Review Request 42279: AMBARI-14657 : Stop All services fails at AMS on cluster with NN HA enabled, with non root Ambari user and AMS in distributed mode.

[Aravindan Vijayan <av...@hortonworks.com>]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42279/
-----------------------------------------------------------

Review request for Ambari, Dmytro Sen, Sumit Mohanty, and Sid Wagle.


Bugs: AMBARI-14657
    https://issues.apache.org/jira/browse/AMBARI-14657


Repository: ambari


Description
-------

PROBLEM
On a cluster with NN HA enabled, non root Ambari user and AMS in distributed mode, STOP ALL services command fails at AMS.

BUG
Wrong role command ordering between Journal Node and AMS caused Journal Node to be stopped before AMS. Hence, graceful shut down of AMS HBase in distributed mode failed. The fall back option to kill the HBase daemon also failed because of insufficient permissions for the non root ambari user.

FIX
Fix the role command ordering and use the ambari sudo helper to give permissions to the fallback kill command.


Diffs
-----

  ambari-server/src/main/resources/common-services/AMBARI_METRICS/0.1.0/package/scripts/ams_service.py 3d1ffda 
  ambari-server/src/main/resources/common-services/AMBARI_METRICS/0.1.0/package/scripts/hbase_service.py 5f03ca0 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/role_command_order.json f5ecd7b 
  ambari-server/src/test/python/stacks/2.0.6/AMBARI_METRICS/test_metrics_collector.py 52bac97 

Diff: https://reviews.apache.org/r/42279/diff/


Testing
-------

Manual testing done to trigger both the issues (Role command ordering and Sudo) sepearately and together.

ambari-server unit tests pass.

Submitted patch through Apache.

Java unit tests running locally.

Manual testing being done with AMS embedded mode to make sure no regression has been caused.


Thanks,

Aravindan Vijayan