[
https://issues.apache.org/jira/browse/HADOOP-5934?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714301#action_12714301
]
Hemanth Yamijala commented on HADOOP-5934:
------------------------------------------
One more minor nit in testClusterBlockingForLackOfMemory.
There's a comment mismatch: submitting a job with 1m, 1r, but comment says 1m,
0r - same mismatch happens for both normal jobs.
> testHighRamJobWithSpeculativeExecution needs some changes
> ---------------------------------------------------------
>
> Key: HADOOP-5934
> URL: https://issues.apache.org/jira/browse/HADOOP-5934
> Project: Hadoop Core
> Issue Type: Bug
> Components: contrib/capacity-sched
> Reporter: Hemanth Yamijala
>
> There are some problems I noticed with
> TestCapacityScheduler.testHighRamJobWithSpeculativeExecution():
> - Comments are out of sync. For e.g. we say "// 2 map and 2 reduce slots" and
> start a cluster with 3 maps and reduces
> - In the test for reduces with speculative execution, the second job
> submitted is submitting high RAM maps
> - In the normal job's reduces, the memory is set as 1 * 104, instead of 1*
> 1024
> - Test should check for pendingReduces instead of pendingMaps in the reduce
> speculation case.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
