fix(rpc): add issue 2815 regression and docker validation (#2828)

2026-07-26 08:18:18 +00:00 · 2026-05-06 20:22:13 +08:00
parent 3898d524fe
commit 41ba34a145
5 changed files with 302 additions and 8 deletions
@@ -0,0 +1 @@
+data/
@@ -0,0 +1,106 @@
+# Issue 2815 Local Docker Verification
+
+## Purpose
+
+This directory contains the local distributed Docker verification assets used to validate issue `#2815` against the current source build.
+
+The target behavior is:
+
+- 4-node distributed cluster starts successfully
+- `/health/ready` becomes reachable on each node
+- logs no longer contain `storage_info failed: Io error: wrong msgpack marker FixArray(1)`
+- internode RPC authentication succeeds with an explicit non-default RPC secret
+
+## Files
+
+- `docker-compose.yml`: 4-node distributed cluster using a locally built image
+
+## Data Directories
+
+Create the bind-mount directories before `docker compose up`:
+
+```bash
+mkdir -p .docker/test/issues-2815/data/rustfs{1..4}-disk{0..3}
+```
+
+## Build
+
+Apple Silicon / arm64 host:
+
+```bash
+docker build --platform linux/arm64 -f Dockerfile.source -t rustfs-issue-2815-local .
+```
+
+If you intentionally want amd64 emulation:
+
+```bash
+docker build --platform linux/amd64 -f Dockerfile.source -t rustfs-issue-2815-local .
+```
+
+## Run
+
+```bash
+docker compose -f .docker/test/issues-2815/docker-compose.yml up -d
+```
+
+If the image platform is not `linux/arm64`, align compose explicitly:
+
+```bash
+RUSTFS_DOCKER_PLATFORM=linux/amd64 docker compose -f .docker/test/issues-2815/docker-compose.yml up -d
+```
+
+## Health Checks
+
+Container-level healthcheck is now included and probes:
+
+```bash
+curl -fsS http://127.0.0.1:9000/health
+```
+
+Manual checks:
+
+```bash
+curl -i http://127.0.0.1:9101/health/ready
+curl -i http://127.0.0.1:9102/health/ready
+curl -i http://127.0.0.1:9103/health/ready
+curl -i http://127.0.0.1:9104/health/ready
+```
+
+## RPC Secret Requirement
+
+The current source build no longer reproduces the original `FixArray(1)` decode error from issue `#2815`.
+
+Earlier local Docker attempts failed during erasure bootstrap with:
+
+```text
+No valid auth token
+store init failed to load formats after 10 retries: erasure read quorum
+```
+
+Root cause:
+
+- RPC authentication rejects the default secret `rustfsadmin`
+- distributed local Docker validation therefore needs an explicit non-default secret
+
+This compose now sets both:
+
+- `RUSTFS_SECRET_KEY=issue-2815-secret`
+- `RUSTFS_RPC_SECRET=issue-2815-rpc-secret`
+
+With those values in place, the current 4-node local Docker cluster reaches healthy state and `/health/ready` returns `200`.
+
+In other words:
+
+- `RUSTFS_ACCESS_KEY` may still be `rustfsadmin` for local service credentials if desired
+- `RUSTFS_SECRET_KEY` can still be used for service credentials
+- but RPC authentication must not resolve to the default secret value `rustfsadmin`
+- if `RUSTFS_RPC_SECRET` is unset, the code falls back to `RUSTFS_SECRET_KEY`
+- so at least one of them must provide a non-default shared secret for internode RPC signing
+
+## Suggested Debug Commands
+
+```bash
+docker compose -f .docker/test/issues-2815/docker-compose.yml ps
+docker compose -f .docker/test/issues-2815/docker-compose.yml logs --no-color --tail=200
+docker compose -f .docker/test/issues-2815/docker-compose.yml down -v
+```
@@ -0,0 +1,120 @@
+services:
+  rustfs1:
+    image: rustfs-issue-2815-local
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/arm64}
+    hostname: rustfs1
+    container_name: rustfs-issue-2815-rustfs1
+    environment:
+      RUSTFS_ADDRESS: "0.0.0.0:9000"
+      RUSTFS_ACCESS_KEY: "rustfsadmin"
+      RUSTFS_SECRET_KEY: "issue-2815-secret"
