[
https://issues.apache.org/jira/browse/CASSANDRA-18741?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17752815#comment-17752815
]
Ekaterina Dimitrova commented on CASSANDRA-18741:
-------------------------------------------------
The maybeCreateHeapDump was added only in 5.0 as part of CASSANDRA-17795 - Add
ability to generate a one-time heap dump output to a file on Unhandled
Exceptions.
While running RandomSchemaTest, a breakpoint points to the
JVMStabilityInspector and the Commit Log Service.
By default dump_heap_on_uncaught_exception is false.
Though in the logs of RandomSchemaTest it is indeed
dump_heap_on_uncaught_exception=true and I can see that CASSANDRA-17795 set the
flag to true for unit tests.
> Some unit tests create a heap dump each time
> --------------------------------------------
>
> Key: CASSANDRA-18741
> URL: https://issues.apache.org/jira/browse/CASSANDRA-18741
> Project: Cassandra
> Issue Type: Bug
> Components: Test/unit
> Reporter: Brandon Williams
> Priority: Normal
> Fix For: 5.x
>
>
> I noticed this by running out of space while working on CASSANDRA-18706.
> Something must be calling HeaupUtils.maybeCreateHeapDump because the
> build/test directory is littered with pidxxxx-epochxxxxxxxx.hprof files
> despite the tests not crashing. I checked a couple other random tests,
> RandomSchemaTest also generates a dump but KeywordSplitTest does not.
> This behavior does not occur on 4.1.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
