[
https://issues.apache.org/jira/browse/MAPREDUCE-2921?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13105858#comment-13105858
]
Ravi Gummadi commented on MAPREDUCE-2921:
-----------------------------------------
The existing behavior has the advantage that it avoids the following issue:
If TT starts up even with single good disk/mapredLocalDir ignoring all other
bad disks, then that node can go into IO contention issues for this single disk
from all tasks --- because we are not reducing the number of slots on this TT
based on bad disks.
> TaskTracker won't start with failed local directory
> ---------------------------------------------------
>
> Key: MAPREDUCE-2921
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-2921
> Project: Hadoop Map/Reduce
> Issue Type: Sub-task
> Components: tasktracker
> Affects Versions: 0.20.204.0
> Reporter: Eli Collins
>
> Chmod'ing one of the mapred local directories so it's not executable will
> cause the TT to fail to start. Doing this after the TT has started will
> result in a TT that is up but can not execute tasks.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
