[
https://issues.apache.org/jira/browse/CASSANDRA-10115?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Joshua McKenzie updated CASSANDRA-10115:
----------------------------------------
Attachment: 10115_v1.txt
Attaching a patch that doubles the rows and cols on the test to try and ensure
deeper levels as I'm getting non-deterministic counts of the levels from local
test runs, although none are off enough to have an empty L2 and thus cause test
failure.
Assuming the invariant that L0 count being == 1 sstable means deeper levels are
completed (waitForLeveling relies on this), the only thing that really makes
sense here is that we're skirting the edge of having enough data for both an L1
and L2.
Added a bit more debug logging surrounding a failure and snapshot the counts so
conditional check and printed results are from (close to) an atomic point in
time.
> Windows utest 2.2: LeveledCompactionStrategyTest.testGrouperLevels flaky
> ------------------------------------------------------------------------
>
> Key: CASSANDRA-10115
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10115
> Project: Cassandra
> Issue Type: Bug
> Reporter: Joshua McKenzie
> Assignee: Joshua McKenzie
> Labels: Windows
> Fix For: 2.2.x
>
> Attachments: 10115_v1.txt
>
>
> {noformat}
> junit.framework.AssertionFailedError
> at
> org.apache.cassandra.db.compaction.LeveledCompactionStrategyTest.testGrouperLevels(LeveledCompactionStrategyTest.java:131)
> {noformat}
> [Test is flaky on
> Windows|http://cassci.datastax.com/view/cassandra-2.2/job/cassandra-2.2_utest_win32/lastCompletedBuild/testReport/org.apache.cassandra.db.compaction/LeveledCompactionStrategyTest/testGrouperLevels/history/]
> [Test is consistent on
> linux|http://cassci.datastax.com/view/cassandra-2.2/job/cassandra-2.2_utest/lastCompletedBuild/testReport/org.apache.cassandra.db.compaction/LeveledCompactionStrategyTest/testGrouperLevels/history/]
> Doesn't repro locally.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
