[
https://issues.apache.org/jira/browse/YARN-8233?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16678045#comment-16678045
]
Tao Yang edited comment on YARN-8233 at 11/7/18 11:06 AM:
----------------------------------------------------------
Hi, [~ajisakaa], [~cheersyang]
I have attached 3 patches:
(1) patch for branch-3.1 just update test case since
SchedulerApplicationAttempt#hasPendingResourceRequest API has changed in 3.2
and trunk.
(2) patch for branch-3.0 includes modification above and drop the modification
in CapacityScheduler#attemptAllocationOnNode since the method is not exist yet.
(3) patch for branch-2 includes modifications above and update UT to add final
keywords for variables which are used in Mockito#doAnswer. branch-2.9 can use
branch-2 patch.
I have applied these patch on my local environment, tried to run UT and did not
found any problems. Just in case, please help to review these new patches
before committing, Thanks!
was (Author: tao yang):
Hi, [~ajisakaa], [~cheersyang]
I have attached 3 patches:
(1) patch for branch-3.1 just update test case since
SchedulerApplicationAttempt#hasPendingResourceRequest API has changed in 3.2
and trunk.
(2) patch for branch-3.0 includes modification above and drop the modification
in CapacityScheduler#attemptAllocationOnNode since the method is not exist yet.
(3) patch for branch-2 includes modifications above and update UT to add final
keywords for variables which are used in Mockito#doAnswer. branch-2.9 can use
branch-2 patch.
Please help to review these new patches before committing, Thanks.
> NPE in CapacityScheduler#tryCommit when handling allocate/reserve proposal
> whose allocatedOrReservedContainer is null
> ---------------------------------------------------------------------------------------------------------------------
>
> Key: YARN-8233
> URL: https://issues.apache.org/jira/browse/YARN-8233
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 3.2.0
> Reporter: Tao Yang
> Assignee: Tao Yang
> Priority: Critical
> Fix For: 3.3.0, 3.2.1
>
> Attachments: YARN-8233.001.branch-2.patch,
> YARN-8233.001.branch-3.0.patch, YARN-8233.001.branch-3.1.patch,
> YARN-8233.001.patch, YARN-8233.002.patch, YARN-8233.003.patch
>
>
> Recently we saw a NPE problem in CapacityScheduler#tryCommit when try to find
> the attemptId by calling {{c.getAllocatedOrReservedContainer().get...}} from
> an allocate/reserve proposal. But got null allocatedOrReservedContainer and
> thrown NPE.
> Reference code:
> {code:java}
> // find the application to accept and apply the ResourceCommitRequest
> if (request.anythingAllocatedOrReserved()) {
> ContainerAllocationProposal<FiCaSchedulerApp, FiCaSchedulerNode> c =
> request.getFirstAllocatedOrReservedContainer();
> attemptId =
> c.getAllocatedOrReservedContainer().getSchedulerApplicationAttempt()
> .getApplicationAttemptId(); //NPE happens here
> } else { ...
> {code}
> The proposal was constructed in
> {{CapacityScheduler#createResourceCommitRequest}} and
> allocatedOrReservedContainer is possibly null in async-scheduling process
> when node was lost or application was finished (details in
> {{CapacityScheduler#getSchedulerContainer}}).
> Reference code:
> {code:java}
> // Allocated something
> List<AssignmentInformation.AssignmentDetails> allocations =
> csAssignment.getAssignmentInformation().getAllocationDetails();
> if (!allocations.isEmpty()) {
> RMContainer rmContainer = allocations.get(0).rmContainer;
> allocated = new ContainerAllocationProposal<>(
> getSchedulerContainer(rmContainer, true), //possibly null
> getSchedulerContainersToRelease(csAssignment),
>
> getSchedulerContainer(csAssignment.getFulfilledReservedContainer(),
> false), csAssignment.getType(),
> csAssignment.getRequestLocalityType(),
> csAssignment.getSchedulingMode() != null ?
> csAssignment.getSchedulingMode() :
> SchedulingMode.RESPECT_PARTITION_EXCLUSIVITY,
> csAssignment.getResource());
> }
> {code}
> I think we should add null check for allocateOrReserveContainer before create
> allocate/reserve proposals. Besides the allocation process has increase
> unconfirmed resource of app when creating an allocate assignment, so if this
> check is null, we should decrease the unconfirmed resource of live app.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
