[
https://issues.apache.org/jira/browse/YARN-9649?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Amithsha updated YARN-9649:
---------------------------
Issue Type: Improvement (was: Bug)
> Display Blacklisted Nodes By AM in RM Log and AM Rest api
> ---------------------------------------------------------
>
> Key: YARN-9649
> URL: https://issues.apache.org/jira/browse/YARN-9649
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: resourcemanager
> Affects Versions: 2.9.0
> Reporter: Amithsha
> Priority: Major
>
> In our prod cluster, blacklisting nodes via AM is enabled by default. All of
> a sudden observed the compute fluctuation. On further debugging found that AM
> is blacklisting the nodes. But no list of blacklisted nodes is logged by RM,
> also curling AM logs are not possible since the number of jobs is huge.
> https://issues.apache.org/jira/browse/YARN-4307
> As described in the above the Jira we can view UI. But not in log or RestApi.
> *It will be helpful if we push the list to RestApi or to RM log.*
> In RM log we found the followings
> 2015-10-28 21:57:35,982 DEBUG
> org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.allocator.RegularContainerAllocator:
> Skipping 'host' localhost for application_1446028825407_0002 since it has
> been blacklisted
> But this doesnt contian the information about the App where its container
> failed.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
