[
https://issues.apache.org/jira/browse/YARN-4697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15251022#comment-15251022
]
Haibo Chen commented on YARN-4697:
----------------------------------
[~vinodkv] Upon NM restart, NM will try to recover all applications and submit
a log aggregation task to the thread pool for each application recovered.
Therefore, a large number of recovered applications plus concurrent
applications can cause the thread pool to increase without a bound.
> NM aggregation thread pool is not bound by limits
> -------------------------------------------------
>
> Key: YARN-4697
> URL: https://issues.apache.org/jira/browse/YARN-4697
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: nodemanager
> Reporter: Haibo Chen
> Assignee: Haibo Chen
> Priority: Critical
> Fix For: 2.9.0
>
> Attachments: yarn4697.001.patch, yarn4697.002.patch,
> yarn4697.003.patch, yarn4697.004.patch
>
>
> In the LogAggregationService.java we create a threadpool to upload logs from
> the nodemanager to HDFS if log aggregation is turned on. This is a cached
> threadpool which based on the javadoc is an ulimited pool of threads.
> In the case that we have had a problem with log aggregation this could cause
> a problem on restart. The number of threads created at that point could be
> huge and will put a large load on the NameNode and in worse case could even
> bring it down due to file descriptor issues.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
