[
https://issues.apache.org/jira/browse/MAPREDUCE-5124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16275375#comment-16275375
]
Jason Lowe commented on MAPREDUCE-5124:
---------------------------------------
Sorry, Konstantin, I should have checked with you first. Please feel free to
revert it from 2.7.5, and apologies for the omitted CHANGES.txt. I'll correct
it when it's recommitted for 2.7.6.
I am pretty confident this is a safe change for 2.7 since it only changes the
status update code path and doesn't affect the task start, killed, failed,
succeeded paths. However I totally understand the logic behind punting this to
2.7.6. The patch will get a chance to run on our 2.8 clusters at scale in the
interim which will increase our confidence in it by then.
> AM lacks flow control for task events
> -------------------------------------
>
> Key: MAPREDUCE-5124
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5124
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mr-am
> Affects Versions: 2.0.3-alpha, 0.23.5
> Reporter: Jason Lowe
> Assignee: Peter Bacsko
> Fix For: 2.7.5, 3.1.0, 3.0.1, 2.10.0, 2.9.1, 2.8.4
>
> Attachments: MAPREDUCE-5124-001.patch, MAPREDUCE-5124-002.patch,
> MAPREDUCE-5124-003.patch, MAPREDUCE-5124-CoalescingPOC-1.patch,
> MAPREDUCE-5124-CoalescingPOC2.patch, MAPREDUCE-5124-CoalescingPOC3.patch,
> MAPREDUCE-5124-branch-2.001.patch, MAPREDUCE-5124-branch-2.002.patch,
> MAPREDUCE-5124-branch-2.7.001.patch, MAPREDUCE-5124-branch-2.7.002.patch,
> MAPREDUCE-5124-branch-2.7.002.patch, MAPREDUCE-5124-branch-2.8.001.patch,
> MAPREDUCE-5124-branch-2.8.001.patch, MAPREDUCE-5124-branch-2.8.002.patch,
> MAPREDUCE-5124-branch-2.8.002.patch, MAPREDUCE-5124-branch-2.9.001.patch,
> MAPREDUCE-5124-branch-2.9.002.patch, MAPREDUCE-5124-proto.2.txt,
> MAPREDUCE-5124-prototype.txt
>
>
> The AM does not have any flow control to limit the incoming rate of events
> from tasks. If the AM is unable to keep pace with the rate of incoming
> events for a sufficient period of time then it will eventually exhaust the
> heap and crash. MAPREDUCE-5043 addressed a major bottleneck for event
> processing, but the AM could still get behind if it's starved for CPU and/or
> handling a very large job with tens of thousands of active tasks.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: mapreduce-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: mapreduce-issues-h...@hadoop.apache.org
