Github user sryza commented on the pull request:
https://github.com/apache/spark/pull/4363#issuecomment-73614405
@andrewor14 do you think it's worthwhile to make a best effort attempt to
kill executors on cluster managers that support it? I know `killExecutors`
doesn't currently work this way, but I think it could be worth either changing
the semantics or adding another method or flag that supports this.
I also think this PR has some general value in centralizing the logic for
executor expiry outside the `BlockManagerMaster`. Currently IIUC we rely on
Akka disassociation for canceling tasks on unreachable executors, but unifying
this with what's currently the blockmanager timeout would make things more
consistent and allow us to be more proactive about rescheduling.
---
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]
