v3.0.8`
+   `docker pull vitalif/vitastor:v3.0.15`
 2. Установите скрипты в хост-систему командой: \
-   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.8 install.sh`
+   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.15 install.sh`
 3. Перезагрузите правила udev: \
    `udevadm control --reload-rules`
 4. Включите сервис vitastor-host: \

docs/installation/source.en.md

@@ -15,9 +15,8 @@
 - gcc and g++ 8 or newer, clang 10 or newer, or other compiler with C++11 plus
   designated initializers support from C++20
 - CMake
-- jerasure, c-ares headers and libraries
+- jerasure headers and libraries
 - ISA-L, libibverbs, librdmacm, libnl3 headers and libraries (optional)
-- tcmalloc (google-perftools-dev)
 
 ## Basic instructions
 

docs/installation/source.ru.md

@@ -15,9 +15,8 @@
 - gcc и g++ >= 8, либо clang >= 10, либо другой компилятор с поддержкой C++11 плюс
   назначенных инициализаторов (designated initializers) из C++20
 - CMake
-- Заголовки и библиотеки jerasure, c-ares
+- Заголовки и библиотеки jerasure
 - Опционально - заголовки и библиотеки ISA-L, libibverbs, librdmacm, libnl3
-- tcmalloc (google-perftools-dev)
 
 ## Базовая инструкция
 

docs/intro/features.en.md

@@ -41,8 +41,6 @@
 - [Built-in Prometheus metric exporter](../config/monitor.en.md#enable_prometheus)
 - [NFS RDMA support](../usage/nfs.en.md#rdma) (probably also usable for GPUDirect)
 - [S3](../installation/s3.en.md)
-- [TLS support for etcd connections](../config/security.en.md)
-- [AES-256-XTS image encryption](../usage/cli.en.md#create) and [Vault support](../config/security.en.md#vault_url) for key storage
 
 ## Plugins and tools
 

docs/intro/features.ru.md

@@ -43,8 +43,6 @@
 - [Встроенный Prometheus-экспортер метрик](../config/monitor.ru.md#enable_prometheus)
 - [Поддержка NFS RDMA](../usage/nfs.ru.md#rdma) (вероятно, также подходящая для GPUDirect)
 - [S3](../installation/s3.ru.md)
-- [Поддержка TLS-соединений с etcd](../config/security.ru.md)
-- [AES-256-XTS шифрование данных](../usage/cli.ru.md#create) и [поддержка Vault](../config/security.ru.md#vault_url) для хранения ключей
 
 ## Драйверы и инструменты
 

docs/usage/cli.en.md

@@ -125,31 +125,18 @@ bench-kaveri         kaveri    10 G  10 G    0 B/s  0     0      0 us  0 B/s  0
 
 ## create
 
-`vitastor-cli create -s|--size SIZE [OPTIONS] <name>`
+`vitastor-cli create -s|--size <size> [-p|--pool <id|name>] [--parent <parent_name>[@<snapshot>]] <name>`
 
-Create an image. Options:
-
-* `-s|--size SIZE` - New image size in bytes or with a K/M/G/T unit suffix.
-* `-p|--pool POOL` - Specify pool for the new image (may be omitted if there is only 1 pool).
-* `--parent PARENT` - Create a copy-on-write image clone based on PARENT (or PARENT@SNAPSHOT).
-  If parent is not a snapshot, it must be a read-only image.
-* `--enc-key random` - Generate a new random AES-256-XTS encryption key for the new image.
-* `--enc-key HEX` - Set a specified AES-256-XTS key (64 bytes in hex) for the new image.
-* `--enc-key vault:ID` - Use an encryption key from an external Vault secret with specified ID.
+Create an image. You may use K/M/G/T suffixes for `<size>`. If `--parent` is specified,
+a copy-on-write image clone is created. Parent must be a snapshot (readonly image).
+Pool must be specified if there is more than one pool.
 
 ```
-vitastor-cli create --snapshot <snapshot> [OPTIONS] <image>
-vitastor-cli snap-create [OPTIONS] <image>@<snapshot>
+vitastor-cli create --snapshot <snapshot> [-p|--pool <id|name>] <image>
+vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>
 ```
 
-Create a snapshot of image `<image>`. May be used live if only a single writer is active.
-
-Options:
-
-* `-p|--pool POOL` - Move image to pool POOL, leaving the snapshot in the old pool.
-* `--enc-key random` - Change image encryption key to a new random AES-256-XTS key.
-* `--enc-key KEY` - Change image encryption key to a specified key, Vault key or to an empty key.
-  By default, the image retains its old encryption key when taking a snapshot.
+Create a snapshot of image `<name>` (either form can be used). May be used live if only a single writer is active.
 
 See also about [how to export snapshots](qemu.en.md#exporting-snapshots).
 
@@ -164,7 +151,6 @@ You should resize file system in the image, if present, before shrinking it.
 * `--deleted 1|0` - Set/clear 'deleted image' flag (set automatically during unfinished deletes).
 * `-f|--force` - Proceed with shrinking or setting readwrite flag even if the image has children.
 * `--down-ok` - Proceed with shrinking even if some data will be left on unavailable OSDs.
-* `--enc-key HEX` - Change image encryption key (allowed only with `--force`).
 
 ## dd
 

docs/usage/cli.ru.md

@@ -127,32 +127,19 @@ bench-kaveri         kaveri    10 G  10 G    0 B/s  0     0      0 us  0 B/s  0
 
 ## create
 
-`vitastor-cli create -s|--size SIZE [ОПЦИИ] <name>`
+`vitastor-cli create -s|--size <size> [-p|--pool <id|name>] [--parent <parent_name>[@<snapshot>]] <name>`
 
-Создать образ. Опции:
-
-* `-s|--size SIZE` - Размер нового образа в байтах или с суффиксом K/M/G/T (кило/мега/гига/терабайт).
-* `-p|--pool POOL` - Создать образ в заданном пуле (можно не указывать, если пул всего один).
-* `--parent PARENT` - Создать легковесный клон на основе образа `PARENT` или снимка `PARENT@SNAP`.
-  Если `PARENT` - не снимок, он должен быть помечен как образ только для чтения.
-* `--enc-key random` - Сгенерировать случайный ключ шифрования AES-256-XTS для нового образа.
-* `--enc-key HEX` - Установить заданный ключ AES-256-XTS (64 байта в hex) для нового образа.
-* `--enc-key vault:ID` - Использовать ключ из внешнего секрета с заданным ID из Vault.
+Создать образ. Для размера `<size>` можно использовать суффиксы K/M/G/T (килобайт-мегабайт-гигабайт-терабайт).
+Если указана опция `--parent`, создаётся клон образа. Родитель `<parent_name>[@<snapshot>]` должен быть
+снимком (или просто немодифицируемым образом). Пул обязательно указывать, если в кластере больше одного пула.
 
 ```
-vitastor-cli create --snapshot <snapshot> [ОПЦИИ] <image>
-vitastor-cli snap-create [ОПЦИИ] <image>@<snapshot>
+vitastor-cli create --snapshot <snapshot> [-p|--pool <id|name>] <image>
+vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>
 ```
 
-Создать снимок образа `<image>` (можно использовать любую форму команды).
-Снимок можно создавать без остановки клиентов, если пишущих клиентов не больше одного.
-
-Опции:
-
-* `-p|--pool POOL` - Переместить образ в пул POOL, оставив снимок в старом пуле.
-* `--enc-key random` - Изменить ключ шифрования образа на новый случайный ключ AES-256-XTS.
-* `--enc-key KEY` - Изменить ключ шифрования образа на заданный ключ, ключ из Vault или пустой ключ.
-  По умолчанию шифрованные образы сохраняют старый ключ при снятии снимка.
+Создать снимок образа `<name>` (можно использовать любую форму команды). Снимок можно создавать без остановки
+клиентов, если пишущий клиент максимум 1.
 
 Смотрите также информацию о том, [как экспортировать снимки](qemu.ru.md#экспорт-снимков).
 
@@ -169,7 +156,6 @@ vitastor-cli snap-create [ОПЦИИ] <image>@<snapshot>
 * `--deleted 1|0` - Установить/снять флаг "образ удалён" (устанавливается при незавершённом удалении).
 * `-f|--force` - Разрешить уменьшение или перевод в чтение-запись образа, у которого есть клоны.
 * `--down-ok` - Разрешить уменьшение, даже если часть данных останется неудалённой на недоступных OSD.
-* `--enc-key HEX` - Изменить ключ шифрования образа (разрешено только с `--force`).
 
 ## dd
 

docs/usage/nfs.en.md

@@ -262,3 +262,4 @@ Options:
 | `--logfile <FILE>` | log to the specified file                                |
 | `--enforce 1`      | enforce permissions at the server side (no by default)   |
 | `--foreground 1`   | stay in foreground, do not daemonize                     |
+| `--trace`          | trace all NFS requests                                   |

docs/usage/nfs.ru.md

@@ -274,3 +274,4 @@ VitastorFS из GPUDirect.
 | `--logfile <FILE>` | записывать логи в заданный файл                         |
 | `--enforce 1`      | проверять права доступа на стороне сервера (по умолчанию нет) |
 | `--foreground 1`   | не уходить в фон после запуска                          |
+| `--trace`          | логгировать все запросы NFS                             |

mon/antietcd_adapter.js

@@ -27,7 +27,17 @@ class AntiEtcdAdapter
             is_local['::'] = true;
             is_local[''] = true;
             // split :, 3 -> <schema>:<//ip>:<port>
-            const selected = cluster.map(s => s.split(':', 3)).filter(ip => is_local[ip[1].substr(2)] && (!cfg_port || ip[2] == cfg_port));
+            const selected = [];
+            for (let i = 0; i < cluster.length; i++)
+            {
+                const m = /^(https?:\/\/)?(?:\[(.*)\]|([^\[\:]+))(?::(\d+))?$/.exec(cluster[i]);
+                if (!m)
+                    continue;
+                const ip = m[3] || m[2];
+                const port = m[4] || 2379;
+                if (is_local[ip] && (!cfg_port || port == cfg_port))
+                    selected.push({ idx: i, ip, port });
+            }
             if (selected.length > 1)
             {
                 console.error('More than 1 etcd_address matches local IPs, please specify port');
@@ -36,30 +46,17 @@ class AntiEtcdAdapter
             else if (selected.length == 1)
             {
                 const antietcd_config = {
-                    ip: selected[0][1].substr(2),
-                    port: selected[0][2],
-                    cert: config.antietcd_cert,
-                    key: config.antietcd_key,
-                    ca: config.etcd_ca,
-                    data: config.antietcd_data_file || ((config.antietcd_data_dir || '/var/lib/vitastor') + '/mon_'+selected[0][2]+'.json.gz'),
+                    ip: selected[0].ip,
+                    port: selected[0].port,
+                    data: config.antietcd_data_file || ((config.antietcd_data_dir || '/var/lib/vitastor') + '/mon_'+selected[0].port+'.json.gz'),
                     persist_filter: vitastor_persist_filter({ vitastor_prefix: config.etcd_prefix || '/vitastor' }),
-                    node_id: selected[0][1].substr(2)+':'+selected[0][2], // node_id = ip:port
+                    node_id: cluster[selected[0].idx].replace(/^(https?:\/\/)/, ''), // same as in <cluster> below
                     cluster: (cluster.length == 1 ? null : cluster.reduce((a, c) => { a[c.replace(/^(https?:\/\/)/, '')] = c; return a; }, {})),
                     cluster_key: (config.etcd_prefix || '/vitastor'),
                     stale_read: 1,
                     log_level: 1,
+                    logs: { cluster: true },
                 };
-                if (config.use_auth)
-                {
-                    antietcd_config.client_cert_auth = true;
-                    antietcd_config.auth_filter = require('./vitastor_auth_filter.js');
-                    antietcd_config.peer_ca = config.antietcd_server_ca;
-                    if (!config.antietcd_server_ca || config.antietcd_server_ca == config.etcd_ca)
-                    {
-                        console.error('Secure setup requires separate antietcd_server_ca (for signing antietcd server certificates) and etcd_ca (for signing client certificates)');
-                        process.exit(1);
-                    }
-                }
                 for (const key in config)
                 {
                     if (key.substr(0, 9) === 'antietcd_')
@@ -184,7 +181,7 @@ class AntiEtcdAdapter
                     await new Promise(ok => setTimeout(ok, timeout-(Date.now()-prev)));
                 }
                 prev = Date.now();
-                const res = await this.antietcd.api(path.replace(/^\/+/, '').replace(/\/+$/, '').replace(/\/+/g, '_'), body, { username: 'root' });
+                const res = await this.antietcd.api(path.replace(/^\/+/, '').replace(/\/+$/, '').replace(/\/+/g, '_'), body);
                 if (res.error)
                 {
                     console.error('Failed to query antietcd '+path+' (retry '+retry+'/'+retries+'): '+res.error);

mon/etcd_adapter.js

@@ -1,9 +1,7 @@
 // Copyright (c) Vitaliy Filippov, 2019+
 // License: VNPL-1.1 (see README.md for details)
 
-const fs = require('fs');
 const http = require('http');
-const https = require('https');
 const WebSocket = require('ws');
 const { b64, local_ips } = require('./utils.js');
 
@@ -17,30 +15,11 @@ class EtcdAdapter
         this.ws = null;
         this.ws_alive = false;
         this.ws_keepalive_timer = null;
-        this.opts = {};
     }
 
     parse_config(config)
     {
         this.parse_etcd_addresses(config.etcd_address||config.etcd_url);
-        if (config.mon_etcd_client_cert || config.etcd_client_cert)
-        {
-            this.opts.cert = config.mon_etcd_client_cert || config.etcd_client_cert;
-            if (this.opts.cert.substr(0, 5) != '-----')
-                this.opts.cert = fs.readFileSync(this.opts.cert, { encoding: 'utf-8' });
-        }
-        if (config.mon_etcd_client_key || config.etcd_client_key)
-        {
-            this.opts.key = config.mon_etcd_client_key || config.etcd_client_key;
-            if (this.opts.key.substr(0, 5) != '-----')
-                this.opts.key = fs.readFileSync(this.opts.key, { encoding: 'utf-8' });
-        }
-        if (config.etcd_ca)
-        {
-            this.opts.ca = config.etcd_ca;
-            if (this.opts.ca.substr(0, 5) != '-----')
-                this.opts.ca = fs.readFileSync(this.opts.ca, { encoding: 'utf-8' });
-        }
     }
 
     parse_etcd_addresses(addrs)
@@ -60,7 +39,7 @@ class EtcdAdapter
         for (let url of addrs)
         {
             let scheme = 'http';
-            url = url.trim().replace(/^(https?):\/\//i, (m, m1) => { scheme = m1.toLowerCase(); return ''; });
+            url = url.trim().replace(/^(https?):\/\//, (m, m1) => { scheme = m1; return ''; });
             const slash = url.indexOf('/');
             const colon = url.indexOf(':');
             const is_local = is_local_ip[colon >= 0 ? url.substr(0, colon) : (slash >= 0 ? url.substr(0, slash) : url)];
@@ -151,7 +130,7 @@ class EtcdAdapter
                     }
                     ok(false);
                 }, this.mon.config.etcd_mon_timeout);
-                this.ws = new WebSocket(base+'/watch', this.opts);
+                this.ws = new WebSocket(base+'/watch');
                 this.ws_used_url = cur_addr;
                 const fail = () =>
                 {
@@ -293,7 +272,7 @@ class EtcdAdapter
             {
                 throw new Error(MON_STOPPED);
             }
-            const res = await POST(base+path, body, timeout, this.opts);
+            const res = await POST(base+path, body, timeout);
             if (this.mon.stopped)
             {
                 throw new Error(MON_STOPPED);
@@ -319,7 +298,7 @@ class EtcdAdapter
     }
 }
 
-function POST(url, body, timeout, opts)
+function POST(url, body, timeout)
 {
     return new Promise(ok =>
     {
@@ -331,10 +310,10 @@ function POST(url, body, timeout, opts)
             req = null;
             ok({ error: 'timeout' });
         }, timeout) : null;
-        let req = (url.substr(0, 5) == 'https' ? https : http).request(url, { method: 'POST', headers: {
+        let req = http.request(url, { method: 'POST', headers: {
             'Content-Type': 'application/json',
             'Content-Length': body_text.length,
-        }, ...(opts||{}) }, (res) =>
+        } }, (res) =>
         {
             if (!req)
             {

mon/etcd_schema.js

@@ -16,7 +16,6 @@ const etcd_allow = new RegExp('^'+[
     'config/pools',
     'config/osd/[1-9]\\d*',
     'config/pgs', // old name
-    'config/user/.*',
     'pg/config',
     'config/inode/[1-9]\\d*/[1-9]\\d*',
     'osd/state/[1-9]\\d*',
@@ -46,14 +45,7 @@ const etcd_tree = {
             config_path: "/etc/vitastor/vitastor.conf",
             etcd_prefix: "/vitastor",
             // etcd connection - configurable online
-            etcd_address: "http://10.0.115.10:2379/v3",
-            etcd_client_cert: "",
-            etcd_client_key: "",
-            osd_etcd_client_cert: "",
-            osd_etcd_client_key: "",
-            mon_etcd_client_cert: "",
-            mon_etcd_client_key: "",
-            etcd_ca: "",
+            etcd_address: "10.0.115.10:2379/v3",
             // mon
             etcd_mon_ttl: 5, // min: 1
             etcd_mon_timeout: 1000, // ms. min: 0
@@ -209,8 +201,6 @@ const etcd_tree = {
                 primary_affinity_tags?: 'nvme' | [ 'nvme', ... ],
                 // scrub interval
                 scrub_interval?: '30d',
-                // users allowed to create images in this pool
-                creator_group?: '',
             },
             ...
         }, */
@@ -227,21 +217,10 @@ const etcd_tree = {
                     parent_id?: <inode_t>,
                     readonly?: boolean,
                     deleted?: boolean,
-                    enc_key?: string,
-                    owner?: string,
-                    owner_group?: string,
-                    reader_group?: string,
                 }
             }
         }, */
         inode: {},
-        /* user: {
-            <username>: {
-                type: 'osd'|'mon'|'admin'|'client',
-                groups: string[],
-            },
-        }, */
-        user: {},
     },
     osd: {
         state: {

mon/lp_optimizer/lp_optimizer.js

@@ -112,9 +112,10 @@ function make_cyclic(pgs, parity_space)
 {
     if (parity_space > 1)
     {
-        for (const pg in pgs)
+        for (const id in pgs)
         {
-            for (let i = 1; i < pg.size; i++)
+            const pg = pgs[id];
+            for (let i = 1; i < pg.length; i++)
             {
                 const cyclic = [ ...pg.slice(i), ...pg.slice(0, i) ];
                 pgs['pg_'+cyclic.join('_')] = cyclic;

mon/mon.js

@@ -627,7 +627,7 @@ class Mon
                 if (this.state.pg.history[pool_id] &&
                     this.state.pg.history[pool_id][pg])
                 {
-                    pg_history[pg-1] = this.state.pg.history[pool_id][pg];
+                    pg_history[pg-1] = JSON.parse(JSON.stringify(this.state.pg.history[pool_id][pg]));
                 }
             }
             const real_prev_pgs = [];
@@ -719,7 +719,7 @@ class Mon
                 this.next_recheck_timer = null;
                 this.next_recheck_at = 0;
                 this.schedule_recheck();
-            }, now-this.next_recheck_at);
+            }, (this.next_recheck_at-now)*1000);
         }
     }
 

mon/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vitastor-mon",
-  "version": "3.0.8",
+  "version": "3.0.15",
   "description": "Vitastor SDS monitor service",
   "main": "mon-main.js",
   "scripts": {
@@ -9,7 +9,7 @@
   "author": "Vitaliy Filippov",
   "license": "UNLICENSED",
   "dependencies": {
-    "antietcd": "^1.2.4",
+    "antietcd": "^1.3.1",
     "sprintf-js": "^1.1.2",
     "ws": "^7.2.5"
   },

mon/pg_utils.js

@@ -84,7 +84,7 @@ function scale_pg_history(prev_pg_history, prev_pgs, new_pgs)
         finish_pg_history(merged_history[1]);
         for (let i = 0; i < new_pg_count; i++)
         {
-            new_pg_history[i] = { ...merged_history[1] };
+            new_pg_history[i] = JSON.parse(JSON.stringify(merged_history[1]));
         }
     }
     // Mark history keys for removed PGs as removed
@@ -102,7 +102,7 @@ function scale_pg_count(prev_pgs, new_pg_count)
     {
         for (let i = prev_pgs.length; i < new_pg_count; i++)
         {
-            prev_pgs[i] = prev_pgs[i % prev_pgs.length];
+            prev_pgs[i] = [ ...prev_pgs[i % prev_pgs.length] ];
         }
     }
     else if (prev_pgs.length > new_pg_count)

mon/stats.js

@@ -37,6 +37,7 @@ function derive_osd_stats(st, prev, prev_diff)
         const n = c.count - BigInt(pr && pr.count||0);
         diff.recovery_stats[op] = { ...c, bps: n > 0 ? b*1000n/timediff : 0n, iops: n > 0 ? n*1000n/timediff : 0n };
     }
+    diff.inode_stats = {};
     for (const pool_id in st.inode_stats||{})
     {
         diff.inode_stats[pool_id] = {};

