Github user JoshRosen commented on the pull request:
https://github.com/apache/spark/pull/3786#issuecomment-104979321
FWIW, it's possible that bugs in Spark itself can lead to infinite fetch
failures and stage recomputations (I introduced one such bug when developing on
my new shuffle write code, but fixed it before submitting my PR). Even though
such a flag might mask the presence of such bugs, I think it's still worthwhile
to have some sort of limit on stage retries in order to keep Spark from going
into an infinite loop, since that's a really bad failure mode. Therefore, we
might want to consider re-opening this PR and adding proper tests.
---
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]
