Bug v2026.2.12] Session .lock file persists and blocks agent - Fixed in 2026.2.19?
Hi OpenClaw community,
We are running OpenClaw 2026.2.12 (Docker, ghcr.io/hostinger/hvps-openclaw:latest, build 2026-02-14) on a Hostinger VPS and experiencing a recurring blocking issue.
Setup:
Docker container, 32GB RAM, CPU load < 0.5 (resources not the cause)
Agent connected via Telegram
The recurring error:
Agent failed before reply: All models failed (2):
google/gemini-3-flash-preview: session file locked (timeout 10000ms):
pid=28 /data/.openclaw/agents/main/sessions/4ac50e50-...jsonl.lock (timeout)
google/gemini-3-pro-preview: session file locked (timeout 10000ms):
pid=28 /data/.openclaw/agents/main/sessions/4ac50e50-...jsonl.lock (timeout)
What happens:
Agent starts typing on Telegram then stops mid-response.
Next message triggers the error above.
Only fix: manually delete the .lock file inside the container.
Happens multiple times per hour after any failed tool call (SSH, git, long tasks) or container restart.
Our questions:
What is the exact purpose of the .jsonl.lock file? Is it a write-lock for the session log? Why doesn't it auto-cleanup when the process ends?
Is this bug fixed in version 2026.2.19? (we're on 2026.2.12, latest stable on GitHub is 2026.2.19)
Is there a configuration option to set a timeout/auto-cleanup for stale lock files?
Current workaround (not ideal):
bash
docker exec openclaw-container rm -f /data/.openclaw/agents/main/sessions/*.lock
Thanks for any insight!
We are running OpenClaw 2026.2.12 (Docker, ghcr.io/hostinger/hvps-openclaw:latest, build 2026-02-14) on a Hostinger VPS and experiencing a recurring blocking issue.
Setup:Docker container, 32GB RAM, CPU load < 0.5 (resources not the cause)
Agent connected via Telegram
The recurring error:Agent failed before reply: All models failed (2):
google/gemini-3-flash-preview: session file locked (timeout 10000ms):
pid=28 /data/.openclaw/agents/main/sessions/4ac50e50-...jsonl.lock (timeout)
google/gemini-3-pro-preview: session file locked (timeout 10000ms):
pid=28 /data/.openclaw/agents/main/sessions/4ac50e50-...jsonl.lock (timeout)
What happens:Agent starts typing on Telegram then stops mid-response.
Next message triggers the error above.
Only fix: manually delete the .lock file inside the container.
Happens multiple times per hour after any failed tool call (SSH, git, long tasks) or container restart.
Our questions:What is the exact purpose of the .jsonl.lock file? Is it a write-lock for the session log? Why doesn't it auto-cleanup when the process ends?
Is this bug fixed in version 2026.2.19? (we're on 2026.2.12, latest stable on GitHub is 2026.2.19)
Is there a configuration option to set a timeout/auto-cleanup for stale lock files?
Current workaround (not ideal):
bash
docker exec openclaw-container rm -f /data/.openclaw/agents/main/sessions/*.lock
Thanks for any insight!
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
134,426Members
Resources
Similar Threads
Was this page helpful?
