Nodemanager dont start and hive client is taking so much time to start.

[#1]

I have a 4 node cluster with 3 nodemanagers. When i am trying to start nodemanagers they are going up and shutting down immediately. Only one node manager is working. When i am trying to start my hive client it is taking so much time to start.

[#2]

when ever you initiate it will try to get min containers thats the reason it will be in hung up state.

[#3]

so how to allocate the minimum container to overcome this issue

[#4]

@R c

This is a good starting point link

[#5]

That did not help can you be more clear. My question is : I have 3 node managers in the cluster of which i am able to start only one. Remaining two node managers come up without an error and shut down immediately. I have followed all the memory allocations as per HDP

[#6]

@R c

what is the value yarn.nodemanager.aux-services in your yarn-site.xml file

[#7]

mapreduce_shuffle,spark_shuffle

[#8 · Accepted solution]

@R c

First make a backup of yarn-site.xml then change it to mapreduce_shuffle and retest. Remember to restart the all the services on the node

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

To size better your memory have a look at HDP Memory Configuration Settings

[#9]

This Worked for me. thank you can you explain the reason for this which would really help.