+      RUSTFS_RPC_SECRET: "issue-2815-rpc-secret"
+      RUSTFS_CONSOLE_ENABLE: "false"
+      RUST_LOG: "info"
+      RUSTFS_UNSAFE_BYPASS_DISK_CHECK: "true"
+      RUSTFS_VOLUMES: "http://rustfs{1...4}:9000/data/rustfs{0...3}"
+    volumes:
+      - ./data/rustfs1-disk0:/data/rustfs0
+      - ./data/rustfs1-disk1:/data/rustfs1
+      - ./data/rustfs1-disk2:/data/rustfs2
+      - ./data/rustfs1-disk3:/data/rustfs3
+    networks: [rustfs-issue-2815-net]
+    ports:
+      - "9101:9000"
+    healthcheck:
+      test: ["CMD", "sh", "-c", "curl -fsS http://127.0.0.1:9000/health || exit 1"]
+      interval: 15s
+      timeout: 5s
+      retries: 8
+      start_period: 30s
+
+  rustfs2:
+    image: rustfs-issue-2815-local
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/arm64}
+    hostname: rustfs2
+    container_name: rustfs-issue-2815-rustfs2
+    environment:
+      RUSTFS_ADDRESS: "0.0.0.0:9000"
+      RUSTFS_ACCESS_KEY: "rustfsadmin"
+      RUSTFS_SECRET_KEY: "issue-2815-secret"
+      RUSTFS_RPC_SECRET: "issue-2815-rpc-secret"
+      RUSTFS_CONSOLE_ENABLE: "false"
+      RUST_LOG: "info"
+      RUSTFS_UNSAFE_BYPASS_DISK_CHECK: "true"
+      RUSTFS_VOLUMES: "http://rustfs{1...4}:9000/data/rustfs{0...3}"
+    volumes:
+      - ./data/rustfs2-disk0:/data/rustfs0
+      - ./data/rustfs2-disk1:/data/rustfs1
+      - ./data/rustfs2-disk2:/data/rustfs2
+      - ./data/rustfs2-disk3:/data/rustfs3
+    networks: [rustfs-issue-2815-net]
+    ports:
+      - "9102:9000"
+    healthcheck:
+      test: ["CMD", "sh", "-c", "curl -fsS http://127.0.0.1:9000/health || exit 1"]
+      interval: 15s
+      timeout: 5s
+      retries: 8
+      start_period: 30s
+
+  rustfs3:
+    image: rustfs-issue-2815-local
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/arm64}
+    hostname: rustfs3
+    container_name: rustfs-issue-2815-rustfs3
+    environment:
+      RUSTFS_ADDRESS: "0.0.0.0:9000"
+      RUSTFS_ACCESS_KEY: "rustfsadmin"
+      RUSTFS_SECRET_KEY: "issue-2815-secret"
+      RUSTFS_RPC_SECRET: "issue-2815-rpc-secret"
+      RUSTFS_CONSOLE_ENABLE: "false"
+      RUST_LOG: "info"
+      RUSTFS_UNSAFE_BYPASS_DISK_CHECK: "true"
+      RUSTFS_VOLUMES: "http://rustfs{1...4}:9000/data/rustfs{0...3}"
+    volumes:
+      - ./data/rustfs3-disk0:/data/rustfs0
+      - ./data/rustfs3-disk1:/data/rustfs1
+      - ./data/rustfs3-disk2:/data/rustfs2
+      - ./data/rustfs3-disk3:/data/rustfs3
+    networks: [rustfs-issue-2815-net]
+    ports:
+      - "9103:9000"
+    healthcheck:
+      test: ["CMD", "sh", "-c", "curl -fsS http://127.0.0.1:9000/health || exit 1"]
+      interval: 15s
+      timeout: 5s
+      retries: 8
+      start_period: 30s
+
+  rustfs4:
+    image: rustfs-issue-2815-local
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/arm64}
+    hostname: rustfs4
+    container_name: rustfs-issue-2815-rustfs4
+    environment:
+      RUSTFS_ADDRESS: "0.0.0.0:9000"
+      RUSTFS_ACCESS_KEY: "rustfsadmin"
+      RUSTFS_SECRET_KEY: "issue-2815-secret"
+      RUSTFS_RPC_SECRET: "issue-2815-rpc-secret"
+      RUSTFS_CONSOLE_ENABLE: "false"
+      RUST_LOG: "info"
+      RUSTFS_UNSAFE_BYPASS_DISK_CHECK: "true"
+      RUSTFS_VOLUMES: "http://rustfs{1...4}:9000/data/rustfs{0...3}"
+    volumes:
+      - ./data/rustfs4-disk0:/data/rustfs0
+      - ./data/rustfs4-disk1:/data/rustfs1
+      - ./data/rustfs4-disk2:/data/rustfs2
+      - ./data/rustfs4-disk3:/data/rustfs3
+    networks: [rustfs-issue-2815-net]
+    ports:
+      - "9104:9000"
+    healthcheck:
+      test: ["CMD", "sh", "-c", "curl -fsS http://127.0.0.1:9000/health || exit 1"]
+      interval: 15s
+      timeout: 5s
+      retries: 8
+      start_period: 30s
+
+networks:
+  rustfs-issue-2815-net:
+    name: rustfs-issue-2815-net