You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Ekaterina Dimitrova (Jira)" <ji...@apache.org> on 2022/09/27 12:54:00 UTC
[jira] [Comment Edited] (CASSANDRA-17884) Test failure: o.a.c.repair.RepairJobTest.testNoTreesRetainedAfterDifference
[ https://issues.apache.org/jira/browse/CASSANDRA-17884?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17610019#comment-17610019 ]
Ekaterina Dimitrova edited comment on CASSANDRA-17884 at 9/27/22 12:53 PM:
---------------------------------------------------------------------------
-I actually tried an unreleased in maven central jamm jar with the mid resources and this time I didn't see the issue. I will try one more run to confirm that the issue was solved and I am not missing anything.-
This is the [branch|https://github.com/apache/cassandra/compare/trunk...ekaterinadimitrova2:cassandra:17884-4.0-hack] where I added a jar including [this commit|https://github.com/jbellis/jamm/commit/33d78c1035211e5f9e69ad9247f3a0f6ef75c923] (this is not the latest work in the repo though) which was supposed to solve that issues with JDK11 according to the author commit message in the jamm repo.
It seems it didn't and we are missing something else to filter. - [https://app.circleci.com/pipelines/github/ekaterinadimitrova2/cassandra/1955/workflows/20b3aa2c-fefd-42df-8ad0-b787add2fbb1/jobs/15494/steps]
I added annoying print in the jamm code and verified it is printed in the CircleCI logs to ensure nothing went wrong and we use the new jar I plugged in.
I also think we need to get back to the repo [TreePrint|https://github.com/ekaterinadimitrova2/jamm/blob/33d78c1035211e5f9e69ad9247f3a0f6ef75c923/src/org/github/jamm/TreePrinter.java] class as I think it is really useful for debugging. Others think in the same way. If we feel it lacks some testing, better to add that than to remove it completely.
was (Author: e.dimitrova):
I actually tried an unreleased in maven central jamm jar with the mid resources and this time I didn't see the issue. I will try one more run to confirm that the issue was solved and I am not missing anything.
This is the [branch|https://github.com/apache/cassandra/compare/trunk...ekaterinadimitrova2:cassandra:17884-4.0-hack] where I added a jar including [this commit|https://github.com/jbellis/jamm/commit/33d78c1035211e5f9e69ad9247f3a0f6ef75c923] (this is not the latest work in the repo though) which was supposed to solve that issues with JDK11 according to the author commit message in the jamm repo.
It seems it didn't and we are missing something else to filter. - https://app.circleci.com/pipelines/github/ekaterinadimitrova2/cassandra/1955/workflows/20b3aa2c-fefd-42df-8ad0-b787add2fbb1/jobs/15494/steps
I added annoying print in the jamm code and verified it is printed in the CircleCI logs to ensure nothing went wrong and we use the new jar I plugged in.
I also think we need to get back to the repo [TreePrint|https://github.com/ekaterinadimitrova2/jamm/blob/33d78c1035211e5f9e69ad9247f3a0f6ef75c923/src/org/github/jamm/TreePrinter.java] class as I think it is really useful for debugging. Others think in the same way. If we feel it lacks some testing, better to add that than to remove it completely.
> Test failure: o.a.c.repair.RepairJobTest.testNoTreesRetainedAfterDifference
> ---------------------------------------------------------------------------
>
> Key: CASSANDRA-17884
> URL: https://issues.apache.org/jira/browse/CASSANDRA-17884
> Project: Cassandra
> Issue Type: Bug
> Components: Test/unit
> Reporter: Andres de la Peña
> Priority: Normal
> Fix For: 4.0.x, 4.1.x, 4.x
>
>
> The unit test {{org.apache.cassandra.repair.RepairJobTest#testNoTreesRetainedAfterDifference}} seems to be slightly flaky at least in 4.0. We haven't seen it failing on Butler, but after [a failure on a patch branch|https://app.circleci.com/pipelines/github/jacek-lewandowski/cassandra/281/workflows/e6094c00-b611-4772-9772-205f4c76ecba/jobs/2126/tests], this repeated run shows a 0.28% flakiness:
> * https://app.circleci.com/pipelines/github/adelapena/cassandra/2076/workflows/8adbfe99-afb5-43af-84ad-43df2a2a86e2/jobs/20816/tests
> * https://app.circleci.com/pipelines/github/adelapena/cassandra/2076/workflows/e550d8c2-7c35-4326-bd95-6909978d344b/jobs/20817/tests
> {code}
> java.lang.reflect.InaccessibleObjectException: Unable to make field private jdk.internal.platform.cgroupv1.SubSystem$MemorySubSystem jdk.internal.platform.cgroupv1.Metrics.memory accessible: module java.base does not "opens jdk.internal.platform.cgroupv1" to unnamed module @157853da
> at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:340)
> at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:280)
> at java.base/java.lang.reflect.Field.checkCanSetAccessible(Field.java:176)
> at java.base/java.lang.reflect.Field.setAccessible(Field.java:170)
> at org.github.jamm.MemoryMeter.addFieldChildren(MemoryMeter.java:330)
> at org.github.jamm.MemoryMeter.measureDeep(MemoryMeter.java:269)
> at org.apache.cassandra.utils.ObjectSizes.measureDeep(ObjectSizes.java:216)
> at org.apache.cassandra.repair.RepairJobTest.testNoTreesRetainedAfterDifference(RepairJobTest.java:271)
> 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}
> Flakiness is really low but the failure is clearly reproducible in j11.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org