mirrors/seaweedfs
1
0
Fork 0
mirror of https://github.com/seaweedfs/seaweedfs.git synced 2026-09-19 22:44:18 +00:00
Code Issues Packages Projects Releases Wiki Activity
13,152 Commits 181 Branches 360 Tags
3da4c19046e06ec7e75cb449e0d9287ef1a3473b
Commit Graph
8 Commits
Author SHA1 Message Date
pingqiuandClaude Opus 4.6 ebe95b6e2e fix: flusher OOM on multi-block writes + testrunner enhancements 
Bug: flusher.go:336 allocated make([]byte, entryLen) per dirty block
instead of per unique WAL entry. A 4MB WriteLBA creates 1024 dirty map
entries (one per 4KB block), all sharing the same WAL offset. The flusher
read the full 4MB WAL entry 1024 times into separate buffers:
1024 × 4MB = 4GB per 4MB write → OOM on mkfs.ext4.

Root cause: flusher assumed 1:1 dirty-block-to-WAL-entry mapping.
WriteLBA supports multi-block writes but the flusher never deduplicated
shared WAL offsets.

Fix: deduplicate WAL reads by WalOffset in flushOnceLocked(). Multiple
dirty blocks from the same WAL entry share one read buffer and one
DecodeWALEntry call. Memory: O(WAL_entries × size) not O(blocks × size).
For a 4MB write: 4GB → 4MB.

Verified on hardware (m01/M02 25Gbps RoCE):
- Before: mkfs.ext4 → VS RSS 100MB→25GB → OOM killed
- After: mkfs.ext4 → VS RSS 129MB stable, mkfs succeeds
- pgbench TPC-B c=4: 1,248 TPS (RF=1, previously blocked by OOM)

Tests added:
- flusher_test.go: flush_multiblock_shared_wal_read (16 blocks share
  one WAL offset, flush dedup verified)
- flusher_test.go: flush_multiblock_data_correct (3 mixed multi-block
  writes, all data correct after flush)
- test/component/large_write_test.go: 7 component tests (single 4MB,
  sequential mkfs sim, concurrent, mixed sizes, production volume,
  flusher throughput 30s sustained)
- iscsi/large_write_mem_test.go: 2 iSCSI session memory tests (4MB
  R2T flow, slow device)

Testrunner enhancements (same commit — all tested on hardware):
- discover_primary action: maps primary IP → topology node name,
  supports alt_ips for multi-NIC (RoCE + management)
- NodeSpec.AltIPs field for multi-NIC node identification
- 5 new YAML scenarios: ec3, ec5, degraded sync_all/best_effort, pgbench
- All 13 hardware-verified scenarios PASS

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-02 14:24:10 -07:00
pingqiuandClaude Opus 4.6 785a7d7efd feat: wire real pinner into flusher retention + real WAL scan executor (Phase 07 P1) 
Pinner wired to real retention:
- NewPinner calls vol.SetV2RetentionFloor(p.MinWALRetentionFloor)
- Flusher.RetentionFloorFn() / SetRetentionFloorFn() exposed
- SetV2RetentionFloor chains with existing shipper retention floor
- Holds actually prevent WAL reclaim (not just tracked state)

Executor uses real WAL scan:
- BlockVol.ScanWALEntries(fromLSN, callback) wraps wal.ScanFrom
  with real fd, walOffset, checkpointLSN
- Executor.StreamWALEntries uses ScanWALEntries (not stub)
- Reads real WAL entries, tracks highest LSN scanned

CommittedLSN mapping:
- Explicitly documented as interim V1 model (committed = checkpointed)
- Will diverge when V2 distributed commit separates from local flush

Carry-forward:
- TransferSnapshot/TransferFullBase/TruncateWAL: stubs (need extent I/O)
- Control intent from confirmed failover: deferred

