SEZ9 commented on PR #12710:
URL: https://github.com/apache/seatunnel/pull/12710#issuecomment-6392870933

   Re-review of head `96069fda` done; since `f7ee365c` the only delta is the 
`dev` merge, so the PR-owned diff is unchanged and my earlier findings still 
apply as follows.
   
   **F1 / F2 — SIGTERM path and shutdown policy (master/worker deployments)**
   The description and the user-facing docs now state clearly that the result 
is a fast member exit via the Hazelcast shutdown hook (`TERMINATE`), not a 
graceful drain of running jobs, and they document the `GRACEFUL` option 
together with the `hazelcast.graceful.shutdown.max.wait` < 
`terminationGracePeriodSeconds` constraint. Given that 
`SeaTunnelServer.shutdown(boolean terminate)` does not act on the terminate 
flag, switching the chart default to `GRACEFUL` would only add 
partition-migration time to every pod stop without making job stop cleaner, so 
keeping `TERMINATE` as the chart default is the right call and preserves 
current behaviour for existing releases. Remaining asks:
   - Add one sentence to the `values.yaml` comment next to 
`terminationGracePeriodSeconds` (master and worker) making that intent 
explicit: default policy is `TERMINATE`, i.e. fast member exit, not graceful 
job stop.
   - Open a small follow-up issue for the graceful-job-stop work 
(`SeaTunnelServer.shutdown` ignoring the terminate flag) and link it from the 
description instead of the current "left for a follow-up" wording. This is 
description-only, so no new push is needed.
   - Please add a one-line note of the shutdown time you observed in the 
`docker stop` A/B so we can confirm the 30s default (same as the Kubernetes 
default) is sufficient on a typical node.
   
   **F3 — Kubernetes graceful-shutdown docs**
   Re-checked the en/zh pages on this head: they lead with the exec-form 
`command` plus `terminationGracePeriodSeconds`, and `preStop` + 
`stop-seatunnel-cluster.sh` is presented as optional for setups that must keep 
`/bin/sh -c`. That resolves my concern; I consider F3 addressed, no further 
change needed there.
   
   Once the `values.yaml` comment and the linked follow-up issue are in, I'm 
happy to approve.
   
   <!-- streview-comment:2017 -->


-- 
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]

Reply via email to