[
https://issues.apache.org/jira/browse/CASSANDRA-21500?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18095422#comment-18095422
]
Benedict Elliott Smith edited comment on CASSANDRA-21500 at 7/10/26 9:06 PM:
-----------------------------------------------------------------------------
I think I already have a fix for this pending (in a larger patch I haven't
completed yet)
was (Author: benedict):
I think I already have a fix for this pending
> Flaky test: AccordCQLTestBase#testCASSimulatorLite
> --------------------------------------------------
>
> Key: CASSANDRA-21500
> URL: https://issues.apache.org/jira/browse/CASSANDRA-21500
> Project: Apache Cassandra
> Issue Type: Bug
> Components: Accord
> Reporter: C. Scott Andreas
> Assignee: Benedict Elliott Smith
> Priority: Normal
>
> {{AccordCQLTestBase.testCASSimulatorLite}} fires N concurrent CAS ({{{}UPDATE
> ... IF EXISTS{}}}) at {{CL=ALL}} on a single key through Accord in a 2-node
> cluster, then asserts the two appended columns stay in lockstep. The ten-way
> contention produced in this test fails pretty often in CI with a
> WriteTimeoutException.
> Reducing the concurrency of this CAS test from 10 to 4 prevents this test
> from flaking in CI.
> Additional details and flakiness stats at varied levels of concurrency
> (model-assisted): [https://github.com/cscotta/cassandra/pull/9]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
