[YARN-8326] Yarn 3.0 seems runs slower than Yarn 2.6

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.0.0
• Fix Version/s: 3.2.0, 3.1.1
• Component/s: yarn

Environment

This is the yarn-site.xml for 3.0. 

 

<configuration>

<property>
<name>hadoop.registry.dns.bind-port</name>
<value>5353</value>
</property>

<property>
<name>hadoop.registry.dns.domain-name</name>
<value>hwx.site</value>
</property>

<property>
<name>hadoop.registry.dns.enabled</name>
<value>true</value>
</property>

<property>
<name>hadoop.registry.dns.zone-mask</name>
<value>255.255.255.0</value>
</property>

<property>
<name>hadoop.registry.dns.zone-subnet</name>
<value>172.17.0.0</value>
</property>

<property>
<name>manage.include.files</name>
<value>false</value>
</property>

<property>
<name>yarn.acl.enable</name>
<value>false</value>
</property>

<property>
<name>yarn.admin.acl</name>
<value>yarn</value>
</property>

<property>
<name>yarn.client.nodemanager-connect.max-wait-ms</name>
<value>60000</value>
</property>

<property>
<name>yarn.client.nodemanager-connect.retry-interval-ms</name>
<value>10000</value>
</property>

<property>
<name>yarn.http.policy</name>
<value>HTTP_ONLY</value>
</property>

<property>
<name>yarn.log-aggregation-enable</name>
<value>false</value>
</property>

<property>
<name>yarn.log-aggregation.retain-seconds</name>
<value>2592000</value>
</property>

<property>
<name>yarn.log.server.url</name>
<value>http://xxxxxx:19888/jobhistory/logs</value>
</property>

<property>
<name>yarn.log.server.web-service.url</name>
<value>http://xxxxxx:8188/ws/v1/applicationhistory</value>
</property>

<property>
<name>yarn.node-labels.enabled</name>
<value>false</value>
</property>

<property>
<name>yarn.node-labels.fs-store.retry-policy-spec</name>
<value>2000, 500</value>
</property>

<property>
<name>yarn.node-labels.fs-store.root-dir</name>
<value>/system/yarn/node-labels</value>
</property>

<property>
<name>yarn.nodemanager.address</name>
<value>0.0.0.0:45454</value>
</property>

<property>
<name>yarn.nodemanager.admin-env</name>
<value>MALLOC_ARENA_MAX=$MALLOC_ARENA_MAX</value>
</property>

<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle,spark2_shuffle,timeline_collector</value>
</property>

<property>
<name>yarn.nodemanager.aux-services.mapreduce_shuffle.class</name>
<value>org.apache.hadoop.mapred.ShuffleHandler</value>
</property>

<property>
<name>yarn.nodemanager.aux-services.spark2_shuffle.class</name>
<value>org.apache.spark.network.yarn.YarnShuffleService</value>
</property>

<property>
<name>yarn.nodemanager.aux-services.spark2_shuffle.classpath</name>
<value>/usr/spark2/aux/*</value>
</property>

<property>
<name>yarn.nodemanager.aux-services.spark_shuffle.class</name>
<value>org.apache.spark.network.yarn.YarnShuffleService</value>
</property>

<property>
<name>yarn.nodemanager.aux-services.timeline_collector.class</name>
<value>org.apache.hadoop.yarn.server.timelineservice.collector.PerNodeTimelineCollectorsAuxService</value>
</property>

<property>
<name>yarn.nodemanager.bind-host</name>
<value>0.0.0.0</value>
</property>

<property>
<name>yarn.nodemanager.container-executor.class</name>
<value>org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor</value>
</property>

<property>
<name>yarn.nodemanager.container-metrics.unregister-delay-ms</name>
<value>60000</value>
</property>

<property>
<name>yarn.nodemanager.container-monitor.interval-ms</name>
<value>3000</value>
</property>

<property>
<name>yarn.nodemanager.delete.debug-delay-sec</name>
<value>0</value>
</property>

