mirrors/seaweedfs
1
0
Fork 0
mirror of https://github.com/seaweedfs/seaweedfs.git synced 2026-08-19 13:46:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs(p15): mark G9A close-ready

Browse Source
This commit is contained in:
pingqiu
2026-05-02 18:38:39 -07:00
parent f5670b5dc4
commit e72ce0d530
1 changed files with 12 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
sw-block/design/v3-phase-15-g9a-ack-reintegration-mini-plan.md
+12 -9
View File
@@ -1,7 +1,7 @@
# V3 Phase 15 — G9A ACK + Reintegration Policy Mini-Plan
**Date**: 2026-05-02
**Status**: ACTIVE; first ACK profile slice pushed on `p15-g9a/ack-reintegration-policy`
**Status**: CLOSE-READY; first ACK/reintegration policy slice pushed on `p15-g9a/ack-reintegration-policy`
**Predecessor**: G8 CLOSED 2026-05-02
**Code repo**: `seaweed_block`
@@ -59,13 +59,13 @@ Rule: recovery progress is not a substitute for synchronous ACK eligibility.
| `154bd96` | subprocess L2 best-effort oracle: same RF=2 daemon/iSCSI shape, secondary down => foreground WRITE still returns GOOD. |
| `da8a321` | authority reintegration oracle: returned/high-evidence replica with `ReadyForPrimary=false` is skipped as failover target until a progress-ready fact exists. |
| `6d4a0e7` | `/status` maps engine `ModeRecovering` to `ReplicationRole=recovering`, distinct from `not_ready` and `replica_ready` (not yet emitted). |
| `f1117ce` | component oracle: best-effort foreground writes succeed while replica is down, then production probe/recovery path catches the lagging replica back up to byte-equal. |
### 3.2 Next red tests
### 3.2 Forward-carry tests
1. `TestG9A_SyncQuorumWriteFailsWhenPeerInRecovery_Process`
- Follow-up variant where peer is explicitly in recovery (not merely down) and the daemon path still fails foreground ACK.
2. `TestG9A_BestEffortStillFeedsLaggingReplica`
- Follow-up variant that proves best-effort success does not disable catch-up/rebuild for the lagging peer.
- Follow-up L2 variant where peer is explicitly in recovery (not merely down) and the daemon path still fails foreground ACK.
- Requires a product-safe way to force/observe peer `ReplicaCatchingUp` in subprocess harness; do not add a test-only production RPC just for this.
---
@@ -81,9 +81,12 @@ G9A first slices do not claim:
---
## 5. Immediate Next Step
## 5. Close-Ready Posture
Continue on `p15-g9a/ack-reintegration-policy`:
G9A first slice can close on current evidence if architect accepts subprocess-secondary-down as the L2 strict ACK oracle and component-level catching-up as the explicit recovery-state oracle.
1. Decide whether first G9A closes here or includes explicit in-recovery daemon ACK variant.
2. If continuing, add best-effort lagging-peer feed/recovery evidence.
Non-claims to carry forward:
1. Product-daemon L2 "peer explicitly in recovery" strict ACK variant.
2. `replica_ready` publication after completed reintegration.
3. RF>=3 quorum/placement policy.