[
https://issues.apache.org/jira/browse/YARN-4479?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15066415#comment-15066415
]
Sunil G commented on YARN-4479:
-------------------------------
Hi [~Naganarasimha]
bq.IUC Sunil G also wanted to say the same ?
I meant in slight different way. With existing approach, any application which
was running earlier to RM restart will also be in RUNNING state. It may starve
as per the scenario which you and Rohith mentioned, but the state of
application will be running.
Also adding to the existing discussion, I would like to point out few things.
This patch tries to activate all applications which were running before RM
restart happened. It may be getting activated with different order, but it is
trying to put all these apps in the activated list of scheduler (App state will
be RUNNING still).
1. After restart with or without ordering, only highest priority app will be
selected for scheduling from activated list. This same behavior is happening
before RM restart also. SO there seems no impact with this. Pls correct me if I
am wrong.
2. All containers which were running earlier will still continue, and all
pending requests will be updated/refreshed and this is from
ApplicationMasterService thread. So if all earlier running apps are
activated,then behavior will be same from scheduler end, correct?
Being said all this points, I also feel that we may need to add more complex
code to keep the same order as you proposed. So if there are no major impacts,
I think the approach taken in this patch looks fine. Thoughts?
> Retrospect app-priority in pendingOrderingPolicy during recovering
> applications
> -------------------------------------------------------------------------------
>
> Key: YARN-4479
> URL: https://issues.apache.org/jira/browse/YARN-4479
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: api, resourcemanager
> Reporter: Rohith Sharma K S
> Assignee: Rohith Sharma K S
> Attachments: 0001-YARN-4479.patch
>
>
> Currently, same ordering policy is used for pending applications and active
> applications. When priority is configured for an applications, during
> recovery high priority application get activated first. It is possible that
> low priority job was submitted and running state.
> This causes low priority job in starvation after recovery
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