mon/vitastor_auth_filter.js

@@ -1,471 +0,0 @@
-// AntiEtcd authentication filter for Vitastor
-// (c) Vitaliy Filippov, 2026
-// License: Mozilla Public License 2.0 or Vitastor Network Public License 1.1
-
-// Permissions are based on:
-// 1. Users.
-//    Stored in /vitastor/config/user/<username>.
-//    Has 2 properties:
-//    - type, one of: osd, mon, admin, client.
-//      osd, mon types should be used by OSDs/monitors.
-//      admin should be used for administrative access from vitastor-cli.
-//      client should be used for regular clients.
-//    - groups, a list of group names the user is included in.
-// 2. Images.
-//    Stored in /vitastor/config/inode/<pool>/<inode>. Has the following properties:
-//    - owner (user name)
-//    - owner_group (group name)
-//    - reader_group
-
-const static_perms = {
-    invalid: {
-        keys: {},
-        prefixes: {},
-    },
-    osd: {
-        keys: { '/pg/config': false },
-        prefixes: { '/osd/': true, '/pg/state/': true, '/pg/history/': true, '/pgstats/': true },
-    },
-    mon: {
-        keys: { '/pg/config': true, '/stats': true, '/history/last_clean_pgs': true },
-        prefixes: {
-            '/config/': false, '/osd/': false, '/mon/': true, '/pg/history/': true,
-            '/pgstats/': false, '/inode/stats/': true, '/pool/stats/': true,
-        },
-    },
-    admin: {
-        keys: { '/stats': false },
-        prefixes: {
-            '/config/': true, '/osd/': true, '/index/': true, '/pg/history/': true,
-            '/mon/': false, '/pg/': false, '/pgstats/': false, '/inode/stats/': false, '/pool/stats/': false,
-        },
-    },
-    client: {
-        keys: { '/config/global': false, '/config/node_placement': false, '/config/pools': false, '/pg/config': false },
-        prefixes: { '/osd/stats/': false, '/pg/state/': false, '/index/maxid/': false },
-    },
-};
-
-const api_perms = {
-    osd: { lease_grant: true, lease_revoke: true, lease_keepalive: true },
-    mon: { lease_grant: true, lease_revoke: true, lease_keepalive: true },
-    admin: { maintenance_status: true },
-    client: {},
-};
-
-class VitastorAuthFilter
-{
-    constructor(antietcd)
-    {
-        this.cfg = antietcd.cfg;
-        this.antietcd = antietcd;
-        this.prefix = this.cfg.vitastor_prefix || '/vitastor';
-        this.prefix_parts = this.prefix.split('/');
-    }
-
-    _get(path, decode)
-    {
-        let cur = this.antietcd.etctree.state;
-        path = path instanceof Array ? path : path.split('/');
-        for (const p of path)
-        {
-            if (!cur.children)
-            {
-                return null;
-            }
-            cur = cur.children[p];
-            if (!cur)
-            {
-                return null;
-            }
-        }
-        if (decode)
-        {
-            return this._decode(path, cur.value);
-        }
-        return cur;
-    }
-
-    _decode(path, cur)
-    {
-        if (!cur)
-        {
-            return null;
-        }
-        if (cur)
-        {
-            try
-            {
-                cur = JSON.parse(cur);
-            }
-            catch (e)
-            {
-                console.warn('Invalid JSON in '+(path instanceof Array ? path.join('/') : path)+': '+e);
-            }
-        }
-        return cur;
-    }
-
-    // userInfo: { name: string, type: string, perms: static_perms[type], groups: { [string]: true } }
-    _check_compare(check, userInfo, checked)
-    {
-        let key = String(check.key);
-        if (key.substr(0, this.prefix.length) !== this.prefix)
-        {
-            return false;
-        }
-        key = key.substr(this.prefix.length);
-        if (key in userInfo.perms.keys)
-        {
-            return true;
-        }
-        for (const pfx in userInfo.perms.prefixes)
-        {
-            if (key.substr(0, pfx.length) == pfx)
-            {
-                return true;
-            }
-        }
-        if (userInfo.type == 'client')
-        {
-            // Image permissions
-            if (key.substr(0, 14) == '/config/inode/')
-            {
-                // Allowed to check that a key does not exist
-                if (check.target == 'VERSION' && check.version == 0)
-                {
-                    checked['M'+key] = true;
-                    return true;
-                }
-                else if (check.target == 'MOD')
-                {
-                    const data = this._get(check.key);
-                    if (!data || data.mod_revision != check.mod_revision)
-                    {
-                        // Break check to trigger CAS failure
-                        check.mod_revision = '18446744073709551615'; // UINT64_MAX
-                        return true;
-                    }
-                    const inode = this._decode(check.key, data.value);
-                    if (inode && (inode.owner_group && userInfo.groups[inode.owner_group] ||
-                        inode.owner === userInfo.name))
-                    {
-                        checked['M'+key] = true;
-                        return true;
-                    }
-                }
-                return false;
-            }
-            if (key.substr(0, 13) == '/index/image/')
-            {
-                // Allowed to check that a key does not exist
-                if (check.target == 'VERSION' && check.version == 0)
-                {
-                    checked['M'+key] = true;
-                    return true;
-                }
-                else if (check.target == 'MOD')
-                {
-                    let data = this._get(check.key);
-                    if (!data || data.mod_revision != check.mod_revision)
-                    {
-                        // Break check to trigger CAS failure
-                        check.mod_revision = '18446744073709551615'; // UINT64_MAX
-                        return true;
-                    }
-                    data = this._decode(check.key, data.value);
-                    if (data)
-                    {
-                        const inode = this._get([ ...this.prefix_parts, 'config', 'inode', data.pool_id, data.id ], true);
-                        if (inode && (inode.owner_group && userInfo.groups[inode.owner_group] ||
-                            inode.owner === userInfo.name))
-                        {
-                            checked['M'+key] = true;
-                            return true;
-                        }
-                    }
-                }
-                return false;
-            }
-            if (key.substr(0, 13) == '/index/maxid/')
-            {
-                const pool_id = key.substr(13);
-                const pool_cfg = this._get([ ...this.prefix_parts, 'config', 'pools' ], true);
-                if (!pool_cfg || !pool_cfg[pool_id] || !pool_cfg[pool_id].creator_group || !userInfo.groups[pool_cfg[pool_id].creator_group])
-                {
-                    return false;
-                }
-                if (check.target == 'VERSION' && check.version == 0)
-                {
-                    checked['I'+parseInt(key.substr(13))+'_0'] = true;
-                    return true;
-                }
-                else if (check.target == 'MOD')
-                {
-                    const data = this._get(check.key);
-                    if (!data || data.mod_revision != check.mod_revision)
-                    {
-                        // Break check to trigger CAS failure
-                        check.mod_revision = '18446744073709551615'; // UINT64_MAX
-                        return true;
-                    }
-                    checked['I'+parseInt(key.substr(13))+'_'+data.value] = true;
-                    return true;
-                }
-                return false;
-            }
-        }
-        return false;
-    }
-
-    _check_read(kv, userInfo)
-    {
-        let key = String(kv.key);
-        if (key.substr(0, this.prefix.length) !== this.prefix)
-        {
-            return false;
-        }
-        key = key.substr(this.prefix.length);
-        if (key in userInfo.perms.keys)
-        {
-            return true;
-        }
-        for (const pfx in userInfo.perms.prefixes)
-        {
-            if (key.substr(0, pfx.length) == pfx)
-            {
-                return true;
-            }
-        }
-        if (userInfo.type == 'client')
-        {
-            // Image permissions
-            if (key.substr(0, 14) == '/config/inode/')
-            {
-                const inode = this._decode(kv.key, kv.value);
-                if (inode && (inode.reader_group && userInfo.groups[inode.reader_group] ||
-                    inode.owner_group && userInfo.groups[inode.owner_group] ||
-                    inode.owner === userInfo.name))
-                {
-                    return true;
-                }
-                return false;
-            }
-            if (key.substr(0, 13) == '/index/image/')
-            {
-                const data = this._decode(kv.key, kv.value);
-                const inode = this._get([ ...this.prefix_parts, 'config', 'inode', data.pool_id, data.id ], true);
-                if (inode && (inode.reader_group && userInfo.groups[inode.reader_group] ||
-                    inode.owner_group && userInfo.groups[inode.owner_group] ||
-                    inode.owner === userInfo.name))
-                {
-                    return true;
-                }
-                return false;
-            }
-        }
-        return false;
-    }
-
-    _check_write(put, userInfo, checked)
-    {
-        let key = String(put.key);
-        if (key.substr(0, this.prefix.length) !== this.prefix)
-        {
-            return false;
-        }
-        key = key.substr(this.prefix.length);
-        if (userInfo.perms.keys[key])
-        {
-            return true;
-        }
-        for (const pfx in userInfo.perms.prefixes)
-        {
-            if (userInfo.perms.prefixes[pfx] && key.substr(0, pfx.length) == pfx)
-            {
-                return true;
-            }
-        }
-        if (checked && userInfo.type == 'client')
-        {
-            if (key.substr(0, 13) == '/index/maxid/' &&
-                checked['I'+parseInt(key.substr(13))+'_'+(put.value-1)])
-            {
-                // Allowed to increment maxid
-                return true;
-            }
-            if (checked['M'+key])
-            {
-                // Allowed to modify known images with CAS checks
-                return true;
-            }
-        }
-        return false;
-    }
-
-    _check_req(req, userInfo, checked)
-    {
-        let r;
-        if ((r = (req.request_range || req.requestRange)))
-        {
-            // All range queries are allowed, but responses are filtered - it's simpler
-        }
-        else if ((r = (req.request_put || req.requestPut)))
-        {
-            if (!this._check_write(r, userInfo, checked))
-                return false;
-        }
-        else if ((r = (req.request_delete_range || req.requestDeleteRange)))
-        {
-            if (!r.range_end || r.range_end === r.key)
-            {
-                if (!this._check_write({ key: r.key }, userInfo))
-                    return false;
-            }
-            else
-            {
-                // All keys in range must satisfy prefix
-                r.range_end = String(r.range_end);
-                if (r.key.length != r.range_end.length ||
-                    r.key[r.key.length-1] != '/' ||
-                    r.range_end[r.range_end.length-1] != '0')
-                {
-                    return false;
-                }
-                let key = r.key.substr(this.prefix.length);
-                let found = false;
-                for (const pfx in userInfo.perms.prefixes)
-                {
-                    if (userInfo.perms.prefixes[pfx] && key.substr(0, pfx.length) == pfx)
-                    {
-                        found = true;
-                        break;
-                    }
-                }
-                if (!found)
-                    return false;
-            }
-        }
-        return true;
-    }
-
-    _get_user(username)
-    {
-        if (!username)
-        {
-            return null;
-        }
-        let userInfo = this._get([ ...this.prefix_parts, 'config', 'user', username ], true);
-        if (!userInfo)
-        {
-            userInfo = { type: 'client' };
-        }
-        userInfo.perms = static_perms[userInfo.type] || static_perms['invalid'];
-        userInfo.name = username;
-        if (userInfo.groups instanceof Array)
-        {
-            userInfo.groups = userInfo.groups.reduce((a, c) => { a[c] = true; return a; }, {});
-        }
-        else
-        {
-            userInfo.groups = {};
-        }
-        return userInfo;
-    }
-
-    filter_api(username, api/*, data*/)
-    {
-        if (username === 'root')
-        {
-            return true;
-        }
-        const userInfo = this._get([ ...this.prefix_parts, 'config', 'user', username ], true);
-        return userInfo && api_perms[userInfo.type] && api_perms[userInfo.type][api];
-    }
-
-    filter_txn(username, txn)
-    {
-        if (username === 'root')
-        {
-            return true;
-        }
-        const userInfo = this._get_user(username);
-        if (!userInfo)
-        {
-            return null;
-        }
-        const checked = {};
-        if (txn.compare)
-        {
-            for (const check of txn.compare)
-            {
-                if (!this._check_compare(check, userInfo, checked))
-                    return null;
-            }
-        }
-        // Special transactions:
-        // 1. create image: create config/inode and index/image, increment index/maxid/<pool> (with CAS)
-        // 2. create snapshot: same as create image but also rename previous to @snap
-        if (txn.success)
-        {
-            for (const req of txn.success)
-            {
-                if (!this._check_req(req, userInfo, checked))
-                    return null;
-            }
-        }
-        if (txn.failure)
-        {
-            for (const req of txn.failure)
-            {
-                if (!this._check_req(req, userInfo, null))
-                    return null;
-            }
-        }
-        return txn;
-    }
-
-    filter_txn_response(username, txn, res)
-    {
-        if (!res.responses || username === 'root')
-        {
-            return;
-        }
-        const userInfo = this._get_user(username);
-        if (!userInfo)
-        {
-            for (const resp of res.responses)
-            {
-                if (resp.response_range && resp.response_range.kvs)
-                {
-                    resp.response_range.kvs = [];
-                }
-            }
-            return;
-        }
-        for (const resp of res.responses)
-        {
-            if (resp.response_range && resp.response_range.kvs)
-            {
-                resp.response_range.kvs = resp.response_range.kvs.filter(kv => this._check_read(kv, userInfo));
-            }
-        }
-    }
-
-    filter_watch_message(username, msg)
-    {
-        if (!msg.result || !msg.result.events || username === 'root')
-        {
-            return;
-        }
-        const userInfo = this._get_user(username);
-        if (!userInfo)
-        {
-            msg.result.events = [];
-            return;
-        }
-        msg.result.events = msg.result.events.filter(ev => this._check_read(ev.kv, userInfo));
-    }
-}
-
-module.exports = VitastorAuthFilter;

node-binding/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vitastor",
-  "version": "3.0.8",
+  "version": "3.0.15",
   "description": "Low-level native bindings to Vitastor client library",
   "main": "index.js",
   "keywords": [

patches/VitastorPlugin.pm

@@ -366,15 +366,38 @@ sub map_volume
     my $prefix = defined $scfg->{vitastor_prefix} ? $scfg->{vitastor_prefix} : 'pve/';
 
     my ($vtype, $img_name, $vmid) = $class->parse_volname($volname);
-    my $name = $img_name;
+    my $name = $prefix.$img_name;
     $name .= '@'.$snapname if $snapname;
 
     my $mapped = run_cli($scfg, [ 'ls' ], binary => '/usr/bin/vitastor-nbd');
-    my ($kerneldev) = grep { $mapped->{$_}->{image} eq $prefix.$name } keys %$mapped;
-    return $kerneldev if $kerneldev && -b $kerneldev; # already mapped
+    my ($kerneldev) = grep {
+        $mapped->{$_} && $mapped->{$_}->{image} && $mapped->{$_}->{image} eq $name
+    } keys %$mapped;
 
-    $kerneldev = run_cli($scfg, [ 'map', '--image', $prefix.$name ], binary => '/usr/bin/vitastor-nbd', json => 0);
-    return $kerneldev;
+    if ($kerneldev && -b $kerneldev)
+    {
+        my $size = `/usr/sbin/blockdev --getsize64 $kerneldev`;
+        return $kerneldev if $size && $size > 0;
+    }
+
+    my $map_out = run_cli($scfg, [ 'map', '--image', $name ], binary => '/usr/bin/vitastor-nbd', json => 0);
+    $map_out =~ s/^\s+|\s+$//gso;
+
+    # Wait until the device is started
+    for (my $i = 0; $i < 100; $i++)
+    {
+        $mapped = run_cli($scfg, [ 'ls' ], binary => '/usr/bin/vitastor-nbd');
+        ($kerneldev) = grep { $mapped->{$_} && $mapped->{$_}->{image} && $mapped->{$_}->{image} eq $name } keys %$mapped;
+        if ($kerneldev && -b $kerneldev)
+        {
+            my $size = `/usr/sbin/blockdev --getsize64 $kerneldev`;
+            return $kerneldev if $size && $size > 0;
+        }
+        select(undef, undef, undef, 0.1);
+    }
+
+    die "Failed to map Vitastor image $name via NBD".
+        ($map_out ? ", vitastor-nbd map returned '$map_out'" : "")."\n";
 }
 
 sub unmap_volume
@@ -383,13 +406,19 @@ sub unmap_volume
     my $prefix = defined $scfg->{vitastor_prefix} ? $scfg->{vitastor_prefix} : 'pve/';
 
     my ($vtype, $name, $vmid) = $class->parse_volname($volname);
+    $name = $prefix.$name;
     $name .= '@'.$snapname if $snapname;
-
     my $mapped = run_cli($scfg, [ 'ls' ], binary => '/usr/bin/vitastor-nbd');
-    my ($kerneldev) = grep { $mapped->{$_}->{image} eq $prefix.$name } keys %$mapped;
-    if ($kerneldev && -b $kerneldev)
+
+    my @kerneldevs = grep {
+        $mapped->{$_} && $mapped->{$_}->{image} && $mapped->{$_}->{image} eq $name
+    } keys %$mapped;
+
+    for my $kerneldev (@kerneldevs)
     {
-        run_cli($scfg, [ 'unmap', $kerneldev ], binary => '/usr/bin/vitastor-nbd', json => 0);
+        next if !$kerneldev || !-b $kerneldev;
+        eval { run_cli($scfg, [ 'unmap', $kerneldev ], binary => '/usr/bin/vitastor-nbd', json => 0); };
+        warn "Failed to unmap Vitastor image $name from $kerneldev: $@" if $@;
     }
 
     return 1;
@@ -405,7 +434,13 @@ sub activate_volume
 sub deactivate_volume
 {
     my ($class, $storeid, $scfg, $volname, $snapname, $cache) = @_;
-    $class->unmap_volume($storeid, $scfg, $volname, $snapname) if $scfg->{vitastor_nbd};
+
+    # Even with vitastor_nbd=0, Proxmox may call map_volume() for special
+    # volumes like tpmstate0 because swtpm needs a local file/block path.
+    # Therefore, always try to unmap an existing NBD mapping here.
+    # unmap_volume() is a no-op if the volume is not currently mapped.
+    $class->unmap_volume($storeid, $scfg, $volname, $snapname);
+
     return 1;
 }
 

patches/cinder-vitastor.py

@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils
 
-VITASTOR_VERSION = '3.0.8'
+VITASTOR_VERSION = '3.0.15'
 
 LOG = logging.getLogger(__name__)
 

rpm/vitastor-el10.Dockerfile

@@ -11,7 +11,7 @@ WORKDIR /root
 RUN sed -i 's/enabled=0/enabled=1/' /etc/yum.repos.d/*.repo
 RUN dnf -y install epel-release dnf-plugins-core
 RUN dnf -y install https://vitastor.io/rpms/centos/10/vitastor-release-1.0-1.el10.noarch.rpm
-RUN dnf -y install gcc-c++ gperftools-devel fio nodejs rpm-build jerasure-devel isa-l-devel gf-complete-devel rdma-core-devel cmake libnl3-devel c-ares-devel
+RUN dnf -y install gcc-c++ gperftools-devel fio nodejs rpm-build jerasure-devel isa-l-devel gf-complete-devel rdma-core-devel cmake libnl3-devel
 RUN dnf download --source fio
 RUN rpm --nomd5 -i fio*.src.rpm
 RUN cd ~/rpmbuild/SPECS && dnf builddep -y --spec fio.spec

rpm/vitastor-el10.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.8
+Version:        3.0.15
 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-3.0.8.el10.tar.gz
+Source0:        vitastor-3.0.15.el10.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-c++
@@ -16,7 +16,6 @@ BuildRequires:  gf-complete-devel
 BuildRequires:  rdma-core-devel
 BuildRequires:  cmake
 BuildRequires:  libnl3-devel
-BuildRequires:  c-ares-devel
 Requires:       vitastor-osd = %{version}-%{release}
 Requires:       vitastor-mon = %{version}-%{release}
 Requires:       vitastor-client = %{version}-%{release}

rpm/vitastor-el7.Dockerfile

@@ -15,7 +15,7 @@ RUN yum -y --enablerepo=extras install centos-release-scl epel-release yum-utils
 RUN perl -i -pe 's!mirrorlist=!#mirrorlist=!s; s!#\s*baseurl=http://mirror.centos.org!baseurl=http://vault.centos.org!' /etc/yum.repos.d/CentOS-SCLo-scl*.repo
 RUN yum -y install https://vitastor.io/rpms/centos/7/vitastor-release-1.0-1.el7.noarch.rpm
 RUN yum -y install devtoolset-9-gcc-c++ devtoolset-9-libatomic-devel gcc make cmake gperftools-devel \
-    fio rh-nodejs12 jerasure-devel libisa-l-devel gf-complete-devel rdma-core-devel libnl3-devel c-ares-devel
+    fio rh-nodejs12 jerasure-devel libisa-l-devel gf-complete-devel rdma-core-devel libnl3-devel
 RUN yumdownloader --disablerepo=centos-sclo-rh --source fio
 RUN rpm --nomd5 -i fio*.src.rpm
 RUN rm -f /etc/yum.repos.d/CentOS-Media.repo

rpm/vitastor-el7.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.8
+Version:        3.0.15
 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-3.0.8.el7.tar.gz
+Source0:        vitastor-3.0.15.el7.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  devtoolset-9-gcc-c++
@@ -17,7 +17,6 @@ BuildRequires:  gf-complete-devel
 BuildRequires:  rdma-core-devel
 BuildRequires:  cmake3
 BuildRequires:  libnl3-devel
-BuildRequires:  c-ares-devel
 Requires:       vitastor-osd = %{version}-%{release}
 Requires:       vitastor-mon = %{version}-%{release}
 Requires:       vitastor-client = %{version}-%{release}

rpm/vitastor-el8.Dockerfile

@@ -13,7 +13,7 @@ RUN dnf -y install centos-release-advanced-virtualization epel-release dnf-plugi
 RUN sed -i 's/^mirrorlist=/#mirrorlist=/; s!#baseurl=.*!baseurl=http://vault.centos.org/centos/8.4.2105/virt/$basearch/$avdir/!; s!^baseurl=.*Source/.*!baseurl=http://vault.centos.org/centos/8.4.2105/virt/Source/advanced-virtualization/!' /etc/yum.repos.d/CentOS-Advanced-Virtualization.repo
 RUN yum -y install https://vitastor.io/rpms/centos/8/vitastor-release-1.0-1.el8.noarch.rpm
 RUN dnf -y install gcc-toolset-9 gcc-toolset-9-gcc-c++ gperftools-devel \
-    fio nodejs rpm-build jerasure-devel libisa-l-devel gf-complete-devel libibverbs-devel libarchive cmake libnl3-devel c-ares-devel
+    fio nodejs rpm-build jerasure-devel libisa-l-devel gf-complete-devel libibverbs-devel libarchive cmake libnl3-devel
 RUN dnf download --source fio
 RUN rpm --nomd5 -i fio*.src.rpm
 RUN cd ~/rpmbuild/SPECS && dnf builddep -y --enablerepo=powertools --spec fio.spec

rpm/vitastor-el8.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.8
+Version:        3.0.15
 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-3.0.8.el8.tar.gz
+Source0:        vitastor-3.0.15.el8.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-toolset-9-gcc-c++
@@ -16,7 +16,6 @@ BuildRequires:  gf-complete-devel
 BuildRequires:  rdma-core-devel
 BuildRequires:  cmake
 BuildRequires:  libnl3-devel
-BuildRequires:  c-ares-devel
 Requires:       vitastor-osd = %{version}-%{release}
 Requires:       vitastor-mon = %{version}-%{release}
 Requires:       vitastor-client = %{version}-%{release}

rpm/vitastor-el9.Dockerfile

@@ -10,7 +10,7 @@ WORKDIR /root
 RUN sed -i 's/enabled=0/enabled=1/' /etc/yum.repos.d/*.repo
 RUN dnf -y install epel-release dnf-plugins-core
 RUN dnf -y install https://vitastor.io/rpms/centos/9/vitastor-release-1.0-1.el9.noarch.rpm
-RUN dnf -y install gcc-c++ gperftools-devel fio nodejs rpm-build jerasure-devel libisa-l-devel gf-complete-devel rdma-core-devel libarchive cmake libnl3-devel c-ares-devel
+RUN dnf -y install gcc-c++ gperftools-devel fio nodejs rpm-build jerasure-devel libisa-l-devel gf-complete-devel rdma-core-devel libarchive cmake libnl3-devel
 RUN dnf download --source fio
 RUN rpm --nomd5 -i fio*.src.rpm
 RUN cd ~/rpmbuild/SPECS && dnf builddep -y --spec fio.spec

rpm/vitastor-el9.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.8
+Version:        3.0.15
 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-3.0.8.el9.tar.gz
+Source0:        vitastor-3.0.15.el9.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-c++
@@ -16,7 +16,6 @@ BuildRequires:  gf-complete-devel
 BuildRequires:  rdma-core-devel
 BuildRequires:  cmake
 BuildRequires:  libnl3-devel
-BuildRequires:  c-ares-devel
 Requires:       vitastor-osd = %{version}-%{release}
 Requires:       vitastor-mon = %{version}-%{release}
 Requires:       vitastor-client = %{version}-%{release}

src/CMakeLists.txt

@@ -20,7 +20,7 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 endif()
 set(ENABLE_COVERAGE false CACHE BOOL "Enable code coverage")
 
-add_definitions(-DVITASTOR_VERSION="3.0.8")
+add_definitions(-DVITASTOR_VERSION="3.0.15")
 add_definitions(-D_GNU_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fdiagnostics-color=always -fno-omit-frame-pointer -fvisibility=hidden -I ${CMAKE_SOURCE_DIR}/src)
 add_link_options(-fno-omit-frame-pointer)
 if (${WITH_ASAN})
@@ -74,14 +74,6 @@ if (RDMACM_LIBRARIES)
 	add_definitions(-DWITH_RDMACM)
 endif (RDMACM_LIBRARIES)
 
-find_package(OpenSSL REQUIRED)
-if (OPENSSL_FOUND)
-	add_definitions(-DWITH_OPENSSL)
-endif (OPENSSL_FOUND)
-
-pkg_check_modules(CARES REQUIRED libcares)
-include_directories(${CARES_INCLUDE_DIRS})
-
 if (${WITH_SYSTEM_LIBURING})
 	pkg_check_modules(LIBURING REQUIRED liburing>=2.10)
 	include_directories(${LIBURING_INCLUDE_DIRS})

src/blockstore/blockstore_disk.cpp

@@ -83,23 +83,20 @@ void blockstore_disk_t::parse_config(std::map<std::string, std::string> & config
     {
         data_csum_type = BLOCKSTORE_CSUM_CRC32C;
     }
-    else if (config["data_csum_type"] == "xxh3_32")
-    {
-        data_csum_type = BLOCKSTORE_CSUM_XXH3_32;
-    }
     else if (config["data_csum_type"] == "" || config["data_csum_type"] == "none")
     {
         data_csum_type = BLOCKSTORE_CSUM_NONE;
     }
     else
     {
-        throw std::runtime_error("data_csum_type="+config["data_csum_type"]+" is unsupported, only \"crc32c\", \"xxh3_32\" and \"none\" are supported");
+        throw std::runtime_error("data_csum_type="+config["data_csum_type"]+" is unsupported, only \"crc32c\" and \"none\" are supported");
     }
     csum_block_size = parse_size(config["csum_block_size"]);
     discard_on_start = config.find("discard_on_start") != config.end() &&
         (config["discard_on_start"] == "true" || config["discard_on_start"] == "1" || config["discard_on_start"] == "yes");
     gc_on_start = config.find("gc_on_start") == config.end() ||
         (config["gc_on_start"] == "true" || config["gc_on_start"] == "1" || config["gc_on_start"] == "yes");
+    skip_double_claim = (config["skip_double_claim"] == "true" || config["skip_double_claim"] == "1" || config["skip_double_claim"] == "yes");
     min_discard_size = parse_size(config["min_discard_size"]);
     if (!min_discard_size)
         min_discard_size = 1024*1024;
@@ -520,7 +517,7 @@ void blockstore_disk_t::close_all()
 
 // Sadly DISCARD only works through ioctl(), but it seems to always block the device queue,
 // so it's not a big deal that we can only run it synchronously.
-int blockstore_disk_t::trim_data(std::function<bool(uint64_t)> is_free)
+int blockstore_disk_t::trim_data(std::function<bool(uint64_t)> is_used)
 {
     if (mock_mode)
     {
@@ -531,7 +528,7 @@ int blockstore_disk_t::trim_data(std::function<bool(uint64_t)> is_free)
     uint64_t discarded = 0;
     for (; i <= block_count; i++)
     {
-        if (i >= block_count || is_free(i))
+        if (i >= block_count || is_used(i))
         {
             if (i > j && (i-j)*data_block_size >= min_discard_size)
             {
@@ -544,17 +541,21 @@ int blockstore_disk_t::trim_data(std::function<bool(uint64_t)> is_free)
                     if (range[0] % discard_granularity)
                         range[0] = range[0] + discard_granularity - (range[0] % discard_granularity);
                     if (range[0] >= range[1])
-                        continue;
-                    range[1] -= range[0];
+                        range[1] = 0;
+                    else
+                        range[1] -= range[0];
                 }
-                r = ioctl(data_fd, BLKDISCARD, &range);
-                if (r != 0)
+                if (range[1] > 0)
                 {
-                    fprintf(stderr, "Failed to execute BLKDISCARD %ju+%ju on %s: %s (code %d)\n",
-                        range[0], range[1], data_device.c_str(), strerror(-r), r);
-                    return -errno;
+                    r = ioctl(data_fd, BLKDISCARD, &range);
+                    if (r != 0)
+                    {
+                        fprintf(stderr, "Failed to execute BLKDISCARD %ju+%ju on %s: %s (code %d)\n",
+                            range[0], range[1], data_device.c_str(), strerror(-r), r);
+                        return -errno;
+                    }
+                    discarded += range[1];
                 }
-                discarded += range[1];
             }
             j = i+1;
         }

src/blockstore/blockstore_disk.h

@@ -16,7 +16,6 @@
 #define BLOCKSTORE_CSUM_NONE 0
 // Lower byte of checksum type is its length
 #define BLOCKSTORE_CSUM_CRC32C 0x104
-#define BLOCKSTORE_CSUM_XXH3_32 0x204
 
 #define MOCK_DATA_FD 1000
 #define MOCK_META_FD 1001
@@ -60,6 +59,8 @@ struct blockstore_disk_t
     bool discard_on_start = false;
     // GC on start (new store)
     bool gc_on_start = true;
+    // Skip double claim conflicts on start (new store, temporary until the bug is found)
+    bool skip_double_claim = false;
     uint64_t min_discard_size = 1024*1024;
     uint64_t discard_granularity = 0;
 
