[
https://issues.apache.org/jira/browse/CASSANDRA-21388?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dmitry Konstantinov updated CASSANDRA-21388:
--------------------------------------------
Fix Version/s: 6.0-alpha2
7.0
(was: 5.0.x)
(was: 6.x)
(was: 7.x)
Source Control Link:
https://github.com/apache/cassandra/commit/1f078c0daf3818c6c9897504af8daa256ef88fde
Resolution: Fixed
Status: Resolved (was: Ready to Commit)
> Reduce microbenchmark smoke tests execution time
> ------------------------------------------------
>
> Key: CASSANDRA-21388
> URL: https://issues.apache.org/jira/browse/CASSANDRA-21388
> Project: Apache Cassandra
> Issue Type: Task
> Components: Test/burn
> Reporter: Dmitry Konstantinov
> Assignee: Dmitry Konstantinov
> Priority: Normal
> Fix For: 6.0-alpha2, 7.0
>
> Attachments: image-2026-05-18-19-37-42-349.png,
> image-2026-05-19-15-38-54-942.png
>
> Time Spent: 40m
> Remaining Estimate: 0h
>
> Currently, when pre-commit profile is executed, microbenchmarks (smoke test
> mode) take too much time and delay an overall job completion.
> Example: [https://pre-ci.cassandra.apache.org/job/cassandra/547/]
> !image-2026-05-18-19-37-42-349.png|width=700!
> We can:
> * set explicitly a duration for an JMH run iteration to 1 sec (/)
> * reduce number of option combinations in tests (all the details are not so
> needed for a regression testing to see if the benchmarks are still alive,
> unfortunately there are no profiles in JMH to control it, so I've not found a
> better way than moving extras to plain comments) (/)
> * fix the tests which do not release resources properly on shutdown and JMH
> has to wait for a timeout to kill a forked JVM (/)
> * support splits for benchmark-test to run tests in parallel in different
> containers (/)
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