<property>
<name>yarn.nodemanager.disk-health-checker.max-disk-utilization-per-disk-percentage</name>
<value>90</value>
</property>

<property>
<name>yarn.nodemanager.disk-health-checker.min-free-space-per-disk-mb</name>
<value>1000</value>
</property>

<property>
<name>yarn.nodemanager.disk-health-checker.min-healthy-disks</name>
<value>0.25</value>
</property>

<property>
<name>yarn.nodemanager.health-checker.interval-ms</name>
<value>135000</value>
</property>

<property>
<name>yarn.nodemanager.health-checker.script.timeout-ms</name>
<value>60000</value>
</property>

<property>
<name>yarn.nodemanager.linux-container-executor.cgroups.strict-resource-usage</name>
<value>false</value>
</property>

<property>
<name>yarn.nodemanager.linux-container-executor.group</name>
<value>hadoop</value>
</property>

<property>
<name>yarn.nodemanager.linux-container-executor.nonsecure-mode.limit-users</name>
<value>false</value>
</property>

<property>
<name>yarn.nodemanager.local-dirs</name>
<value>/hadoop/yarn/local</value>
</property>

<property>
<name>yarn.nodemanager.log-aggregation.compression-type</name>
<value>gz</value>
</property>

<property>
<name>yarn.nodemanager.log-aggregation.debug-enabled</name>
<value>false</value>
</property>

<property>
<name>yarn.nodemanager.log-aggregation.num-log-files-per-app</name>
<value>30</value>
</property>

<property>
<name>yarn.nodemanager.log-aggregation.roll-monitoring-interval-seconds</name>
<value>3600</value>
</property>

<property>
<name>yarn.nodemanager.log-dirs</name>
<value>/hadoop/yarn/log</value>
</property>

<property>
<name>yarn.nodemanager.log.retain-seconds</name>
<value>604800</value>
</property>

<property>
<name>yarn.nodemanager.pmem-check-enabled</name>
<value>false</value>
</property>

<property>
<name>yarn.nodemanager.recovery.dir</name>
<value>/var/log/hadoop-yarn/nodemanager/recovery-state</value>
</property>

<property>
<name>yarn.nodemanager.recovery.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.nodemanager.recovery.supervised</name>
<value>true</value>
</property>

<property>
<name>yarn.nodemanager.remote-app-log-dir</name>
<value>/app-logs</value>
</property>

<property>
<name>yarn.nodemanager.remote-app-log-dir-suffix</name>
<value>logs</value>
</property>

<property>
<name>yarn.nodemanager.resource-plugins</name>
<value></value>
</property>

<property>
<name>yarn.nodemanager.resource-plugins.gpu.allowed-gpu-devices</name>
<value>auto</value>
</property>

<property>
<name>yarn.nodemanager.resource-plugins.gpu.docker-plugin</name>
<value>nvidia-docker-v1</value>
</property>

<property>
<name>yarn.nodemanager.resource-plugins.gpu.docker-plugin.nvidiadocker-
v1.endpoint</name>
<value>http://localhost:3476/v1.0/docker/cli</value>
</property>

<property>
<name>yarn.nodemanager.resource-plugins.gpu.path-to-discovery-executables</name>
<value></value>
</property>

<property>
<name>yarn.nodemanager.resource.cpu-vcores</name>
<value>6</value>
</property>

<property>
<name>yarn.nodemanager.resource.memory-mb</name>
<value>12288</value>
</property>

<property>
<name>yarn.nodemanager.resource.percentage-physical-cpu-limit</name>
<value>80</value>
</property>

<property>
<name>yarn.nodemanager.runtime.linux.allowed-runtimes</name>
<value>default,docker</value>
</property>

<property>
<name>yarn.nodemanager.runtime.linux.docker.allowed-container-networks</name>
<value>host,none,bridge</value>
</property>

