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 "Jason Lowe (JIRA)" <ji...@apache.org> on 2016/08/11 21:48:20 UTC

[jira] [Updated] (YARN-1529) Add Localization overhead metrics to NM

     [ https://issues.apache.org/jira/browse/YARN-1529?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jason Lowe updated YARN-1529:
-----------------------------
    Attachment: YARN-1529.v04.patch

I've attached a version 4 of the patch upmerged to trunk which is what we're using internally.  It's heavily derived from Gera's patch.

I agree that writing the metrics to ATS would be interesting and useful, but I'm not sure we should tie NM-level localization metrics and container-level metrics together in one JIRA.  We've found the node-level aggregated metrics very useful on their own.  As such I'm thinking we might want to proceed in this JIRA with the aggregated container localization metrics in the NM and move the per-container metrics in ATS to a separate JIRA.  That way we can get some of the benefits sooner (and on clusters that don't have ATS configured or prepared to handle the extra load).

> Add Localization overhead metrics to NM
> ---------------------------------------
>
>                 Key: YARN-1529
>                 URL: https://issues.apache.org/jira/browse/YARN-1529
>             Project: Hadoop YARN
>          Issue Type: Improvement
>          Components: nodemanager
>            Reporter: Gera Shegalov
>            Assignee: Chris Trezzo
>         Attachments: YARN-1529.v01.patch, YARN-1529.v02.patch, YARN-1529.v03.patch, YARN-1529.v04.patch
>
>
> Users are often unaware of localization cost that their jobs incur. To measure effectiveness of localization caches it is necessary to expose the overhead in the form of metrics.
> We propose addition of the following metrics to NodeManagerMetrics.
> When a container is about to launch, its set of LocalResources has to be fetched from a central location, typically on HDFS, that results in a number of download requests for the files missing in caches.
> LocalizedFilesMissed: total files (requests) downloaded from DFS. Cache misses.
> LocalizedFilesCached: total localization requests that were served from local caches. Cache hits.
> LocalizedBytesMissed: total bytes downloaded from DFS due to cache misses.
> LocalizedBytesCached: total bytes satisfied from local caches.
> Localized(Files|Bytes)CachedRatio: percentage of localized (files|bytes) that were served out of cache: ratio = 100 * caches / (caches + misses)
> LocalizationDownloadNanos: total elapsed time in nanoseconds for a container to go from ResourceRequestTransition to LocalizedTransition



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org