Susheel Gupta created YARN-11801:
------------------------------------
Summary: NPE in FifoCandidatesSelector.selectCandidates when
preempting resources for an auto-created queue without child queues
Key: YARN-11801
URL: https://issues.apache.org/jira/browse/YARN-11801
Project: Hadoop YARN
Issue Type: Bug
Components: yarn
Affects Versions: 3.4.0, 3.5.0
Reporter: Susheel Gupta
When enabling the ProportionalCapacityPreemptionPolicy in the YARN
SchedulingMonitor, the system encounters a NullPointerException in
{{{}FifoCandidatesSelector.selectCandidates{}}}. This happens when an
auto-created queue exists but does not have any child queues.
NullPointer stack trace:
{code:java}
2025-03-24 08:36:12,593 ERROR monitor.SchedulingMonitor: Exception raised while
executing preemption checker, skip this run..., exception=
java.lang.NullPointerException
at
org.apache.hadoop.yarn.server.resourcemanager.monitor.capacity.FifoCandidatesSelector.selectCandidates(FifoCandidatesSelector.java:104)
at
org.apache.hadoop.yarn.server.resourcemanager.monitor.capacity.ProportionalCapacityPreemptionPolicy.containerBasedPreemptOrKill(ProportionalCapacityPreemptionPolicy.java:515)
at
org.apache.hadoop.yarn.server.resourcemanager.monitor.capacity.ProportionalCapacityPreemptionPolicy.editSchedule(ProportionalCapacityPreemptionPolicy.java:344)
at
org.apache.hadoop.yarn.server.resourcemanager.monitor.SchedulingMonitor.invokePolicy(SchedulingMonitor.java:100)
at
org.apache.hadoop.yarn.server.resourcemanager.monitor.SchedulingMonitor$PolicyInvoker.run(SchedulingMonitor.java:112)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748){code}
Capacity-scheduler used:
{noformat}
<configuration>
<property>
<name>yarn.scheduler.capacity.mapping-rule-json</name>
<value/>
</property>
<property>
<name>yarn.scheduler.capacity.root.client.capacity</name>
<value>50</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.client.leaf-queue-template.capacity</name>
<value>0</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.client.maximum-capacity</name>
<value>100</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.client.auto-create-child-queue.enabled</name>
<value>true</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.default.capacity</name>
<value>50</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.client.leaf-queue-template.maximum-capacity</name>
<value>100</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.queues</name>
<value>default,client</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.capacity</name>
<value>100</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.default.maximum-capacity</name>
<value>100</value>
</property>
<property>
<name>yarn.scheduler.capacity.schedule-asynchronously.enable</name>
<value>true</value>
</property>
<property>
<name>yarn.webservice.mutation-api.version</name>
<value>1742806178771</value>
</property>
<property>
<name>yarn.scheduler.capacity.root.default.maximum-am-resource-percent</name>
<value>0.2</value>
</property>
<property>
<name>yarn.scheduler.capacity.mapping-rule-format</name>
<value>json</value>
</property>
</configuration>{noformat}
Add this in yarn-site.xml:
{code:java}
<property>
<name>yarn.resourcemanager.scheduler.monitor.enable</name>
<value>true</value>
</property>
<property>
<name>yarn.resourcemanager.scheduler.monitor.policies</name>
<value>org.apache.hadoop.yarn.server.resourcemanager.monitor.capacity.ProportionalCapacityPreemptionPolicy,org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.AutoCreatedQueueDeletionPolicy,org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.QueueManagementDynamicEditPolicy</value>
</property>{code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-dev-h...@hadoop.apache.org
