smengcl opened a new pull request, #11664: URL: https://github.com/apache/ozone/pull/11664
Generated-by: Claude Code (Opus 5.5) ## What changes were proposed in this pull request? `acceptance (secure)` intermittently fails in `ozonesecure/public-key-cert-recovery-test.sh` with `Timed out waiting on recon 9888 to become available`. Recon shuts down cleanly on `docker-compose stop recon`, but never starts again: its container log has no second `STARTUP_MSG` and the `recon` hostname stays unresolvable, while `docker-compose start recon` still exits 0. `docker compose start` lists the project containers once and skips those reported as running. The Docker daemon wakes up the `stop` caller (`SetStopped`) before it updates the container list view (`CheckpointTo(containersReplica)`), so a `start` issued right after `stop` can still see the container as running and do nothing. This PR replaces `stop` + `start` with `docker-compose restart` for both Recon and datanodes. `restart` calls `ContainerRestart` for each container unconditionally, so there is no stale state check between stop and start. ## What is the link to the Apache JIRA https://issues.apache.org/jira/browse/HDDS-17236 ## How was this patch tested? The same failure, with the identical log pattern, also occurred on master on 2026-07-13, 2026-07-15 and 2026-08-31. A single green run cannot prove the race is gone. The fix removes the race by construction, since nothing reads the container state between stopping and starting. - [CI on fork](https://github.com/smengcl/hadoop-ozone/actions/runs/39582983383), including `acceptance (secure)`: in progress 🤖 Generated with [Claude Code](https://claude.com/claude-code) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
