[
https://issues.apache.org/jira/browse/HADOOP-4831?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vinod K V updated HADOOP-4831:
------------------------------
Attachment: HADOOP-4831-20090113-git.txt
Attaching a new patch addressing the review comments.
- Added the tests testJobInitInMultipleQs and jobKillWhileSetupIsRunning to
respectively test job initialization in the presence of multiple queues and job
kill.
- Have done some refactoring of methods. Moved waitTillSetupTaskStartsRunning,
waitTillSetupTaskFinishes and waitTillCleanupTaskFinishes to
ClusterWithCapacityScheduler as it is the right place for these methods.
- Also, added some code so as to handle exceptions in individual tests and to
proceed to the following tests without failing. Also, added tear-down code for
each test in a 'finally' block - to kill all jobs from previous test. This is
needed as the testcase tries to reuse the the cluster for various tests.
The total run time for TestJobInitialization hovers around 7 minutes. To test
this, comment out the line "this.reclaimCapacityThread.start()"
(CapacityTaskScheduler.java +1239), otherwise it is very highly likely that
HADOOP-4977 is hit.
> Have end to end tests based on MiniMRCluster to verify the correct behaviour
> of job initialization.
> ---------------------------------------------------------------------------------------------------
>
> Key: HADOOP-4831
> URL: https://issues.apache.org/jira/browse/HADOOP-4831
> Project: Hadoop Core
> Issue Type: Bug
> Components: contrib/capacity-sched
> Reporter: Vinod K V
> Assignee: Vinod K V
> Attachments: HADOOP-4831-20081230-git-final.txt,
> HADOOP-4831-20090107-svn-1.txt, HADOOP-4831-20090113-git.txt
>
>
> Write tests to verify that the right number of jobs in the right order are
> initialized and that the limits on the number of jobs that can stay
> initialized at any time are honoured.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
