[
https://issues.apache.org/jira/browse/HADOOP-7298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13071261#comment-13071261
]
Joep Rottinghuis commented on HADOOP-7298:
------------------------------------------
Todd, in the waitFor method why do you keep a count of finished and then wait
in a loop?
Wouldn't a CountDownLatch work at least as well?
A problem I'm seeing in many tests is that they sleep for some # seconds in a
loop.
For one single test that does not seem to matter too much, but adding it all up
that would result in a lot of unnecessary waiting. If one waits x seconds in a
loop, then on average 1/2 x of time is wasted. The total quite of tests runs in
about 4 hours (on 0.20-security) even when I throw bigger hardware at it.
When a CountDownLatch is used to coordinate between threads (either to signal
that work can start, or that threads have finished) then there is no such
additional waiting.
> Add test utility for writing multi-threaded tests
> -------------------------------------------------
>
> Key: HADOOP-7298
> URL: https://issues.apache.org/jira/browse/HADOOP-7298
> Project: Hadoop Common
> Issue Type: Test
> Components: test
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.22.0, 0.23.0
>
> Attachments: hadoop-7298.txt, hadoop-7298.txt, hadoop-7298.txt
>
>
> A lot of our tests spawn off multiple threads in order to check various
> synchronization issues, etc. It's often tedious to write these kinds of tests
> because you have to manually propagate exceptions back to the main thread,
> etc.
> In HBase we have developed a testing utility which makes writing these kinds
> of tests much easier. I'd like to copy that utility into Hadoop so we can use
> it here as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
