[
https://issues.apache.org/jira/browse/CASSANDRA-18710?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17786300#comment-17786300
]
Jacek Lewandowski commented on CASSANDRA-18710:
-----------------------------------------------
Now I'm able to reproduce the extra flush consistently. Dropping tables after
test is executed asynchronously, so in some rare cases it may get executed
while we are adding rows to our test table, thus causing extra flush (or even
flushes).
I think that regardless whether it is correct to flush everything on each
schema change or not, I think that changing schema and dropping sstables
asynchronously after each test can cause flakiness. Also I don't see much
justification for doing that async as it take usually less than a second per
test case.
So, summing up, I propose to fix this by calculating moving average as an
expected value and create another ticket to convert that async processing
(which should be applied to other branches as well) - I suppose that we do not
need to remove tables - only KEYSPACE_PER_TEST is expected to be empty
> Test failure:
> org.apache.cassandra.io.DiskSpaceMetricsTest.testFlushSize-.jdk17 (from
> org.apache.cassandra.io.DiskSpaceMetricsTest-.jdk17)
> ------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-18710
> URL: https://issues.apache.org/jira/browse/CASSANDRA-18710
> Project: Cassandra
> Issue Type: Bug
> Components: Test/unit
> Reporter: Ekaterina Dimitrova
> Assignee: Jacek Lewandowski
> Priority: Normal
> Fix For: 4.1.x, 5.0-beta, 5.0.x, 5.x
>
> Attachments: org.apache.cassandra.io.DiskSpaceMetricsTest.txt
>
>
> Seen here:
> [https://ci-cassandra.apache.org/job/Cassandra-trunk/1644/testReport/org.apache.cassandra.io/DiskSpaceMetricsTest/testFlushSize__jdk17/]
> h3.
> {code:java}
> Error Message
> expected:<7200.0> but was:<1367.83970468544>
> Stacktrace
> junit.framework.AssertionFailedError: expected:<7200.0> but
> was:<1367.83970468544> at
> org.apache.cassandra.io.DiskSpaceMetricsTest.testFlushSize(DiskSpaceMetricsTest.java:119)
> at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native
> Method) at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
> at
> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> {code}
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
