Github user cloud-fan commented on the issue:
https://github.com/apache/spark/pull/18388
if it's ok to break shuffle service backward compatibility(by default this
config is off), I think we should introduce a new response type to tell the
client that, the shuffle service is still up but just in memory shortage,
please do not give up and keep trying.
Currently we just close the connection, so the client has no idea what's
going on and may mistakenly report FetchFailure and fail the stage/job.
---
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]
