[
https://issues.apache.org/jira/browse/CASSANDRA-17298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17880037#comment-17880037
]
Stefan Miklosovic commented on CASSANDRA-17298:
-----------------------------------------------
I run this branch (1) in the multiplexer (2). It is basically your stuff + mine
you just merged and all squashed into 1 commit. Multiplexer works in such a
way that it will run all tests which we modified or added 500 times (by
default) to rule out any flakes.
You see that some jobs passed and some not. I have 25 containers and 500 tests,
that means that each container runs all tests 20 times. These failed (3).
{code}
junit.framework.AssertionFailedError: Expected heap usage close to 85.836MiB,
got 78.878MiB, 6.957MiB difference. Delta per partition: 112.24 bytes
at
org.apache.cassandra.db.memtable.MemtableSizeTestBase.testSize(MemtableSizeTestBase.java:178)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
{code}
It is always org.apache.cassandra.db.memtable.MemtableSizeOffheapBuffersTest.
(1) https://github.com/instaclustr/cassandra/commits/CASSANDRA-17298-4.0/
(2)
https://app.circleci.com/pipelines/github/instaclustr/cassandra/4675/workflows/a548e71f-605c-4acf-bb78-f42d59a6e48d/jobs/269136/steps
(3)
https://app.circleci.com/pipelines/github/instaclustr/cassandra/4675/workflows/a548e71f-605c-4acf-bb78-f42d59a6e48d/jobs/269136/tests
> Test Failure: org.apache.cassandra.cql3.MemtableSizeTest
> --------------------------------------------------------
>
> Key: CASSANDRA-17298
> URL: https://issues.apache.org/jira/browse/CASSANDRA-17298
> Project: Cassandra
> Issue Type: Bug
> Components: Test/unit
> Reporter: Josh McKenzie
> Assignee: Dmitry Konstantinov
> Priority: Normal
> Fix For: 4.0.x, 4.1.x
>
> Attachments: analyzed_objects.svg, structure_example.svg
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> [https://ci-cassandra.apache.org/job/Cassandra-4.0/313/testReport/org.apache.cassandra.cql3/MemtableSizeTest/testTruncationReleasesLogSpace_2/]
> Failed 4 times in the last 30 runs. Flakiness: 27%, Stability: 86%
> Error Message
> Expected heap usage close to 49.930MiB, got 41.542MiB.
> {code}
> Stacktrace
> junit.framework.AssertionFailedError: Expected heap usage close to 49.930MiB,
> got 41.542MiB.
> at
> org.apache.cassandra.cql3.MemtableSizeTest.testSize(MemtableSizeTest.java:130)
> at org.apache.cassandra.Util.runCatchingAssertionError(Util.java:644)
> at org.apache.cassandra.Util.flakyTest(Util.java:669)
> at
> org.apache.cassandra.cql3.MemtableSizeTest.testTruncationReleasesLogSpace(MemtableSizeTest.java:61)
> at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
> at
> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> {code}
> *UPDATE:* It was discovered that unit tests were running with
> memtable_allocation_type: offheap_objects when we ship C* with heap_buffers.
> So we changed that in CASSANDRA-19326, now we test with
> memtable_allocation_type: heap_buffers. As a result, this test now fails all
> the time on 4.0 and 4.1.
--
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
