Rabbitmq HAProxy高可用集群部署

RabbitMQ　HAProxy高可用配置

1. 修改三台服务器 PC, node1, node2 的hosts文件，可以解析hostname到ip

2. 设置三台服务器的 erlang Cookie(文件地址：/var/lib/rabbitmq/.erlang.cookie)相同，只有这样不同服务器的RabbitMQ才可以通信

3. 在RabbitMQ的配置文件中设置集群配置，这样在RabbitMQ启动时就会自动加入进去，不需要每次手动设置

如node1节点的配置 cat /etc/rabbitmq/rabbitmq.conf

[
  {rabbit,
    [{cluster_nodes, {['rabbit@PC', 'rabbit@node2'],disc}}]
  }
].

配置完成后，启动RabbimtMQ，查看集群状态

root@PC:~# rabbitmqctl cluster_status
Cluster status of node rabbit@PC
[{nodes,[{disc,[rabbit@PC,rabbit@node1,rabbit@node2]}]},
 {running_nodes,[rabbit@node1,rabbit@node2,rabbit@PC]},
 {cluster_name,<<"rabbit@PC">>},
 {partitions,[]},
 {alarms,[{rabbit@node1,[]},{rabbit@node2,[]},{rabbit@PC,[]}]}]

4. 配置成RabbitMQ启动后自动加入集群，RabbitMQ集群中各个节点除了队列，数据/状态都是各自保存相同的副本。消息队列只存储在创建该队列的节点上。一旦该服务器宕掉，就无法再读取该队列的消息，因此这样的RabbitMQ集群并不具备高可用性。为了高可用，还需要创建镜像队列。

创建镜像队列

rabbitmqctl set_policy pigeon_policy "^pigeon" '{"ha-mode":"nodes","ha-params":["rabbit@node1", "rabbit@node2"]}'

5. 安装配置HAProxy,并修改配置文件如下：

cat /etc/haproxy/haproxy.cfg

global
  log 127.0.0.1 local0 notice

  maxconn 10000
  user haproxy
  daemon

defaults
  timeout connect 5s
  timeout client 100s
  timeout server 100s

listen rabbitmq
  bind :5673
  mode tcp
  balance roundrobin
  server PC 172.16.20.157:5672 check inter 5s rise 2 fall 3
  server node2 172.16.20.61:5672 check inter 5s rise 2 fall 3


frontend front_rabbitmq_management
  bind :15672
  default_backend back_rabbitmq_management

backend back_rabbitmq_management
  balance source
  server PC PC:15672 check
  server node2 node2:15672 check

如果连接HAProxy失败，看下是否是端口不通,如果不通则开通5673端口

firewall-cmd --permanent --add-port=5673/tcp
firewall-cmd --reload

---

１．如果服务器安装后修改了hostname，出现了　unable to connect to epmd (port 4369) on bogon: nxdomain (non-existing domain)　这个问题，请看下是否 /etc/hosts 文件、/etc/hostname文件和环境变量$HOSTNAME　是否一致。

2. 如果启动HAProxy出现　/etc/rc.d/init.d/haproxy: line 26: [: =: unary operator expected　这个错误，修改/etc/init.d/haproxy 文件的26行 [ ${NETWORKING} = "no" ] && exit 0 为　[ "${NETWORKING}" = "no" ] && exit 0

3. 如果Ubuntu 出现如下错误

    rabbit@PC:
      * connected to epmd (port 4369) on PC
      * epmd reports: node 'rabbit' not running at all
                      no other nodes on PC
      * suggestion: start the node

    current node details:
    - node name: 'rabbitmq-cli-15@PC'
    - home dir: /var/lib/rabbitmq
    - cookie hash: 3N+Ub3klGzwQE0bRAMczGA==

则 rm /var/log/rabbitmq/*, ps axu|grep rabbimtq　删掉rabbitmq进程，运行rabbitmq-server restart

4. 如果出现如下错误:

    Error: {cannot_start_mnesia,
               {{shutdown,{failed_to_start_child,mnesia_kernel_sup,killed}},
                {mnesia_app,start,[normal,[]]}}}

则 rm -rf /var/lib/rabbitmq/mnesia, rm -rf /var/log/rabbitmq/* 然后重启rabbitmq server

5.如果出现如下错误:

[root@node2 ~]# rabbitmqctl status
Status of node rabbit@node2
Error: unable to connect to node rabbit@node2: nodedown

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

attempted to contact: [rabbit@node2]

rabbit@node2:
  * unable to connect to epmd (port 4369) on node2: address (cannot connect to host/port)

在/etc/hosts文件里面添加解析到本地hostname的条目；如 192.168.22.55 myPC