[
https://issues.apache.org/jira/browse/GOBBLIN-1069?focusedWorklogId=398748&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-398748
]
ASF GitHub Bot logged work on GOBBLIN-1069:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 05/Mar/20 22:56
Start Date: 05/Mar/20 22:56
Worklog Time Spent: 10m
Work Description: sv2000 commented on pull request #2908: [GOBBLIN-1069]
Add NPE check in handleContainerCompletion method
URL: https://github.com/apache/incubator-gobblin/pull/2908#discussion_r388613538
##########
File path: gobblin-yarn/src/main/java/org/apache/gobblin/yarn/YarnService.java
##########
@@ -620,7 +620,10 @@ private boolean shouldStickToTheSameNode(int
containerExitStatus) {
*/
protected void handleContainerCompletion(ContainerStatus containerStatus) {
Map.Entry<Container, String> completedContainerEntry =
this.containerMap.remove(containerStatus.getContainerId());
-
+ if (completedContainerEntry == null) {
Review comment:
@ZihanLi58 shouldn't we get a replacement container for this? One reason why
we are hitting NPE is that container allocation/completion are asynchronous
calls and there is no guarantee that the callbacks are processed in the same
order. In other words, handleContainerCompletion() callback may happen before
onContainersAllocated(), which can cause NPE. In this scenario, we should be
requesting a replacement container.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 398748)
Time Spent: 1h 20m (was: 1h 10m)
> Add NPE check in handleContainerCompletion method
> -------------------------------------------------
>
> Key: GOBBLIN-1069
> URL: https://issues.apache.org/jira/browse/GOBBLIN-1069
> Project: Apache Gobblin
> Issue Type: Task
> Reporter: Zihan Li
> Priority: Major
> Time Spent: 1h 20m
> Remaining Estimate: 0h
>
> Add NPE check in handleContainerCompletion method to make sure call the
> method twice for the same container will not fail the job
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
