[
https://issues.apache.org/jira/browse/SPARK-19698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15881564#comment-15881564
]
Kay Ousterhout commented on SPARK-19698:
----------------------------------------
I see -- I agree that everything in your description is correct. The driver
will allow all tasks to finish if it's still running (e.g., if other tasks are
being submitted), but you're right it will shut down the workers while some
tasks are still in progress if the Driver shuts down.
To think about how to fix this, let me ask you a question about your workload:
suppose a task is in the middle of manipulating some external state (as you
described in the JIRA description) and it gets killed suddenly because the JVM
runs out of memory (e.g., because another concurrently running task used up all
of the memory). In that case, the job listener won't be told about the failed
task, and it will be re-tried. Does that pose a problem in the same way that
the behavior described in the PR is problematic?
> Race condition in stale attempt task completion vs current attempt task
> completion when task is doing persistent state changes
> ------------------------------------------------------------------------------------------------------------------------------
>
> Key: SPARK-19698
> URL: https://issues.apache.org/jira/browse/SPARK-19698
> Project: Spark
> Issue Type: Bug
> Components: Mesos, Spark Core
> Affects Versions: 2.0.0
> Reporter: Charles Allen
>
> We have encountered a strange scenario in our production environment. Below
> is the best guess we have right now as to what's going on.
> Potentially, the final stage of a job has a failure in one of the tasks (such
> as OOME on the executor) which can cause tasks for that stage to be
> relaunched in a second attempt.
> https://github.com/apache/spark/blob/v2.1.0/core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala#L1155
> keeps track of which tasks have been completed, but does NOT keep track of
> which attempt those tasks were completed in. As such, we have encountered a
> scenario where a particular task gets executed twice in different stage
> attempts, and the DAGScheduler does not consider if the second attempt is
> still running. This means if the first task attempt succeeded, the second
> attempt can be cancelled part-way through its run cycle if all other tasks
> (including the prior failed) are completed successfully.
> What this means is that if a task is manipulating some state somewhere (for
> example: a upload-to-temporary-file-location, then delete-then-move on an
> underlying s3n storage implementation) the driver can improperly shutdown the
> running (2nd attempt) task between state manipulations, leaving the
> persistent state in a bad state since the 2nd attempt never got to complete
> its manipulations, and was terminated prematurely at some arbitrary point in
> its state change logic (ex: finished the delete but not the move).
> This is using the mesos coarse grained executor. It is unclear if this
> behavior is limited to the mesos coarse grained executor or not.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
