[jira] [Updated] (YARN-7368) Yarn Work-Preserving Better Handling Failed Disk

["BELUGA BEHR (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/YARN-7368?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

BELUGA BEHR updated YARN-7368:
------------------------------
    Affects Version/s: 3.0.0

> Yarn Work-Preserving Better Handling Failed Disk
> ------------------------------------------------
>
>                 Key: YARN-7368
>                 URL: https://issues.apache.org/jira/browse/YARN-7368
>             Project: Hadoop YARN
>          Issue Type: Improvement
>          Components: nodemanager, yarn
>    Affects Versions: 2.8.1, 3.0.0
>            Reporter: BELUGA BEHR
>
> If the drive that hosts the {{yarn.nodemanager.recovery.dir}} is broken then the entire NodeManager will not start.  Please improve this so that if the directory is not able to be created/accessed then the recovery portion of the NM is simply skipped and the NM continues to operate as normal.
> It may also be beneficial to be able to define multiple directories, like YARN logging directories, so that if one drive fails, not all of the recovery data is lost.
> https://hadoop.apache.org/docs/r2.7.2/hadoop-yarn/hadoop-yarn-site/NodeManagerRestart.html



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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