Github user kayousterhout commented on a diff in the pull request:
https://github.com/apache/spark/pull/8887#discussion_r43670488
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala ---
@@ -468,11 +468,18 @@ private[spark] class TaskSchedulerImpl(
removeExecutor(executorId, reason)
failedExecutor = Some(executorId)
} else {
- // We may get multiple executorLost() calls with different loss
reasons. For example, one
- // may be triggered by a dropped connection from the slave while
another may be a report
- // of executor termination from Mesos. We produce log messages
for both so we eventually
- // report the termination reason.
- logError("Lost an executor " + executorId + " (already removed):
" + reason)
+ executorIdToHost.get(executorId) match {
+ case Some(_) =>
+ // We may need to update rootPool in case the executor was
pending a real loss reason.
--- End diff --
What about changing this comment to something like:
"If the executor is still in executorIdToHost, it means we haven't gotten
the real loss reason yet, so still call removeExecutor to fail all of the tasks
running on that executor (with the real loss reason)." (as-is, it's not clear
what it means to update rootPool)
---
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]
