Symptom
After upgrading to 2026.10.5, the Gateway exits before HTTP readiness with Gateway failed to start: database is locked.
Evidence what you'll see
- The shared-database transaction failed at
BEGIN, with SQLite primary error 5 and zero service calls. - The first start timed out waiting for readiness. The second printed that lock error. systemd started a replacement until the operator stopped the unit.
- The report does not name the competing lock owner.
Why it happens
Startup recovery fails immediately when worker placement recovery hits a transient SQLite write lock. Pull request 178326 retries only SQLITE_BUSY and SQLITE_LOCKED around that awaited recovery, up to five times with capped backoff.
Affected versions
OpenClaw 2026.10.5 (1b58cb9a1b), Raspberry Pi 500, Debian GNU/Linux 13 (trixie), Linux 6.18.50+rpt-rpi-2712, arm64, Node v24.21.0, npm 11.19.0 global under ~/.npm-global, systemd user service. The last version the report started from is 2026.10.4.
Fix steps
0 of 2 done
Your check marks stay in this browser only.
How to check it worked
- A start that hits a brief SQLite lock during placement recovery reaches HTTP readiness, instead of exiting with
Gateway failed to start: database is locked.
If it still fails
- No published release tag contains pull request 178326. Tag
v2026.10.5was compared against merge73dc02db0aand does not contain it. - A lock that lasts through all five retries still fails startup. The pull request says that remaining failure names another process or connection.
Sources and credit
Summarized in our own words from these public reports. Credit goes to their authors.
Provided as is, without warranty. Try changes in a non-production environment and back up first. Security-related steps: verify against the official docs. Terms · Request removal or attribution change
Hit an error we don't cover? Submit it