Github user davies commented on the pull request:
https://github.com/apache/spark/pull/10045#issuecomment-160746417
@kayousterhout I'd like to increase the timeout to a little bit larger, for
example, one minute.
Considering that we have N executors and M tasks (M > N), if there is one
bad executor (bad host), all other executors will be busy with computing, other
idle tasks will be scheduled to this bad executor and fail. If the timeout * 3
is smaller than the time used to finished all the task in (N-1) executor, the
job will fail.
The disadvantage of having larger timeout is that you have only one
executor and hit some temporary failure (could succeed after retry), it will
slow down the 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]
