[
https://issues.apache.org/jira/browse/CASSANDRA-12283?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15414157#comment-15414157
]
Benjamin Lerer commented on CASSANDRA-12283:
--------------------------------------------
In my opinion the condition should be either: {{current millis - start millis <
timeout millis}} or {{current millis < start millis + timeout millis}} but the
current one does not make any sense to me (it was already changed by [~blambov]
in trunk).
The test seems to timeout randomly. I could not reproduce it on my Windows
machine so I was looking at what could introduce some random delays. According
to [this|http://www.javamex.com/tutorials/threads/yield.shtml]
{{Thread.yield()}} seems to be able to result into some random delay. By
consequence I though that it might be good to remove it anyway.
> CommitLogSegmentManagerTest.testCompressedCommitLogBackpressure is flaky
> ------------------------------------------------------------------------
>
> Key: CASSANDRA-12283
> URL: https://issues.apache.org/jira/browse/CASSANDRA-12283
> Project: Cassandra
> Issue Type: Test
> Reporter: Joshua McKenzie
> Assignee: Benjamin Lerer
> Priority: Minor
> Labels: unittest
>
> Failed 3 of the last 38 runs.
> [Failure|http://cassci.datastax.com/job/cassandra-3.9_testall/lastCompletedBuild/testReport/org.apache.cassandra.db.commitlog/CommitLogSegmentManagerTest/testCompressedCommitLogBackpressure/]
> Details:
> Error Message
> Timeout occurred. Please note the time in the report does not reflect the
> time until the timeout.
> Stacktrace
> junit.framework.AssertionFailedError: Timeout occurred. Please note the time
> in the report does not reflect the time until the timeout.
> at java.lang.Thread.run(Thread.java:745)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