@@ -82,7 +83,7 @@ struct blockstore_disk_t
     void calc_lengths(bool skip_meta_check = false);
     void check_lengths();
     void close_all();
-    int trim_data(std::function<bool(uint64_t)> is_free);
+    int trim_data(std::function<bool(uint64_t)> is_used);
 
     inline uint64_t dirty_dyn_size(uint64_t offset, uint64_t len)
     {

src/blockstore/blockstore_flush.cpp

@@ -174,14 +174,18 @@ bool journal_flusher_co::loop()
     else if (wait_state == 19) goto resume_19;
     else if (wait_state == 20) goto resume_20;
     else if (wait_state == 21) goto resume_21;
+    else if (wait_state == 22) goto resume_22;
+    else if (wait_state == 23) goto resume_23;
+    else if (wait_state == 24) goto resume_24;
+    else if (wait_state == 25) goto resume_25;
 resume_0:
     wait_state = 0;
     wait_count = 0;
     cur_oid = {};
     res = bs->heap->get_next_compact(cur_oid);
-    // Advance fsynced_lsn every <journal_trim_interval> intent writes
     if ((bs->intent_write_counter >= bs->journal_trim_interval) && co_id == 0)
     {
+        // Advance fsynced_lsn every <journal_trim_interval> intent writes
         bs->intent_write_counter = 0;
 resume_17:
 resume_18:
@@ -196,6 +200,7 @@ resume_21:
     if (res == ENOENT && flusher->force_start > 0 && co_id == 0 &&
         (!bs->dsk.disable_journal_fsync || !bs->dsk.disable_meta_fsync || !bs->dsk.disable_data_fsync))
     {
+        // When under pressure, do an additional fsync to force entries to be marked compactable
         flusher->active_flushers++;
 resume_14:
 resume_15:
@@ -259,11 +264,9 @@ resume_1:
         if (wr->type() == BS_HEAP_SMALL_WRITE ||
             wr->type() == BS_HEAP_INTENT_WRITE && bs->dsk.csum_block_size > bs->dsk.bitmap_granularity)
         {
-            auto res = bs->prepare_read(read_vec, cur_obj, wr, 0, bs->dsk.data_block_size,
+            bs->prepare_read(read_vec, cur_obj, wr, 0, bs->dsk.data_block_size,
                 wr->type() == BS_HEAP_INTENT_WRITE && bs->dsk.csum_block_size > bs->dsk.bitmap_granularity && !bs->perfect_csum_update
                 ? COPY_BUF_SKIP_CSUM : 0);
-            if (res > 0)
-                copy_count++;
         }
     });
     if (!compact_info.compact_lsn)
@@ -273,30 +276,53 @@ resume_1:
         bs->heap->unlock_entry(cur_oid);
         goto resume_0;
     }
-    mem_or(new_bmp, compact_info.clean_wr->get_int_bitmap(bs->heap), bs->dsk.clean_entry_bitmap_size);
-    if (!bitmap_copied)
-    {
-        memcpy(new_ext_bmp, compact_info.clean_wr->get_ext_bitmap(bs->heap), bs->dsk.clean_entry_bitmap_size);
-        bitmap_copied = true;
-    }
-    if (bs->dsk.csum_block_size && bs->dsk.csum_block_size <= bs->dsk.bitmap_granularity)
-    {
-        memcpy(new_csums, compact_info.clean_wr->get_checksums(bs->heap), bs->dsk.data_block_size/bs->dsk.csum_block_size * (bs->dsk.data_csum_type & 0xFF));
-        for (size_t i = csum_copy.size(); i > 0; i--)
-        {
-            auto wr = csum_copy[i-1];
-            memcpy(new_csums + wr->small().offset/bs->dsk.csum_block_size*(bs->dsk.data_csum_type & 0xFF),
-                wr->get_checksums(bs->heap), wr->small().len/bs->dsk.csum_block_size*(bs->dsk.data_csum_type & 0xFF));
-        }
-        csum_copy.clear();
-    }
-    clean_loc = compact_info.clean_wr->big_location(bs->heap);
     flusher->active_flushers++;
-    if (bs->log_level > 10)
+    for (i = 0; i < read_vec.size(); i++)
     {
-        printf("Compacting %jx:%jx v%ju..v%ju / l%ju..l%ju (%d writes)\n", cur_oid.inode, cur_oid.stripe,
-            compact_info.clean_wr->version, compact_info.compact_version,
-            compact_info.clean_wr->lsn, compact_info.compact_lsn, copy_count);
+        if ((read_vec[i].copy_flags & COPY_BUF_JOURNAL) &&
+            !(read_vec[i].copy_flags & COPY_BUF_COALESCED))
+        {
+            copy_count++;
+        }
+    }
+    if (copy_count > 0 && !bs->dsk.disable_data_fsync)
+    {
+        init_fsync_data();
+    }
+    if (compact_info.do_delete)
+    {
+        if (bs->log_level > 10)
+        {
+            printf("Compacting %jx:%jx up to l%ju (delete)\n", cur_oid.inode, cur_oid.stripe, compact_info.compact_lsn);
+        }
+        clean_loc = UINT64_MAX;
+    }
+    else
+    {
+        if (bs->log_level > 10)
+        {
+            printf("Compacting %jx:%jx v%ju..v%ju / l%ju..l%ju (%d writes)\n", cur_oid.inode, cur_oid.stripe,
+                compact_info.clean_wr->version, compact_info.compact_version,
+                compact_info.clean_wr->lsn, compact_info.compact_lsn, copy_count);
+        }
+        mem_or(new_bmp, compact_info.clean_wr->get_int_bitmap(bs->heap), bs->dsk.clean_entry_bitmap_size);
+        if (!bitmap_copied)
+        {
+            memcpy(new_ext_bmp, compact_info.clean_wr->get_ext_bitmap(bs->heap), bs->dsk.clean_entry_bitmap_size);
+            bitmap_copied = true;
+        }
+        if (bs->dsk.csum_block_size && bs->dsk.csum_block_size <= bs->dsk.bitmap_granularity)
+        {
+            memcpy(new_csums, compact_info.clean_wr->get_checksums(bs->heap), bs->dsk.data_block_size/bs->dsk.csum_block_size * (bs->dsk.data_csum_type & 0xFF));
+            for (size_t i = csum_copy.size(); i > 0; i--)
+            {
+                auto wr = csum_copy[i-1];
+                memcpy(new_csums + wr->small().offset/bs->dsk.csum_block_size*(bs->dsk.data_csum_type & 0xFF),
+                    wr->get_checksums(bs->heap), wr->small().len/bs->dsk.csum_block_size*(bs->dsk.data_csum_type & 0xFF));
+            }
+            csum_copy.clear();
+        }
+        clean_loc = compact_info.clean_wr->big_location(bs->heap);
     }
     overwrite_start = overwrite_end = 0;
     if (read_vec.size() > 0)
@@ -336,6 +362,13 @@ resume_3:
     if (res == ENOENT || res == EDOM)
     {
         // Abort compaction
+abort_compact:
+        if (copy_count > 0 && !bs->dsk.disable_data_fsync)
+        {
+            cur_sync->member_count--;
+            if (cur_sync->member_count > 0)
+                bs->ringloop->wakeup();
+        }
         flusher->flushing.erase(cur_oid);
         bs->heap->unlock_entry(cur_oid);
         flusher->active_flushers--;
@@ -349,10 +382,7 @@ resume_4:
         if (res == ENOENT)
         {
             // Abort compaction
-            flusher->flushing.erase(cur_oid);
-            bs->heap->unlock_entry(cur_oid);
-            flusher->active_flushers--;
-            goto resume_0;
+            goto abort_compact;
         }
         if (res == EAGAIN)
         {
@@ -381,14 +411,14 @@ resume_9:
     for (i = 0; i < read_vec.size(); i++)
     {
         if ((read_vec[i].copy_flags & COPY_BUF_JOURNAL) &&
-            !(read_vec[i].copy_flags & COPY_BUF_COALESCED) ||
-            (read_vec[i].copy_flags & COPY_BUF_PADDED)) // FIXME Shit, simplify these flags
+            !(read_vec[i].copy_flags & COPY_BUF_COALESCED))
         {
             assert(read_vec[i].buf);
             await_sqe(10);
             data->iov = (struct iovec){ read_vec[i].buf + (read_vec[i].copy_flags & COPY_BUF_PADDED
                 ? read_vec[i].offset - read_vec[i].disk_offset : 0), (size_t)read_vec[i].len };
             data->callback = simple_callback_w;
+            assert(clean_loc + read_vec[i].offset + data->iov.iov_len <= bs->dsk.block_count*bs->dsk.data_block_size);
             io_uring_prep_writev(sqe, bs->dsk.data_fd, &data->iov, 1, bs->dsk.data_offset + clean_loc + read_vec[i].offset);
             wait_count++;
         }
@@ -399,6 +429,17 @@ resume_11:
         wait_state = 11;
         return false;
     }
+    if (copy_count > 0 && !bs->dsk.disable_data_fsync)
+    {
+resume_22:
+resume_23:
+resume_24:
+resume_25:
+        if (!fsync_data(22))
+        {
+            return false;
+        }
+    }
     // Lock is only needed to prevent freeing the big_write because we overwrite it...
     bs->heap->unlock_entry(cur_oid);
     // Mark the object compacted, but don't free and remove small_writes
@@ -408,12 +449,14 @@ resume_11:
     if (!cur_obj)
     {
         // Abort compaction
+        flusher->active_flushers--;
         flusher->flushing.erase(cur_oid);
         goto resume_0;
     }
     if (!calc_block_checksums())
     {
         // Abort compaction
+        flusher->active_flushers--;
         flusher->flushing.erase(cur_oid);
         goto resume_0;
     }
@@ -422,6 +465,7 @@ resume_11:
     if (res == EBUSY)
     {
         // Abort compaction, object is already overwritten by something else
+        flusher->active_flushers--;
         flusher->flushing.erase(cur_oid);
         goto resume_0;
     }
@@ -586,13 +630,13 @@ int journal_flusher_co::check_and_punch_checksums()
             bs->heap->calc_block_checksums((uint32_t*)(new_csums+csum_off), vec.buf, punch_bmp, vec.offset, vec.offset+vec.len, true, NULL);
         }
     }
-    // Modified, we should add_punch_holes and then write the block to disk
+    // Modified, we should punch_holes and then write the block to disk
     return EBUSY;
 }
 
 bool journal_flusher_co::calc_block_checksums()
 {
-    if (bs->dsk.csum_block_size <= bs->dsk.bitmap_granularity)
+    if (bs->dsk.csum_block_size <= bs->dsk.bitmap_granularity || compact_info.do_delete)
     {
         return true;
     }
@@ -699,6 +743,67 @@ resume_1:
     return true;
 }
 
+void journal_flusher_co::init_fsync_data()
+{
+    cur_sync = flusher->data_syncs.begin();
+    if (cur_sync == flusher->data_syncs.end() || cur_sync->ready_count > 0)
+    {
+        cur_sync = flusher->data_syncs.emplace(cur_sync);
+    }
+    cur_sync->member_count++;
+}
+
+bool journal_flusher_co::fsync_data(int wait_base)
+{
+    if (wait_state == wait_base)
+        goto resume_0;
+    else if (wait_state == wait_base+1)
+        goto resume_1;
+    else if (wait_state == wait_base+2)
+        goto resume_2;
+    else if (wait_state == wait_base+3)
+        goto resume_3;
+    cur_sync->ready_count++;
+resume_0:
+    if (cur_sync->ready_count < cur_sync->member_count)
+    {
+        wait_state = wait_base;
+        return false;
+    }
+    if (!cur_sync->sent)
+    {
+        // Sync batch is ready. Do it.
+        await_sqe(1);
+        data->iov = { 0 };
+        data->callback = simple_callback_w;
+        io_uring_prep_fsync(sqe, bs->dsk.data_fd, IORING_FSYNC_DATASYNC);
+        cur_sync->sent = true;
+        wait_count++;
+resume_2:
+        if (wait_count > 0)
+        {
+            wait_state = wait_base+2;
+            return false;
+        }
+        cur_sync->done = true;
+        // Wake up other flushers
+        bs->ringloop->wakeup();
+    }
+resume_3:
+    if (!cur_sync->done)
+    {
+        wait_state = wait_base+3;
+        return false;
+    }
+    cur_sync->done_count++;
+    if (cur_sync->done_count >= cur_sync->member_count)
+    {
+        flusher->data_syncs.erase(cur_sync);
+        cur_sync = flusher->data_syncs.end();
+    }
+    return true;
+}
+
 bool journal_flusher_co::fsync_meta(int wait_base)
 {
     if (wait_state == wait_base)        goto resume_0;

src/blockstore/blockstore_flush.h

@@ -25,6 +25,15 @@ struct flusher_meta_write_t
     std::map<uint64_t, meta_sector_t>::iterator it;
 };
 
+struct flusher_data_sync_t
+{
+    int member_count = 0;
+    int ready_count = 0;
+    int done_count = 0;
+    bool sent = false;
+    bool done = false;
+};
+
 class journal_flusher_t;
 
 // Journal flusher coroutine
@@ -58,6 +67,7 @@ class journal_flusher_co
     int i, res;
     bool read_to_fill_incomplete;
     int copy_count;
+    std::list<flusher_data_sync_t>::iterator cur_sync;
 
     friend class journal_flusher_t;
 
@@ -68,6 +78,8 @@ class journal_flusher_co
     bool calc_block_checksums();
     bool write_meta_block(int wait_base);
     bool read_buffered(int wait_base);
+    void init_fsync_data();
+    bool fsync_data(int wait_base);
     bool fsync_meta(int wait_base);
     bool fsync_buffer(int wait_base);
     bool trim_lsn(int wait_base);
@@ -88,6 +100,7 @@ class journal_flusher_t
 
     robin_hood::unordered_flat_set<object_id> flushing;
     int active_flushers = 0;
+    std::list<flusher_data_sync_t> data_syncs;
     int wanting_meta_fsync = 0;
     bool fsyncing_meta = false;
     int syncing_buffer = 0;

src/blockstore/blockstore_heap.h

@@ -43,7 +43,7 @@ struct __attribute__((__packed__)) heap_entry_t
 {
     uint16_t size;
     uint16_t entry_type;
-    uint32_t checksum;
+    uint32_t crc32c;
     uint64_t lsn;
     uint64_t inode;
     uint64_t stripe;
@@ -57,11 +57,11 @@ struct __attribute__((__packed__)) heap_entry_t
     inline heap_small_write_t& small() { return *(heap_small_write_t*)this; }
     inline heap_big_write_t& big() { return *(heap_big_write_t*)this; }
     inline heap_big_intent_t& big_intent() { return *(heap_big_intent_t*)this; }
-    bool is_garbage();
+    bool is_garbage() const;
     void set_garbage();
-    bool is_overwrite();
-    bool is_compactable();
-    bool is_before(heap_entry_t *other);
+    bool is_overwrite() const;
+    bool is_compactable() const;
+    bool is_before(const heap_entry_t *other) const;
     uint32_t get_size(blockstore_heap_t *heap);
     uint8_t *get_ext_bitmap(blockstore_heap_t *heap);
     uint8_t *get_int_bitmap(blockstore_heap_t *heap);
@@ -69,8 +69,7 @@ struct __attribute__((__packed__)) heap_entry_t
     uint32_t *get_checksum(blockstore_heap_t *heap);
     uint64_t big_location(blockstore_heap_t *heap);
     void set_big_location(blockstore_heap_t *heap, uint64_t location);
-    uint32_t calc_checksum(blockstore_heap_t *heap);
-    uint32_t calc_checksum(blockstore_disk_t *dsk);
+    uint32_t calc_crc32c();
 };
 
 struct __attribute__((__packed__)) heap_small_write_t
@@ -81,7 +80,7 @@ struct __attribute__((__packed__)) heap_small_write_t
     uint32_t offset;
     uint32_t len;
 
-    // Also includes 1 bitmap and 1 checksum after the bitmap if block checksums are disabled
+    // Also includes 1 bitmap and 1 crc32c after the bitmap if checksums are disabled
 };
 
 struct __attribute__((__packed__)) heap_big_write_t
@@ -99,7 +98,7 @@ struct __attribute__((__packed__)) heap_big_intent_t
     uint32_t offset;
     uint32_t len;
 
-    // Also includes 2 bitmaps and 1 checksums if block checksums are disabled
+    // Also includes 2 bitmaps and 1 crc32c if checksums are disabled
 };
 
 struct __attribute__((__packed__)) heap_list_item_t
@@ -159,6 +158,16 @@ struct heap_li_equal
     }
 };
 
+struct heap_recheck_state_t
+{
+    heap_entry_t *obj = NULL;
+    heap_entry_t *next_wr = NULL;
+    size_t total_reads = 0;
+    size_t sent_reads = 0;
+    size_t checked_reads = 0;
+    heap_entry_t *bad_wr = NULL;
+};
+
 using i64hash_t = robin_hood::hash<uint64_t>;
 using heap_inode_map_t = robin_hood::unordered_flat_set<heap_list_item_t*, heap_li_hash, heap_li_equal, 88>;
 using heap_block_index_t = robin_hood::unordered_flat_map<uint64_t,
@@ -209,9 +218,12 @@ class blockstore_heap_t
 
     bool marked_used_blocks = false;
     bool recheck_queue_filled = false;
-    std::vector<heap_list_item_t*> loaded_list_items;
+    std::vector<heap_list_item_t*> postponed_items;
+    std::vector<heap_list_item_t*> init_erase_items;
     std::set<uint32_t> recheck_modified_blocks;
     std::deque<heap_entry_t*> recheck_queue;
+    std::map<heap_entry_t*, heap_recheck_state_t> recheck_states;
+    size_t recheck_pending_reads = 0;
     int recheck_in_progress = 0;
     bool in_recheck = false;
     std::function<void(bool is_data, uint64_t offset, uint64_t len, uint8_t* buf, std::function<void()>)> recheck_cb;
@@ -220,7 +232,12 @@ class blockstore_heap_t
     uint64_t get_pg_id(inode_t inode, uint64_t stripe);
     bool validate_object(heap_entry_t *obj);
     void fill_recheck_queue();
+    void recheck_drop_entries(heap_entry_t *obj, heap_entry_t *bad_wr);
+    void recheck_start_reads(heap_recheck_state_t *st);
     int mark_used_blocks();
+    void init_free_bad_entry(heap_entry_t *wr);
+    void init_erase_bad_entry(heap_list_item_t *li);
+    bool init_erase_double_claim(heap_list_item_t *prev_li, heap_list_item_t *cur_li);
     void recheck_full_gc();
     void recheck_buffer(heap_entry_t *cwr, uint8_t *buf);
     void defragment_block(uint32_t block_num);
@@ -228,8 +245,9 @@ class blockstore_heap_t
 
     void gc_block(heap_block_info_t & inf);
     int allocate_entry(uint32_t entry_size, uint32_t *block_num, bool allow_last_free);
-    void insert_list_item(heap_list_item_t *li);
+    void insert_list_items(heap_list_item_t** v, size_t count, bool postpone);
     void remove_list_item(heap_list_item_t *li);
+    void unlink_list_item(heap_list_item_t *li);
     int add_entry(uint32_t wr_size, uint32_t *modified_block, bool allow_last_free,
         bool explicit_complete, std::function<void(heap_entry_t *wr)> fill_entry);
     int add_simple(heap_entry_t *obj, uint64_t version, uint32_t *modified_block, uint32_t entry_type);
@@ -344,7 +362,7 @@ public:
 
     // get metadata block data buffer and used space
     void get_meta_block(uint32_t block_num, uint8_t *buffer);
-    void fill_block_empty_space(uint8_t *buffer, uint32_t pos);
+    void fill_block_empty_space(uint8_t *buffer, uint64_t pos);
     uint32_t get_meta_block_used_space(uint32_t block_num);
 
     // get space usage statistics

src/blockstore/blockstore_impl.cpp

@@ -101,6 +101,7 @@ void blockstore_impl_t::loop()
         unsigned initial_ring_space = ringloop->space_left();
         int op_idx = 0, new_idx = 0;
         bool has_unfinished_writes = false;
+        bool has_unfinished_sync = false;
         for (; op_idx < submit_queue.size(); op_idx++, new_idx++)
         {
             auto op = submit_queue[op_idx];
@@ -138,7 +139,13 @@ void blockstore_impl_t::loop()
             else if (op->opcode == BS_OP_SYNC)
             {
                 // syncs only completed writes, so doesn't have to be blocked by anything
-                wr_st = continue_sync(op);
+                if (!has_unfinished_sync)
+                {
+                    wr_st = continue_sync(op);
+                    has_unfinished_sync = (wr_st != 2);
+                }
+                else
+                    wr_st = 0;
             }
             else if (op->opcode == BS_OP_STABLE || op->opcode == BS_OP_ROLLBACK)
             {
@@ -154,9 +161,7 @@ void blockstore_impl_t::loop()
                     wr_st = 2;
                 }
                 else
-                {
                     wr_st = 0;
-                }
             }
             if (wr_st == 2)
             {

src/blockstore/blockstore_impl.h

@@ -117,9 +117,12 @@ public:
 
     journal_flusher_t *flusher;
     int write_iodepth = 0;
-    int inflight_big = 0;
     int intent_write_counter = 0;
-    bool fsyncing_data = false;
+    uint64_t data_fsync_next = 0;
+    uint64_t data_fsync_cur = 0;
+    uint64_t data_fsync_sent = 0;
+    uint64_t data_fsync_done = 0;
+    std::deque<bool> data_fsyncs;
 
     bool live = false, queue_stall = false;
     ring_loop_i *ringloop = NULL;

src/blockstore/blockstore_init.cpp

@@ -10,7 +10,6 @@
 #define INIT_META_EMPTY 0
 #define INIT_META_READING 1
 #define INIT_META_READ_DONE 2
-#define INIT_META_WRITING 3
 
 #define GET_SQE() \
     sqe = bs->get_sqe();\
@@ -23,14 +22,15 @@ blockstore_init_meta::blockstore_init_meta(blockstore_impl_t *bs)
     this->bs = bs;
 }
 
