Sunil G commented on YARN-3136:

Hi [~jlowe]
Yes. I also had a concern on same, hence was planning for scheduler level lock 
to get object from *applications* map.

FiFoScheduler is initializing with *ConcurrentSkipListMap*, CS and Fair 
Scheduler are using *ConcurrentHashMap*.
So if we can have *applications* as *ConcurrentMap*, it looks fine. This does 
enforce other pluggable schedulers to keep *applications* as a concurrentmap. 
Also this will break any such schedulers which are present now once this change 
is made.

> getTransferredContainers can be a bottleneck during AM registration
> -------------------------------------------------------------------
>                 Key: YARN-3136
>                 URL: https://issues.apache.org/jira/browse/YARN-3136
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: scheduler
>    Affects Versions: 2.6.0
>            Reporter: Jason Lowe
>            Assignee: Sunil G
>         Attachments: 0001-YARN-3136.patch
> While examining RM stack traces on a busy cluster I noticed a pattern of AMs 
> stuck waiting for the scheduler lock trying to call getTransferredContainers. 
>  The scheduler lock is highly contended, especially on a large cluster with 
> many nodes heartbeating, and it would be nice if we could find a way to 
> eliminate the need to grab this lock during this call.  We've already done 
> similar work during AM allocate calls to make sure they don't needlessly grab 
> the scheduler lock, and it would be good to do so here as well, if possible.

This message was sent by Atlassian JIRA

Reply via email to