[
https://issues.apache.org/jira/browse/SPARK-57204?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated SPARK-57204:
-----------------------------------
Labels: pull-request-available (was: )
> capture Spark driver stderr to Kubernetes termination log on failure
> --------------------------------------------------------------------
>
> Key: SPARK-57204
> URL: https://issues.apache.org/jira/browse/SPARK-57204
> Project: Spark
> Issue Type: Improvement
> Components: Kubernetes
> Affects Versions: 4.2.0
> Reporter: Fei Wang
> Priority: Trivial
> Labels: pull-request-available
>
> * On non-zero exit, write the last 4KB of the Spark driver's stderr to
> {{/dev/termination-log}} so that {{kubectl describe pod}} surfaces the error
> in the pod's *Last State > Message* field
> * Driver stderr is streamed via a named pipe + awk ring buffer — fully
> visible in {{kubectl logs}} with no unbounded {{/tmp}} disk growth
> * SIGTERM is forwarded to tini so Kubernetes graceful shutdown still reaches
> the Spark driver
> * Only the {{driver}} entrypoint mode is affected; {{executor}} and
> pass-through modes are unchanged
> * The {{TERMINATION_LOG}} env var can override the path (e.g. for local
> testing)
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