-void blockstore_init_meta::handle_event(ring_data_t *data, int buf_num)
+void blockstore_init_meta::handle_event(ring_data_t *data, int buf_num, const char *op)
 {
-    if (data->res < 0)
+    if (data->res != data->iov.iov_len)
     {
-        throw std::runtime_error(
-            std::string("read metadata failed at offset ") + std::to_string(buf_num >= 0 ? bufs[buf_num].offset : last_read_offset) +
-            std::string(": ") + strerror(-data->res)
-        );
+        throw std::runtime_error(strprintf(
+            "%s failed at offset %ju: got %s (code %d), but expected %zu",
+            op, (buf_num >= 0 ? bufs[buf_num].offset : last_read_offset), strerror(-data->res),
+            data->res, data->iov.iov_len
+        ));
     }
     if (buf_num >= 0)
     {
@@ -60,7 +60,7 @@ int blockstore_init_meta::loop()
     GET_SQE();
     last_read_offset = 0;
     data->iov = { bs->meta_superblock, (size_t)bs->dsk.meta_block_size };
-    data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
+    data->callback = [this](ring_data_t *data) { handle_event(data, -1, "read metadata header"); };
     io_uring_prep_readv(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset);
     bs->ringloop->submit();
     submitted++;
@@ -72,25 +72,19 @@ resume_1:
     }
     if (is_zero((uint64_t*)bs->meta_superblock, bs->dsk.meta_block_size))
     {
-        {
-            blockstore_meta_header_v3_t *hdr = (blockstore_meta_header_v3_t *)bs->meta_superblock;
-            hdr->zero = 0;
-            hdr->magic = BLOCKSTORE_META_MAGIC_V1;
-            hdr->version = bs->dsk.meta_format;
-            hdr->meta_block_size = bs->dsk.meta_block_size;
-            hdr->data_block_size = bs->dsk.data_block_size;
-            hdr->bitmap_granularity = bs->dsk.bitmap_granularity;
-            if (bs->dsk.meta_format >= BLOCKSTORE_META_FORMAT_V2)
-            {
-                hdr->data_csum_type = bs->dsk.data_csum_type;
-                hdr->csum_block_size = bs->dsk.csum_block_size;
-            }
-            if (bs->dsk.meta_format >= BLOCKSTORE_META_FORMAT_HEAP)
-            {
-                hdr->meta_area_size = bs->dsk.meta_area_size;
-            }
-            hdr->set_crc32c();
-        }
+        assert(bs->dsk.meta_format == BLOCKSTORE_META_FORMAT_HEAP);
+        blockstore_meta_header_v3_t *hdr = (blockstore_meta_header_v3_t *)bs->meta_superblock;
+        hdr->zero = 0;
+        hdr->magic = BLOCKSTORE_META_MAGIC_V1;
+        hdr->version = bs->dsk.meta_format;
+        hdr->meta_block_size = bs->dsk.meta_block_size;
+        hdr->data_block_size = bs->dsk.data_block_size;
+        hdr->bitmap_granularity = bs->dsk.bitmap_granularity;
+        hdr->completed_lsn = 0;
+        hdr->data_csum_type = bs->dsk.data_csum_type;
+        hdr->csum_block_size = bs->dsk.csum_block_size;
+        hdr->meta_area_size = bs->dsk.meta_area_size;
+        hdr->set_crc32c();
         if (bs->readonly)
         {
             printf("Skipping metadata initialization because blockstore is readonly\n");
@@ -98,21 +92,8 @@ resume_1:
         else
         {
             printf("Initializing metadata area\n");
-            GET_SQE();
-            last_read_offset = 0;
-            data->iov = (struct iovec){ bs->meta_superblock, (size_t)bs->dsk.meta_block_size };
-            data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
-            io_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset);
-            bs->ringloop->submit();
-            submitted++;
-        resume_2:
-            if (submitted > 0)
-            {
-                wait_state = 2;
-                return 1;
-            }
-            zero_on_init = true;
         }
+        zero_on_init = true;
     }
     else
     {
@@ -153,9 +134,17 @@ resume_1:
             );
             exit(1);
         }
+        uint32_t csum = hdr->header_csum;
+        hdr->header_csum = 0;
+        if (crc32c(0, hdr, sizeof(*hdr)) != csum)
+        {
+            printf("Metadata header is corrupt (checksum mismatch).\n");
+            exit(1);
+        }
+        hdr->header_csum = csum;
     }
     bs->heap->start_load(((blockstore_meta_header_v3_t *)bs->meta_superblock)->completed_lsn);
-    if (bs->dsk.inmemory_journal)
+    if (bs->dsk.inmemory_journal && !zero_on_init)
     {
         // Read buffer area
         printf("Reading buffered data\n");
@@ -167,7 +156,7 @@ resume_1:
                 bs->buffer_area + md_offset,
                 (size_t)(bs->dsk.journal_len - md_offset < bs->metadata_buf_size ? bs->dsk.journal_len - md_offset : bs->metadata_buf_size),
             };
-            data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
+            data->callback = [this](ring_data_t *data) { handle_event(data, -1, "read buffer area"); };
             io_uring_prep_readv(sqe, bs->dsk.journal_fd, &data->iov, 1, bs->dsk.journal_offset + md_offset);
             md_offset += data->iov.iov_len;
             submitted++;
@@ -186,7 +175,7 @@ resume_3:
     next_offset = md_offset;
     // Read the rest of the metadata
 resume_4:
-    if (next_offset < bs->dsk.meta_area_size && submitted == 0)
+    if (next_offset < bs->dsk.meta_area_size && submitted == 0 && (!zero_on_init || !bs->readonly))
     {
         // Submit one read
         for (int i = 0; i < 2; i++)
@@ -203,12 +192,15 @@ resume_4:
                 GET_SQE();
                 assert(bufs[i].size <= 0x7fffffff);
                 data->iov = { bufs[i].buf, (size_t)bufs[i].size };
-                data->callback = [this, i](ring_data_t *data) { handle_event(data, i); };
                 if (!zero_on_init)
+                {
+                    data->callback = [this, i](ring_data_t *data) { handle_event(data, i, "read metadata"); };
                     io_uring_prep_readv(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bufs[i].offset);
+                }
                 else
                 {
                     // Fill metadata with empty block pattern
+                    data->callback = [this, i](ring_data_t *data) { handle_event(data, i, "clear metadata"); };
                     memset(bufs[i].buf, 0, bufs[i].size);
                     for (uint64_t o = 0; o < bufs[i].size; o += bs->dsk.meta_block_size)
                         bs->heap->fill_block_empty_space(bufs[i].buf + o, 0);
@@ -224,11 +216,14 @@ resume_4:
         if (bufs[i].state == INIT_META_READ_DONE)
         {
             // Handle result
-            uint64_t loaded = 0;
-            int r = bs->heap->load_blocks(bufs[i].offset-bs->dsk.meta_block_size, bufs[i].size, bufs[i].buf, bs->skip_corrupted_meta_entries, loaded);
-            if (r != 0)
-                exit(1);
-            entries_loaded += loaded;
+            if (!zero_on_init)
+            {
+                uint64_t loaded = 0;
+                int r = bs->heap->load_blocks(bufs[i].offset-bs->dsk.meta_block_size, bufs[i].size, bufs[i].buf, bs->skip_corrupted_meta_entries, loaded);
+                if (r != 0)
+                    exit(1);
+                entries_loaded += loaded;
+            }
             bufs[i].state = 0;
             bs->ringloop->wakeup();
         }
@@ -238,25 +233,9 @@ resume_4:
         wait_state = 4;
         return 1;
     }
-    // metadata read finished
+    // metadata read/clear finished
     bs->heap->finish_load();
     printf("Metadata entries loaded: %ju, rechecking unfinished writes and garbage entries\n", entries_loaded);
-    if (zero_on_init && !bs->dsk.disable_meta_fsync)
-    {
-        GET_SQE();
-        io_uring_prep_fsync(sqe, bs->dsk.meta_fd, IORING_FSYNC_DATASYNC);
-        last_read_offset = 0;
-        data->iov = { 0 };
-        data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
-        submitted++;
-        bs->ringloop->submit();
-    resume_5:
-        if (submitted > 0)
-        {
-            wait_state = 5;
-            return 1;
-        }
-    }
     // asynchronous recheck
 resume_6:
     wait_state = 6;
@@ -311,7 +290,7 @@ resume_8:
         uint32_t block_num = recheck_mod[i];
         uint64_t block_offset = bs->dsk.meta_offset + (uint64_t)(block_num+1) * bs->dsk.meta_block_size;
         data = ((ring_data_t*)sqe->user_data);
-        uint8_t *buf = (uint8_t*)malloc_or_die(bs->dsk.meta_block_size);
+        uint8_t *buf = (uint8_t*)memalign_or_die(MEM_ALIGNMENT, bs->dsk.meta_block_size);
         bs->heap->get_meta_block(block_num, buf);
         data->iov = { buf, bs->dsk.meta_block_size };
         data->callback = [this, buf, block_offset](ring_data_t *data)
@@ -337,6 +316,44 @@ resume_9:
     }
     free(metadata_buffer);
     metadata_buffer = NULL;
+do_fsync:
+    if (!bs->dsk.disable_meta_fsync && !bs->readonly)
+    {
+        GET_SQE();
+        io_uring_prep_fsync(sqe, bs->dsk.meta_fd, IORING_FSYNC_DATASYNC);
+        last_read_offset = 0;
+        data->iov = { 0 };
+        data->callback = [this](ring_data_t *data) { handle_event(data, -1, "fsync metadata"); };
+        submitted++;
+        bs->ringloop->submit();
+    resume_5:
+        if (submitted > 0)
+        {
+            wait_state = 5;
+            return 1;
+        }
+    }
+    if (zero_on_init && !header_written && !bs->readonly)
+    {
+        GET_SQE();
+        header_written = true;
+        last_read_offset = 0;
+        data->iov = (struct iovec){ bs->meta_superblock, (size_t)bs->dsk.meta_block_size };
+        data->callback = [this](ring_data_t *data) { handle_event(data, -1, "write metadata header"); };
+        io_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset);
+        bs->ringloop->submit();
+        submitted++;
+    resume_2:
+        if (submitted > 0)
+        {
+            wait_state = 2;
+            return 1;
+        }
+        if (!bs->dsk.disable_meta_fsync)
+        {
+            goto do_fsync;
+        }
+    }
     printf("Loading finished. Data used: %ju / %ju bytes (%s / %s)\n",
         bs->heap->get_data_used_space(), bs->dsk.block_count * bs->dsk.data_block_size,
         format_size(bs->heap->get_data_used_space()).c_str(),

src/blockstore/blockstore_init.h

@@ -17,6 +17,7 @@ class blockstore_init_meta
     int wait_state = 0;
     int wait_count = 0;
     bool zero_on_init = false;
+    bool header_written = false;
     void *metadata_buffer = NULL;
     blockstore_init_meta_buf bufs[2] = {};
     int submitted = 0;
@@ -29,7 +30,7 @@ class blockstore_init_meta
     std::vector<uint32_t> recheck_mod;
     int i = 0, j = 0;
     bool handle_meta_block(uint8_t *buf, uint64_t count, uint64_t done_cnt);
-    void handle_event(ring_data_t *data, int buf_num);
+    void handle_event(ring_data_t *data, int buf_num, const char *op);
 public:
     blockstore_init_meta(blockstore_impl_t *bs);
     int loop();

src/blockstore/blockstore_mock.cpp

@@ -0,0 +1,113 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
+#include "blockstore_mock.h"
+
+blockstore_mock_t::blockstore_mock_t(const blockstore_config_t & config)
+{
+}
+
+void blockstore_mock_t::parse_config(blockstore_config_t & config)
+{
+}
+
+void* blockstore_mock_t::reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit)
+{
+    return NULL;
+}
+
+bool blockstore_mock_t::reshard_continue(void *reshard_state, uint64_t chunk_limit)
+{
+    return true;
+}
+
+void blockstore_mock_t::loop()
+{
+}
+
+bool blockstore_mock_t::is_started()
+{
+    return true;
+}
+
+bool blockstore_mock_t::is_stalled()
+{
+    return false;
+}
+
+bool blockstore_mock_t::is_safe_to_stop()
+{
+    return true;
+}
+
+void blockstore_mock_t::enqueue_op(blockstore_op_t *op)
+{
+}
+
+int blockstore_mock_t::read_bitmap(object_id oid, uint64_t target_version, void *bitmap, uint64_t *result_version)
+{
+    return -EIO;
+}
+
+const std::map<uint64_t, uint64_t> & blockstore_mock_t::get_inode_space_stats()
+{
+    return inode_space;
+}
+
+void blockstore_mock_t::set_no_inode_stats(const std::vector<uint64_t> & pool_ids)
+{
+}
+
+void blockstore_mock_t::dump_diagnostics()
+{
+}
+
+std::string blockstore_mock_t::get_op_diag(blockstore_op_t *op)
+{
+    return "";
+}
+
+uint32_t blockstore_mock_t::get_block_size()
+{
+    return block_size;
+}
+
+uint64_t blockstore_mock_t::get_block_count()
+{
+    return block_count;
+}
+
+uint64_t blockstore_mock_t::get_free_block_count()
+{
+    return block_count;
+}
+
+uint64_t blockstore_mock_t::get_journal_size()
+{
+    return 32*1024*1024;
+}
+
+uint32_t blockstore_mock_t::get_bitmap_granularity()
+{
+    return bitmap_granularity;
+}
+
+uint64_t blockstore_mock_t::get_live_entries()
+{
+    return 0;
+}
+
+uint64_t blockstore_mock_t::get_live_memory()
+{
+    return 0;
+}
+
+uint64_t blockstore_mock_t::get_garbage_entries()
+{
+    return 0;
+}
+
+uint64_t blockstore_mock_t::get_garbage_memory()
+{
+    return 0;
+}

src/blockstore/blockstore_mock.h

@@ -0,0 +1,39 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
+#pragma once
+
+#include "blockstore.h"
+
+class blockstore_mock_t: public blockstore_i
+{
+public:
+    uint32_t block_size = 128*1024;
+    uint32_t bitmap_granularity = 4096;
+    uint64_t block_count = 100*1024*8;
+    std::map<uint64_t, uint64_t> inode_space;
+
+    blockstore_mock_t(const blockstore_config_t & config);
+    void parse_config(blockstore_config_t & config) override;
+    void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit) override;
+    bool reshard_continue(void *reshard_state, uint64_t chunk_limit) override;
+    void loop() override;
+    bool is_started() override;
+    bool is_stalled() override;
+    bool is_safe_to_stop() override;
+    void enqueue_op(blockstore_op_t *op) override;
+    int read_bitmap(object_id oid, uint64_t target_version, void *bitmap, uint64_t *result_version = NULL) override;
+    const std::map<uint64_t, uint64_t> & get_inode_space_stats() override;
+    void set_no_inode_stats(const std::vector<uint64_t> & pool_ids) override;
+    void dump_diagnostics() override;
+    std::string get_op_diag(blockstore_op_t *op) override;
+    uint32_t get_block_size() override;
+    uint64_t get_block_count() override;
+    uint64_t get_free_block_count() override;
+    uint64_t get_journal_size() override;
+    uint32_t get_bitmap_granularity() override;
+    uint64_t get_live_entries() override;
+    uint64_t get_live_memory() override;
+    uint64_t get_garbage_entries() override;
+    uint64_t get_garbage_memory() override;
+};

src/blockstore/blockstore_stable.cpp

@@ -16,6 +16,7 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
     else if (priv->op_state == 5) goto resume_5;
     assert(!priv->op_state);
     op->retval = 0;
+    PRIV(op)->lsn = 0;
     priv->modified_block = priv->modified_block2 = UINT32_MAX;
     for (priv->stab_pos = 0; priv->stab_pos < op->len; priv->stab_pos++)
     {
@@ -27,6 +28,7 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
                 FINISH_OP(op);
                 return 2;
             }
+            priv->modified_block2 = UINT32_MAX;
             int res = op->opcode == BS_OP_STABLE
                 ? heap->add_commit(obj, v[priv->stab_pos].version, &priv->modified_block2)
                 : heap->add_rollback(obj, v[priv->stab_pos].version, &priv->modified_block2);
@@ -36,6 +38,12 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
                 FINISH_OP(op);
                 return 2;
             }
+            if (res == ENOENT)
+            {
+                op->retval = -ENOENT;
+                FINISH_OP(op);
+                return 2;
+            }
             if (res == ENOSPC)
             {
                 if (!heap->get_to_compact_count())
@@ -45,11 +53,6 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
                     FINISH_OP(op);
                     return 2;
                 }
-                if (priv->modified_block2 != UINT32_MAX)
-                {
-                    priv->stab_pos--;
-                    goto resume_1;
-                }
                 priv->wait_for = WAIT_COMPACTION;
                 priv->wait_detail = heap->get_compacted_count();
                 flusher->request_trim();

src/blockstore/blockstore_sync.cpp

@@ -9,6 +9,7 @@ int blockstore_impl_t::continue_sync(blockstore_op_t *op)
     if (!PRIV(op)->op_state)
     {
         op->retval = 0;
+        PRIV(op)->lsn = 0;
     }
     int res = do_sync(op, 0);
     if (res == 2)
@@ -28,9 +29,12 @@ bool blockstore_impl_t::has_unsynced()
 
 bool blockstore_impl_t::submit_fsyncs(int & wait_count)
 {
-    int n = (unsynced_meta_write_count > 0 && !dsk.disable_meta_fsync) +
-        (unsynced_buffer_write_count > 0 && !dsk.disable_journal_fsync && dsk.journal_fd != dsk.meta_fd) +
-        (unsynced_data_write_count > 0 && !dsk.disable_data_fsync && dsk.data_fd != dsk.meta_fd && dsk.data_fd != dsk.journal_fd);
+    int n = (unsynced_meta_write_count > 0 && !dsk.disable_meta_fsync ? 1 : 0) +
+        (unsynced_buffer_write_count > 0 && !dsk.disable_journal_fsync &&
+            (!unsynced_meta_write_count || dsk.journal_fd != dsk.meta_fd) ? 1 : 0) +
+        (unsynced_data_write_count > 0 && !dsk.disable_data_fsync &&
+            (!unsynced_meta_write_count || dsk.data_fd != dsk.meta_fd) &&
+            (!unsynced_buffer_write_count || dsk.data_fd != dsk.journal_fd) ? 1 : 0);
     if (ringloop->space_left() < n)
     {
         return false;
@@ -59,7 +63,8 @@ bool blockstore_impl_t::submit_fsyncs(int & wait_count)
         data->callback = cb;
         wait_count++;
     }
-    if (unsynced_buffer_write_count > 0 && !dsk.disable_journal_fsync && dsk.meta_fd != dsk.journal_fd)
+    if (unsynced_buffer_write_count > 0 && !dsk.disable_journal_fsync &&
+        (!unsynced_meta_write_count || dsk.journal_fd != dsk.meta_fd))
     {
         // fsync buffer
         io_uring_sqe *sqe = get_sqe();
@@ -70,7 +75,9 @@ bool blockstore_impl_t::submit_fsyncs(int & wait_count)
         data->callback = cb;
         wait_count++;
     }
-    if (unsynced_data_write_count > 0 && !dsk.disable_data_fsync && dsk.data_fd != dsk.meta_fd && dsk.data_fd != dsk.journal_fd)
+    if (unsynced_data_write_count > 0 && !dsk.disable_data_fsync &&
+        (!unsynced_meta_write_count || dsk.data_fd != dsk.meta_fd) &&
+        (!unsynced_buffer_write_count || dsk.data_fd != dsk.journal_fd))
     {
         // fsync data
         io_uring_sqe *sqe = get_sqe();
@@ -104,9 +111,11 @@ int blockstore_impl_t::do_sync(blockstore_op_t *op, int base_state)
         unsynced_data_write_count = unsynced_buffer_write_count = unsynced_meta_write_count = 0;
         return 2;
     }
-    PRIV(op)->modified_block = heap->get_completed_lsn();
+    assert(!PRIV(op)->lsn);
+    PRIV(op)->lsn = heap->get_completed_lsn();
     if (!submit_fsyncs(PRIV(op)->pending_ops))
     {
+        PRIV(op)->lsn = 0;
         PRIV(op)->wait_detail = 1;
         PRIV(op)->wait_for = WAIT_SQE;
         return 0;
@@ -118,6 +127,6 @@ resume_1:
         return 1;
     }
 resume_2:
-    heap->mark_lsn_fsynced(PRIV(op)->modified_block);
+    heap->mark_lsn_fsynced(PRIV(op)->lsn);
     return 2;
 }

src/blockstore/blockstore_write.cpp

@@ -37,6 +37,7 @@ void blockstore_impl_t::prepare_meta_block_write(uint32_t modified_block)
         heap->complete_block_write(modified_block);
         ringloop->wakeup();
     };
+    assert(((uint64_t)modified_block+2)*dsk.meta_block_size <= dsk.meta_area_size);
     io_uring_prep_writev(
         sqe, dsk.meta_fd, &data->iov, 1, dsk.meta_offset + ((uint64_t)modified_block+1)*dsk.meta_block_size
     );
@@ -177,14 +178,18 @@ enospc:
         ring_data_t *data = ((ring_data_t*)sqe->user_data);
         data->iov = (struct iovec){ op->buf, op->len };
         data->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
+        assert(loc+op->offset+op->len <= dsk.block_count*dsk.data_block_size);
         io_uring_prep_writev(sqe, dsk.data_fd, &data->iov, 1, dsk.data_offset + loc + op->offset);
+        if (!dsk.disable_data_fsync)
+        {
+            // use PRIV->lsn for fsync_data_id
+            PRIV(op)->lsn = ++data_fsync_next;
+            data_fsyncs.push_back(false);
+        }
         PRIV(op)->pending_ops++;
         write_iodepth++;
         if (PRIV(op)->write_type == BS_HEAP_BIG_WRITE)
-        {
             PRIV(op)->op_state = 1;
-            inflight_big++;
-        }
         else
             PRIV(op)->op_state = 3;
     }
@@ -264,6 +269,7 @@ enospc:
             BS_SUBMIT_GET_SQE(sqe2, data2);
             data2->iov = (struct iovec){ op->buf, op->len };
             data2->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
+            assert(loc+op->len <= dsk.journal_len);
             io_uring_prep_writev(sqe2, dsk.journal_fd, &data2->iov, 1, dsk.journal_offset + loc);
             PRIV(op)->pending_ops++;
         }
@@ -294,8 +300,6 @@ again:
         goto resume_10;
     else if (op_state == 11)
         goto resume_11;
-    else if (op_state == 12)
-        goto resume_12;
     else
     {
         // In progress
@@ -314,38 +318,44 @@ again:
 resume_2:
     // We must fsync all big writes to avoid complex write workflows
     // It's OK for all HDDs and for server SSDs, but slightly worse for desktop SSDs
-    inflight_big--;
     if (!dsk.disable_data_fsync)
     {
-        // fsync data in a batch
-resume_11:
-        if (inflight_big > 0)
+        // Mark our data write as completed and advance data_fsync_cur
+        data_fsyncs[PRIV(op)->lsn - data_fsync_cur - 1] = true;
+        while (data_fsyncs.size() > 0 && data_fsyncs.front())
+        {
+            data_fsyncs.pop_front();
+            data_fsync_cur++;
+        }
+        PRIV(op)->op_state = 11;
+        // Then wait for all other data writes currently in progress to do less fsync calls
+        // I.e. to fsync data in batches
+        PRIV(op)->lsn = data_fsync_cur + data_fsyncs.size();
+resume_11:
+        if (data_fsync_cur < PRIV(op)->lsn)
         {
-            PRIV(op)->op_state = 11;
             return 1;
         }
-        if (fsyncing_data)
+        if (PRIV(op)->lsn > data_fsync_sent)
         {
-resume_12:
-            if (fsyncing_data)
+            BS_SUBMIT_GET_SQE(sqe, data);
+            io_uring_prep_fsync(sqe, dsk.data_fd, IORING_FSYNC_DATASYNC);
+            data->iov = { 0 };
+            data->callback = [this, op, fs = data_fsync_cur](ring_data_t *data)
             {
-                PRIV(op)->op_state = 12;
-                return 1;
-            }
-            goto resume_4;
+                if (fs > data_fsync_done)
+                {
+                    data_fsync_done = fs;
+                    ringloop->wakeup();
+                }
+            };
+            data_fsync_sent = data_fsync_cur;
         }
-        fsyncing_data = true;
-        BS_SUBMIT_GET_SQE(sqe, data);
-        io_uring_prep_fsync(sqe, dsk.data_fd, IORING_FSYNC_DATASYNC);
-        data->iov = { 0 };
-        data->callback = [this, op](ring_data_t *data)
+        if (PRIV(op)->lsn > data_fsync_done)
         {
-            fsyncing_data = false;
-            handle_write_event(data, op);
-        };
-        PRIV(op)->pending_ops++;
-        PRIV(op)->op_state = 3;
-        return 1;
+            return 1;
+        }
+        PRIV(op)->lsn = 0;
     }
 resume_4:
     {
@@ -453,6 +463,7 @@ resume_10:
     BS_SUBMIT_GET_SQE(sqe, data);
     data->iov = (struct iovec){ op->buf, op->len };
     data->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
+    assert(PRIV(op)->location + op->offset <= dsk.block_count*dsk.data_block_size);
     io_uring_prep_writev(sqe, dsk.data_fd, &data->iov, 1, dsk.data_offset + PRIV(op)->location + op->offset);
     if (dsk.use_atomic_flag)
         sqe->rw_flags = RWF_ATOMIC;

src/blockstore/multilist.cpp

@@ -12,7 +12,7 @@ multilist_alloc_t::multilist_alloc_t(uint32_t count, uint32_t maxn):
     count(count), maxn(maxn)
 {
     // not-so-memory-efficient: 16 MB memory per 1 GB buffer space, but buffer spaces are small, so OK
-    assert(count > 1 && count < 0x80000000);
+    assert(count > 1 && count < 0x80000000 && count >= maxn);
     sizes.resize(count);
     nexts.resize(count); // nexts[i] = 0 -> area is used; nexts[i] = 1 -> no next; nexts[i] >= 2 -> next item
     prevs.resize(count);
@@ -171,7 +171,7 @@ void multilist_alloc_t::print()
     printf("\n");
 }
 
-void multilist_alloc_t::use(uint32_t pos, uint32_t size)
+bool multilist_alloc_t::use(uint32_t pos, uint32_t size)
 {
     assert(pos+size <= count && size > 0);
     if (sizes[pos] <= 0)
@@ -182,7 +182,8 @@ void multilist_alloc_t::use(uint32_t pos, uint32_t size)
         else
             while (start > 0 && !sizes[start])
                 start--;
-        assert(sizes[start] >= size);
+        if (sizes[start] < size+(pos-start))
+            return false;
         use_full(start);
         uint32_t full = sizes[start];
         sizes[pos-1] = -pos+start;
@@ -199,7 +200,8 @@ void multilist_alloc_t::use(uint32_t pos, uint32_t size)
     }
     else
     {
-        assert(sizes[pos] >= size);
+        if (sizes[pos] < size)
+            return false;
         use_full(pos);
         if (sizes[pos] > size)
         {
@@ -214,12 +216,13 @@ void multilist_alloc_t::use(uint32_t pos, uint32_t size)
 #ifdef MULTILIST_TRACE
     print();
 #endif
+    return true;
 }
 
 void multilist_alloc_t::use_full(uint32_t pos)
 {
     uint32_t prevsize = sizes[pos];
-    assert(prevsize);
+    assert(prevsize > 0);
     assert(nexts[pos]);
     uint32_t pi = (prevsize < maxn ? prevsize : maxn)-1;
     if (heads[pi] == pos+1)

src/blockstore/multilist.h

@@ -17,7 +17,7 @@ struct multilist_alloc_t
     bool is_free(uint32_t pos);
     uint32_t find(uint32_t size);
     void use_full(uint32_t pos);
-    void use(uint32_t pos, uint32_t size);
+    bool use(uint32_t pos, uint32_t size);
     void do_free(uint32_t pos);
     void free(uint32_t pos);
     void verify();

src/blockstore/ondisk_formats.h

@@ -141,7 +141,7 @@ struct __attribute__((__packed__)) journal_entry
 inline uint32_t je_crc32(journal_entry *je)
 {
     // 0x48674bc7 = crc32(4 zero bytes)
-    return crc32c(0x48674bc7, ((uint8_t*)je)+4, je->size-4);
+    return je->size < 4 ? 0 : crc32c(0x48674bc7, ((uint8_t*)je)+4, je->size-4);
 }
 
 // "VITAstor"

src/blockstore/v1/flush.cpp

@@ -71,6 +71,11 @@ bool journal_flusher_t::is_active()
     return active_flushers > 0 || dequeuing;
 }
 
