{
"api": "1.0",
"astute_sha": "aa5aed61035a8dc4035ab1619a8bb540a7430a95",
"auth_required": true,
"build_id": "2014-07-28_10-30-48",
"build_number": "367",
"feature_groups": [
"mirantis",
"experimental"
],
"fuellib_sha": "d1c7f67b3cf51978d3178c8666ea398f2477dcb5",
"fuelmain_sha": "9adfbf5a52cedbdd16ec1a74f6c44c5b3419b87c",
"nailgun_sha": "83cc9ed44ebc8dd97248483b6d414ebbc4cff3c0",
"ostf_sha": "8c328521b1444f22c50463b9432193e20ed33813",
"production": "docker",
"release": "5.1"
}
Steps to reproduce:
1) Deploy any HA env (ubuntu or centos)
2) Run the following commands on any controller after deployment:
rabbitmq-plugins enable rabbitmq_management
rabbitmq-plugins list
crm resource restart p_rabbitmq-server
Expected result:
Plugins are loaded, rabbitmq listens 15672 and 55672 extra ports. "grep plugins /var/log/rabbitmq/ -r" shows something like this:
Server startup complete; 6 plugins started
Actual result:
Plugins are not loaded, rabbitmq listens only 41055 and 5673 ports. "grep plugins /var/log/rabbitmq/ -r" shows zero plugins:
Server startup complete; 0 plugins started.
It looks like RABBITMQ_NODE_ONLY environment variable is the problem here.
RABBIT* environment vairables for the "beam" process started via init script (with working plugins): PID_FILE= /var/run/ rabbitmq/ pid NODE_PORT= 5673
RABBITMQ_
RABBITMQ_
RABBIT* environment vairables for the "beam" process started via OCF (with not working plugins): NODE_ONLY= 1 NODENAME= rabbit@ node-4 PID_FILE= /var/run/ rabbitmq/ pid NODE_PORT= 5673
RABBITMQ_
RABBITMQ_
RABBITMQ_
RABBITMQ_
For the experiment purpose I've remove RABBITMQ_ NODE_ONLY= 1 from OCF script and executed "crm resource restart p_rabbitmq-server" - rabbit has started with plugins:
/<email address hidden>: Server startup complete; 6 plugins started.