[
https://issues.apache.org/jira/browse/YARN-4685?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rohith Sharma K S updated YARN-4685:
------------------------------------
Description:
AM blacklist addition or removal is updated only when RMAppAttempt is scheduled
i.e {{RMAppAttemptImpl#ScheduleTransition#transition}}. But once attempt is
scheduled if there is any removeNode/addNode in cluster then this is not
updated to {{BlackListManager#refreshNodeHostCount}}. This leads
BlackListManager to operate on stale NM's count. And application is in ACCEPTED
state and wait forever even if blacklisted nodes are reconnected with clearing
disk space.
was:
AM blacklist addition or removal is updated only when RMAppAttempt is scheduled
i.e {{RMAppAttemptImpl#ScheduleTransition#transition}}. But once attempt is
scheduled if there is any removeNode/addNode in cluster then this is not
updated to {{BlackListManager#refreshNodeHostCount}}. This leads
BlackListManager to operate on stale NM's count. And application is in ACCEPTED
state and wait forever even if we add more nodes to cluster.
Solution is update BlacklistManager for every
{{RMAppAttemptImpl#AMContainerAllocatedTransition#transition}} call. This
ensures if there is any addition/removal in nodes, this will be updated to
BlacklistManager
> AM blacklisting result in application to get hanged
> ---------------------------------------------------
>
> Key: YARN-4685
> URL: https://issues.apache.org/jira/browse/YARN-4685
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 2.8.0
> Reporter: Rohith Sharma K S
> Assignee: Rohith Sharma K S
> Priority: Critical
>
> AM blacklist addition or removal is updated only when RMAppAttempt is
> scheduled i.e {{RMAppAttemptImpl#ScheduleTransition#transition}}. But once
> attempt is scheduled if there is any removeNode/addNode in cluster then this
> is not updated to {{BlackListManager#refreshNodeHostCount}}. This leads
> BlackListManager to operate on stale NM's count. And application is in
> ACCEPTED state and wait forever even if blacklisted nodes are reconnected
> with clearing disk space.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