+size_t journal_flusher_t::get_queue_size()
+{
+    return flush_queue.size();
+}
+
 void journal_flusher_t::loop()
 {
     target_flusher_count = bs->write_iodepth*2;
@@ -384,6 +389,7 @@ stop_flusher:
         wait_state = 0;
         return true;
     }
+    copy_count = 0;
     try_trim = true;
     cur.oid = flusher->flush_queue.front();
     cur.version = flusher->flush_versions[cur.oid];
@@ -511,6 +517,31 @@ resume_2:
             {
                 uo_it->second.was_changed = true;
             }
+            if (!bs->journal.inmemory)
+            {
+                // Verify journaled data checksums (but not COALESCED)
+                for (it = v.begin(); it != v.end(); it++)
+                {
+                    if (it->copy_flags == COPY_BUF_JOURNAL)
+                    {
+                        iovec iov = { .iov_base = it->buf, .iov_len = it->len };
+                        bs->verify_journal_checksums(
+                            it->csum_buf, it->offset, &iov, 1,
+                            [&](uint32_t bad_block, uint32_t calc_csum, uint32_t stored_csum)
+                            {
+                                printf(
+                                    "Checksum mismatch in object %jx:%jx v%ju in journal at 0x%jx, checksum block #%u: got %08x, expected %08x\n",
+                                    cur.oid.inode, cur.oid.stripe, cur.version, it->disk_offset,
+                                    bad_block / bs->dsk.csum_block_size, calc_csum, stored_csum
+                                );
+                                bad_block += it->offset;
+                                assert(!(bad_block % bs->dsk.csum_block_size) && bad_block < bs->dsk.data_block_size);
+                                mangle_csum_blocks.insert(bad_block);
+                            }
+                        );
+                    }
+                }
+            }
         }
         // Submit data writes
         for (it = v.begin(); it != v.end(); it++)
@@ -520,6 +551,7 @@ resume_2:
                 await_sqe(15);
                 data->iov = (struct iovec){ it->buf, (size_t)it->len };
                 data->callback = simple_callback_w;
+                assert(clean_loc+it->offset+it->len <= bs->dsk.block_count*bs->dsk.data_block_size);
                 io_uring_prep_writev(
                     sqe, bs->dsk.data_fd, &data->iov, 1, bs->dsk.data_offset + clean_loc + it->offset
                 );
@@ -633,6 +665,7 @@ resume_2:
         }
         // All done
         flusher->active_flushers--;
+        copy_count = 0; // used by is_mutated()...
         wait_state = 0;
         goto resume_0;
     }
@@ -749,6 +782,7 @@ bool journal_flusher_co::write_meta_block(flusher_meta_write_t & meta_block, int
     await_sqe(0);
     data->iov = (struct iovec){ meta_block.buf, (size_t)bs->dsk.meta_block_size };
     data->callback = simple_callback_w;
+    assert(bs->dsk.meta_block_size + meta_block.sector + bs->dsk.meta_block_size <= bs->dsk.meta_area_size);
     io_uring_prep_writev(
         sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bs->dsk.meta_block_size + meta_block.sector
     );
@@ -813,35 +847,21 @@ bool journal_flusher_co::clear_incomplete_csum_block_bits(int wait_base)
                 bs->verify_padded_checksums(new_clean_bitmap, new_clean_bitmap + 2*bs->dsk.clean_entry_bitmap_size,
                     v[i].offset, &iov, 1, [&](uint32_t bad_block, uint32_t calc_csum, uint32_t stored_csum)
                 {
-                    printf("Checksum mismatch in object %jx:%jx v%ju in data area at offset 0x%jx+0x%x: got %08x, expected %08x\n",
+                    printf("Checksum mismatch in object %jx:%jx v%ju in data area at offset 0x%jx+0x%x during flush: got %08x, expected %08x\n",
                         cur.oid.inode, cur.oid.stripe, old_clean_ver, old_clean_loc, bad_block, calc_csum, stored_csum);
-                    for (uint32_t j = 0; j < bs->dsk.csum_block_size; j += bs->dsk.bitmap_granularity)
-                    {
-                        // Simplest method of mangling: flip one byte in every sector
-                        ((uint8_t*)v[i].buf)[j+bad_block-v[i].offset] ^= 0xff;
-                    }
+                    assert(!(bad_block % bs->dsk.csum_block_size) && bad_block < bs->dsk.data_block_size);
+                    mangle_csum_blocks.insert(bad_block);
                 });
             }
             else
             {
                 bs->verify_journal_checksums(v[i].csum_buf, v[i].offset, &iov, 1, [&](uint32_t bad_block, uint32_t calc_csum, uint32_t stored_csum)
                 {
-                    printf("Checksum mismatch in object %jx:%jx v%ju in journal at offset 0x%jx+0x%x (block offset 0x%jx): got %08x, expected %08x\n",
+                    printf("Checksum mismatch in object %jx:%jx v%ju in journal at offset 0x%jx+0x%x (block offset 0x%jx) during flush: got %08x, expected %08x\n",
                         cur.oid.inode, cur.oid.stripe, old_clean_ver,
                         v[i].disk_offset, bad_block, v[i].offset, calc_csum, stored_csum);
-                    bad_block += (v[i].offset/bs->dsk.csum_block_size) * bs->dsk.csum_block_size;
-                    uint32_t bad_block_end = bad_block + bs->dsk.csum_block_size + (v[i].offset/bs->dsk.csum_block_size) * bs->dsk.csum_block_size;
-                    if (bad_block < v[i].offset)
-                        bad_block = v[i].offset;
-                    if (bad_block_end > v[i].offset+v[i].len)
-                        bad_block_end = v[i].offset+v[i].len;
-                    bad_block -= v[i].offset;
-                    bad_block_end -= v[i].offset;
-                    for (uint32_t j = bad_block; j < bad_block_end; j += bs->dsk.bitmap_granularity)
-                    {
-                        // Simplest method of mangling: flip one byte in every sector
-                        ((uint8_t*)v[i].buf)[j] ^= 0xff;
-                    }
+                    assert(!(bad_block % bs->dsk.csum_block_size) && bad_block < bs->dsk.data_block_size);
+                    mangle_csum_blocks.insert(bad_block);
                 });
             }
         }
@@ -950,6 +970,11 @@ void journal_flusher_co::calc_block_checksums(uint32_t *new_data_csums, bool ski
     }
     // `v` should contain aligned items, possibly split into pieces
     assert(!block_done);
+    for (uint32_t mangle_block: mangle_csum_blocks)
+    {
+        // Flip 1 bit
+        new_data_csums[mangle_block / bs->dsk.csum_block_size] ^= 1;
+    }
 }
 
 void journal_flusher_co::scan_dirty()
@@ -1086,7 +1111,8 @@ void journal_flusher_co::scan_dirty()
             last--;
         read_to_fill_incomplete = bs->fill_partial_checksum_blocks(
             v, fulfilled, bmp_ptr, NULL, false, NULL, v[0].offset/bs->dsk.csum_block_size * bs->dsk.csum_block_size,
-            ((v[last].offset+v[last].len-1) / bs->dsk.csum_block_size + 1) * bs->dsk.csum_block_size
+            ((v[last].offset+v[last].len-1) / bs->dsk.csum_block_size + 1) * bs->dsk.csum_block_size,
+            0, bs->dsk.data_block_size
         );
     }
     else if (fill_incomplete && clean_init_bitmap)
@@ -1116,6 +1142,7 @@ bool journal_flusher_co::read_dirty(int wait_base)
     if (wait_state == wait_base)        goto resume_0;
     else if (wait_state == wait_base+1) goto resume_1;
     wait_count = wait_journal_count = 0;
+    mangle_csum_blocks.clear();
     if (bs->journal.inmemory && !read_to_fill_incomplete)
     {
         // Happy path: nothing to read :)
@@ -1347,7 +1374,7 @@ bool journal_flusher_co::fsync_batch(bool fsync_meta, int wait_base)
         cur_sync->ready_count++;
         flusher->syncing_flushers++;
     resume_1:
-        if (!cur_sync->state)
+        if (cur_sync->state == 0)
         {
             if (flusher->syncing_flushers >= flusher->active_flushers || !flusher->flush_queue.size())
             {
@@ -1375,6 +1402,12 @@ bool journal_flusher_co::fsync_batch(bool fsync_meta, int wait_base)
                 return false;
             }
         }
+        else if (cur_sync->state == 1)
+        {
+            // Wait for fsync completion
+            wait_state = wait_base+1;
+            return false;
+        }
         flusher->syncing_flushers--;
         cur_sync->ready_count--;
         if (cur_sync->ready_count == 0)

src/blockstore/v1/flush.h

@@ -66,6 +66,7 @@ class journal_flusher_co
     uint64_t clean_bitmap_offset, clean_bitmap_len;
     uint8_t *clean_init_dyn_ptr;
     uint8_t *new_clean_bitmap;
+    std::unordered_set<uint32_t> mangle_csum_blocks;
 
     uint64_t new_trim_pos;
 
@@ -123,6 +124,7 @@ public:
     void loop();
     bool is_trim_wanted() { return trim_wanted; }
     bool is_active();
+    size_t get_queue_size();
     void mark_trim_possible();
     void request_trim();
     void release_trim();

src/blockstore/v1/impl.cpp

@@ -6,11 +6,12 @@
 
 namespace v1 {
 
-blockstore_impl_t::blockstore_impl_t(blockstore_config_t & config, ring_loop_i *ringloop, timerfd_manager_t *tfd)
+blockstore_impl_t::blockstore_impl_t(blockstore_config_t & config, ring_loop_i *ringloop, timerfd_manager_t *tfd, bool mock_mode)
 {
     assert(sizeof(blockstore_op_private_t) <= BS_OP_PRIVATE_DATA_SIZE);
     this->tfd = tfd;
     this->ringloop = ringloop;
+    dsk.mock_mode = mock_mode;
     ring_consumer.loop = [this]() { loop(); };
     ringloop->register_consumer(&ring_consumer);
     initialized = 0;
@@ -35,6 +36,11 @@ blockstore_impl_t::blockstore_impl_t(blockstore_config_t & config, ring_loop_i *
 
 blockstore_impl_t::~blockstore_impl_t()
 {
+    for (auto& obj: dirty_db)
+    {
+        if (obj.second.dyn_data)
+            free(obj.second.dyn_data);
+    }
     delete data_alloc;
     delete flusher;
     if (zero_object)
@@ -877,7 +883,7 @@ uint64_t blockstore_impl_t::get_garbage_entries()
 
 uint64_t blockstore_impl_t::get_garbage_memory()
 {
-    return sizeof(blockstore_dirty_db_t::node_type) * dirty_db.size();
+    return (sizeof(obj_ver_id) + sizeof(dirty_entry) + 32) * dirty_db.size();
 }
 
 } // namespace v1

src/blockstore/v1/impl.h

@@ -30,6 +30,8 @@
 
 //#define BLOCKSTORE_DEBUG
 
+struct bs_test_t;
+
 namespace v1 {
 
 #include "journal.h"
@@ -96,7 +98,7 @@ struct blockstore_op_private_t
     int op_state;
 
     // Read
-    uint64_t clean_block_used;
+    uint64_t clean_loc_used;
     std::vector<copy_buffer_t> read_vec;
 
     // Sync, write
@@ -122,6 +124,7 @@ typedef uint64_t pool_pg_id_t;
 
 class blockstore_impl_t: public blockstore_i
 {
+    friend struct ::bs_test_t;
     blockstore_disk_t dsk;
 
     /******* OPTIONS *******/
@@ -220,6 +223,7 @@ class blockstore_impl_t: public blockstore_i
 
     // Read
     int dequeue_read(blockstore_op_t *read_op);
+    void release_clean(blockstore_op_t *op);
     void find_holes(std::vector<copy_buffer_t> & read_vec, uint32_t item_start, uint32_t item_end,
         std::function<int(int, bool, uint32_t, uint32_t)> callback);
     int fulfill_read(blockstore_op_t *read_op,
@@ -230,7 +234,8 @@ class blockstore_impl_t: public blockstore_i
         uint8_t *clean_entry_bitmap, int *dyn_data,
         uint32_t item_start, uint32_t item_end, uint64_t clean_loc, uint64_t clean_ver);
     int fill_partial_checksum_blocks(std::vector<copy_buffer_t> & rv, uint64_t & fulfilled,
-        uint8_t *clean_entry_bitmap, int *dyn_data, bool from_journal, uint8_t *read_buf, uint64_t read_offset, uint64_t read_end);
+        uint8_t *clean_entry_bitmap, int *dyn_data, bool from_journal, uint8_t *read_buf,
+        uint32_t read_offset, uint32_t read_end, uint32_t item_start, uint32_t item_end);
     int pad_journal_read(std::vector<copy_buffer_t> & rv, copy_buffer_t & cp,
         uint64_t dirty_offset, uint64_t dirty_end, uint64_t dirty_loc, uint8_t *csum_ptr, int *dyn_data,
         uint64_t offset, uint64_t submit_len, uint64_t & blk_begin, uint64_t & blk_end, uint8_t* & blk_buf);
@@ -281,7 +286,7 @@ class blockstore_impl_t: public blockstore_i
 
 public:
 
-    blockstore_impl_t(blockstore_config_t & config, ring_loop_i *ringloop, timerfd_manager_t *tfd);
+    blockstore_impl_t(blockstore_config_t & config, ring_loop_i *ringloop, timerfd_manager_t *tfd, bool mock_mode = false);
     ~blockstore_impl_t();
 
     void parse_config(blockstore_config_t & config);

src/blockstore/v1/init.cpp

@@ -1,6 +1,7 @@
 // Copyright (c) Vitaliy Filippov, 2019+
 // License: VNPL-1.1 (see README.md for details)
 
+#include "str_util.h"
 #include "impl.h"
 #include "internal.h"
 
@@ -30,14 +31,15 @@ blockstore_init_meta::blockstore_init_meta(blockstore_impl_t *bs)
     this->bs = bs;
 }
 
-void blockstore_init_meta::handle_event(ring_data_t *data, int buf_num)
+void blockstore_init_meta::handle_event(ring_data_t *data, int buf_num, const char *op)
 {
-    if (data->res < 0)
+    if (data->res != data->iov.iov_len)
     {
-        throw std::runtime_error(
-            std::string("read metadata failed at offset ") + std::to_string(buf_num >= 0 ? bufs[buf_num].offset : last_read_offset) +
-            std::string(": ") + strerror(-data->res)
-        );
+        throw std::runtime_error(strprintf(
+            "%s failed at offset %ju: got %s (code %d), but expected %zu",
+            op, (buf_num >= 0 ? bufs[buf_num].offset : last_read_offset), strerror(-data->res),
+            data->res, data->iov.iov_len
+        ));
     }
     if (buf_num >= 0)
     {
@@ -65,10 +67,11 @@ int blockstore_init_meta::loop()
     if (!metadata_buffer)
         throw std::runtime_error("Failed to allocate metadata read buffer");
     // Read superblock
+    hdr = (blockstore_meta_header_v2_t *)memalign_or_die(MEM_ALIGNMENT, bs->dsk.meta_block_size);
     GET_SQE();
     last_read_offset = 0;
-    data->iov = { metadata_buffer, (size_t)bs->dsk.meta_block_size };
-    data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
+    data->iov = { hdr, (size_t)bs->dsk.meta_block_size };
+    data->callback = [this](ring_data_t *data) { handle_event(data, -1, "read metadata header"); };
     io_uring_prep_readv(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset);
     bs->ringloop->submit();
     submitted++;
@@ -78,24 +81,8 @@ resume_1:
         wait_state = 1;
         return 1;
     }
-    if (iszero((uint64_t*)metadata_buffer, bs->dsk.meta_block_size / sizeof(uint64_t)))
+    if (iszero((uint64_t*)hdr, bs->dsk.meta_block_size / sizeof(uint64_t)))
     {
-        {
-            blockstore_meta_header_v2_t *hdr = (blockstore_meta_header_v2_t *)metadata_buffer;
-            hdr->zero = 0;
-            hdr->magic = BLOCKSTORE_META_MAGIC_V1;
-            hdr->version = bs->dsk.meta_format;
-            hdr->meta_block_size = bs->dsk.meta_block_size;
-            hdr->data_block_size = bs->dsk.data_block_size;
-            hdr->bitmap_granularity = bs->dsk.bitmap_granularity;
-            if (bs->dsk.meta_format >= BLOCKSTORE_META_FORMAT_V2)
-            {
-                hdr->data_csum_type = bs->dsk.data_csum_type;
-                hdr->csum_block_size = bs->dsk.csum_block_size;
-                hdr->header_csum = 0;
-                hdr->header_csum = crc32c(0, hdr, sizeof(*hdr));
-            }
-        }
         if (bs->readonly)
         {
             printf("Skipping metadata initialization because blockstore is readonly\n");
@@ -103,25 +90,11 @@ resume_1:
         else
         {
             printf("Initializing metadata area\n");
-            GET_SQE();
-            last_read_offset = 0;
-            data->iov = (struct iovec){ metadata_buffer, (size_t)bs->dsk.meta_block_size };
-            data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
-            io_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset);
-            bs->ringloop->submit();
-            submitted++;
-        resume_3:
-            if (submitted > 0)
-            {
-                wait_state = 3;
-                return 1;
-            }
-            zero_on_init = true;
         }
+        zero_on_init = true;
     }
     else
     {
-        blockstore_meta_header_v2_t *hdr = (blockstore_meta_header_v2_t *)metadata_buffer;
         if (hdr->zero != 0 || hdr->magic != BLOCKSTORE_META_MAGIC_V1 || hdr->version < BLOCKSTORE_META_FORMAT_V1)
         {
             printf(
@@ -223,12 +196,15 @@ resume_2:
                 GET_SQE();
                 assert(bufs[i].size <= 0x7fffffff);
                 data->iov = { bufs[i].buf, (size_t)bufs[i].size };
-                data->callback = [this, i](ring_data_t *data) { handle_event(data, i); };
                 if (!zero_on_init)
+                {
+                    data->callback = [this, i](ring_data_t *data) { handle_event(data, i, "read metadata"); };
                     io_uring_prep_readv(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bufs[i].offset);
+                }
                 else
                 {
                     // Fill metadata with zeroes
+                    data->callback = [this, i](ring_data_t *data) { handle_event(data, i, "clear metadata"); };
                     memset(data->iov.iov_base, 0, data->iov.iov_len);
                     io_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bufs[i].offset);
                 }
@@ -256,7 +232,7 @@ resume_2:
                 GET_SQE();
                 assert(bufs[i].size <= 0x7fffffff);
                 data->iov = { bufs[i].buf, (size_t)bufs[i].size };
-                data->callback = [this, i](ring_data_t *data) { handle_event(data, i); };
+                data->callback = [this, i](ring_data_t *data) { handle_event(data, i, "write metadata"); };
                 io_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bufs[i].offset);
                 bs->ringloop->submit();
                 bufs[i].state = INIT_META_WRITING;
@@ -285,7 +261,7 @@ resume_2:
             GET_SQE();
             last_read_offset = (1+next_offset)*bs->dsk.meta_block_size;
             data->iov = { metadata_buffer, (size_t)bs->dsk.meta_block_size };
-            data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
+            data->callback = [this](ring_data_t *data) { handle_event(data, -1, "read metadata"); };
             io_uring_prep_readv(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + (1+next_offset)*bs->dsk.meta_block_size);
             bs->ringloop->submit();
             submitted++;
@@ -302,7 +278,7 @@ resume_5:
             }
             GET_SQE();
             data->iov = { metadata_buffer, (size_t)bs->dsk.meta_block_size };
-            data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
+            data->callback = [this](ring_data_t *data) { handle_event(data, -1, "write metadata"); };
             io_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + (1+next_offset)*bs->dsk.meta_block_size);
             bs->ringloop->submit();
             submitted++;
@@ -317,27 +293,64 @@ resume_6:
     }
     // metadata read finished
     printf("Metadata entries loaded: %ju, free blocks: %ju / %ju\n", entries_loaded, bs->data_alloc->get_free_count(), bs->dsk.block_count);
+    if (zero_on_init && !bs->readonly)
+    {
+do_fsync:
+        if (!bs->disable_meta_fsync)
+        {
+            GET_SQE();
+            io_uring_prep_fsync(sqe, bs->dsk.meta_fd, IORING_FSYNC_DATASYNC);
+            last_read_offset = 0;
+            data->iov = { 0 };
+            data->callback = [this](ring_data_t *data) { handle_event(data, -1, "fsync metadata"); };
+            submitted++;
+            bs->ringloop->submit();
+resume_4:
+            if (submitted > 0)
+            {
+                wait_state = 4;
+                return 1;
+            }
+        }
+        if (!header_written)
+        {
+            GET_SQE();
+            hdr->zero = 0;
+            hdr->magic = BLOCKSTORE_META_MAGIC_V1;
+            hdr->version = bs->dsk.meta_format;
+            hdr->meta_block_size = bs->dsk.meta_block_size;
+            hdr->data_block_size = bs->dsk.data_block_size;
+            hdr->bitmap_granularity = bs->dsk.bitmap_granularity;
+            if (bs->dsk.meta_format >= BLOCKSTORE_META_FORMAT_V2)
+            {
+                hdr->data_csum_type = bs->dsk.data_csum_type;
+                hdr->csum_block_size = bs->dsk.csum_block_size;
+                hdr->header_csum = 0;
+                hdr->header_csum = crc32c(0, hdr, sizeof(*hdr));
+            }
+            header_written = true;
+            last_read_offset = 0;
+            data->iov = (struct iovec){ hdr, (size_t)bs->dsk.meta_block_size };
+            data->callback = [this](ring_data_t *data) { handle_event(data, -1, "write metadata header"); };
+            io_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset);
+            bs->ringloop->submit();
+            submitted++;
+resume_3:
+            if (submitted > 0)
+            {
+                wait_state = 3;
+                return 1;
+            }
+            goto do_fsync;
+        }
+    }
     if (!bs->inmemory_meta)
     {
         free(metadata_buffer);
         metadata_buffer = NULL;
     }
-    if (zero_on_init && !bs->disable_meta_fsync)
-    {
-        GET_SQE();
-        io_uring_prep_fsync(sqe, bs->dsk.meta_fd, IORING_FSYNC_DATASYNC);
-        last_read_offset = 0;
-        data->iov = { 0 };
-        data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
-        submitted++;
-        bs->ringloop->submit();
-    resume_4:
-        if (submitted > 0)
-        {
-            wait_state = 4;
-            return 1;
-        }
-    }
+    free(hdr);
+    hdr = NULL;
     return 0;
 }
 
@@ -345,6 +358,8 @@ bool blockstore_init_meta::handle_meta_block(uint8_t *buf, uint64_t entries_per_
 {
     bool updated = false;
     uint64_t max_i = entries_per_block;
+    if (done_cnt > bs->dsk.block_count)
+        return false;
     if (max_i > bs->dsk.block_count-done_cnt)
         max_i = bs->dsk.block_count-done_cnt;
     for (uint64_t i = 0; i < max_i; i++)
@@ -455,21 +470,21 @@ blockstore_init_journal::blockstore_init_journal(blockstore_impl_t *bs)
     };
 }
 
