[
https://issues.apache.org/jira/browse/PHOENIX-4110?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16138885#comment-16138885
]
James Taylor edited comment on PHOENIX-4110 at 8/23/17 7:03 PM:
----------------------------------------------------------------
Nice analysis, [~samarthjain]! Alternate idea: how about we just get rid of
PhoenixTestDriver and the modifications we do to HBase Configuration in
BaseTest? Then we can just turn on class-level parallelization (FYI, the prior
issue with endless logging to stdout has been fixed). This has the advantages
of our tests better mimicing how they'll run on a real cluster.
To do this, you'd just need to uncomment these lines in the top level pom.xml
(and change methods to classes, at least as an initial test):
{code}
<parallel>classes</parallel>
<threadCount>20</threadCount>
{code}
and then set numForkedIT to 1 as we shouldn't need multiple JVMs.
We can play around with the best value for threadCount.
We also need to convert all ClientManagedTimeTests due to PHOENIX-4096
was (Author: jamestaylor):
Nice analysis, [~samarthjain]! Alternate idea: how about we just get rid of
PhoenixTestDriver and the modifications we do to HBase Configuration in
BaseTest? Then we can just turn on class-level parallelization (FYI, the prior
issue with endless logging to stdout has been fixed). This has the advantages
of our tests better mimicing how they'll run on a real cluster.
We also need to convert all ClientManagedTimeTests due to PHOENIX-4096
> ParallelRunListener should monitor number of tables and not number of tests
> ---------------------------------------------------------------------------
>
> Key: PHOENIX-4110
> URL: https://issues.apache.org/jira/browse/PHOENIX-4110
> Project: Phoenix
> Issue Type: Bug
> Reporter: Samarth Jain
> Assignee: Samarth Jain
> Attachments: PHOENIX-4110.patch,
> PHOENIX-4110_V2_4.x-HBase-0.98.patch, PHOENIX-4110_v3.patch
>
>
> ParallelRunListener today monitors the number of tests that have been run to
> determine when mini cluster should be shut down. This helps prevent our test
> JVM forks running in OOM. A better heuristic would be to instead check the
> number of tables that were created by tests. This way when a particular test
> class has created lots of tables, we can shut down the mini cluster sooner.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
