[
https://issues.apache.org/jira/browse/SOLR-13370?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16810244#comment-16810244
]
Gus Heck commented on SOLR-13370:
---------------------------------
Well, I had hoped that turning down the number of nodes in the embedded cluster
for the test would make a difference, but I ran this suite in my ide under a
profiler which showed it was allowed to ask for up to 10GB of mem and the run
with only 1 node wound up asking for more memory than the run with 4 (3.9G vs
2.4G) so this isn't an instant obvious win, but I have no idea what the memory
availability on the test machines looks like and so these runs are likely to be
highly unrealistic to begin with. The logs don't seem too report memory
settings. I also didn't see anything immediately obvious in the build.xml or
common-build.xml though I certainly could have missed it.
Has anyone already dug up the answer to the question "how much memory is
available for my test" on the build machines?
> Investigate Memory Footprint of CategorRoutedAliasUpdateProcessorTest
> ---------------------------------------------------------------------
>
> Key: SOLR-13370
> URL: https://issues.apache.org/jira/browse/SOLR-13370
> Project: Solr
> Issue Type: Sub-task
> Security Level: Public(Default Security Level. Issues are Public)
> Components: Tests
> Affects Versions: master (9.0)
> Reporter: Gus Heck
> Priority: Major
>
> The test is failing too frequently, usually with OOM on the build servers.
> This sub task will track changes/investigation/discussion to improve that
> situation.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
