Github user tgravescs commented on the pull request:
https://github.com/apache/spark/pull/5663#issuecomment-96748667
@vanzin The question was basically trying to get at fully was is happening.
I did not say we should fix the reporting, as I agree with you, that is a
separate issue.
You are also right that retrying to connect to RM won't help fix things
that have already gone away, but it can give you the real state. Just because
some network connects are down doesn't mean all of them are. Reconnecting to
RM could get you the real state of the application on the cluster. We
shouldn't just look at the exact case reported here, you have to look at other
possibilities and how this patch would affect them. For instance, if the RM
goes down and we can't get the application status, it doesn't mean the
application master has disconnected from the client/driver or that the
executors have died.
---
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]
