Github user sitalkedia commented on the pull request:
https://github.com/apache/spark/pull/11254#issuecomment-186052095
@markhamstra Ofcourse we want the stage to fail as soon as we detect fetch
failure so the recomputation can be done quickly. But the configuration
parameter in discussion here controls the maximum number of times a stage can
incur fetch failure before it is aborted (and hence the corresponding job is
failed). Currently the value is hardcode to 4, which means that when we have
more than 4 instances of fetch failure due to machine reboot or any other
error, we abort the stage and the job fails. This is an issue for our long
running Spark jobs which typically see several planned/unplanned machine reboot.
---
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]
