RabbitMQ doesn't start after master node reset
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Invalid
|
High
|
Matthew Mosesohn |
Bug Description
After hard reset of master node, RabbitMQ looks started but doesn't work properly:
[root@d075c6249142 rabbitmq]# rabbitmqctl list_queues
Listing queues ...
Error: unable to connect to node rabbit@fuel: nodedown
DIAGNOSTICS
===========
nodes in question: [rabbit@fuel]
hosts, their running nodes and ports:
- unable to connect to epmd on fuel: nxdomain (non-existing domain)
current node details:
- node name: rabbitmqctl1149
- home dir: /var/lib/rabbitmq
- cookie hash: 2xPc5aEDV5CIGSl
[root@d075c6249142 rabbitmq]# hostname
d075c6249142
[root@d075c6249142 rabbitmq]# exit
[root@fuel ~]# hostname
fuel.domain.tld
If I try to restart it:
[root@d075c6249142 ~]# /etc/init.
Restarting rabbitmq-server: RabbitMQ is not running
RabbitMQ server startup failed on retry 1 of 60!
RabbitMQ server startup failed on retry 2 of 60!
RabbitMQ server startup failed on retry 3 of 60!
RabbitMQ server startup failed on retry 4 of 60!
RabbitMQ server startup failed on retry 5 of 60!
RabbitMQ server startup failed on retry 6 of 60!
RabbitMQ server startup failed on retry 7 of 60!
....
[root@d075c6249142 ~]# cat /var/log/
ERROR: epmd error for host "fuel": nxdomain (non-existing domain) [FAILED]
I guess it can be related to hostname.
{"build_id": "2014-05-
description: | updated |
tags: | added: ha |
tags: | removed: ha |
Changed in fuel: | |
status: | Confirmed → Triaged |
importance: | Undecided → High |
Changed in fuel: | |
status: | Incomplete → Invalid |
it is 100% related to the hostname. you should use hostname as -n argument value when running rabbitmqctl and you should not use init script as it is not how rabbitmq should be restarted.