1

I have a Spark job which trains a model using Spark ML's logistic regression. In the Spark UI under the stage details page for a tree aggregation stage I see a few tasks with a status of "GET RESULT".

What does this status mean? What causes a task to be in this status?

Daniel Siegmann
  • 287
  • 1
  • 3
  • 5

1 Answers1

1

If you check TaskInfo you'll see that "GET RESULT" status requires gettingResult to be set to true and this happens when gettingResultTime is not zero.

Finally gettingResultTime represents:

The time when the task started remotely getting the result. Will not be set if the task result was sent immediately when the task finished (as opposed to sending an IndirectTaskResult and later fetching the result from the block manager).

So to summarize it means that task have finished and result is fetched from the block manager.

zero323
  • 322,348
  • 103
  • 959
  • 935