Github user tgravescs commented on a diff in the pull request:
https://github.com/apache/spark/pull/10951#discussion_r52172877
--- Diff: core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala
---
@@ -1153,7 +1153,16 @@ class DAGScheduler(
}
if (!stageIdToStage.contains(task.stageId)) {
- // Skip all the actions if the stage has been cancelled.
+ logInfo("skip normal actions as stage cancelled")
+ // Need to handle tasks coming in late (speculative and jobs killed)
+ // post a task end event so accounting for things manually tracking
tasks work.
+ // This really should be something other then success since the
other speculative task
+ // finished first.
--- End diff --
I think that is fine for now since its combining failed vs success tasks.
I do think its a bit weird that Spark marks all speculative tasks as Success
even when both obviously don't commit. That is part of the other jira I was
going to file though and if needed it can be split back apart at that point.
It does seem a bit odd to throw SPARK-13054 in with the other changes in
the same pr though.
---
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]