-void blockstore_init_journal::handle_event(ring_data_t *data1)
+void blockstore_init_journal::handle_event(ring_data_t *data)
 {
-    if (data1->res <= 0)
+    if (data->res != data->iov.iov_len)
     {
-        throw std::runtime_error(
-            std::string("read journal failed at offset ") + std::to_string(journal_pos) +
-            std::string(": ") + strerror(-data1->res)
-        );
+        throw std::runtime_error(strprintf(
+            "read journal failed at offset %ju: got %s (code %d), but expected %zu",
+            journal_pos, strerror(-data->res), data->res, data->iov.iov_len
+        ));
     }
     done.push_back({
         .buf = submitted_buf,
         .pos = journal_pos,
-        .len = (uint64_t)data1->res,
+        .len = (uint64_t)data->res,
     });
-    journal_pos += data1->res;
+    journal_pos += data->res;
     if (journal_pos >= bs->journal.len)
     {
         // Continue from the beginning

src/blockstore/v1/init.h

@@ -16,7 +16,9 @@ class blockstore_init_meta
     blockstore_impl_t *bs;
     int wait_state = 0;
     bool zero_on_init = false;
+    bool header_written = false;
     void *metadata_buffer = NULL;
+    blockstore_meta_header_v2_t *hdr = NULL;
     blockstore_init_meta_buf bufs[2] = {};
     int submitted = 0;
     struct io_uring_sqe *sqe;
@@ -29,7 +31,7 @@ class blockstore_init_meta
     int i = 0, j = 0;
     std::vector<uint64_t> entries_to_zero;
     bool handle_meta_block(uint8_t *buf, uint64_t count, uint64_t done_cnt);
-    void handle_event(ring_data_t *data, int buf_num);
+    void handle_event(ring_data_t *data, int buf_num, const char *op);
 public:
     blockstore_init_meta(blockstore_impl_t *bs);
     int loop();

src/blockstore/v1/journal.cpp

@@ -193,6 +193,7 @@ void blockstore_impl_t::prepare_journal_sector_write(int cur_sector, blockstore_
             (size_t)journal.block_size
         };
         data->callback = [this, flush_id = journal.submit_id](ring_data_t *data) { handle_journal_write(data, flush_id); };
+        assert(journal.sector_info[cur_sector].offset+journal.block_size <= dsk.journal_len);
         io_uring_prep_writev(
             sqe, dsk.journal_fd, &data->iov, 1, journal.offset + journal.sector_info[cur_sector].offset
         );

src/blockstore/v1/read.cpp

@@ -101,8 +101,8 @@ int blockstore_impl_t::fulfill_read(blockstore_op_t *read_op,
                     .copy_flags = COPY_BUF_JOURNAL|COPY_BUF_CSUM_FILL,
                     .offset = blk_begin,
                     .len = blk_end-blk_begin,
-                    .csum_buf = (csum + (blk_begin/dsk.csum_block_size -
-                        item_start/dsk.csum_block_size) * (dsk.data_csum_type & 0xFF)),
+                    .csum_buf = (!csum ? NULL : (csum + (blk_begin/dsk.csum_block_size -
+                        item_start/dsk.csum_block_size) * (dsk.data_csum_type & 0xFF))),
                     .dyn_data = dyn_data,
                 });
                 if (dyn_data)
@@ -134,7 +134,7 @@ int blockstore_impl_t::fulfill_read(blockstore_op_t *read_op,
                 // If we don't track it then we may IN THEORY read another object's data:
                 // submit read -> remove the object -> flush remove -> overwrite with another object -> finish read
                 // Very improbable, but possible
-                PRIV(read_op)->clean_block_used = 1;
+                PRIV(read_op)->clean_loc_used = UINT64_MAX;
             }
             rv.insert(rv.begin() + pos, el);
             fulfilled += el.len;
@@ -167,7 +167,8 @@ uint8_t* blockstore_impl_t::get_clean_entry_bitmap(uint64_t block_loc, int offse
 }
 
 int blockstore_impl_t::fill_partial_checksum_blocks(std::vector<copy_buffer_t> & rv, uint64_t & fulfilled,
-    uint8_t *clean_entry_bitmap, int *dyn_data, bool from_journal, uint8_t *read_buf, uint64_t read_offset, uint64_t read_end)
+    uint8_t *clean_entry_bitmap, int *dyn_data, bool from_journal, uint8_t *read_buf,
+    uint32_t read_offset, uint32_t read_end, uint32_t item_start, uint32_t item_end)
 {
     if (read_end == read_offset)
         return 0;
@@ -175,10 +176,38 @@ int blockstore_impl_t::fill_partial_checksum_blocks(std::vector<copy_buffer_t> &
     read_buf -= read_offset;
     uint32_t last_block = (read_end-1)/dsk.csum_block_size;
     uint32_t start_block = read_offset/dsk.csum_block_size;
+    uint32_t item_start_block = item_start/dsk.csum_block_size;
     uint32_t end_block = 0;
+    auto zero_range = [&](int pos, bool alloc, uint32_t cur_start, uint32_t cur_end)
+    {
+        if (alloc)
+            return 0;
+        copy_buffer_t el = {
+            .copy_flags = COPY_BUF_ZERO,
+            .offset = cur_start,
+            .len = cur_end-cur_start,
+        };
+        rv.insert(rv.begin() + pos, el);
+        if (read_buf)
+            memset(read_buf + el.offset - read_offset, 0, el.len);
+        fulfilled += el.len;
+        return 1;
+    };
+    if (read_offset < item_start)
+    {
+        // Zero-fill the beginning
+        find_holes(rv, read_offset, item_start, zero_range);
+        read_offset = item_start;
+    }
+    if (read_end > item_end)
+    {
+        // Zero-fill the end
+        find_holes(rv, item_end, read_end, zero_range);
+        read_end = item_end;
+    }
     while (start_block <= last_block)
     {
-        if (read_range_fulfilled(rv, fulfilled, read_buf, clean_entry_bitmap,
+        if (read_range_fulfilled(rv, fulfilled, read_buf, from_journal ? NULL : clean_entry_bitmap,
             start_block*dsk.csum_block_size < read_offset ? read_offset : start_block*dsk.csum_block_size,
             (start_block+1)*dsk.csum_block_size > read_end ? read_end : (start_block+1)*dsk.csum_block_size))
         {
@@ -190,7 +219,7 @@ int blockstore_impl_t::fill_partial_checksum_blocks(std::vector<copy_buffer_t> &
             // Find a sequence of checksum blocks required to be read
             end_block = start_block;
             while ((end_block+1)*dsk.csum_block_size < read_end &&
-                !read_range_fulfilled(rv, fulfilled, read_buf, clean_entry_bitmap,
+                !read_range_fulfilled(rv, fulfilled, read_buf, from_journal ? NULL : clean_entry_bitmap,
                     (end_block+1)*dsk.csum_block_size < read_offset ? read_offset : (end_block+1)*dsk.csum_block_size,
                     (end_block+2)*dsk.csum_block_size > read_end ? read_end : (end_block+2)*dsk.csum_block_size))
             {
@@ -202,8 +231,10 @@ int blockstore_impl_t::fill_partial_checksum_blocks(std::vector<copy_buffer_t> &
                 .copy_flags = COPY_BUF_CSUM_FILL | (from_journal ? COPY_BUF_JOURNALED_BIG : 0),
                 .offset = start_block*dsk.csum_block_size,
                 .len = (end_block-start_block)*dsk.csum_block_size,
-                // save clean_entry_bitmap if we're reading clean data from the journal
-                .csum_buf = from_journal ? clean_entry_bitmap : NULL,
+                // save checksum reference if we're reading clean data from the journal
+                .csum_buf = from_journal
+                    ? clean_entry_bitmap + dsk.clean_entry_bitmap_size + (start_block-item_start_block)*(dsk.data_csum_type & 0xFF)
+                    : NULL,
                 .dyn_data = dyn_data,
             });
             if (dyn_data)
@@ -226,6 +257,11 @@ bool blockstore_impl_t::read_range_fulfilled(std::vector<copy_buffer_t> & rv, ui
     {
         if (alloc)
             return 0;
+        if (!clean_entry_bitmap)
+        {
+            all_done = false;
+            return 0;
+        }
         int diff = 0;
         uint32_t bmp_start = cur_start/dsk.bitmap_granularity;
         uint32_t bmp_end = cur_end/dsk.bitmap_granularity;
@@ -323,7 +359,7 @@ bool blockstore_impl_t::read_checksum_block(blockstore_op_t *op, int rv_pos, uin
             {
                 iov[n_iov++] = (struct iovec){ (uint8_t*)op->buf+cur_start-op->offset, lim_end-cur_start };
                 rv.insert(rv.begin() + pos, (copy_buffer_t){
-                    .copy_flags = COPY_BUF_DATA,
+                    .copy_flags = COPY_BUF_DATA|COPY_BUF_COALESCED,
                     .offset = cur_start,
                     .len = lim_end-cur_start,
                 });
@@ -361,10 +397,10 @@ bool blockstore_impl_t::read_checksum_block(blockstore_op_t *op, int rv_pos, uin
         PRIV(op)->pending_ops++;
         io_uring_prep_readv(sqe, submit_fd, iov + n_pos, n_cur, submit_offset + clean_loc + item_start + d_pos);
         data->callback = [this, op](ring_data_t *data) { handle_read_event(data, op); };
-        if (n_pos > 0 || n_pos + IOV_MAX < n_iov)
+        if (n_pos > 0 || n_iov > IOV_MAX)
         {
             uint32_t d_len = 0;
-            for (int i = 0; i < IOV_MAX; i++)
+            for (int i = 0; i < n_cur; i++)
                 d_len += iov[n_pos+i].iov_len;
             data->iov.iov_len = d_len;
             d_pos += d_len;
@@ -376,7 +412,7 @@ bool blockstore_impl_t::read_checksum_block(blockstore_op_t *op, int rv_pos, uin
     {
         // Reads running parallel to flushes of the same clean block may read
         // a mixture of old and new data. So we don't verify checksums for such blocks.
-        PRIV(op)->clean_block_used = 1;
+        PRIV(op)->clean_loc_used = UINT64_MAX;
     }
     return true;
 }
@@ -402,7 +438,7 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *read_op)
     }
     uint64_t fulfilled = 0;
     PRIV(read_op)->pending_ops = 0;
-    PRIV(read_op)->clean_block_used = 0;
+    PRIV(read_op)->clean_loc_used = 0;
     auto & rv = PRIV(read_op)->read_vec;
     uint64_t result_version = 0;
     if (dirty_found)
@@ -515,26 +551,50 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *read_op)
     return 2;
 undo_read:
     // need to wait. undo added requests, don't dequeue op
-    if (dsk.csum_block_size > dsk.bitmap_granularity)
+    release_clean(read_op);
+    for (auto & vec: rv)
     {
-        for (auto & vec: rv)
+        if ((vec.copy_flags & COPY_BUF_CSUM_FILL) && vec.buf)
         {
-            if ((vec.copy_flags & COPY_BUF_CSUM_FILL) && vec.buf)
-            {
-                free(vec.buf);
-                vec.buf = NULL;
-            }
-            if (vec.dyn_data && --(*vec.dyn_data) == 0) // refcount
-            {
-                free(vec.dyn_data);
-                vec.dyn_data = NULL;
-            }
+            free(vec.buf);
+            vec.buf = NULL;
+        }
+        if (vec.dyn_data && --(*vec.dyn_data) == 0) // refcount
+        {
+            free(vec.dyn_data);
+            vec.dyn_data = NULL;
         }
     }
     rv.clear();
     return 0;
 }
 
+void blockstore_impl_t::release_clean(blockstore_op_t *op)
+{
+    if (PRIV(op)->clean_loc_used == UINT64_MAX)
+    {
+        PRIV(op)->clean_loc_used = 0;
+    }
+    if (PRIV(op)->clean_loc_used)
+    {
+        // Release clean data block
+        auto uo_it = used_clean_objects.find(PRIV(op)->clean_loc_used - 1);
+        if (uo_it != used_clean_objects.end())
+        {
+            uo_it->second.refs--;
+            if (uo_it->second.refs <= 0)
+            {
+                if (uo_it->second.was_freed)
+                {
+                    data_alloc->set((PRIV(op)->clean_loc_used - 1) / dsk.data_block_size, false);
+                }
+                used_clean_objects.erase(uo_it);
+            }
+        }
+        PRIV(op)->clean_loc_used = 0;
+    }
+}
+
 int blockstore_impl_t::pad_journal_read(std::vector<copy_buffer_t> & rv, copy_buffer_t & cp,
     // FIXME Passing dirty_entry& would be nicer
     uint64_t dirty_offset, uint64_t dirty_end, uint64_t dirty_loc, uint8_t *csum_ptr, int *dyn_data,
@@ -598,11 +658,15 @@ bool blockstore_impl_t::fulfill_clean_read(blockstore_op_t *read_op, uint64_t &
     {
         auto & rv = PRIV(read_op)->read_vec;
         int req = fill_partial_checksum_blocks(rv, fulfilled, clean_entry_bitmap, dyn_data, from_journal,
-            (uint8_t*)read_op->buf, read_op->offset, read_op->offset+read_op->len);
+            (uint8_t*)read_op->buf, read_op->offset, read_op->offset+read_op->len, item_start, item_end);
         if (!inmemory_meta && !from_journal && req > 0)
         {
             // Read checksums from disk
             uint8_t *csum_buf = read_clean_meta_block(read_op, clean_loc, rv.size()-req);
+            if (!csum_buf)
+            {
+                return false;
+            }
             for (int i = req; i > 0; i--)
             {
                 rv[rv.size()-i].csum_buf = csum_buf;
@@ -615,7 +679,7 @@ bool blockstore_impl_t::fulfill_clean_read(blockstore_op_t *read_op, uint64_t &
                 return false;
             }
         }
-        PRIV(read_op)->clean_block_used = req > 0;
+        PRIV(read_op)->clean_loc_used = req > 0 ? UINT64_MAX : 0;
     }
     else if (from_journal)
     {
@@ -665,6 +729,10 @@ bool blockstore_impl_t::fulfill_clean_read(blockstore_op_t *read_op, uint64_t &
                 {
                     // Read checksums from disk
                     csum_buf = read_clean_meta_block(read_op, clean_loc, PRIV(read_op)->read_vec.size());
+                    if (!csum_buf)
+                    {
+                        return false;
+                    }
                     csum_done = true;
                 }
                 uint8_t *csum = !dsk.csum_block_size ? 0 : (csum_buf + 2*dsk.clean_entry_bitmap_size + bmp_start*(dsk.data_csum_type & 0xFF));
@@ -679,13 +747,13 @@ bool blockstore_impl_t::fulfill_clean_read(blockstore_op_t *read_op, uint64_t &
         }
     }
     // Increment reference counter if clean data is being read from the disk
-    if (PRIV(read_op)->clean_block_used)
+    if (PRIV(read_op)->clean_loc_used == UINT64_MAX)
     {
         auto & uo = used_clean_objects[clean_loc];
         uo.refs++;
         if (dsk.csum_block_size && flusher->is_mutated(clean_loc))
             uo.was_changed = true;
-        PRIV(read_op)->clean_block_used = clean_loc;
+        PRIV(read_op)->clean_loc_used = clean_loc + 1;
     }
     return true;
 }
@@ -725,12 +793,18 @@ bool blockstore_impl_t::verify_padded_checksums(uint8_t *clean_entry_bitmap, uin
         while (pos < iov[i].iov_len)
         {
             uint32_t start = pos;
-            uint8_t bit = (clean_entry_bitmap[bmp_pos >> 3] >> (bmp_pos & 0x7)) & 1;
-            while (pos < iov[i].iov_len && ((clean_entry_bitmap[bmp_pos >> 3] >> (bmp_pos & 0x7)) & 1) == bit)
+            uint8_t bit = 1;
+            if (clean_entry_bitmap)
             {
-                pos += dsk.bitmap_granularity;
-                bmp_pos++;
+                bit = (clean_entry_bitmap[bmp_pos >> 3] >> (bmp_pos & 0x7)) & 1;
+                while (pos < iov[i].iov_len && ((clean_entry_bitmap[bmp_pos >> 3] >> (bmp_pos & 0x7)) & 1) == bit)
+                {
+                    pos += dsk.bitmap_granularity;
+                    bmp_pos++;
+                }
             }
+            else
+                pos = iov[i].iov_len;
             uint32_t len = pos-start;
             auto buf = (uint8_t*)iov[i].iov_base+start;
             while (block_done+len >= dsk.csum_block_size)
@@ -807,7 +881,7 @@ bool blockstore_impl_t::verify_clean_padded_checksums(blockstore_op_t *op, uint6
 {
     uint32_t offset = clean_loc % dsk.data_block_size;
     if (from_journal)
-        return verify_padded_checksums(dyn_data, dyn_data + dsk.clean_entry_bitmap_size, offset, iov, n_iov, bad_block_cb);
+        return verify_padded_checksums(NULL, dyn_data, offset, iov, n_iov, bad_block_cb);
     clean_loc = (clean_loc / dsk.data_block_size) * dsk.data_block_size;
     if (!dyn_data)
     {
@@ -835,7 +909,7 @@ void blockstore_impl_t::handle_read_event(ring_data_t *data, blockstore_op_t *op
             void *meta_block = NULL;
             if (dsk.csum_block_size > dsk.bitmap_granularity)
             {
-                for (int i = rv.size()-1; i >= 0 && (rv[i].copy_flags & COPY_BUF_CSUM_FILL); i--)
+                for (int i = 0; i < rv.size(); i++)
                 {
                     if (rv[i].copy_flags & COPY_BUF_META_BLOCK)
                     {
@@ -845,8 +919,41 @@ void blockstore_impl_t::handle_read_event(ring_data_t *data, blockstore_op_t *op
                         rv[i].buf = NULL;
                         continue;
                     }
-                    struct iovec *iov = (struct iovec*)((uint8_t*)rv[i].buf + (rv[i].len & 0xFFFFFFFF));
-                    int n_iov = rv[i].len >> 32;
+                    if (rv[i].copy_flags & COPY_BUF_ZERO)
+                    {
+                        // Zero read
+                        continue;
+                    }
+                    if (rv[i].copy_flags & COPY_BUF_COALESCED)
+                    {
+                        // Sub-block shared with another read. Skip
+                        continue;
+                    }
+                    if ((rv[i].copy_flags & COPY_BUF_JOURNAL) && journal.inmemory)
+                    {
+                        // Do not check journal checksums in-memory
+                        continue;
+                    }
+                    iovec single_iov = {};
+                    iovec *iov = NULL;
+                    int n_iov = 0;
+                    if (rv[i].copy_flags & COPY_BUF_CSUM_FILL)
+                    {
+                        // Padded, buffer list passed using a 'creepy way'
+                        iov = (struct iovec*)((uint8_t*)rv[i].buf + (rv[i].len & 0xFFFFFFFF));
+                        n_iov = rv[i].len >> 32;
+                    }
+                    else
+                    {
+                        // Not padded, buffer is fully within the input buffer
+                        assert(op->buf);
+                        assert(rv[i].csum_buf);
+                        iov = &single_iov;
+                        n_iov = 1;
+                        assert(rv[i].offset >= op->offset);
+                        assert(rv[i].offset + rv[i].len <= op->offset + op->len);
+                        single_iov = { .iov_base = op->buf + rv[i].offset - op->offset, .iov_len = rv[i].len };
+                    }
                     bool ok = true;
                     if (rv[i].copy_flags & COPY_BUF_JOURNAL)
                     {
@@ -944,23 +1051,7 @@ void blockstore_impl_t::handle_read_event(ring_data_t *data, blockstore_op_t *op
                 meta_block = NULL;
             }
         }
-        if (PRIV(op)->clean_block_used)
-        {
-            // Release clean data block
-            auto uo_it = used_clean_objects.find(PRIV(op)->clean_block_used);
-            if (uo_it != used_clean_objects.end())
-            {
-                uo_it->second.refs--;
-                if (uo_it->second.refs <= 0)
-                {
-                    if (uo_it->second.was_freed)
-                    {
-                        data_alloc->set(PRIV(op)->clean_block_used, false);
-                    }
-                    used_clean_objects.erase(uo_it);
-                }
-            }
-        }
+        release_clean(op);
         if (!journal.inmemory)
         {
             // Release journal sector usage

src/blockstore/v1/stable.cpp

@@ -491,7 +491,7 @@ void blockstore_impl_t::mark_stable(obj_ver_id v, bool forget_dirty)
                     if (!exists)
                     {
                         uint64_t space_id = dirty_it->first.oid.inode;
-                        if (no_inode_stats[dirty_it->first.oid.inode >> (64-POOL_ID_BITS)])
+                        if (no_inode_stats.find(dirty_it->first.oid.inode >> (64-POOL_ID_BITS)) != no_inode_stats.end())
                             space_id = space_id & ~(((uint64_t)1 << (64-POOL_ID_BITS)) - 1);
                         inode_space_stats[space_id] += dsk.data_block_size;
                         used_blocks++;
@@ -501,7 +501,7 @@ void blockstore_impl_t::mark_stable(obj_ver_id v, bool forget_dirty)
                 else if (IS_DELETE(dirty_it->second.state))
                 {
                     uint64_t space_id = dirty_it->first.oid.inode;
-                    if (no_inode_stats[dirty_it->first.oid.inode >> (64-POOL_ID_BITS)])
+                    if (no_inode_stats.find(dirty_it->first.oid.inode >> (64-POOL_ID_BITS)) != no_inode_stats.end())
                         space_id = space_id & ~(((uint64_t)1 << (64-POOL_ID_BITS)) - 1);
                     auto & sp = inode_space_stats[space_id];
                     if (sp > dsk.data_block_size)

src/blockstore/v1/write.cpp

@@ -368,9 +368,9 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
         }
         data->iov.iov_len = op->len + stripe_offset + stripe_end; // to check it in the callback
         data->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
-        io_uring_prep_writev(
-            sqe, dsk.data_fd, PRIV(op)->iov_zerofill, vcnt, dsk.data_offset + (loc * dsk.data_block_size) + op->offset - stripe_offset
-        );
+        const uint64_t write_offset = (loc * dsk.data_block_size) + op->offset - stripe_offset;
+        assert(write_offset+op->len+stripe_offset+stripe_end <= dsk.block_count*dsk.data_block_size);
+        io_uring_prep_writev(sqe, dsk.data_fd, PRIV(op)->iov_zerofill, vcnt, dsk.data_offset + write_offset);
         PRIV(op)->pending_ops = 1;
         if (!(dirty_it->second.state & BS_ST_INSTANT))
         {
@@ -495,9 +495,8 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
                 .op = op,
             });
             data2->callback = [this, flush_id = journal.submit_id](ring_data_t *data) { handle_journal_write(data, flush_id); };
-            io_uring_prep_writev(
-                sqe2, dsk.journal_fd, &data2->iov, 1, journal.offset + journal.next_free
-            );
+            assert(journal.next_free+op->len <= dsk.journal_len);
+            io_uring_prep_writev(sqe2, dsk.journal_fd, &data2->iov, 1, journal.offset + journal.next_free);
             PRIV(op)->pending_ops++;
         }
         else

src/client/CMakeLists.txt

@@ -3,6 +3,22 @@ cmake_minimum_required(VERSION 2.8...3.30)
 project(vitastor)
 
 # libvitastor_common.a
+add_library(vitastor_common STATIC
+	etcd_state_client.cpp
+	msgr_stop.cpp
+	msgr_op.cpp
+	../../json11/json11.cpp
+	osd_ops.cpp
+	pg_states.cpp
+	../util/allocator.cpp
+	../util/addr_util.cpp
+	../util/timerfd_manager.cpp
+	../util/str_util.cpp
+	../util/json_util.cpp
+)
+target_compile_options(vitastor_common PUBLIC -fPIC)
+
+# libvitastor_net.a
 set(MSGR_RDMA "")
 if (IBVERBS_LIBRARIES)
 	set(MSGR_RDMA "msgr_rdma.cpp")
@@ -11,30 +27,36 @@ set(MSGR_RDMACM "")
 if (RDMACM_LIBRARIES)
 	set(MSGR_RDMACM "msgr_rdmacm.cpp")
 endif (RDMACM_LIBRARIES)
-add_library(vitastor_common STATIC
-	../util/epoll_manager.cpp etcd_state_client.cpp messenger.cpp ../util/addr_util.cpp ../util/xxh_x86dispatch.c
-	msgr_encrypt.cpp msgr_stop.cpp msgr_op.cpp msgr_send.cpp msgr_receive.cpp ../util/ringloop.cpp ../../json11/json11.cpp
-	http_client.cpp osd_ops.cpp pg_states.cpp ../util/timerfd_manager.cpp ../util/str_util.cpp ../util/json_util.cpp ${MSGR_RDMA} ${MSGR_RDMACM}
+add_library(vitastor_net STATIC
+	../util/epoll_manager.cpp
+	etcd_state_client_http.cpp
+	messenger.cpp
+	msgr_iothread.cpp
+	msgr_send.cpp
+	msgr_receive.cpp
+	../util/ringloop.cpp
+	http_client.cpp
+	${MSGR_RDMA}
+	${MSGR_RDMACM}
 )
-target_link_libraries(vitastor_common pthread ${OPENSSL_LIBRARIES} ${CARES_LIBRARIES})
-target_compile_options(vitastor_common PUBLIC -fPIC)
+target_link_libraries(vitastor_net pthread vitastor_common)
+target_compile_options(vitastor_net PUBLIC -fPIC)
 
 # libvitastor_client.so
 add_library(vitastor_client SHARED
 	cluster_client.cpp
+	cluster_client_real.cpp
 	cluster_client_list.cpp
 	cluster_client_wb.cpp
-	cluster_client_icache.cpp
 	vitastor_c.cpp
 )
 set_target_properties(vitastor_client PROPERTIES PUBLIC_HEADER "client/vitastor_c.h")
 target_link_libraries(vitastor_client
-	vitastor_common
+	vitastor_net
 	vitastor_cli
 	${LIBURING_LIBRARIES}
 	${IBVERBS_LIBRARIES}
 	${RDMACM_LIBRARIES}
-	${OPENSSL_LIBRARIES}
 )
 set_target_properties(vitastor_client PROPERTIES VERSION ${VITASTOR_VERSION} SOVERSION 0)
 configure_file(vitastor.pc.in vitastor.pc @ONLY)
@@ -54,9 +76,6 @@ if (${WITH_FIO})
 		../util/rw_blocking.cpp
 		../util/addr_util.cpp
 	)
-	target_link_libraries(fio_vitastor_sec
-		tcmalloc_minimal
-	)
 endif (${WITH_FIO})
 
 # vitastor-nbd
@@ -100,11 +119,13 @@ endif (${WITH_QEMU})
 add_executable(test_cluster_client
 	EXCLUDE_FROM_ALL
 	../test/test_cluster_client.cpp
-	pg_states.cpp osd_ops.cpp cluster_client.cpp cluster_client_list.cpp cluster_client_wb.cpp cluster_client_icache.cpp msgr_op.cpp ../test/mock/messenger.cpp msgr_stop.cpp msgr_encrypt.cpp
-	etcd_state_client.cpp ../util/timerfd_manager.cpp ../util/addr_util.cpp ../util/str_util.cpp ../util/json_util.cpp ../util/xxh_x86dispatch.c ../../json11/json11.cpp
+	cluster_client.cpp
+	cluster_client_list.cpp
+	cluster_client_wb.cpp
+	../test/mock/messenger.cpp
+	etcd_state_client_mock.cpp
 )
-target_link_libraries(test_cluster_client ${OPENSSL_LIBRARIES})
-target_compile_definitions(test_cluster_client PUBLIC -D__MOCK__)
+target_link_libraries(test_cluster_client vitastor_common ${LIBURING_LIBRARIES})
 target_include_directories(test_cluster_client BEFORE PUBLIC ${CMAKE_SOURCE_DIR}/src/test/mock)
 add_dependencies(build_tests test_cluster_client)
 add_test(NAME test_cluster_client COMMAND test_cluster_client)

src/client/cluster_client.cpp

@@ -11,7 +11,7 @@
 #define TRY_SEND_CONNECTING 1
 #define TRY_SEND_OK 2
 
-cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json config)
+cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json config, std::unique_ptr<etcd_state_client_t> st_cli_ptr)
 {
     wb = new writeback_cache_t();
 
@@ -53,29 +53,30 @@ cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd
     };
     msgr.parse_config(config);
 
-    st_cli.tfd = tfd;
-    st_cli.on_load_config_hook = [this](json11::Json::object & cfg) { on_load_config_hook(cfg); };
-    st_cli.on_change_osd_state_hook = [this](uint64_t peer_osd) { on_change_osd_state_hook(peer_osd); };
-    st_cli.on_change_pool_config_hook = [this]() { on_change_pool_config_hook(); };
-    st_cli.on_change_pg_config_hook = [this]() { on_change_pool_config_hook(); };
-    st_cli.on_change_pg_state_hook = [this](pool_id_t pool_id, pg_num_t pg_num, osd_num_t prev_primary) { on_change_pg_state_hook(pool_id, pg_num, prev_primary); };
-    st_cli.on_change_node_placement_hook = [this]() { on_change_node_placement_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(); };
-    st_cli.on_inode_change_hook = [this](uint64_t inode, bool removed) { on_change_inode_hook(inode, removed); };
+    st_cli = std::move(st_cli_ptr);
+    st_cli->on_load_config_hook = [this](json11::Json::object & cfg) { on_load_config_hook(cfg); };
+    st_cli->on_change_osd_state_hook = [this](uint64_t peer_osd) { on_change_osd_state_hook(peer_osd); };
+    st_cli->on_change_pool_config_hook = [this]() { on_change_pool_config_hook(); };
+    st_cli->on_change_pg_config_hook = [this]() { on_change_pool_config_hook(); };
+    st_cli->on_change_pg_state_hook = [this](pool_id_t pool_id, pg_num_t pg_num, osd_num_t prev_primary) { on_change_pg_state_hook(pool_id, pg_num, prev_primary); };
+    st_cli->on_change_node_placement_hook = [this]() { on_change_node_placement_hook(); };
+    st_cli->on_load_pgs_hook = [this](bool success) { on_load_pgs_hook(success); };
+    st_cli->on_reload_hook = [this]() { this->st_cli->load_global_config(); };
 
