You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Weiwei Yang (JIRA)" <ji...@apache.org> on 2018/03/14 10:11:00 UTC
[jira] [Commented] (YARN-7636) Re-reservation count may overflow
when cluster resource exhausted for a long time
[ https://issues.apache.org/jira/browse/YARN-7636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16398349#comment-16398349 ]
Weiwei Yang commented on YARN-7636:
-----------------------------------
Thanks [~Tao Yang] for catching this bug and providing the fix, it looks good to me. Pending on jenkins. Only thing is do we need to get this into 3.1.0? Ping [~leftnoteasy] for his opinion.
> Re-reservation count may overflow when cluster resource exhausted for a long time
> ----------------------------------------------------------------------------------
>
> Key: YARN-7636
> URL: https://issues.apache.org/jira/browse/YARN-7636
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 3.1.0, 2.9.1
> Reporter: Tao Yang
> Assignee: Tao Yang
> Priority: Major
> Attachments: YARN-7636.001.patch, YARN-7636.002.patch
>
>
> This happens on our production cluster twice, when a request cannot be satisfied for a long time, it continually triggers the re-reservation and eventually caused the overflow. This will crash the scheduler.
> Exception stack:
> {noformat}
> java.lang.IllegalArgumentException: Overflow adding 1 occurrences to a count of 2147483647
> at com.google.common.collect.ConcurrentHashMultiset.add(ConcurrentHashMultiset.java:246)
> at com.google.common.collect.AbstractMultiset.add(AbstractMultiset.java:80)
> at com.google.common.collect.ConcurrentHashMultiset.add(ConcurrentHashMultiset.java:51)
> at org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt.addReReservation(SchedulerApplicationAttempt.java:406)
> at org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt.reserve(SchedulerApplicationAttempt.java:555)
> at org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerApp.reserve(FiCaSchedulerApp.java:1076)
> at org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerApp.apply(FiCaSchedulerApp.java:795)
> at org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler.tryCommit(CapacityScheduler.java:2770)
> at org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler$ResourceCommitterService.run(CapacityScheduler.java:546)
> {noformat}
> Refer to handling of SchedulerApplicationAttempt#addSchedulingOpportunity, we can ignore this exception to avoid this problem.
> This problem may happens in SchedulerApplicationAttempt#addMissedNonPartitionedRequestSchedulingOpportunity, fix it in the same way.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org