Github user jinxing64 commented on the issue:
https://github.com/apache/spark/pull/16876
@kayousterhout
It's great to give a definition of `pendingPartitions` in `ShuffleMapStage`.
May I ask a question and make my understanding about `pendingPartitions`
clear ? It means:
1. PartitionId gets removed from `pendingPartitions` because of real
success(not including ones from failed executors);
2. `pendingPartitions` is the exact mirror(in reverse) of `ShuffleMapStage`
's output locations. Thus when `pendingPartition` is empty, all output
locations are available;
3. When executor lost, we should check if there is need to add partitionId
back to `pendingPartitions` accordingly.
4. No need to consider resubmit stage when `pendingPartitions` is empty.
Is my understanding correct ? It'd great if you can help rectify my wrong
points.
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
