How Ambari configration-recommend works?

[#1]

Hey all:

I'm working on the configuration tuning of some services like hive,hdfs,yarn,etc.

I found Ambari would set some configs automatically .For example ,when I change 'yarn-nodemanager.resource.memory-mb',several configs changed ,which is very useful .

I'd like to understand the mechanism of auto-recommend for services, but I found nowhere. Could someone help me ?Thanks a lot!

[#2 · Accepted solution]

@Henry Luo,

This is done by the ambari stack advisor. The stack advisor is invoked whenever a new service is added/deleted or any config change is made. The stack advisor recommends you the set of configs need to be used. If you want to tweak something than what is needed , you can do it in stack_advisor.py

https://github.com/apache/ambari/blob/trunk/ambari-server/src/main/resources/stacks/HDP/2.6/services...

.

If you are using HDP , then the path for stack_advisor.py is /var/lib/ambari-server/resources/stacks/HDP/{HDP-version}/services/stack_advisor.py.

You can make the changes in the file and it should work. You need to restart ambari server for the new changes to be reflected. Also ambari agents caches these files , so you need to restart agents as well for the latest changes to be picked up.

.

Additional reference : https://community.hortonworks.com/questions/141855/stack-advisor-how-to-use-it.html?childToView=1418...

Please "Accept" the answer if you find this helpful or revert back if you need help 🙂 .

.

-Aditya

[#3]

It's what I want.Thank you very much !