[
https://issues.apache.org/jira/browse/FLINK-18828?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17186532#comment-17186532
]
Ufuk Celebi edited comment on FLINK-18828 at 8/28/20, 1:19 PM:
---------------------------------------------------------------
[~fly_in_gis] Thanks for the pointers. The Flink job would only transition to
FAILED when the Flink-level restart strategy has been exhausted. In your
example for fixed-delay with 3 attempts, the first three restarts would _not_
result in the container to exit, but only on the 4th failure would the job
transition to FAILED and the container exit.
I think the bigger problem with my proposal to set the policy to Never is that
it would not restart in other failure scenarios (e.g. OOM killed). So overall,
I don't think it's a viable option.
So overall, I don't see a good way around this problem without your proposed
change.
---
Maybe as a follow-up we want to resurrect
https://issues.apache.org/jira/browse/FLINK-10948 ? That way, users would at
least be able to determine the final Flink job status.
was (Author: uce):
[~fly_in_gis] Thanks for the pointers. The Flink job would only transition to
FAILED when the Flink-level restart strategy has been exhausted. In your
example for fixed-delay with 3 attempts, the first three restarts would _not_
result in the container to exit, but only on the 4th failure would the job
transition to FAILED and the container exit.
I think the bigger problem with my proposal to set the policy to Never is that
it would not restart in other failure scenarios (e.g. OOM killed). So overall,
I don't think it's a viable option.
So overall, I don't see a good way around this problem without your proposed
change.
---
Maybe as a follow-up we want to resurrect
https://issues.apache.org/jira/browse/FLINK-10948? That way, users would at
least be able to determine the final Flink job status.
> Terminate jobmanager process with zero exit code to avoid unexpected
> restarting by K8s
> --------------------------------------------------------------------------------------
>
> Key: FLINK-18828
> URL: https://issues.apache.org/jira/browse/FLINK-18828
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Coordination
> Affects Versions: 1.10.1, 1.12.0, 1.11.1
> Reporter: Yang Wang
> Priority: Major
> Fix For: 1.12.0, 1.11.2, 1.10.3
>
>
> Currently, Flink jobmanager process terminates with a non-zero exit code if
> the job reaches the {{ApplicationStatus.FAILED}}. It is not ideal in K8s
> deployment, since non-zero exit code will cause unexpected restarting. Also
> from a framework's perspective, a FAILED job does not mean that Flink has
> failed and, hence, the return code could still be 0.
> > Note:
> This is a special case for standalone K8s deployment. For
> standalone/Yarn/Mesos/native K8s, terminating with non-zero exit code is
> harmless. And a non-zero exit code could help to check the job result quickly.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
