Github user jerryshao commented on a diff in the pull request:
https://github.com/apache/spark/pull/9684#discussion_r44753574
--- Diff:
yarn/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocator.scala ---
@@ -538,6 +549,11 @@ private[yarn] class YarnAllocator(
if (executorIdToContainer.contains(eid)) {
pendingLossReasonRequests
.getOrElseUpdate(eid, new ArrayBuffer[RpcCallContext]) += context
+ } else if (releasedExecutorLossReasons.get(eid) != null) {
+ // Executor is already released explicitly before getting the loss
reason, so directly send
+ // the pre-stored lost reason
+ context.reply(releasedExecutorLossReasons.get(eid))
+ releasedExecutorLossReasons.remove(eid)
} else {
logWarning(s"Tried to get the loss reason for non-existent executor
$eid")
--- End diff --
Note: here it would also be better to response a message, otherwise driver
will pending on waiting the message until timeout. This may happens when issue
a loss reason query, AM already processed this completed containers (since they
asynchronously).
---
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
