openclaw - 💡(How to fix) Fix session file locked when gateway times out and falls back to embedded runner [1 participants]

Gateway agent failed; falling back to embedded: Error: gateway timeout after 90000ms session file locked (timeout 10000ms): pid=47321 ...9e051a52.jsonl.lock FallbackSummaryError: All models failed (3): qiniu/deepseek-v3: session file locked ... (timeout) | qiniu/kimi-k2: Request was aborted (timeout) | qiniu/glm-4.5: Request was aborted (timeout)

TL;DR

Generate a new sessionId for the embedded runner when falling back from the gateway to avoid session lock conflicts.

Guidance

• Identify the sessionId generation logic and modify it to create a new session ID when the gateway times out and the embedded runner takes over.
• Verify that the gateway releases the session lock after timing out to ensure the embedded runner can acquire the lock with the new sessionId.
• Consider implementing a retry mechanism with a backoff strategy for the embedded runner to handle cases where the gateway still holds the lock.
• Review the agentCommand() function to ensure it can handle a new sessionId and update the session file accordingly.

Example

} catch (err) {
  // Generate a new sessionId for the embedded runner
  const newSessionId = generateNewSessionId();
  return await agentCommand({ ...localOpts, sessionId: newSessionId }, runtime, deps);
}

Notes

This solution assumes that generating a new sessionId is feasible and does not interfere with the existing session management logic. Additionally, the generateNewSessionId() function is not defined in the provided code snippet and should be implemented according to the existing session ID generation logic.

Recommendation

Apply workaround: Generate a new sessionId for the embedded runner to avoid session lock conflicts, as this is a more straightforward solution than modifying the gateway's lock release mechanism.