<property>
<name>yarn.nodemanager.runtime.linux.docker.capabilities</name>
<value>
CHOWN,DAC_OVERRIDE,FSETID,FOWNER,MKNOD,NET_RAW,SETGID,SETUID,SETFCAP,
SETPCAP,NET_BIND_SERVICE,SYS_CHROOT,KILL,AUDIT_WRITE</value>
</property>

<property>
<name>yarn.nodemanager.runtime.linux.docker.default-container-network</name>
<value>host</value>
</property>

<property>
<name>yarn.nodemanager.runtime.linux.docker.privileged-containers.acl</name>
<value></value>
</property>

<property>
<name>yarn.nodemanager.runtime.linux.docker.privileged-containers.allowed</name>
<value>false</value>
</property>

<property>
<name>yarn.nodemanager.vmem-check-enabled</name>
<value>false</value>
</property>

<property>
<name>yarn.nodemanager.vmem-pmem-ratio</name>
<value>2.1</value>
</property>

<property>
<name>yarn.nodemanager.webapp.cross-origin.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.resourcemanager.address</name>
<value>xxxxxxx:8050</value>
</property>

<property>
<name>yarn.resourcemanager.admin.address</name>
<value>xxxxxx:8141</value>
</property>

<property>
<name>yarn.resourcemanager.am.max-attempts</name>
<value>2</value>
</property>

<property>
<name>yarn.resourcemanager.bind-host</name>
<value>0.0.0.0</value>
</property>

<property>
<name>yarn.resourcemanager.connect.max-wait.ms</name>
<value>900000</value>
</property>

<property>
<name>yarn.resourcemanager.connect.retry-interval.ms</name>
<value>30000</value>
</property>

<property>
<name>yarn.resourcemanager.fs.state-store.retry-policy-spec</name>
<value>2000, 500</value>
</property>

<property>
<name>yarn.resourcemanager.fs.state-store.uri</name>
<value> </value>
</property>

<property>
<name>yarn.resourcemanager.ha.enabled</name>
<value>false</value>
</property>

<property>
<name>yarn.resourcemanager.hostname</name>
<value>xxxxxxxx</value>
</property>

<property>
<name>yarn.resourcemanager.monitor.capacity.preemption.monitoring_interval</name>
<value>15000</value>
</property>

<property>
<name>yarn.resourcemanager.monitor.capacity.preemption.natural_termination_factor</name>
<value>1</value>
</property>

<property>
<name>yarn.resourcemanager.monitor.capacity.preemption.total_preemption_per_round</name>
<value>0.25</value>
</property>

<property>
<name>yarn.resourcemanager.nodes.exclude-path</name>
<value>/etc/hadoop/conf/yarn.exclude</value>
</property>

<property>
<name>yarn.resourcemanager.recovery.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.resourcemanager.resource-tracker.address</name>
<value>xxxxxxx:8025</value>
</property>

<property>
<name>yarn.resourcemanager.scheduler.address</name>
<value>xxxxxxxx:8030</value>
</property>

<property>
<name>yarn.resourcemanager.scheduler.class</name>
<value>org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler</value>
</property>

<property>
<name>yarn.resourcemanager.scheduler.monitor.enable</name>
<value>false</value>
</property>

<property>
<name>yarn.resourcemanager.state-store.max-completed-applications</name>
<value>${yarn.resourcemanager.max-completed-applications}</value>
</property>

<property>
<name>yarn.resourcemanager.store.class</name>
<value>org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore</value>
</property>

<property>
<name>yarn.resourcemanager.system-metrics-publisher.dispatcher.pool-size</name>
<value>10</value>
</property>

<property>
<name>yarn.resourcemanager.system-metrics-publisher.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.resourcemanager.webapp.address</name>
<value>xxxxxx:8088</value>
</property>

<property>
<name>yarn.resourcemanager.webapp.cross-origin.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.resourcemanager.webapp.delegation-token-auth-filter.enabled</name>
<value>false</value>
</property>

