[
https://issues.apache.org/jira/browse/HBASE-19956?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stack updated HBASE-19956:
--------------------------
Description:
Appy suggestion from parent issue.
Notion is that during transition measuring timeout on a class-basis instead of
on a test method basis, categories only mess us up. The only important regard
when it comes to timeout is that our timeout cuts in before the surefire one
does; it just does a savage kill w/o regard for who is about. If we kill,
there'll be info on culprit. Thats what we want.
We used to have timeout per test method. It was changed for hbase2 to be class
based. Easier for devs to grok. For now, we purge changing test class timeout
based off category and just paint all tests w/ the one 10-minute max brush.
was:Appy suggestion from parent issue.
> Remove category as a consideration timing out tests; set all test to timeout
> at 10minutes regardless
> ----------------------------------------------------------------------------------------------------
>
> Key: HBASE-19956
> URL: https://issues.apache.org/jira/browse/HBASE-19956
> Project: HBase
> Issue Type: Sub-task
> Reporter: stack
> Priority: Major
>
> Appy suggestion from parent issue.
> Notion is that during transition measuring timeout on a class-basis instead
> of on a test method basis, categories only mess us up. The only important
> regard when it comes to timeout is that our timeout cuts in before the
> surefire one does; it just does a savage kill w/o regard for who is about. If
> we kill, there'll be info on culprit. Thats what we want.
> We used to have timeout per test method. It was changed for hbase2 to be
> class based. Easier for devs to grok. For now, we purge changing test class
> timeout based off category and just paint all tests w/ the one 10-minute max
> brush.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
