[
https://issues.apache.org/jira/browse/YARN-3933?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16356599#comment-16356599
]
stefanlee edited comment on YARN-3933 at 2/8/18 7:37 AM:
---------------------------------------------------------
[~yufeigu] thanks, yesterday, i found in our cluster the utilization rate of
resource is very low , but there is a lot of pending applications in it, and
RM has no exception, then i found a queue has negative-usage and also has
pending resource, so i doubt Whether a queue has negative-usage resource can
lead to FairScheduler do not assign containers to any other queues. thanks for
this jira[YARN-3933|https://issues.apache.org/jira/browse/YARN-3933] it seems
as same as my scenario.
was (Author: imstefanlee):
[~yufeigu] thanks, yesterday, i found in our cluster the utilization rate of
resource is very low , but there is a lot of pending applications in it, and
RM has no exception, then i found a queue has negative-usage and also has
pending resource, so i doubt Whether a queue has negative-usage resource can
lead to FairScheduler do not assign containers to any other queues. thanks for
this jira[link title|https://issues.apache.org/jira/browse/YARN-3933] it seems
as same as my scenario.
> FairScheduler: Multiple calls to completedContainer are not safe
> ----------------------------------------------------------------
>
> Key: YARN-3933
> URL: https://issues.apache.org/jira/browse/YARN-3933
> Project: Hadoop YARN
> Issue Type: Bug
> Components: fairscheduler
> Affects Versions: 2.7.1
> Reporter: Lavkesh Lahngir
> Assignee: Shiwei Guo
> Priority: Major
> Labels: oct16-medium
> Fix For: 2.8.0, 3.0.0-alpha4
>
> Attachments: YARN-3933.001.patch, YARN-3933.002.patch,
> YARN-3933.003.patch, YARN-3933.004.patch, YARN-3933.005.patch,
> YARN-3933.006.patch, yarn-3933-branch-2.8.patch
>
>
> In our cluster we are seeing available memory and cores being negative.
> Initial inspection:
> Scenario no. 1:
> In capacity scheduler the method allocateContainersToNode() checks if
> there are excess reservation of containers for an application, and they are
> no longer needed then it calls queue.completedContainer() which causes
> resources being negative. And they were never assigned in the first place.
> I am still looking through the code. Can somebody suggest how to simulate
> excess containers assignments ?
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