<property>
<name>yarn.resourcemanager.webapp.https.address</name>
<value>wxxxxxx:8090</value>
</property>

<property>
<name>yarn.resourcemanager.work-preserving-recovery.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.resourcemanager.work-preserving-recovery.scheduling-wait-ms</name>
<value>10000</value>
</property>

<property>
<name>yarn.resourcemanager.zk-acl</name>
<value>world:anyone:rwcda</value>
</property>

<property>
<name>yarn.resourcemanager.zk-address</name>
<value>xxxxxx:2181,xxxxxx:2181,xxxxxx:2181</value>
</property>

<property>
<name>yarn.resourcemanager.zk-num-retries</name>
<value>1000</value>
</property>

<property>
<name>yarn.resourcemanager.zk-retry-interval-ms</name>
<value>1000</value>
</property>

<property>
<name>yarn.resourcemanager.zk-state-store.parent-path</name>
<value>/rmstore</value>
</property>

<property>
<name>yarn.resourcemanager.zk-timeout-ms</name>
<value>10000</value>
</property>

<property>
<name>yarn.rm.system-metricspublisher.emit-container-events</name>
<value>true</value>
</property>

<property>
<name>yarn.scheduler.capacity.ordering-policy.priority-utilization.underutilized-preemption.enabled</name>
<value>false</value>
</property>

<property>
<name>yarn.scheduler.maximum-allocation-mb</name>
<value>12288</value>
</property>

<property>
<name>yarn.scheduler.maximum-allocation-vcores</name>
<value>6</value>
</property>

<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>64</value>
</property>

<property>
<name>yarn.scheduler.minimum-allocation-vcores</name>
<value>1</value>
</property>

<property>
<name>yarn.service.framework.path</name>
<value>/yarn/service-dep.tar.gz</value>
</property>

<property>
<name>yarn.system-metricspublisher.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.timeline-service.address</name>
<value>xxxxxx:10200</value>
</property>

<property>
<name>yarn.timeline-service.bind-host</name>
<value>0.0.0.0</value>
</property>

<property>
<name>yarn.timeline-service.client.max-retries</name>
<value>30</value>
</property>

<property>
<name>yarn.timeline-service.client.retry-interval-ms</name>
<value>1000</value>
</property>

<property>
<name>yarn.timeline-service.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.active-dir</name>
<value>/ats/active/</value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.app-cache-size</name>
<value>10</value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.cleaner-interval-seconds</name>
<value>3600</value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.done-dir</name>
<value>/ats/done/</value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.group-id-plugin-classes</name>
<value></value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.group-id-plugin-classpath</name>
<value></value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.retain-seconds</name>
<value>604800</value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.scan-interval-seconds</name>
<value>60</value>
</property>

<property>
<name>yarn.timeline-service.entity-group-fs-store.summary-store</name>
<value>org.apache.hadoop.yarn.server.timeline.RollingLevelDBTimelineStore</value>
</property>

<property>
<name>yarn.timeline-service.generic-application-history.store-class</name>
<value>org.apache.hadoop.yarn.server.applicationhistoryservice.NullApplicationHistoryStore</value>
</property>

<property>
<name>yarn.timeline-service.hbase-schema.prefix</name>
<value>prod.</value>
</property>

<property>
<name>yarn.timeline-service.hbase.configuration.file</name>
<value>file:///etc/yarn-hbase/conf/hbase-site.xml</value>
</property>

<property>
<name>yarn.timeline-service.hbase.coprocessor.jar.hdfs.location</name>
<value>file:///hadoop-yarn-client/timelineservice/hadoop-yarn-server-timelineservice-hbase-coprocessor.jar</value>
</property>

<property>
<name>yarn.timeline-service.http-authentication.simple.anonymous.allowed</name>
<value>true</value>
</property>

<property>
<name>yarn.timeline-service.http-authentication.type</name>
<value>simple</value>
</property>

<property>
<name>yarn.timeline-service.http-cross-origin.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.timeline-service.leveldb-state-store.path</name>
<value>/hadoop/yarn/timeline</value>
</property>

