fix: memory leaks and hanging issues #10846
Closed
Conversation
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
added 7 commits
January 27, 2026 14:04
- Add server-side error handling that sends rpc.error messages to client - Add client-side handling of rpc.error to reject pending promises - Add rejectAll() method to reject all pending requests on worker crash - Add 60s timeout for RPC requests to prevent infinite hangs - Add MAX_PENDING_REQUESTS limit (1000) as a safety guard This fixes hangs where worker-side exceptions were never propagated back to the client, causing await calls to hang indefinitely.
- Call client.rejectAll() on worker error to fail pending RPCs - Add terminal cleanup on worker crash to prevent leaked handles - Add SIGINT/SIGTERM handlers to cleanup before exit
- Add error logging and consecutive error tracking for event stream - Add exponential backoff on repeated failures - Cleanup GlobalBus listener on worker shutdown
- Use single stdin listener instead of creating new ones per message - Properly cleanup listener on process exit
- Remove stdout/stderr listeners when subprocess closes - Prevents listener accumulation on repeated MCP restarts
- Track and remove listeners when mDNS service stops - Prevents listener leaks on repeated start/stop cycles
- Wrap SDK calls with 60s timeout to prevent indefinite hangs - Applies to agent.ts, session.ts, and run.ts - Uses withTimeout utility for consistent timeout handling
Contributor
|
Thanks for your contribution! This PR doesn't have a linked issue. All PRs must reference an existing issue. Please:
See CONTRIBUTING.md for details. |
Contributor
|
The following comment was made by an LLM, it may be inaccurate: Found a potential related PR: PR #10392: fix(opencode): memory leaks in session handling Why it might be related:
Please review PR #10392 to determine if it overlaps with the current PR's session-related memory leak fixes. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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
This PR fixes several memory leaks and hanging issues in OpenCode:
RPC Error Propagation (Hang Prevention)
rpc.errormessagesrejectAll()method to reject all pending requests on worker crashrejectAll()to fail all pending RPCsMemory Leak Fixes
Event Stream Resilience
SDK Timeout Protection
Root Cause
The main hanging issue was caused by RPC having NO error propagation - when worker-side code threw an exception, the error was never sent back to the client, causing
awaitcalls to hang indefinitely.Testing