[
https://issues.apache.org/jira/browse/YARN-1166?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Zhijie Shen updated YARN-1166:
------------------------------
Attachment: YARN-1166.4.patch
[~ajisakaa], thanks!
I've uploaded a new patch to do the following two things:
1. Change appsFailed to be a counter
2. Expose RMContext to AppSchedulerInfo, such that QueueMetrics can use the app
info to determine whether it is a last attempt or not. The counter only
increase at the last attempt.
Modified the test cases to verify the logic.
It's a compromise to do the trick here. I've considered to correct the logic to
only the increment on last attempt failure, but it turns out to be a lot
changes on the path of the APP_REMOVE event from RMApp/RMAppAttempt to
QueueMetrics. IMHO, I'm conservative to do such kind of change when release
2.2.0 is coming.
> YARN 'appsFailed' metric should be of type 'counter'
> ----------------------------------------------------
>
> Key: YARN-1166
> URL: https://issues.apache.org/jira/browse/YARN-1166
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 2.1.0-beta
> Reporter: Srimanth Gunturi
> Assignee: Zhijie Shen
> Priority: Blocker
> Attachments: YARN-1166.2.patch, YARN-1166.3.patch, YARN-1166.4.patch,
> YARN-1166.patch
>
>
> Currently in YARN's queue metrics, the cumulative metric 'appsFailed' is of
> type 'guage' - which means the exact value will be reported.
> All other cumulative queue metrics (AppsSubmitted, AppsCompleted, AppsKilled)
> are all of type 'counter' - meaning Ganglia will use slope to provide deltas
> between time-points.
> To be consistent, AppsFailed metric should also be of type 'counter'.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
