Problem
Container failed, exitCode=-1000. No space available in any of the local directories.
Cause
Out of local disc space on executing datanodes.
Collect Details:
- Locate 'core-site.xml' to determine value for 'hadoop.tmp.dir'.
- Locate 'yarn-site.xml' to determine value for 'yarn.nodemanager.local-dirs'.
- Identify which node(s) within your cluster executed the failed job.
- Identify which Hadoop user ID executed the failed job.
Solution
Ensure the fully interpolated path for 'yarn.nodemanager.local-dirs' has sufficient space to store intermediate execute data.
If purging that path is required, you must bring that datanode offline before doing so.
Comments
0 comments
Please sign in to leave a comment.