Issue with Yarn jobs (accepted state due to AM container)

[#1]

Hi,

We have an issue with yarn jobs and the number that can be run concurrently . There is a case that 6 TEZ jobs needs to be run concurrently but they all go to accepted state.

I uploaded the following information for your reference. I tried different configuration but still no luck. Please advise me

hive-site.xml under /usr/hdp/current/hive-server2/conf

Tex-site.xml under /usr/hdp/current/tez-client

Yarn-site.xml and Mapr-site.xml under /usr/hdp/current/hadoop-yarn-resourcemanager/etc/hadoop

Also

Yarn queue manager

Thanks

SJ

[#2]

Also please find the following attatch

acceptedjob.png

accpeted-application-in-rm-ui.png

[#3 · Accepted solution]

Hi @Sanaz Janbakhsh,

Please try increasing the number of node managers and try running the jobs.

Also look for the parameters under Scheduler section of Yarn (yarn.scheduler.capacity.maximum-applications / yarn.scheduler.capacity.<queue-path>.maximum-application) and (yarn.scheduler.capacity.maximum-am-resource-percent / yarn.scheduler.capacity.<queue-path>.maximum-am-resource-percent) which will restrict the number of concurrent running and pending apps.

Please look at the link for more info.

https://hadoop.apache.org/docs/r2.7.3/hadoop-yarn/hadoop-yarn-site/CapacityScheduler.html

Thanks,

Aditya

[#4]

Hi Aditya,

It was really helpful. Thanks.