-
-
Notifications
You must be signed in to change notification settings - Fork 80.7k
[Bug]: Session stuck in permanent busy state after mid-turn gateway restart — no auto-recovery, only /new or /reset clears it #92519
Copy link
Copy link
Closed
Labels
P1High-priority user-facing bug, regression, or broken workflow.High-priority user-facing bug, regression, or broken workflow.clawsweeper:needs-live-reproClawSweeper needs live local, crabbox, or manual validation to confirm this issue.ClawSweeper needs live local, crabbox, or manual validation to confirm this issue.clawsweeper:needs-maintainer-reviewClawSweeper marked this issue as needing maintainer review before automation.ClawSweeper marked this issue as needing maintainer review before automation.clawsweeper:no-new-fix-prClawSweeper does not recommend queueing a new automated fix PR for this issue.ClawSweeper does not recommend queueing a new automated fix PR for this issue.impact:crash-loopCrash, hang, restart loop, or process-level availability failure.Crash, hang, restart loop, or process-level availability failure.impact:message-lossChannel message delivery can be lost, duplicated, or misrouted.Channel message delivery can be lost, duplicated, or misrouted.impact:session-stateSession, memory, transcript, context, or agent state can drift or corrupt.Session, memory, transcript, context, or agent state can drift or corrupt.issue-rating: 🐚 platinum hermitGood issue quality with a plausible reproduction path needing some confirmation.Good issue quality with a plausible reproduction path needing some confirmation.staleMarked as stale due to inactivityMarked as stale due to inactivity
Description
Metadata
Metadata
Assignees
Labels
P1High-priority user-facing bug, regression, or broken workflow.High-priority user-facing bug, regression, or broken workflow.clawsweeper:needs-live-reproClawSweeper needs live local, crabbox, or manual validation to confirm this issue.ClawSweeper needs live local, crabbox, or manual validation to confirm this issue.clawsweeper:needs-maintainer-reviewClawSweeper marked this issue as needing maintainer review before automation.ClawSweeper marked this issue as needing maintainer review before automation.clawsweeper:no-new-fix-prClawSweeper does not recommend queueing a new automated fix PR for this issue.ClawSweeper does not recommend queueing a new automated fix PR for this issue.impact:crash-loopCrash, hang, restart loop, or process-level availability failure.Crash, hang, restart loop, or process-level availability failure.impact:message-lossChannel message delivery can be lost, duplicated, or misrouted.Channel message delivery can be lost, duplicated, or misrouted.impact:session-stateSession, memory, transcript, context, or agent state can drift or corrupt.Session, memory, transcript, context, or agent state can drift or corrupt.issue-rating: 🐚 platinum hermitGood issue quality with a plausible reproduction path needing some confirmation.Good issue quality with a plausible reproduction path needing some confirmation.staleMarked as stale due to inactivityMarked as stale due to inactivity
Type
Fields
Priority
None yet
Bug type
Crash (process/app exits or hangs)
Summary
When a gateway restart (launchctl kickstart, crash, manual /restart, or config-triggered restart) interrupts a mid-turn LLM completion, the session enters "busy" state and never recovers. No new messages are processed. Only
/newor/resetclears it.Repro Steps
Expected Behavior
On gateway startup, sessions with no active LLM completion should have their busy flag cleared. The gateway should detect orphaned busy states (completion PID/promise no longer exists) and auto-recover within a reasonable timeout (e.g., 30 seconds).
Actual Behavior
Session stays busy indefinitely. The
assemble()function continues to fire (301 calls observed over a 12-hour span) but dispatches zero LLM completions. The session is effectively dead until the user manually runs/newor/reset.Evidence
assemble()calls with 0 LLM dispatches across a 12-hour windowRelated Issues
Environment