[jira] [Updated] (MAPREDUCE-6960) Shuffle Handler prints disk error stack traces for every read failure.

["Kuhu Shukla (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/MAPREDUCE-6960?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Kuhu Shukla updated MAPREDUCE-6960:
-----------------------------------
    Attachment: MAPREDUCE-6960.001.patch

v1 patch that logs the exception without the stack trace only in the case of DiskErrorException. We can do the same for all IOExceptions but that would make any other kinds of IOException debugging difficult.

> Shuffle Handler prints disk error stack traces for every read failure.
> ----------------------------------------------------------------------
>
>                 Key: MAPREDUCE-6960
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-6960
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>            Reporter: Kuhu Shukla
>            Assignee: Kuhu Shukla
>         Attachments: MAPREDUCE-6960.001.patch
>
>
> {code}
>  } catch (IOException e) {
>           LOG.error("Shuffle error :", e);
> {code}
> In cases where the read from a disk fails and throws a DiskErrorException, the shuffle handler prints the entire stack trace for each and every one of the failures causing the nodemanager logs to quickly fill up the disk. 



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

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