strelok89 commented on PR #3854: URL: https://github.com/apache/celeborn/pull/3854#issuecomment-6532893590
Thanks @zaynt4606. Rebased onto `main` with #3855 in, and the conflicts are resolved: - Per-zone StatefulSets, autoscaling and drain are preserved. `worker.podManagementPolicy` and the worker probes apply to every zone's StatefulSet, and the probes sit alongside the drain `preStop` hook. - One behaviour change from the earlier revision: the worker headless service no longer sets `publishNotReadyAddresses` unconditionally. It sets it when `worker.startupProbe` or `worker.readinessProbe` is set, or when `worker.drain` is enabled as before, so default installs leave the worker service unchanged. The master service still always sets it. - `helm unittest charts/celeborn --file "tests/**/*_test.yaml" --strict` passes (152 tests) and `helm lint` is clean. I've also updated the PR description to match: all probes are opt-in, and both master startup and readiness probes require `Parallel` with more than one master. -- 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]
