miki3421 opened a new pull request, #162: URL: https://github.com/apache/openserverless-operator/pull/162
Resume-time S3 IAM reconciliation can run before the private SeaweedFS Filer is reachable, even when Kubernetes reports the pod Ready. Share the startup readiness wait with resume, retry connection/read timeouts and lost startup tunnels, and add HTTP startup/readiness probes for the Filer. Existing StatefulSets receive only missing probes through a strategic merge patch, followed by a rollout wait before IAM operations and ingress publication. Existing custom probes, images and storage settings are preserved. Persisted IAM verification still fails on policy errors. Fixes apache/openserverless#286. ## Validation - 166 Python unit tests passed, including 12 dedicated readiness tests. - Isolated 0.9.0 laboratory with SeaweedFS 3.97: resume/restart checks, 12 synthetic S3 checks across Alpha and Beta, and both frontends passed. Synthetic objects were removed; all 22 PVC UIDs were preserved. Repeated reconciliation did not restart the unchanged SeaweedFS pod. - WSL2 with K3s 1.37.2 and SeaweedFS 3.97: the final candidate passed resume and a coordinated SeaweedFS/operator restart without a handler failure. All 18 persistent pods were Ready afterwards; all 11 PVC UIDs were preserved. The running module was verified against the tested source hash. This was focused readiness/IAM validation, not full application acceptance. Installing missing probes causes a rolling restart and a temporary storage interruption. An OnDelete StatefulSet requiring probes needs an operator-managed rollout; this change does not force pod deletion. Validation was single agent; independent review is not claimed. -- 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]
