[
https://issues.apache.org/jira/browse/YARN-2983?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14257424#comment-14257424
]
Jian He commented on YARN-2983:
-------------------------------
thanks [~varun_saxena], I filed YARN-2987 while looking at this.
I think this NPE is very rare to happen, because so far,
scheduler.getAppsInQueue returns only running applications. If the app is
present in scheduler, it is very unlikely it's not present in the rmContext, as
RM is always evicting the oldest app first from the rmContext. do you also see
other scenarios ?
> NPE possible in ClientRMService#getQueueInfo
> --------------------------------------------
>
> Key: YARN-2983
> URL: https://issues.apache.org/jira/browse/YARN-2983
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 2.5.1
> Reporter: Varun Saxena
> Assignee: Varun Saxena
> Attachments: YARN-2983.patch
>
>
> While going through code for checking YARN-2978 , found one issue.
> During construction of {{GetQueueInfoResponse}} in
> {{ClientRMService#getQueueInfo}}, we first collect application attempts from
> scheduler and then get apps from a {{ConcurrentHashMap}} in {{RMContext}}.
> Although the operation(get/put/remove,etc) itself on a ConcurrentHashMap is
> thread-safe, but a series of multiple {{ConcurrentHashMap#get}} (say, in a
> for loop) are not.
> For instance, in code below, we are calling rmContext.getRMApps()#get in a
> loop. Now a ConcurrentHashMap#get can return null if the key doesnt exist.
> But there is no null check inside this for loop before dereferencing the
> value returned i.e. rmApp. Although all the applicationattempts have been
> fetched for the queue just above the for loop, but as this block of code is
> not synchronized, there is a possibility that another thread may delete RMApp
> from the ConcurrentHashMap at the same time. This can happen when an app
> finishes/completes and number of completed apps exceed the config
> {{yarn.resourcemanager.max-completed-applications}}.
> I think there should be a null check inside this for loop, otherwise a NPE
> can occur.
> {code:title=ClientRMService#getQueueInfo}
> public GetQueueInfoResponse getQueueInfo(GetQueueInfoRequest request)
> throws YarnException {
> .....
> if (request.getIncludeApplications()) {
> List<ApplicationAttemptId> apps =
> scheduler.getAppsInQueue(request.getQueueName());
> appReports = new ArrayList<ApplicationReport>(apps.size());
> for (ApplicationAttemptId app : apps) {
> RMApp rmApp = rmContext.getRMApps().get(app.getApplicationId());
> appReports.add(rmApp.createAndGetApplicationReport(null, true));
> }
> }
> ......
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