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-03-30 20:01:46 -07:00
Ping QiuandClaude Opus 4.6 8b2b5f6f66 feat: Phase 6 CP6-3 -- failover + rebuild in Kubernetes, 126 tests 
Wire low-level fencing primitives to master/VS control plane and CSI:

- Proto: replica/rebuild address fields on assignment/info/response messages
- Assignment queue: retain-until-confirmed (Peek+Confirm), stale epoch pruning
- VS assignment receiver: processes assignments from HeartbeatResponse
- BlockService replication: ProcessAssignments, deterministic ports (FNV hash)
- Registry replica tracking: SetReplica/ClearReplica/SwapPrimaryReplica
- CreateBlockVolume: primary + replica, enqueues assignments, single-copy mode
- Failover: lease-aware promotion, deferred timers with cancellation on reconnect
- ControllerPublish: returns fresh primary iSCSI address after failover
- Recovery: recoverBlockVolumes drains pendingRebuilds, enqueues Rebuilding
- Real integration tests on M02: failover address switch, rebuild data
  consistency, full lifecycle failover+rebuild (3 tests, all PASS)

Review fixes (12 findings, 5 High, 5 Medium, 2 Low):
- R1-1: AllocateBlockVolume returns replication ports
- R1-2: setupPrimaryReplication starts rebuild server
- R1-3: VS sends periodic block heartbeat for assignment confirmation
- R2-F1: LastLeaseGrant set before Register (no stale-lease race)
- R2-F2: Deferred promotion timers cancelled on VS reconnect
- R2-F3: SwapPrimaryReplica uses RoleToWire instead of uint32(1)
- R2-F4: DeleteBlockVolume deletes replica (best-effort)
- R2-F5: SwapPrimaryReplica computes epoch atomically under lock
- QA: SetReplica removes old replica from byServer index (BUG-QA-CP63-1)

126 CP6-3 tests (67 dev + 48 QA + 8 integration + 3 real).
Cumulative Phase 6: 352 tests. All PASS.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-05 00:52:05 -08:00
Ping QiuandClaude Opus 4.6 5a9a52f2d0 feat: Phase 6 CP6-2 -- CSI control-plane integration + csi-sanity/k3s validation 
CP6-2 wires the CSI driver to SeaweedFS master/volume-server control plane:
- Proto: block volume messages in master.proto/volume_server.proto, codegen
- Master registry: in-memory BlockVolumeRegistry with Pending->Active status,
  full/delta heartbeat, inflight lock, placement (fewest volumes)
- VS gRPC: AllocateBlockVolume/DeleteBlockVolume handlers, shared naming
- Master RPCs: CreateBlockVolume (retry up to 3 servers), Delete, Lookup
- Heartbeat: block volume fields wired into bidirectional stream
- CSI Controller: VolumeBackend interface (Local + Master), returns volume_context
- CSI Node: reads volume_context for remote targets, staged map + IQN derivation
- Mode flag: --mode=controller/node/all, --master for control-plane
- K8s manifests: csi-driver.yaml, csi-controller.yaml, csi-node.yaml

csi-sanity conformance (33 pass, 58 skip) found 6 bugs:
- BUG-SANITY-1/2/3: missing VolumeCapabilities/VolumeCapability validation
- BUG-SANITY-4: NodePublish used mount instead of bind mount
- BUG-SANITY-5: NodeUnpublish didn't remove target path
- BUG-SANITY-6: NodeUnpublish failed on unmounted path

k3s Level 4 (PVC->Pod data persistence) found 1 bug:
- BUG-K3S-1: IsLoggedIn didn't handle iscsiadm exit code 21

226 CSI tests + 54 server tests = 280 new tests, all passing.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-04 11:01:08 -08:00
Ping QiuandClaude Opus 4.6 98d0e9e631 feat: Phase 5 CP5-1 -- ALUA + multipath failover, 28 tests 
Add ALUA (Asymmetric Logical Unit Access) support to the iSCSI target,
enabling dm-multipath on Linux to automatically detect path state changes
and reroute I/O during HA failover without initiator-side intervention.

