[
https://issues.apache.org/jira/browse/MAPREDUCE-6622?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15157804#comment-15157804
]
Robert Kanter commented on MAPREDUCE-6622:
------------------------------------------
Looks good overall, some minor things:
- The release note recommendations for
{{mapreduce.jobhistory.loadedtasks.cache.size}} are good, but I think we should
also put them into mapred-default.xml.
- Undo the {{blah.blah.*}} imports
- Add a comment explaining why you set the concurrency levels to 1. I can
easily see someone coming along later and saying, "hey, more concurrency sounds
good" without realizing the funny side effects that you showed me offline
- Add a comment explaining why you do the -1 in
{{Math.min(loadedTasksCacheSize-1, value.getTotalMaps() +
value.getTotalReduces());}} for the weights
- Use {{conf.setInt(..., #)}} instead of {{conf.set(..., "#")}} in the tests.
> Add capability to set JHS job cache to a task-based limit
> ---------------------------------------------------------
>
> Key: MAPREDUCE-6622
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6622
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: jobhistoryserver
> Affects Versions: 2.7.2
> Reporter: Ray Chiang
> Assignee: Ray Chiang
> Labels: supportability
> Attachments: MAPREDUCE-6622.001.patch, MAPREDUCE-6622.002.patch,
> MAPREDUCE-6622.003.patch
>
>
> When setting the property mapreduce.jobhistory.loadedjobs.cache.size the jobs
> can be of varying size. This is generally not a problem when the jobs sizes
> are uniform or small, but when the job sizes can be very large (say greater
> than 250k tasks), then the JHS heap size can grow tremendously.
> In cases, where multiple jobs are very large, then the JHS can lock up and
> spend all its time in GC. However, since the cache is holding on to all the
> jobs, not much heap space can be freed up.
> By setting a property that sets a cap on the number of tasks allowed in the
> cache and since the total number of tasks loaded is directly proportional to
> the amount of heap used, this should help prevent the JHS from locking up.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
