[
https://issues.apache.org/jira/browse/SPARK-61388?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hyukjin Kwon resolved SPARK-61388.
----------------------------------
Resolution: Fixed
Issue resolved by pull request 477
[https://github.com/apache/spark-connect-rust/pull/477]
> Interrupt only the execution that an interrupted call waits for, in the
> Python client as PySpark does
> -----------------------------------------------------------------------------------------------------
>
> Key: SPARK-61388
> URL: https://issues.apache.org/jira/browse/SPARK-61388
> Project: Spark
> Issue Type: Bug
> Components: Connect
> Affects Versions: connect-rust-4.3.0
> Reporter: Hyukjin Kwon
> Assignee: Hyukjin Kwon
> Priority: Major
> Labels: pull-request-available
> Fix For: connect-rust-4.3.0
>
>
> A follow-up of SPARK-61259 and SPARK-61353, found in a post-merge review. The
> in-flight registry that a SIGINT interrupts is keyed by thread only, and its
> innermost entry wins, so SIGINT during next() of one toLocalIterator
> interrupts another half-consumed iterator of the thread instead. With such an
> iterator or a streaming listener alive on the thread, SIGINT during
> SparkSession.stop() also interrupts that execution and then waits for the
> release. The watch also latches the first signal it sees when no execution is
> in flight (for example during the plan compression Config RPC), so an
> execution that the same call starts afterwards runs to its end before
> KeyboardInterrupt is raised.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]