-    st_cli.parse_config(config);
-    st_cli.infinite_start = false;
+    st_cli->parse_config(config);
+    st_cli->infinite_start = false;
     if (!config["client_infinite_start"].is_null())
     {
-        st_cli.infinite_start = config["client_infinite_start"].bool_value();
+        st_cli->infinite_start = config["client_infinite_start"].bool_value();
     }
-    st_cli.load_global_config();
+    st_cli->load_global_config();
+
+    scrap_buffer_size = SCRAP_BUFFER_SIZE;
+    scrap_buffer = malloc_or_die(scrap_buffer_size);
 }
 
 cluster_client_t::~cluster_client_t()
 {
-    vault_destroy();
     if (retry_timeout_id >= 0)
     {
         tfd->clear_timer(retry_timeout_id);
@@ -93,6 +94,7 @@ cluster_client_t::~cluster_client_t()
     {
         ringloop->unregister_consumer(&consumer);
     }
+    free(scrap_buffer);
     delete wb;
     wb = NULL;
 }
@@ -467,7 +469,7 @@ void cluster_client_t::on_load_config_hook(json11::Json::object & etcd_global_co
         auto etcd_report_interval = config["etcd_report_interval"].uint64_value();
         if (!etcd_report_interval)
             etcd_report_interval = 5;
-        client_wait_up_timeout = 1+etcd_report_interval+(st_cli.max_etcd_attempts*(2*st_cli.etcd_quick_timeout)+999)/1000;
+        client_wait_up_timeout = 1+etcd_report_interval+(st_cli->max_etcd_attempts*(2*st_cli->etcd_quick_timeout)+999)/1000;
     }
     // log_level
     log_level = config["log_level"].uint64_value();
@@ -479,11 +481,9 @@ void cluster_client_t::on_load_config_hook(json11::Json::object & etcd_global_co
         self_tree_metrics.clear();
         client_hostname = new_hostname;
     }
-    // vault
-    vault_parse_config();
     msgr.parse_config(config);
-    st_cli.parse_config(config);
-    st_cli.load_pgs();
+    st_cli->parse_config(config);
+    st_cli->load_pgs();
 }
 
 osd_num_t cluster_client_t::select_random_osd(const std::vector<osd_num_t> & osds)
@@ -492,7 +492,7 @@ osd_num_t cluster_client_t::select_random_osd(const std::vector<osd_num_t> & osd
     int alive_count = 0;
     for (auto & osd_num: osds)
     {
-        if (!st_cli.peer_states[osd_num].is_null())
+        if (!st_cli->peer_states[osd_num].is_null())
             alive_set[alive_count++] = osd_num;
     }
     if (!alive_count)
@@ -509,7 +509,7 @@ osd_num_t cluster_client_t::select_nearest_osd(const std::vector<osd_num_t> & os
         while (self_tree_metrics.find(cur_id) == self_tree_metrics.end())
         {
             self_tree_metrics[cur_id] = metric++;
-            json11::Json cur_placement = st_cli.node_placement[cur_id];
+            json11::Json cur_placement = st_cli->node_placement[cur_id];
             cur_id = cur_placement["parent"].string_value();
         }
         if (cur_id != "")
@@ -529,7 +529,7 @@ osd_num_t cluster_client_t::select_nearest_osd(const std::vector<osd_num_t> & os
         }
         else
         {
-            auto & peer_state = st_cli.peer_states[osd_num];
+            auto & peer_state = st_cli->peer_states[osd_num];
             if (!peer_state.is_null())
             {
                 metric = self_tree_metrics[""];
@@ -539,7 +539,7 @@ osd_num_t cluster_client_t::select_nearest_osd(const std::vector<osd_num_t> & os
                 while (seen.find(cur_id) == seen.end())
                 {
                     seen.insert(cur_id);
-                    json11::Json cur_placement = st_cli.node_placement[cur_id];
+                    json11::Json cur_placement = st_cli->node_placement[cur_id];
                     std::string cur_parent = cur_placement["parent"].string_value();
                     cur_id = (!first || cur_parent != "" ? cur_parent : peer_state["host"].string_value());
                     first = false;
@@ -564,7 +564,7 @@ osd_num_t cluster_client_t::select_nearest_osd(const std::vector<osd_num_t> & os
 
 void cluster_client_t::on_load_pgs_hook(bool success)
 {
-    for (auto & pool_item: st_cli.pool_config)
+    for (auto & pool_item: st_cli->pool_config)
     {
         pg_counts[pool_item.first] = pool_item.second.real_pg_count;
     }
@@ -584,13 +584,13 @@ void cluster_client_t::on_load_pgs_hook(bool success)
 
 void cluster_client_t::on_change_pool_config_hook()
 {
-    for (auto & pool_item: st_cli.pool_config)
+    for (auto & pool_item: st_cli->pool_config)
     {
         if (pg_counts[pool_item.first] != pool_item.second.real_pg_count)
         {
             if (log_level > 2 && pg_counts[pool_item.first])
             {
-                printf("Pool %u (%s) PG count changed from %lu to %lu\n", pool_item.first, pool_item.second.name.c_str(),
+                fprintf(stderr, "Pool %u (%s) PG count changed from %lu to %lu\n", pool_item.first, pool_item.second.name.c_str(),
                     pg_counts[pool_item.first], pool_item.second.real_pg_count);
             }
             // At this point, all pool operations should have been suspended
@@ -607,15 +607,12 @@ void cluster_client_t::on_change_pool_config_hook()
             pg_counts[pool_item.first] = pool_item.second.real_pg_count;
         }
     }
-    inode_cache.clear();
-    inode_cache_children.clear();
-    vault_keys.clear();
     continue_ops();
 }
 
 void cluster_client_t::on_change_pg_state_hook(pool_id_t pool_id, pg_num_t pg_num, osd_num_t prev_primary)
 {
-    auto & pg_cfg = st_cli.pool_config[pool_id].pg_config[pg_num];
+    auto & pg_cfg = st_cli->pool_config[pool_id].pg_config[pg_num];
     if (pg_cfg.cur_primary != prev_primary)
     {
         // Repeat this PG operations because an OSD which stopped being primary may not fsync operations
@@ -633,8 +630,8 @@ bool cluster_client_t::get_immediate_commit(uint64_t inode)
     pool_id_t pool_id = INODE_POOL(inode);
     if (!pool_id)
         return true;
-    auto pool_it = st_cli.pool_config.find(pool_id);
-    if (pool_it == st_cli.pool_config.end())
+    auto pool_it = st_cli->pool_config.find(pool_id);
+    if (pool_it == st_cli->pool_config.end())
         return true;
     return pool_it->second.immediate_commit == IMMEDIATE_ALL;
 }
@@ -644,7 +641,7 @@ void cluster_client_t::on_change_osd_state_hook(uint64_t peer_osd)
     osd_tree_metrics.erase(peer_osd);
     if (msgr.wanted_peers.find(peer_osd) != msgr.wanted_peers.end())
     {
-        msgr.connect_peer(peer_osd, st_cli.peer_states[peer_osd]);
+        msgr.connect_peer(peer_osd, st_cli->peer_states[peer_osd]);
         continue_lists();
     }
 }
@@ -676,10 +673,6 @@ bool cluster_client_t::flush()
 {
     if (!ringloop)
     {
-        if (vault_loading)
-        {
-            return false;
-        }
         if (wb->writeback_queue.size())
         {
             wb->start_writebacks(this, 0);
@@ -702,7 +695,7 @@ bool cluster_client_t::flush()
         sync_done = true;
     };
     execute(sync);
-    while (!sync_done || vault_loading)
+    while (!sync_done)
     {
         ringloop->loop();
         if (!sync_done)
@@ -943,8 +936,8 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
         cb(op);
         return false;
     }
-    auto pool_it = st_cli.pool_config.find(pool_id);
-    if (pool_it == st_cli.pool_config.end() || pool_it->second.real_pg_count == 0)
+    auto pool_it = st_cli->pool_config.find(pool_id);
+    if (pool_it == st_cli->pool_config.end() || pool_it->second.real_pg_count == 0)
     {
         // Pools are loaded, but this one is unknown
         op->retval = -EINVAL;
@@ -965,40 +958,10 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
     {
         op->flags |= OP_IMMEDIATE_COMMIT;
     }
-    bool searched = false;
-    std::shared_ptr<inode_cache_t> icache;
-    if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_WRITE)
-    {
-        if (!searched)
-        {
-            icache = inode_cache_get(op->inode);
-            searched = true;
-        }
-        if (icache && icache->has_parent_loop && op->opcode == OSD_OP_READ)
-        {
-            op->retval = -EINVAL;
-            auto cb = std::move(op->callback);
-            cb(op);
-            return false;
-        }
-        if (icache && icache->op_enc)
-        {
-            // Use shared_ptr aliasing to attach op_enc to the inode cache entry
-            op->enc = std::shared_ptr<osd_op_enc_t>(icache, icache->op_enc);
-        }
-        else
-            op->enc.reset();
-    }
-    else
-        op->enc.reset();
     if ((op->opcode == OSD_OP_WRITE || op->opcode == OSD_OP_DELETE) && !(op->flags & OSD_OP_IGNORE_READONLY))
     {
-        if (!searched)
-        {
-            icache = inode_cache_get(op->inode);
-            searched = true;
-        }
-        if (icache && icache->readonly)
+        auto ino_it = st_cli->inode_config.find(op->inode);
+        if (ino_it != st_cli->inode_config.end() && ino_it->second.readonly)
         {
             op->retval = -EROFS;
             auto cb = std::move(op->callback);
@@ -1009,39 +972,33 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
     op->deoptimise_snapshot = false;
     if (enable_writeback && (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP))
     {
-        if (!searched)
+        auto ino_it = st_cli->inode_config.find(op->inode);
+        if (ino_it != st_cli->inode_config.end())
         {
-            icache = inode_cache_get(op->inode);
-            searched = true;
-        }
-        if (icache)
-        {
-            for (auto & parent: icache->chain)
+            int chain_size = 0;
+            while (ino_it != st_cli->inode_config.end() && ino_it->second.parent_id)
             {
-                if (INODE_POOL(parent) == INODE_POOL(op->inode) && wb->has_inode(parent))
+                // Check for loops - FIXME check it in etcd_state_client
+                if (ino_it->second.parent_id == op->inode ||
+                    chain_size > st_cli->inode_config.size())
+                {
+                    op->retval = -EINVAL;
+                    auto cb = std::move(op->callback);
+                    cb(op);
+                    return false;
+                }
+                if (INODE_POOL(ino_it->second.parent_id) == INODE_POOL(ino_it->first) &&
+                    wb->has_inode(ino_it->second.parent_id))
                 {
                     // Deoptimise reads - we have dirty data for one of the parent layer(s).
                     op->deoptimise_snapshot = true;
                     break;
                 }
+                chain_size++;
+                ino_it = st_cli->inode_config.find(ino_it->second.parent_id);
             }
         }
     }
-    if (icache && icache->err_code)
-    {
-        if (icache->err_code == EPERM)
-        {
-            op->retval = -EPERM;
-            auto cb = std::move(op->callback);
-            cb(op);
-            return false;
-        }
-        else if (icache->err_code == EAGAIN)
-        {
-            key_wait_ops.push_back(op);
-            return false;
-        }
-    }
     return true;
 }
 
@@ -1057,7 +1014,7 @@ void cluster_client_t::execute_raw(osd_num_t osd_num, osd_op_t *op)
     else
     {
         if (msgr.wanted_peers.find(osd_num) == msgr.wanted_peers.end())
-            msgr.connect_peer(osd_num, st_cli.peer_states[osd_num]);
+            msgr.connect_peer(osd_num, st_cli->peer_states[osd_num]);
         raw_ops.emplace(osd_num, op);
     }
 }
@@ -1162,35 +1119,40 @@ resume_2:
         // Finished successfully
         // Even if the PG count has changed in meanwhile we treat it as success
         // because if some operations were invalid for the new PG count we'd get errors
+        if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
+        {
+            // Copy part bitmaps only after finishing all part reads
+            for (auto & part: op->parts)
+                if ((part.flags & (PART_SENT|PART_DONE|PART_VALID)) == (PART_SENT|PART_DONE|PART_VALID))
+                    copy_part_bitmap(op, &part);
+        }
         if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
         {
-            uint64_t next_inode = 0;
-            auto icache = inode_cache_get(op->cur_inode);
-            if (icache)
+            // Check parent inode
+            auto ino_it = st_cli->inode_config.find(op->cur_inode);
+            // Skip parents from the same pool
+            int skipped = 0;
+            while (!op->deoptimise_snapshot &&
+                ino_it != st_cli->inode_config.end() && ino_it->second.parent_id &&
+                INODE_POOL(ino_it->second.parent_id) == INODE_POOL(op->cur_inode))
             {
-                if (icache->has_parent_loop)
+                // Check for loops - FIXME check it in etcd_state_client
+                if (ino_it->second.parent_id == op->inode ||
+                    skipped > st_cli->inode_config.size())
                 {
                     op->retval = -EINVAL;
                     erase_op(op);
                     return 1;
                 }
-                if (op->deoptimise_snapshot)
-                {
-                    if (icache->chain.size() > 1)
-                        next_inode = icache->chain[1];
-                }
-                else
-                {
-                    if (icache->other_pool_parent_id)
-                        next_inode = icache->other_pool_parent_id;
-                }
+                skipped++;
+                ino_it = st_cli->inode_config.find(ino_it->second.parent_id);
             }
-            if (next_inode)
+            if (ino_it != st_cli->inode_config.end() &&
+                ino_it->second.parent_id &&
+                ino_it->second.parent_id != op->inode)
             {
                 // Continue reading from the parent inode
-                icache = inode_cache_get(next_inode);
-                op->cur_inode = next_inode;
-                op->enc = (icache && icache->op_enc ? std::shared_ptr<osd_op_enc_t>(icache, icache->op_enc) : nullptr);
+                op->cur_inode = ino_it->second.parent_id;
                 op->parts.clear();
                 op->done_count = 0;
                 goto resume_0;
@@ -1199,7 +1161,7 @@ resume_2:
         op->retval = op->len;
         if (op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
         {
-            auto & pool_cfg = st_cli.pool_config.at(INODE_POOL(op->inode));
+            auto & pool_cfg = st_cli->pool_config.at(INODE_POOL(op->inode));
             op->retval = op->len / pool_cfg.bitmap_granularity;
         }
         if (op->flush_id)
@@ -1209,7 +1171,7 @@ resume_2:
         erase_op(op);
         return 1;
     }
-    else if (op->retval != 0 && !(op->flags & OP_FLUSH_BUFFER) &&
+    else if (op->retval != 0 && op->opcode != OSD_OP_SYNC && !(op->flags & OP_FLUSH_BUFFER) &&
         op->retval != -EPIPE && (op->retval != -EIO || !client_eio_retry_interval) && (op->retval != -ENOSPC || !client_retry_enospc))
     {
         // Fatal error (neither -EPIPE, -EIO nor -ENOSPC)
@@ -1241,7 +1203,7 @@ resume_2:
     return 0;
 }
 
-static void add_iov(int size, int skip, cluster_op_t *op, int &iov_idx, size_t &iov_pos, osd_op_buf_list_t &iov)
+static void add_iov(int size, bool skip, cluster_op_t *op, int &iov_idx, size_t &iov_pos, osd_op_buf_list_t &iov, void *scrap, int scrap_len)
 {
     int left = size;
     while (left > 0 && iov_idx < op->iov.count)
@@ -1249,7 +1211,7 @@ static void add_iov(int size, int skip, cluster_op_t *op, int &iov_idx, size_t &
         int cur_left = op->iov.buf[iov_idx].iov_len - iov_pos;
         if (cur_left < left)
         {
-            if (skip == 0)
+            if (!skip)
             {
                 iov.push_back((uint8_t*)op->iov.buf[iov_idx].iov_base + iov_pos, cur_left);
             }
@@ -1259,7 +1221,7 @@ static void add_iov(int size, int skip, cluster_op_t *op, int &iov_idx, size_t &
         }
         else
         {
-            if (skip == 0)
+            if (!skip)
             {
                 iov.push_back((uint8_t*)op->iov.buf[iov_idx].iov_base + iov_pos, left);
             }
@@ -1268,10 +1230,16 @@ static void add_iov(int size, int skip, cluster_op_t *op, int &iov_idx, size_t &
         }
     }
     assert(left == 0);
-    if (skip == 1)
+    if (skip && scrap_len > 0)
     {
-        // data read into a NULL buffer will be discarded by messenger
-        iov.push_back(NULL, size);
+        // All skipped ranges are read into the same useless buffer
+        left = size;
+        while (left > 0)
+        {
+            int cur_left = scrap_len < left ? scrap_len : left;
+            iov.push_back(scrap, cur_left);
+            left -= cur_left;
+        }
     }
 }
 
@@ -1279,7 +1247,7 @@ void cluster_client_t::slice_rw(cluster_op_t *op)
 {
     // Slice the request into individual object stripe requests
     // Primary OSDs still operate individual stripes, but their size is multiplied by PG minsize in case of EC
-    auto & pool_cfg = st_cli.pool_config.at(INODE_POOL(op->cur_inode));
+    auto & pool_cfg = st_cli->pool_config.at(INODE_POOL(op->cur_inode));
     uint32_t pg_data_size = (pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks);
     uint64_t pg_block_size = pool_cfg.data_block_size * pg_data_size;
     uint64_t first_stripe = (op->offset / pg_block_size) * pg_block_size;
@@ -1291,11 +1259,7 @@ void cluster_client_t::slice_rw(cluster_op_t *op)
         // Allocate memory for the bitmap
         unsigned object_bitmap_size = ((op->len / pool_cfg.bitmap_granularity + 7) / 8);
         object_bitmap_size = (object_bitmap_size < 8 ? 8 : object_bitmap_size);
-        unsigned bitmap_mem = object_bitmap_size +
-            op->parts.size() * pg_data_size *
-            (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8
-            // read chain info - 1 byte per block
-            + (op->enc ? op->len/pool_cfg.bitmap_granularity : 0));
+        unsigned bitmap_mem = object_bitmap_size + (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8 * pg_data_size) * op->parts.size();
         if (!op->bitmap_buf || op->bitmap_buf_size < bitmap_mem)
         {
             op->bitmap_buf = realloc_or_die(op->bitmap_buf, bitmap_mem);
@@ -1337,10 +1301,10 @@ void cluster_client_t::slice_rw(cluster_op_t *op)
                         {
                             begin = cur;
                             // Just advance iov_idx & iov_pos
-                            add_iov(cur-prev, 2, op, iov_idx, iov_pos, op->parts[i].iov);
+                            add_iov(cur-prev, true, op, iov_idx, iov_pos, op->parts[i].iov, NULL, 0);
                         }
                         else
-                            add_iov(cur-prev, skip_prev ? 1 : 0, op, iov_idx, iov_pos, op->parts[i].iov);
+                            add_iov(cur-prev, skip_prev, op, iov_idx, iov_pos, op->parts[i].iov, scrap_buffer, scrap_buffer_size);
                     }
                     skip_prev = skip;
                     prev = cur;
@@ -1351,11 +1315,11 @@ void cluster_client_t::slice_rw(cluster_op_t *op)
             if (skip_prev)
             {
                 // Just advance iov_idx & iov_pos
-                add_iov(end-prev, 2, op, iov_idx, iov_pos, op->parts[i].iov);
+                add_iov(end-prev, true, op, iov_idx, iov_pos, op->parts[i].iov, NULL, 0);
                 end = prev;
             }
             else
-                add_iov(cur-prev, skip_prev ? 1 : 0, op, iov_idx, iov_pos, op->parts[i].iov);
+                add_iov(cur-prev, skip_prev, op, iov_idx, iov_pos, op->parts[i].iov, scrap_buffer, scrap_buffer_size);
             if (end == begin)
             {
                 op->done_count++;
@@ -1364,7 +1328,7 @@ void cluster_client_t::slice_rw(cluster_op_t *op)
         }
         else if (op->opcode != OSD_OP_READ_BITMAP && op->opcode != OSD_OP_READ_CHAIN_BITMAP && op->opcode != OSD_OP_DELETE)
         {
-            add_iov(end-begin, 0, op, iov_idx, iov_pos, op->parts[i].iov);
+            add_iov(end-begin, false, op, iov_idx, iov_pos, op->parts[i].iov, NULL, 0);
         }
         op->parts[i].parent = op;
         op->parts[i].offset = begin;
@@ -1382,7 +1346,7 @@ bool cluster_client_t::affects_pg(uint64_t inode, uint64_t offset, uint64_t len,
     {
         return false;
     }
-    auto & pool_cfg = st_cli.pool_config.at(INODE_POOL(inode));
+    auto & pool_cfg = st_cli->pool_config.at(INODE_POOL(inode));
     uint32_t pg_data_size = (pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks);
     uint64_t pg_block_size = pool_cfg.data_block_size * pg_data_size;
     uint64_t first_stripe = (offset / pg_block_size) * pg_block_size;
@@ -1401,7 +1365,7 @@ bool cluster_client_t::affects_pg(uint64_t inode, uint64_t offset, uint64_t len,
 
 bool cluster_client_t::affects_osd(uint64_t inode, uint64_t offset, uint64_t len, osd_num_t osd)
 {
-    auto & pool_cfg = st_cli.pool_config.at(INODE_POOL(inode));
+    auto & pool_cfg = st_cli->pool_config.at(INODE_POOL(inode));
     uint32_t pg_data_size = (pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks);
     uint64_t pg_block_size = pool_cfg.data_block_size * pg_data_size;
     uint64_t first_stripe = (offset / pg_block_size) * pg_block_size;
@@ -1425,7 +1389,7 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
         init_msgr();
     }
     auto part = &op->parts[i];
-    auto & pool_cfg = st_cli.pool_config.at(INODE_POOL(op->cur_inode));
+    auto & pool_cfg = st_cli->pool_config.at(INODE_POOL(op->cur_inode));
     auto pg_it = pool_cfg.pg_config.find(part->pg_num);
     if (pg_it != pool_cfg.pg_config.end() &&
         !pg_it->second.pause && pg_it->second.cur_primary &&
@@ -1450,14 +1414,14 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
             osd_client_t *cl = peer_it->second;
             part->flags |= PART_SENT|PART_VALID;
             op->inflight_count++;
-            uint32_t pg_data_size = (pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks);
-            uint64_t pg_bitmap_size = pg_data_size * (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8
-                + (op->opcode == OSD_OP_READ && op->enc ? pool_cfg.data_block_size/pool_cfg.bitmap_granularity : 0));
+            uint64_t pg_bitmap_size = (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8) * (
+                pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks
+            );
             uint64_t meta_rev = 0;
             if (op->opcode != OSD_OP_READ_BITMAP && op->opcode != OSD_OP_DELETE && !op->deoptimise_snapshot)
             {
-                auto ino_it = st_cli.inode_config.find(op->cur_inode);
-                if (ino_it != st_cli.inode_config.end())
+                auto ino_it = st_cli->inode_config.find(op->cur_inode);
+                if (ino_it != st_cli->inode_config.end())
                     meta_rev = ino_it->second.mod_revision;
             }
             part->op = (osd_op_t){
@@ -1471,7 +1435,6 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
                     .inode = op->cur_inode,
                     .offset = part->offset,
                     .len = part->len,
-                    .flags = op->opcode == OSD_OP_READ && op->enc && !op->deoptimise_snapshot ? OSD_OP_RETURN_CHAIN : 0,
                     .meta_revision = meta_rev,
                     .version = op->opcode == OSD_OP_WRITE || op->opcode == OSD_OP_DELETE ? op->version : 0,
                 } },
@@ -1479,7 +1442,6 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
                     ? (uint8_t*)op->part_bitmaps + pg_bitmap_size*i : NULL),
                 .bitmap_len = (unsigned)(op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP
                     ? pg_bitmap_size : 0),
-                .enc = op->enc,
                 .callback = cb ? cb : [this, part](osd_op_t *op_part)
                 {
                     handle_op_part(part);
@@ -1491,7 +1453,7 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
         }
         else if (msgr.wanted_peers.find(primary_osd) == msgr.wanted_peers.end())
         {
-            msgr.connect_peer(primary_osd, st_cli.peer_states[primary_osd]);
+            msgr.connect_peer(primary_osd, st_cli->peer_states[primary_osd]);
             return TRY_SEND_CONNECTING;
         }
     }
@@ -1675,13 +1637,6 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
         }
         if (op->inflight_count == 0 && !op->retry_after)
         {
-            // Copy part bitmaps only after finishing all part reads
-            if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
-            {
-                for (auto & part: op->parts)
-                    if ((part.flags & (PART_SENT|PART_VALID|PART_DONE)) == (PART_SENT|PART_VALID|PART_DONE))
-                        copy_part_bitmap(op, &part);
-            }
             if (op->opcode == OSD_OP_SYNC)
                 continue_sync(op);
             else
@@ -1693,7 +1648,7 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
 void cluster_client_t::copy_part_bitmap(cluster_op_t *op, cluster_op_part_t *part)
 {
     // Copy (OR) bitmap
-    auto & pool_cfg = st_cli.pool_config.at(INODE_POOL(op->cur_inode));
+    auto & pool_cfg = st_cli->pool_config.at(INODE_POOL(op->cur_inode));
     uint32_t pg_block_size = pool_cfg.data_block_size * (
         pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks
     );

src/client/cluster_client.h

@@ -4,8 +4,7 @@
 #pragma once
 
 #include "messenger.h"
-#include "etcd_state_client.h"
-#include "../util/robin_hood.h"
+#include "etcd_state_client_http.h"
 
