Github user tgravescs commented on the issue:
https://github.com/apache/spark/pull/14887
so we have had this come up multiple times now and really the shuffle
service should be using the yarn recovery specific disk. We made this change
under SPARK-14963. That disk is supposed to be critical for the nodemanager to
run and if that goes bad its expected the nodemanager will not run at all. That
is the location the nodemanager itself uses for recovery, rolling upgrades,
etc. As well as the mapreduce shuffle handler.
I propose to just close this as dup of that one.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