- ALUAProvider interface with implicit ALUA (TPGS=0x01)
- INQUIRY byte 5 TPGS bits, VPD 0x83 with NAA+TPG+RTP descriptors
- REPORT TARGET PORT GROUPS handler (MAINTENANCE IN SA=0x0A)
- MAINTENANCE OUT rejection (implicit-only, no SET TPG)
- Standby write rejection (NOT_READY ASC=04h ASCQ=0Bh)
- RoleNone maps to Active/Optimized (standalone single-node compatibility)
- NAA-6 device identifier derived from volume UUID
- -tpg-id flag with [1,65535] validation
- dm-multipath config + setup script (group_by_tpg, ALUA prio)
- 12 unit tests + 16 QA adversarial tests + 4 integration tests

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-02 23:29:58 -08:00
Ping QiuandClaude Opus 4.6 7940e6b7c9 feat: Phase 4A CP4b-4 Windows iSCSI + instrumentation + QA tests 
Windows iSCSI Initiator compatibility:
- Add TargetPortalGroupTag to login response (RFC 7143 S13.9)
- Add REQUEST_SENSE, START_STOP_UNIT, MODE_SELECT(6/10) handlers
- Add PERSISTENT_RESERVE_IN/OUT, MAINTENANCE_IN (REPORT SUPPORTED OPCODES)
- Implement MODE SENSE caching (page 0x08) and control (page 0x0A) pages
- Fix Data-In residual underflow/overflow flags (U/O bits on final PDU)
- Rename ScsiReadCapacity16 -> ScsiServiceActionIn16 for correctness

Instrumentation and tooling:
- Add instrumentedAdapter with periodic PERF stats logging
- Add pprof endpoints on admin HTTP server (/debug/pprof/*)
- Add blockbench CLI tool for standalone block device benchmarking
- Add SCSI CDB debug logging in session dispatch

HA integration fixes:
- Move HA test replica ports to 9011-9014 to avoid conflicts
- Add QA adversarial tests for Phase 4A CP4b-4 (755 lines)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-02 21:17:57 -08:00
Ping QiuandClaude Opus 4.6 44da35faf6 test: add integration test infrastructure for blockvol iSCSI 
Test harness for running blockvol iSCSI tests on WSL2 and remote nodes
(m01/M02). Includes Node (SSH/local exec), ISCSIClient (discover/login/
logout), WeedTarget (weed volume server lifecycle), and test suites for
smoke, stress, crash recovery, chaos, perf benchmarks, and apps (fio/dd).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-02 15:05:29 -08:00
Ping QiuandClaude Opus 4.6 c39080ceaa feat: Phase 4A CP4b-4 -- HA integration tests, admin HTTP, 5 bug fixes 
Add HTTP admin server to iscsi-target binary (POST /assign, GET /status,
POST /replica, POST /rebuild) and 7 HA integration tests validating
failover, split-brain prevention, epoch fencing, and demote-under-IO.

New files:
- admin.go: HTTP admin endpoint with input validation
- ha_target.go: HATarget helper wrapping Target + admin HTTP calls
- ha_test.go: 7 HA tests (all PASS on WSL2, 67.7s total)

Bug fixes:
- BUG-CP4B4-1: CmdSN init (expCmdSN=0 not 1, first SCSI cmd was dropped)
- BUG-CP4B4-2: RoleNone->RoleReplica missing SetEpoch (WAL rejected)
- BUG-CP4B4-3: replica applyEntry didn't update vol.nextLSN (status=0)
- BUG-CP4B4-4: PID discovery killed primary instead of replica (shared
  binPath; fixed by grepping volFile)
- BUG-CP4B4-5: artifact collector overwrote primary log with replica log
  (added CollectLabeled method)

Also: 3s write deadline on WAL shipper data connection to avoid 120s TCP
retransmission timeout when replica is dead.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-02 15:04:59 -08:00