fix: prevent message loss when container times out after pipe#694
Closed
neocode24 wants to merge 1 commit intonanocoai:mainfrom
Closed
fix: prevent message loss when container times out after pipe#694neocode24 wants to merge 1 commit intonanocoai:mainfrom
neocode24 wants to merge 1 commit intonanocoai:mainfrom
Conversation
When follow-up messages are piped to an active container via IPC, the message loop was advancing lastAgentTimestamp immediately. If the container then timed out before processing the piped messages, those messages were permanently skipped — the cursor had already moved past them. Three changes fix this: 1. Separate lastPipedTimestamp from lastAgentTimestamp: piping only advances a temporary cursor to prevent re-piping the same messages. Only processGroupMessages advances the authoritative cursor. On container exit, lastPipedTimestamp is cleared so the messages get re-processed. 2. Set pendingMessages=true on pipe: ensures the drain loop re-checks for unprocessed messages after the container exits. 3. Add onPipeCallback to reset the idle timer: when messages are piped, the container needs time to process them. Without resetting the idle timer, the container could time out immediately after receiving input. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Collaborator
|
Message loss on container timeout is a real issue — glad someone is looking at this. @gavrielc would be good to review the approach here. |
Andy-NanoClaw-AI
added
PR: Fix
klapom
added a commit
to klapom/nanoclaw
that referenced
this pull request
Mar 5, 2026
…res) Bug fixes applied: - nanocoai#636: task-scheduler recalculates next_run before enqueue - nanocoai#655: LIMIT 200 on message queries to prevent OOM - nanocoai#670: rateLimitResetAt field in ContainerOutput interface - nanocoai#694: ANTHROPIC_MODEL passthrough to container env - nanocoai#700: session rotation at 5MB JSONL threshold - nanocoai#701: session retry on corrupted resume (clear + retry) - nanocoai#708: update_task MCP tool in ipc-mcp-stdio - nanocoai#719: outputChain .catch() to prevent group hang - nanocoai#729: fix send_message description (remove incorrect scheduled-task note) - nanocoai#735: datePrefix() injects current date/time into all agent prompts - nanocoai#738: ANTHROPIC_MODEL from .env passed to agent container - nanocoai#746: systemd OnFailure restart prevention logic (container hardening) - nanocoai#751: DM-with-bot JID normalization - nanocoai#754: setOnPipeCallback to reset idle timer on piped messages - nanocoai#756: cursorBeforePipe rollback on container crash Features added: - nanocoai#723: streaming infrastructure (STREAM_TEXT markers, onStreamDelta) - nanocoai#742: container hardening (entrypoint.sh privilege drop, env sanitize) - nanocoai#680: add-cli skill (CLI send binary) - nanocoai#727: add-memory skill extracted to .claude/skills/add-memory/ - nanocoai#744: add-s3-storage skill extracted to .claude/skills/add-s3-storage/ Test fixes: - Mock fs.promises in container-runner.test.ts to prevent real I/O - Add ANTHROPIC_MODEL to config mock - Fix cpSync expectation: { recursive: true, force: true } - Fix isActive() to use state.active instead of state.process - Fix container-runtime error message: Docker → Container runtime Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Andy-NanoClaw-AI
added
Status: Blocked
Contributor
Author
|
Closing for now — upstream structure has changed significantly since this PR. The fix is maintained in our fork. Will revisit if the issue resurfaces upstream. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
lastPipedTimestampfromlastAgentTimestamp: piping follow-up messages to an active container no longer advances the authoritative cursor. OnlyprocessGroupMessagesdoes. If the container times out before processing piped messages, they are re-queued on the next cycle.pendingMessages = trueon pipe: ensures the drain loop re-checks for unprocessed messages after container exit.onPipeCallbackto reset idle timer: piped messages now reset the container idle timer, preventing premature timeout immediately after receiving input.Problem
When a user sends follow-up messages while the agent is processing, the message loop pipes them to the active container via IPC. Previously,
lastAgentTimestampwas advanced immediately on pipe. If the container then timed out (e.g., due to network instability), those piped messages were permanently lost — the cursor had already moved past them.Test plan
npx tsc --noEmitpassesnpx vitest run— no regressions🤖 Generated with Claude Code