Support Questions
Find answers, ask questions, and share your expertise

Roles and Flume agent setup

Ambari supports users with roles like cluster level cluster user, service operator, ambari level administrator etc (see https://issues.apache.org/jira/browse/AMBARI-11350)

It should be possible for a cluster user to setup flume agents since it seems more like a "data worker" task.

I'm assuming that a typical data worker (submits MR jobs, sets up data ingestion etc) would belong to the "Cluster User" category.

Currently to use flume, the flume agent needs to be setup which is a configuration change (flume.conf) kind of operation.

Cluster users and Service Operators cannot perform configuration changes and hence cannot setup flume agents.

Can someone clarify the following -

1) What role is typically assigned to the data workers ?

2) Currently flume agent setup can only be done by operators or administrator. Is there a plan to change this?

3) How is flume typically setup in the field across different users and roles ?

2 REPLIES 2

@KRISHNANAND THOMMAND

1) What role is typically assigned to the data workers ?

Datanode

Nodemanager

Supervisor in case of storm

Region Servers

check out http://www.slideshare.net/hortonworks/ops-workshop-asrunon20150112 "slide 10"

2) Currently flume agent setup can only be done by operators or administrator. Is there a plan to change this?

Not sure why developers or DS needs to setup Flume agents. Personally, I don't want end users to take control of setting up agents and start running data ingestion jobs "Security and can cause human errors/chaos"

3) How is flume typically setup in the field across different users and role

Admin has access to setup agents and help end users on data ingestion.

http://hortonworks.com/hadoop/flume

Mentor

@KRISHNANAND THOMMAND are you still having issues with this? Can you accept best answer or provide your own solution?