Comment 0 for bug 1512001

This affects stable/kilo.

ceilometer notification agent is running on 3 nodes with workload_partitioning as enabled, this is the configuration used:

[oslo_messaging_rabbit]
rabbit_hosts = ccp-c1-m1-mgmt:5672,ccp-c1-m2-mgmt:5672,ccp-c1-m3-mgmt:5672
rabbit_userid = ceilometer
rabbit_password = password
rabbit_use_ssl = False

[coordination]
backend_url = kazoo://ccp-c1-m1-mgmt:2181
check_watchers = 10.0
heartbeat = 1.0

[notification]
workload_partitioning=True
messaging_urls=rabbit://guest:password@ccp-c1-m1-mgmt:5672,ccp-c1-m2-mgmt:5672,ccp-c1-m3-mgmt:5672

notification agent log files from 3 nodes attached

generated notifiications from cinder service using cinder create command,
and verified rabbit connectivity as ok.

stack@ccp-c1-m2-mgmt:~$ sudo rabbitmqctl list_queues | grep notification
notifications.error 0
notifications.info 0

stack@ccp-c1-m2-mgmt:~$ sudo rabbitmqctl list_queues | grep ceilometer-pipe*
ceilometer-pipe-image_source:meter_sink.sample 6
ceilometer-pipe-meter_source:meter_sink.sample 6
ceilometer-pipe-volume_source:meter_sink.sample 6

stack@ccp-c1-m2-mgmt:~$ ceilometer meter-list
+------+------+------+-------------+---------+------------+
| Name | Type | Unit | Resource ID | User ID | Project ID |
+------+------+------+-------------+---------+------------+
+------+------+------+-------------+---------+------------+