[jira] [Commented] (CASSANDRA-21431) Jenkins k8s deployed needs the agent garbageCollection to avoid zombie pods

["Michael Semb Wever (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-21431?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18086576#comment-18086576 ] 

Michael Semb Wever commented on CASSANDRA-21431:
------------------------------------------------

https://github.com/apache/cassandra/compare/cassandra-5.0...thelastpickle:cassandra:mck/21431/5.0

> Jenkins k8s deployed needs the agent garbageCollection to avoid zombie pods
> ---------------------------------------------------------------------------
>
>                 Key: CASSANDRA-21431
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-21431
>             Project: Apache Cassandra
>          Issue Type: Improvement
>          Components: CI
>            Reporter: Michael Semb Wever
>            Assignee: Michael Semb Wever
>            Priority: Normal
>
> k8s nodes (created for jenkins agents) can be left unaccounted for (zombied) and left running.  these can accumulate to significant costs for the k8s owner.
> the jenkins helm chart has a plugin garbageCollection that can be enabled, and the pod templates should be setting an appropriate activeDeadlineSeconds value.
> this complements the `.build/run-ci nly-node-cleaner` functionality, reducing the need for it (which was also flawed as nodes with the agent dind/jlnp pods still running were considered still legitimate).



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org