[
https://issues.apache.org/jira/browse/SPARK-57656?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hyukjin Kwon resolved SPARK-57656.
----------------------------------
Fix Version/s: 4.3.0
Resolution: Fixed
Issue resolved by pull request 56728
[https://github.com/apache/spark/pull/56728]
> Stabilize flaky SparkSessionE2ESuite interrupt-all tests
> --------------------------------------------------------
>
> Key: SPARK-57656
> URL: https://issues.apache.org/jira/browse/SPARK-57656
> Project: Spark
> Issue Type: Test
> Components: Connect
> Affects Versions: 5.0.0
> Reporter: Hyukjin Kwon
> Assignee: Hyukjin Kwon
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.3.0
>
>
> {{SparkSessionE2ESuite}} intermittently fails in CI (master push Build and
> Maven JDK 21/25). The first interrupt-all test occasionally fails with a
> {{RemoteClassLoaderError}} (for the map closure's TypeTag artifact class)
> instead of {{OPERATION_CANCELED}}, because an interrupt can land while the
> executor is doing its first-time remote fetch of the closure classes. The
> failed assertion then leaves the background interruptor running, which
> cancels subsequent tests and cascades into ~7 failures across the suite.
> Fix: warm up the map closure once before interrupting so its classes are
> already loaded on the executor, and wrap the test body in try/finally so the
> interruptor is always stopped.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
