DanielLeens commented on PR #12710: URL: https://github.com/apache/seatunnel/pull/12710#issuecomment-6210080494
Thanks @SEZ9 for the follow-up. I checked both questions against the current head `bd58ded09c` so they can be closed with concrete pointers (most of this is also in @ricky2129's 2026-10-14 reply): 1. Shutdown behavior. `terminationGracePeriodSeconds` is wired in `deploy/kubernetes/seatunnel/templates/deployment-seatunnel-master.yaml:43-45` and `deployment-seatunnel-worker.yaml:43-45`, driven by `master/worker.terminationGracePeriodSeconds` in `values.yaml:48` and `values.yaml:114` (default 30, the Kubernetes default). No `hazelcast.shutdownhook.policy` is set anywhere in the chart (the only mentions are the comments in `values.yaml` and the docs), so the stop is Hazelcast's default `TERMINATE`: the member leaves within seconds and its tasks recover from the last checkpoint. That applies identically to master and worker because both Deployments share the same command and the same value. The wording in the description and docs matches this (`hybrid-cluster-mode.md:315` and `separated-cluster-mode.md:480`: a fast, clean exit from the cluster, not a graceful stop of running jobs). 2. Docs. Both K8s pages (en and zh) have the "Health Check and Shutdown" section leading with the exec-form `command` plus `terminationGracePeriodSeconds`, and `preStop` + `stop-seatunnel-cluster.sh` is now an optional tip for setups that must keep `/bin/sh -c`; `operations.md` "Rolling Updates" and `tuning-guide.md` link to it. The PR-owned files are unchanged since your last pass, so nothing new to re-read. The one non-blocking item I suggested earlier is still open: @ricky2129, a small tracking issue for the `SeaTunnelServer.shutdown` terminate-flag follow-up, linked from the description, would make sure it is not lost. CI on `bd58ded09c` is still red on the fork `Build`; my read is the unrelated `SplitClusterFaultToleranceIT` engine flake noted in my last review, so re-running only the failed jobs is the next step. -- 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]
