Skip to content

DeployRabbitMQCluster.markdown

Latest commit

 

History

History
132 lines (91 loc) · 3.53 KB

DeployRabbitMQCluster.markdown

File metadata and controls

132 lines (91 loc) · 3.53 KB

Install and Basic Configuration

Reference: http://www.rabbitmq.com/clustering.html

Our cookbook installs and configures RabbitMQ on 3 nodes basically. But the 2nd and 3rd nodes didn't start up in cluster properly.

On 1st RabbitMQ node

[root@ctrlr0 rabbitmq]# rabbitmq-server --detached
[root@ctrlr0 rabbitmq]# rabbitmqctl cluster_status
Cluster status of node rabbit@ctrlr0 ...
[{nodes,[{disc,[rabbit@ctrlr0]},{ram,[rabbit@ctrlr1]}]},
 {running_nodes,[rabbit@ctrlr0]},
 {partitions,[]}]
...done.

    [root@ctrlr0 ~]# rabbitmqctl stop_app
    Stopping node rabbit@ctrlr0 ...
    ...done.
    [root@ctrlr0 ~]# rabbitmqctl start_app

On 2nd and 3rd RabbitMQ node

[root@ctrlr1 rabbitmq]# rabbitmqctl status
Status of node rabbit@ctrlr1 ...
Error: unable to connect to node rabbit@ctrlr1: nodedown

DIAGNOSTICS
===========

nodes in question: [rabbit@ctrlr1]

hosts, their running nodes and ports:
- ctrlr1: [{rabbitmqctl11004,60256}]

Stop all Existing Processes and Restart them

Do the following commands on 2nd and 3rd RabbitMQ nodes

[root@ctrlr1 ~]# sudo -u rabbitmq rabbitmqctl stop
Stopping and halting node rabbit@ctrlr1 ...
...done.

or

[root@ctrlr2 ~]# kill -9 `ps -ef | grep rabbitmq | awk '{print $2}'`
-bash: kill: (13951) - No such process

[root@ctrlr2 ~]# rabbitmq-server --detached

              RabbitMQ 3.1.5. Copyright (C) 2007-2013 GoPivotal, Inc.
  ##  ##      Licensed under the MPL.  See http://www.rabbitmq.com/
  ##  ##
  ##########  Logs: /var/log/rabbitmq/rabbit@ctrlr2.log
  ######  ##        /var/log/rabbitmq/rabbit@ctrlr2-sasl.log
  ##########
              Starting broker... completed with 7 plugins.
^Z
[1]+  Stopped                 rabbitmq-server --detached
[root@ctrlr2 ~]# bg
[1]+ rabbitmq-server --detached &

[root@ctrlr2 ~]# rabbitmqctl stop_app
Stopping node rabbit@ctrlr2 ...
...done.
[root@ctrlr2 ~]# rabbitmqctl join_cluster --ram rabbit@ctrlr0
Clustering node rabbit@ctrlr2 with rabbit@ctrlr0 ...
...done.
[root@ctrlr2 ~]# rabbitmqctl start_app

Check the Status

[root@ctrlr0 rabbitmq]# rabbitmqctl cluster_status
Cluster status of node rabbit@ctrlr0 ...
[{nodes,[{disc,[rabbit@ctrlr0]},{ram,[rabbit@ctrlr2,rabbit@ctrlr1]}]},
 {running_nodes,[rabbit@ctrlr0]},
 {partitions,[]}]
...done.

Restart RabbitMQ Cluster

rabbitmqctl eval 'application:stop(rabbitmq_management),application:start(rabbitmq_management).'

Deploy RabbitMQ Cluster in HAProxy

On 172.16.0.33 (ctrlr0), edit /etc/haproxy/haproxy.cfg

    listen rabbitmq_management
      bind 0.0.0.0:25672
      mode http
      log global
      balance  source
      #option httpchk
      option tcpka
      server ctrlr0 172.16.0.33:15672 check inter 2000 rise 2 fall 5
      server ctrlr1 172.16.0.34:15672 check inter 2000 rise 2 fall 5
      server ctrlr2 172.16.0.35:15672 check inter 2000 rise 2 fall 5

Restart HAProxy

    /etc/init.d/haproxy restart

Repeat the above on 34(ctrlr1)/ 35(ctrlr2)

Access on Web

    http://9.110.178.27/haproxy

Deploy RabbitMQ Cluster in Nginx

On 9.110.178.27 (management VM host), edit /etc/nginx/nginx.conf

    ...
            }
            location /elasticsearch {
              proxy_pass http://172.16.0.42:9200/;
            }
            location /rabbitmq/ {
              proxy_pass http://172.16.0.37:25672/;
    ...

Restart Nginx

    systemctl restart nginx

Access Management on Web

    http://9.110.178.27/rabbitmq/