ewilliams28/openclaw
1
0
Fork 0
mirror of https://github.com/openclaw/openclaw.git synced 2026-08-25 20:05:46 -06:00
Code Issues Packages Projects Releases Wiki Activity

fix(qa): restore paired-node worker crash recovery proof (#129241)

Browse Source
This commit is contained in:
Peter Steinberger
2026-08-25 03:54:50 -07:00
committed by GitHub
parent 65b566a29b
commit 1ba243c88e
3 changed files with 73 additions and 296 deletions
Download Patch File Download Diff File Expand all files Collapse all files
qa/scenarios/runtime/cloud-worker-midturn-loss.yaml
+6 -6
View File
@@ -7,14 +7,14 @@ scenario:
coverage:
secondary:
- gateway.session-apis-sessions-list
objective: Prove a static-SSH worker can disappear during a streamed turn without losing or duplicating its already committed transcript prefix, silently hanging the turn, or breaking redispatch context; prove that real worker disk-space projections are evicted after reclaim.
objective: Prove a paired-node worker can disappear during a streamed turn without losing or duplicating its already committed transcript prefix, silently hanging the turn, or breaking redispatch context; prove that real worker disk-space projections are evicted after reclaim.
successCriteria:
- A managed-worktree qa-channel session dispatches to a real static-SSH worker through an isolated Gateway.
- A managed-worktree qa-channel session dispatches to a real paired-node worker through an isolated Gateway.
- The mock model persists two assistant/tool-result checkpoints, then pauses during a fifth streamed message.
- Killing the proof-owned SSH and worker process tree leaves exactly the four completed checkpoint messages in Gateway history.
- Killing the proof-owned worker process leaves exactly the four completed checkpoint messages in Gateway history before its paired node disconnects.
- The volatile streamed message is absent from durable history while the turn emits a visible chat error and the placement records a bounded terminal reason.
- Restarting the static-SSH host and redispatching the same session produces one successful recovery turn whose inference context contains each checkpoint exactly once.
- The real static-SSH worker runs the disk-space probe and exposes its projection through `sessions.list`.
- Reconnecting the same paired node and redispatching the session produces one successful recovery turn whose inference context contains each checkpoint exactly once.
- The real paired-node worker runs the disk-space probe and exposes its projection through `sessions.list`.
- Reclaiming the worker causes a later `sessions.list` read to return the reclaimed placement without a stale disk-space projection.
docsRefs:
- docs/gateway/cloud-workers.md
@@ -30,7 +30,7 @@ scenario:
execution:
kind: script
path: test/e2e/qa-lab/runtime/cloud-worker-midturn-loss-proof.ts
summary: Dispatches a managed-worktree qa-channel session to a proof-owned static-SSH worker, kills its process tree during a streamed message, and verifies the durable cutoff, visible failure, and redispatch context.
summary: Dispatches a managed-worktree qa-channel session to a proof-owned paired-node worker, kills its process and disconnects its node during a streamed message, and verifies the durable cutoff, visible failure, and redispatch context.
timeoutMs: 900000
args:
- --artifact-base