sarutak commented on pull request #33253: URL: https://github.com/apache/spark/pull/33253#issuecomment-885405263
> Ah I see, currently we are populating killedTasksSummary in SpeculationStageSummary only if the killed reason is another attempt succeeded. For other reasons, we are not populating it since this is mainly specific to speculation to understand how many speculated tasks killed vs original tasks killed. So yeah killedTasksSummary can be empty. What do you think should be the right behavior? Also Even if another attempt succeeds, `killedTasksSummary` can be empty when a speculative task is the last task in a job. In this case, the killed reason is `Stage finished`. https://github.com/apache/spark/blob/2fe12a75206d4dbef6d7678b876c16876136cdd0/core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala#L1661-L1675 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