<property>
<name>yarn.timeline-service.leveldb-timeline-store.path</name>
<value>/hadoop/yarn/timeline</value>
</property>

<property>
<name>yarn.timeline-service.leveldb-timeline-store.read-cache-size</name>
<value>104857600</value>
</property>

<property>
<name>yarn.timeline-service.leveldb-timeline-store.start-time-read-cache-size</name>
<value>10000</value>
</property>

<property>
<name>yarn.timeline-service.leveldb-timeline-store.start-time-write-cache-size</name>
<value>10000</value>
</property>

<property>
<name>yarn.timeline-service.leveldb-timeline-store.ttl-interval-ms</name>
<value>300000</value>
</property>

<property>
<name>yarn.timeline-service.reader.webapp.address</name>
<value>xxxxxx:8198</value>
</property>

<property>
<name>yarn.timeline-service.reader.webapp.https.address</name>
<value>xxxxxx:8199</value>
</property>

<property>
<name>yarn.timeline-service.recovery.enabled</name>
<value>true</value>
</property>

<property>
<name>yarn.timeline-service.state-store-class</name>
<value>org.apache.hadoop.yarn.server.timeline.recovery.LeveldbTimelineStateStore</value>
</property>

<property>
<name>yarn.timeline-service.store-class</name>
<value>org.apache.hadoop.yarn.server.timeline.EntityGroupFSTimelineStore</value>
</property>

<property>
<name>yarn.timeline-service.ttl-enable</name>
<value>true</value>
</property>

<property>
<name>yarn.timeline-service.ttl-ms</name>
<value>2678400000</value>
</property>

<property>
<name>yarn.timeline-service.version</name>
<value>2.0</value>
</property>

<property>
<name>yarn.timeline-service.versions</name>
<value>1.5f,2.0f</value>
</property>

<property>
<name>yarn.timeline-service.webapp.address</name>
<value>xxxxxx:8188</value>
</property>

<property>
<name>yarn.timeline-service.webapp.https.address</name>
<value>xxxxxx:8190</value>
</property>

<property>
<name>yarn.webapp.api-service.enable</name>
<value>true</value>
</property>

<property>
<name>yarn.webapp.ui2.enable</name>
<value>true</value>
</property>

</configuration>

• Assignee: Shane Kumpf
• Reporter: Hsin-Liang Huang
• Created: 18/May/18 22:32
• Updated: 29/Jun/18 15:48
• Resolved: 22/Jun/18 23:28

Description

Hi,  I am running testcases on Yarn 2.6 and Yarn 3.0 and found out the performance seems like twice slower on Yarn 3.0, and the performance would get even slower if we acquire more containers.   I looked at the node manager logs on 2.6 vs 3.0.   Here is what I find below.  

On 2.6 ,  this is a life cycle of a specific container,  from beginning to end, it takes about  8 seconds (9:53:50 to 9:53:58). 

On 3.0: the life cycle of a specific container looks like this,  it takes 20 seconds to finish the same job.  (9:51:44 to 9:52:04)

 It seems like on 3.0, it spends an extra 5 seconds on monitor.ContinaerMonitorImpl  (marked in red) which doesn't happen in 2.6,  and also after the job is done, and the container is exiting,  on 3.0, it took 5 seconds to do that (9:51:59 to 9:52:04)  which on 2.6, it only took less than 1/.2 of the time. (9: 53:56 to 9:53:58).  

   Since we are running the same unit testcases and usually acquire more than 4 containers,  therefore, when it addess up all these extra seconds, it became a huge performance issue.  On 2.6, the unittest runs 7 hours whilc on 3.0, the same unitests runs 11 hours.  I was told this performance delay might be caused by Hadoop’s new monitoring system Timeline service v2.  Could someone take a look of this?   Thanks for any help on this!!

Issue Links

• is caused by: YARN-5366 Improve handling of the Docker container life cycle