 #define DEFAULT_CLIENT_MAX_DIRTY_BYTES 32*1024*1024
 #define DEFAULT_CLIENT_MAX_DIRTY_OPS 1024
@@ -72,7 +71,6 @@ protected:
     cluster_op_t *prev = NULL, *next = NULL;
     int prev_wait = 0;
     uint64_t flush_id = 0;
-    std::shared_ptr<osd_op_enc_t> enc;
     friend class cluster_client_t;
     friend class writeback_cache_t;
 };
@@ -82,36 +80,14 @@ struct inode_list_osd_t;
 struct inode_list_pg_t;
 class writeback_cache_t;
 
-struct inode_cache_t
-{
-    std::vector<inode_t> chain;
-    uint8_t *key_data = NULL;
-    osd_op_enc_t *op_enc = NULL;
-    bool readonly = false;
-    bool has_parent_loop = false;
-    inode_t other_pool_parent_id = 0;
-    int err_code = 0;
-
-    ~inode_cache_t();
-};
-
-struct vault_load_key_t
-{
-    int key_state = 0;
-    std::string key;
-};
-
 // FIXME: Split into public and private interfaces
 class __attribute__((visibility("default"))) cluster_client_t
 {
-#ifdef __MOCK__
-public:
-#endif
     timerfd_manager_t *tfd = NULL;
     ring_loop_t *ringloop = NULL;
 
-    // config:
-
+    std::map<pool_id_t, uint64_t> pg_counts;
+    std::map<pool_pg_num_t, osd_num_t> pg_primary;
     // client_max_dirty_* is actually "max unsynced", for the case when immediate_commit is off
     uint64_t client_max_dirty_bytes = 0;
     uint64_t client_max_dirty_ops = 0;
@@ -123,23 +99,12 @@ public:
     uint64_t client_max_writeback_iodepth = 0;
     std::string conf_hostname;
 
-    std::string vault_url;
-    std::string vault_client_cert;
-    std::string vault_client_key;
-    std::string vault_ca;
-    std::string vault_secret_api_path;
-    uint64_t vault_timeout_ms = 0;
-    uint64_t vault_error_timeout_sec = 0;
-    uint64_t vault_refresh_leeway_sec = 0;
-
     int log_level = 0;
     int client_retry_interval = 50; // ms
     int client_eio_retry_interval = 1000; // ms
     bool client_retry_enospc = true;
     int client_wait_up_timeout = 16; // sec (for listings)
 
-    // state:
-
     std::string client_hostname;
     std::map<std::string, int> self_tree_metrics;
     std::map<osd_num_t, int> osd_tree_metrics;
@@ -147,28 +112,15 @@ public:
     int retry_timeout_id = -1;
     int retry_timeout_duration = 0;
     std::vector<cluster_op_t*> offline_ops;
-    std::vector<cluster_op_t*> key_wait_ops;
     cluster_op_t *op_queue_head = NULL, *op_queue_tail = NULL;
     writeback_cache_t *wb = NULL;
     std::set<osd_num_t> dirty_osds;
     uint64_t dirty_bytes = 0, dirty_ops = 0;
 
-    // inodes require some extra state for read/write, it's stored here.
-    // moreover, robin_hood access is slightly faster than std::map :)
-    robin_hood::unordered_flat_map<inode_t, std::shared_ptr<inode_cache_t>> inode_cache;
-    std::set<std::pair<inode_t, inode_t>> inode_cache_children;
-
-    http_context_t *vault_http_ctx = NULL;
-    http_co_t *vault_http_cli = NULL;
-    bool vault_loading = false;
-    std::string vault_token;
-    bool vault_auth_error = false;
-    timespec vault_token_expire = {};
-    std::vector<std::string> vault_key_load_queue;
-    std::map<std::string, vault_load_key_t> vault_keys;
+    void *scrap_buffer = NULL;
+    unsigned scrap_buffer_size = 0;
 
     bool pgs_loaded = false;
-    std::map<pool_id_t, uint64_t> pg_counts;
     ring_consumer_t consumer;
     std::vector<std::function<void(void)>> on_ready_hooks;
     int list_retry_timeout_id = -1;
@@ -179,7 +131,7 @@ public:
     bool msgr_initialized = false;
 
 public:
-    etcd_state_client_t st_cli;
+    std::unique_ptr<etcd_state_client_t> st_cli;
 
     osd_messenger_t msgr;
     void init_msgr();
@@ -187,7 +139,8 @@ public:
     json11::Json::object cli_config, file_config, etcd_global_config;
     json11::Json::object config;
 
-    cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json config);
+    static cluster_client_t* create(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json config);
+    cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json config, std::unique_ptr<etcd_state_client_t> st_cli);
     ~cluster_client_t();
     void execute(cluster_op_t *op);
     void execute_raw(osd_num_t osd_num, osd_op_t *op);
@@ -200,22 +153,9 @@ public:
     void list_inode(inode_t inode, uint64_t min_offset, uint64_t max_offset, int max_parallel_pgs, std::function<void(
         int status, int pgs_left, pg_num_t pg_num, std::set<object_id>&& objects)> pg_callback);
 
-    //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; }
-
-#ifndef __MOCK__
 protected:
-#endif
     void continue_ops(int time_passed = 0);
 
-    std::shared_ptr<inode_cache_t> inode_cache_get(inode_t ino);
-    void vault_parse_config();
-    bool vault_check_token();
-    void vault_load_keys();
-    void vault_destroy();
-    void vault_parse_secret(const std::string & key_id, const std::string & err, json11::Json data);
-
-protected:
     bool affects_osd(uint64_t inode, uint64_t offset, uint64_t len, osd_num_t osd);
     bool affects_pg(uint64_t inode, uint64_t offset, uint64_t len, pool_id_t pool_id, pg_num_t pg_num);
 
@@ -225,7 +165,6 @@ protected:
     void on_change_pg_state_hook(pool_id_t pool_id, pg_num_t pg_num, osd_num_t prev_primary);
     void on_change_osd_state_hook(uint64_t peer_osd);
     void on_change_node_placement_hook();
-    void on_change_inode_hook(uint64_t inode, bool removed);
 
     void execute_internal(cluster_op_t *op);
     void execute_cas(cluster_op_t *op);
@@ -242,7 +181,6 @@ protected:
     void erase_op(cluster_op_t *op);
     void calc_wait(cluster_op_t *op);
     void inc_wait(uint64_t opcode, uint64_t flags, cluster_op_t *next, int inc);
-
     void continue_lists();
     bool continue_listing(inode_list_t *lst);
     bool restart_listing(inode_list_t* lst);
@@ -258,4 +196,5 @@ protected:
     osd_num_t select_nearest_osd(const std::vector<osd_num_t> & osds);
 
     friend class writeback_cache_t;
+    friend class cluster_client_test_t;
 };

src/client/cluster_client_icache.cpp

@@ -1,367 +0,0 @@
-// Copyright (c) Vitaliy Filippov, 2019+
-// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
-
-#include <stdexcept>
-#include <assert.h>
-#include "cluster_client_impl.h"
-#include "http_client.h"
-#include "str_util.h"
-
-#define VAULT_KEY_NOT_LOADED 0
-#define VAULT_KEY_LOADING 1
-#define VAULT_KEY_LOADED 2
-#define VAULT_KEY_ERROR 3
-
-inode_cache_t::~inode_cache_t()
-{
-    if (key_data)
-    {
-        free(key_data);
-        key_data = NULL;
-        op_enc = NULL;
-    }
-}
-
-void cluster_client_t::vault_destroy()
-{
-    if (vault_http_ctx)
-    {
-#ifndef __MOCK__
-        http_destroy(vault_http_cli);
-        http_context_destroy(vault_http_ctx);
-        vault_http_cli = NULL;
-        vault_http_ctx = NULL;
-#endif
-    }
-}
-
-void cluster_client_t::vault_parse_config()
-{
-    vault_url = config["vault_url"].string_value();
-    vault_client_cert = config["vault_client_cert"].string_value();
-    vault_client_key = config["vault_client_key"].string_value();
-    vault_ca = config["vault_ca"].string_value();
-    vault_secret_api_path = "/v1/secret/";
-    if (config["vault_secret_api_path"].is_string())
-        vault_secret_api_path = config["vault_secret_api_path"].string_value();
-    vault_timeout_ms = config["vault_timeout_ms"].uint64_value();
-    if (!vault_timeout_ms)
-        vault_timeout_ms = 5000;
-    vault_error_timeout_sec = config["vault_error_timeout_sec"].uint64_value();
-    if (!vault_error_timeout_sec)
-        vault_error_timeout_sec = 60;
-    vault_refresh_leeway_sec = config["vault_refresh_leeway_sec"].uint64_value();
-    if (!vault_refresh_leeway_sec)
-        vault_refresh_leeway_sec = 60;
-}
-
-// FIXME: Rework client API by adding open/close and cache inode information in the "FD" (maybe)
-void cluster_client_t::on_change_inode_hook(uint64_t inode, bool removed)
-{
-    std::vector<inode_t> children = { inode };
-    for (size_t i = 0; i < children.size(); i++)
-    {
-        auto it = inode_cache_children.lower_bound(std::make_pair(children[i], (inode_t)0));
-        while (it != inode_cache_children.end() && it->first == children[i])
-        {
-            children.push_back(it->second);
-            it++;
-        }
-    }
-    for (auto & inode: children)
-    {
-        auto it = inode_cache.find(inode);
-        if (it != inode_cache.end())
-        {
-            auto icache = it->second;
-            for (auto & parent: icache->chain)
-            {
-                inode_cache_children.erase(std::make_pair(parent, inode));
-            }
-            inode_cache.erase(it);
-        }
-    }
-}
-
-std::shared_ptr<inode_cache_t> cluster_client_t::inode_cache_get(inode_t ino)
-{
-    auto icache_it = inode_cache.find(ino);
-    if (icache_it != inode_cache.end())
-    {
-        return icache_it->second;
-    }
-    // Fill inode cache
-    auto ino_it = st_cli.inode_config.find(ino);
-    if (ino_it == st_cli.inode_config.end())
-    {
-        inode_cache[ino] = NULL;
-        return NULL;
-    }
-    auto pool_it = st_cli.pool_config.find(INODE_POOL(ino));
-    if (pool_it == st_cli.pool_config.end())
-    {
-        inode_cache[ino] = NULL;
-        return NULL;
-    }
-    auto & inode_cfg = ino_it->second;
-    auto & pool_cfg = pool_it->second;
-    std::shared_ptr<inode_cache_t> icache = std::make_shared<inode_cache_t>();
-    icache->readonly = inode_cfg.readonly;
-    icache->chain.push_back(ino);
-    std::vector<inode_config_t*> chain_cfg;
-    // FIXME: Allow unencrypted read & write when all chain is encrypted with the same key
-    int enc_key_count = !inode_cfg.enc_key.empty() ? 1 : 0;
-    if (inode_cfg.parent_id)
-    {
-        // Check for loops and cache the chain
-        robin_hood::unordered_flat_set<inode_t> seen;
-        seen.insert(ino);
-        uint64_t parent_id = inode_cfg.parent_id;
-        while (parent_id)
-        {
-            if (seen.find(parent_id) != seen.end())
-            {
-                icache->has_parent_loop = true;
-                break;
-            }
-            seen.insert(parent_id);
-            ino_it = st_cli.inode_config.find(parent_id);
-            if (INODE_POOL(parent_id) == INODE_POOL(ino))
-            {
-                icache->chain.push_back(parent_id);
-                if (ino_it == st_cli.inode_config.end())
-                    chain_cfg.push_back(NULL);
-                else
-                {
-                    chain_cfg.push_back(&ino_it->second);
-                    if (!ino_it->second.enc_key.empty())
-                        enc_key_count++;
-                }
-            }
-            else if (!icache->other_pool_parent_id)
-                icache->other_pool_parent_id = parent_id;
-            if (ino_it == st_cli.inode_config.end())
-                break;
-            parent_id = ino_it->second.parent_id;
-        }
-    }
-    // Check external keys and wait for loading, if required
-    if (enc_key_count)
-    {
-        for (size_t i = 0; i <= chain_cfg.size(); i++)
-        {
-            inode_config_t *cfg = !i ? &inode_cfg : chain_cfg[i-1];
-            if (cfg && cfg->enc_key.substr(0, strlen(VAULT_KEY_PREFIX)) == VAULT_KEY_PREFIX)
-            {
-                auto & ik = vault_keys[inode_cfg.enc_key];
-                if (ik.key_state == VAULT_KEY_ERROR || vault_url.empty())
-                {
-                    icache->err_code = EPERM;
-                    enc_key_count = 0;
-                }
-                else if (ik.key_state == VAULT_KEY_NOT_LOADED)
-                {
-                    ik.key_state = VAULT_KEY_LOADING;
-                    vault_key_load_queue.push_back(inode_cfg.enc_key);
-                    vault_load_keys();
-                    icache->err_code = EAGAIN;
-                    enc_key_count = 0;
-                }
-                else if (ik.key_state == VAULT_KEY_LOADING)
-                {
-                    icache->err_code = EAGAIN;
-                    enc_key_count = 0;
-                }
-                else
-                {
-                    assert(ik.key_state == VAULT_KEY_LOADED);
-                }
-            }
-        }
-    }
-    // Generate encryption key chain, if applicable
-    if (enc_key_count)
-    {
-        uint8_t *key_data = (uint8_t*)malloc_or_die(
-            AES_256_XTS_KEY_SIZE * enc_key_count +
-            sizeof(uint8_t*) * icache->chain.size() +
-            sizeof(osd_op_enc_t)
-        );
-        uint8_t **keys = (uint8_t**)(key_data + AES_256_XTS_KEY_SIZE * enc_key_count);
-        osd_op_enc_t *enc = (osd_op_enc_t*)((uint8_t*)keys + sizeof(uint8_t*)*icache->chain.size());
-        size_t key_pos = 0;
-        for (size_t i = 0; i <= chain_cfg.size(); i++)
-        {
-            inode_config_t *cfg = !i ? &inode_cfg : chain_cfg[i-1];
-            if (cfg && !cfg->enc_key.empty())
-            {
-                const auto & key = cfg->enc_key.substr(0, strlen(VAULT_KEY_PREFIX)) == VAULT_KEY_PREFIX
-                    ? vault_keys.at(cfg->enc_key).key
-                    : cfg->enc_key;
-                assert(key_pos < AES_256_XTS_KEY_SIZE * enc_key_count);
-                assert(key.size() == 2*AES_256_XTS_KEY_SIZE);
-                keys[i] = key_data + key_pos;
-                fromhexstr(key, AES_256_XTS_KEY_SIZE, key_data + key_pos);
-                key_pos += AES_256_XTS_KEY_SIZE;
-            }
-            else
-                keys[i] = NULL;
-        }
-        enc->key_chain = keys;
-        enc->chain_size = icache->chain.size();
-        enc->read_chain_bitmap_pos = pool_cfg.data_block_size/pool_cfg.bitmap_granularity/8;
-        enc->bitmap_granularity = pool_cfg.bitmap_granularity;
-        icache->key_data = key_data;
-        icache->op_enc = enc;
-    }
-    inode_cache[ino] = icache;
-    for (auto & parent: icache->chain)
-    {
-        if (parent != ino)
-            inode_cache_children.insert(std::make_pair(parent, ino));
-    }
-    return icache;
-}
-
-#ifndef __MOCK__
-bool cluster_client_t::vault_check_token()
-{
-    timespec now;
-    clock_gettime(CLOCK_REALTIME, &now);
-    if (!vault_token_expire.tv_sec || vault_token_expire.tv_sec < now.tv_sec)
-    {
-        vault_loading = true;
-        http_json_post(
-            vault_http_cli, vault_url+"/v1/auth/cert/login", json11::Json::object{}, "",
-            (http_options_t){ .timeout = (int)vault_timeout_ms, .keepalive = true },
-            [this](http_message_t *response)
-            {
-                clock_gettime(CLOCK_REALTIME, &vault_token_expire);
-                vault_loading = false;
-                std::string err;
-                json11::Json data;
-                response->parse_json_response(err, data);
-                if (err != "")
-                {
-                    vault_token_expire.tv_sec += vault_error_timeout_sec;
-                    fprintf(stderr, "Vault request failed: %s\n", err.c_str());
-                }
-                else
-                {
-                    uint64_t ttl = data["auth"]["lease_duration"].uint64_value();
-                    vault_token = data["auth"]["client_token"].string_value();
-                    if (vault_token.empty() || !ttl)
-                    {
-                        vault_token_expire.tv_sec += vault_error_timeout_sec;
-                        fprintf(stderr, "No token or lease_duration in Vault response: %s\n", data.dump().c_str());
-                    }
-                    else
-                    {
-                        if (ttl < vault_refresh_leeway_sec)
-                            vault_token_expire.tv_sec += ttl/2;
-                        else
-                            vault_token_expire.tv_sec += ttl - vault_refresh_leeway_sec;
-                    }
-                }
-                vault_load_keys();
-            }
-        );
-        return false;
-    }
-    if (vault_token.empty())
-    {
-        // Auth error happened, mark all loads as failed
-        for (auto & key_id: vault_key_load_queue)
-        {
-            auto & k = vault_keys[key_id];
-            k.key_state = VAULT_KEY_ERROR;
-        }
-        vault_key_load_queue.clear();
-        auto ops = std::move(key_wait_ops);
-        for (cluster_op_t *op: ops)
-            inode_cache.erase(op->inode);
-        for (cluster_op_t *op: ops)
-            execute_internal(op);
-        return false;
-    }
-    return true;
-}
-#endif
-
-void cluster_client_t::vault_load_keys()
-{
-    if (vault_loading || !vault_key_load_queue.size())
-    {
-        return;
-    }
-#ifdef __MOCK__
-    vault_loading = true;
-#else
-    if (!vault_http_ctx)
-    {
-        std::string error;
-        vault_http_ctx = http_context_init(tfd, vault_client_cert, vault_client_key, vault_ca, true, error);
-        if (!vault_http_ctx)
-        {
-            fprintf(stderr, "Failed to initialize HTTP context for Vault: %s\n", error.c_str());
-            exit(1);
-        }
-        vault_http_cli = http_init(vault_http_ctx);
-    }
-    if (!vault_check_token())
-    {
-        return;
-    }
-    std::string key_id = vault_key_load_queue[0];
-    vault_key_load_queue.erase(vault_key_load_queue.begin());
-    vault_loading = true;
-    http_get(
-        vault_http_cli, vault_url+vault_secret_api_path+key_id.substr(strlen(VAULT_KEY_PREFIX)), "X-Vault-Token: "+vault_token+"\r\n",
-        (http_options_t){ .timeout = (int)vault_timeout_ms, .keepalive = true },
-        [this, key_id](http_message_t *response)
-        {
-            vault_loading = false;
-            std::string err;
-            json11::Json data;
-            response->parse_json_response(err, data);
-            vault_parse_secret(key_id, err, data);
-        }
-    );
-#endif
-}
-
-void cluster_client_t::vault_parse_secret(const std::string & key_id, const std::string & err, json11::Json data)
-{
-    vault_loading = false;
-    auto & k = vault_keys[key_id];
-    if (err != "")
-    {
-        k.key_state = VAULT_KEY_ERROR;
-        fprintf(stderr, "Vault %s%s%s request failed: %s\n", vault_url.c_str(),
-            vault_secret_api_path.c_str(), key_id.c_str()+strlen(VAULT_KEY_PREFIX), err.c_str());
-    }
-    else
-    {
-        auto hexkey = data["data"]["key"].string_value();
-        if (hexkey.empty() || !ishexstr(hexkey) || hexkey.size() != 2*AES_256_XTS_KEY_SIZE)
-        {
-            k.key_state = VAULT_KEY_ERROR;
-            fprintf(stderr, "Vault /v1/secret/%s request failed: 'key' is empty or has invalid format\n", key_id.c_str());
-        }
-        else
-        {
-            k.key_state = VAULT_KEY_LOADED;
-            k.key = hexkey;
-        }
-    }
-    if (vault_key_load_queue.empty())
-    {
-        auto ops = std::move(key_wait_ops);
-        for (cluster_op_t *op: ops)
-            inode_cache.erase(op->inode);
-        for (cluster_op_t *op: ops)
-            execute_internal(op);
-    }
-    else
-        vault_load_keys();
-}

src/client/cluster_client_impl.h

@@ -5,6 +5,7 @@
 
 #include "cluster_client.h"
 
+#define SCRAP_BUFFER_SIZE 4*1024*1024
 #define PART_SENT 1
 #define PART_DONE 2
 #define PART_ERROR 4

src/client/cluster_client_list.cpp

@@ -63,14 +63,14 @@ void cluster_client_t::list_inode(inode_t inode, uint64_t min_offset, uint64_t m
 {
     init_msgr();
     pool_id_t pool_id = INODE_POOL(inode);
-    if (!pool_id || st_cli.pool_config.find(pool_id) == st_cli.pool_config.end())
+    if (!pool_id || st_cli->pool_config.find(pool_id) == st_cli->pool_config.end())
     {
         if (log_level > 0)
             fprintf(stderr, "Pool %u does not exist\n", pool_id);
         pg_callback(-EINVAL, 0, 0, std::set<object_id>());
         return;
     }
-    auto pg_stripe_size = st_cli.pool_config.at(pool_id).pg_stripe_size;
+    auto pg_stripe_size = st_cli->pool_config.at(pool_id).pg_stripe_size;
     if (min_offset)
         min_offset = (min_offset/pg_stripe_size) * pg_stripe_size;
     inode_list_t *lst = new inode_list_t();
@@ -110,13 +110,13 @@ bool cluster_client_t::continue_listing(inode_list_t *lst)
 
 bool cluster_client_t::restart_listing(inode_list_t* lst)
 {
-    auto pool_it = st_cli.pool_config.find(lst->pool_id);
+    auto pool_it = st_cli->pool_config.find(lst->pool_id);
     // We want listing to be consistent. To achieve it we should:
     // 1) retry listing of each PG if its state changes
     // 2) abort listing if PG count changes during listing
     // 3) ideally, only talk to the primary OSD - this will be done separately
     // So first we add all PGs without checking their state
-    if (pool_it == st_cli.pool_config.end() ||
+    if (pool_it == st_cli->pool_config.end() ||
         lst->real_pg_count != pool_it->second.real_pg_count)
     {
         for (auto pg: lst->pgs)
@@ -136,7 +136,7 @@ bool cluster_client_t::restart_listing(inode_list_t* lst)
             fprintf(stderr, "PG count in pool %u changed during listing\n", lst->pool_id);
         }
         lst->pgs.clear();
-        if (pool_it == st_cli.pool_config.end())
+        if (pool_it == st_cli->pool_config.end())
         {
             // Unknown pool
             lst->callback(-EINVAL, 0, 0, std::set<object_id>());
@@ -248,7 +248,7 @@ void cluster_client_t::set_list_retry_timeout(int ms, timespec new_time)
 
 int cluster_client_t::start_pg_listing(inode_list_pg_t *pg)
 {
-    auto & pool_cfg = st_cli.pool_config.at(pg->lst->pool_id);
+    auto & pool_cfg = st_cli->pool_config.at(pg->lst->pool_id);
     auto pg_it = pool_cfg.pg_config.find(pg->pg_num);
     assert(pg->lst->real_pg_count == pool_cfg.real_pg_count);
     if (pg_it == pool_cfg.pg_config.end() ||
@@ -277,7 +277,7 @@ int cluster_client_t::start_pg_listing(inode_list_pg_t *pg)
         for (auto peer_it = all_peers.begin(); peer_it != all_peers.end(); )
         {
             if (*peer_it != pg_it->second.cur_primary &&
-                st_cli.peer_states[*peer_it].is_null())
+                st_cli->peer_states[*peer_it].is_null())
             {
                 pg->inactive_osds.push_back(*peer_it);
                 all_peers.erase(peer_it++);
@@ -298,11 +298,11 @@ int cluster_client_t::start_pg_listing(inode_list_pg_t *pg)
         if (msgr.osd_peers.find(peer_osd) == msgr.osd_peers.end())
         {
             // Initiate connection
-            if (st_cli.peer_states[peer_osd].is_null())
+            if (st_cli->peer_states[peer_osd].is_null())
             {
                 return LIST_PG_WAIT_ACTIVE;
             }
-            msgr.connect_peer(peer_osd, st_cli.peer_states[peer_osd]);
+            msgr.connect_peer(peer_osd, st_cli->peer_states[peer_osd]);
             conn = false;
         }
     }
@@ -336,7 +336,7 @@ void cluster_client_t::send_list(inode_list_osd_t *cur_list)
     if (!cur_list->pg->inflight_ops)
         cur_list->pg->lst->inflight_pgs++;
     cur_list->pg->inflight_ops++;
-    auto & pool_cfg = st_cli.pool_config[cur_list->pg->lst->pool_id];
+    auto & pool_cfg = st_cli->pool_config[cur_list->pg->lst->pool_id];
     osd_op_t *op = new osd_op_t();
     op->op_type = OSD_OP_OUT;
     // Already checked that it exists above, but anyway

src/client/cluster_client_real.cpp

@@ -0,0 +1,11 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
+
+#include "cluster_client.h"
+#include "etcd_state_client_http.h"
+
+cluster_client_t* cluster_client_t::create(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json config)
+{
+    auto st_cli = new etcd_state_client_http_t(tfd);
+    return new cluster_client_t(ringloop, tfd, config, std::unique_ptr<etcd_state_client_t>(st_cli));
+}