RabbitMQ windows2016 镜像模式 haproxy+keepalived

节点信息

RabbitMQ-节点     ip  类型  系统
mq1    10.5.9.168 RAM WINDOWS2016
mq2    10.5.9.169 DISK  WINDOWS2016    
整体架构：mq1+mq2 两节点镜像 +haproxy（2节点+keepalived）

先安装otp_win64_22.2.exe,C:Program Fileserl10.6

设置环境变量

此电脑-->鼠标右键“属性”-->高级系统设置-->环境变量-->“新建”系统环境变量
添加环境变量：ERLANG_HOME  C:Program Fileserl10.6
然后双击系统变量path，点击“新建”，将将"%ERLANG_HOME%in"加入到path中。
然后在cmd窗口中运行，能看到版本号，说明erlang安装成功

C:UsersAdministrator>erl
Eshell V10.6  (abort with ^G)
1>
C:UsersAdministrator>echo %path%
C:Windowssystem32;C:Windows;C:WindowsSystem32Wbem;C:WindowsSystem32WindowsPowerShellv1.0;C:Program Fileserl10.6in;C:UsersAdministratorAppDataLocalMicrosoftWindowsApps;
C:UsersAdministrator>echo %path%
C:Windowssystem32;C:Windows;C:WindowsSystem32Wbem;C:WindowsSystem32WindowsPowerShellv1.0;C:Program Fileserl10.6in;C:UsersAdministratorAppDataLocalMicrosoftWindowsApps;
如果存在path不生效，看不到erl，可以手动执行set path=
set path=

再安装rabbitmq-server-3.7.13.exe（默认安装路径）

修改主机名字：C:WindowsSystem32driversetc
10.5.9.168 mq1
10.5.9.169 mq2

在cmd命令中执行ipconfig /flushdns
C:WindowsSystem32configsystemprofile.erlang.cookie 把第一台的.erlang.cookie拷贝到另外的相同路径
C:UsersAdministrator 同样

确认2台机器的4个文件的内容是一致的
将各节点的rabbitmq服务开启: rabbitmq-service start

C:UsersAdministrator>cd C:Program FilesRabbitMQ Server
abbitmq_server-3.7.13sbin
C:Program FilesRabbitMQ Server
abbitmq_server-3.7.13sbin>rabbitmq-service start
RabbitMQ 服务正在启动 .
RabbitMQ 服务已经启动成功。
选择其中一个节点（10.5.9.169）将其停止: rabbitmqctl stop_app
C:Program FilesRabbitMQ Server
abbitmq_server-3.7.13sbin>rabbitmqctl stop_app
Stopping rabbit application on node rabbit@mq2 ...
将10.15.9.169机器加入集群: rabbitmqctl join_cluster --ram rabbit@hostname
(ram 为内存节点, 默认情况下为disc磁盘节点) 注意此时的node在windows机器下面是大写的。
C:Program FilesRabbitMQ Server
abbitmq_server-3.7.13sbin>hostname ##在node1节点上查看主机名
mq1
C:Program FilesRabbitMQ Server
abbitmq_server-3.7.13sbin>rabbitmqctl join_cluster --ram rabbit@mq1
Clustering node rabbit@mq2 with rabbit@mq1
开启rabbitmq服务: rabbitmqctl start_app
C:Program FilesRabbitMQ Server
abbitmq_server-3.7.13sbin>rabbitmqctl start_app
Starting node rabbit@mq2 ...
 completed with 3 plugins.
查看集群状况: rabbitmqctl cluster_status
C:Program FilesRabbitMQ Server
abbitmq_server-3.7.13sbin>rabbitmqctl cluster_status
Cluster status of node rabbit@mq2 ...
[{nodes,[{disc,[rabbit@mq1]},{ram,[rabbit@mq2]}]},
 {running_nodes,[rabbit@mq1,rabbit@mq2]},
 {cluster_name,<<"rabbit@mq2">>},
 {partitions,[]},
 {alarms,[{rabbit@mq1,[]},{rabbit@mq2,[]}]}]
在节点1上查看集群状态
C:Program FilesRabbitMQ Server
abbitmq_server-3.7.13sbin>rabbitmqctl cluster_status
Cluster status of node rabbit@mq1 ...
[{nodes,[{disc,[rabbit@mq1]},{ram,[rabbit@mq2]}]},
 {running_nodes,[rabbit@mq2,rabbit@mq1]},
 {cluster_name,<<"rabbit@mq2">>},
 {partitions,[]},
 {alarms,[{rabbit@mq2,[]},{rabbit@mq1,[]}]}]

---以上是单个节点加入集群的方式，只要一个节点加入到集群中的任何一个节点，该节点就算是加入到了集群中.

配置镜像集群

http://10.5.9.168:15672
http://localhost:15672/#/

guest,guest登录管理端
镜像模式可以在web上配置，登录mq：http://localhost:15672/#/
选择“admin”，选择“policies”，在add/update a policy中添加：
name:ha-all
pattern：^
apply to: Exchanges and queues
priority:0
definition: ha-mode=all
选择“add policy”创建镜像模式
模拟创建一个队列queue ，看队列是否会自动的复制到另外一个节点上面
在web界面操作：模拟在node1节点上面创建一个队列queue
选择“queues”，选择“add a new queue”
name:hq-test
durability:durable
node:rabbit@mq1
auto delete:no
选择“add queue”，在queues页面查看，添加的队列存在2个节点上面

创建queues并查看，2个节点都存在

===haproxy+keepalived

#节点1
[root@localhost ~]# ifconfig
eth0      Link encap:Ethernet  HWaddr 00:50:56:88:F1:B9 
[root@localhost ~]# rpm -qa|grep haproxy
haproxy-1.5.18-1.el6.x86_64
[root@localhost ~]# rpm -qa|grep keepalived
keepalived-1.2.13-5.el6_6.x86_64

[root@localhost haproxy]# vim haproxy.cfg
### haproxy 监控页面地址是：http://10.5.9.171:8100/haproxy_status
listen admin_stats
    bind *:8100
    mode http
    log 127.0.0.1 local3 err
    stats refresh 60s
    stats uri /haproxy_status
    stats realm welcome login Haproxy
    stats auth admin:admin123456
    stats hide-version
    stats admin if TRUE

listen rabbitmq_admin
    bind 0.0.0.0:15672
    server mq1 10.5.9.168:15672
    server mq2 10.5.9.169:15672
    
listen rabbitmq_cluster
    bind 0.0.0.0:5672
    mode tcp
    balance roundrobin
    server mq1 10.5.9.168:5672 check inter 2000 rise 2 fall 3 weight 1
    server mq2 10.5.9.169:5672 check inter 2000 rise 2 fall 3 weight 1

[root@localhost haproxy]# ip a
[root@localhost haproxy]#  ps -ef|grep haproxy
[root@localhost haproxy]# haproxy -f /etc/haproxy/haproxy.cfg 
[WARNING] 077/133130 (20245) : parsing [/etc/haproxy/haproxy.cfg:45] : 'option httplog' not usable with proxy 'rabbitmq_cluster' (needs 'mode http'). Falling back to 'option tcplog'.
[WARNING] 077/133130 (20245) : config : 'option forwardfor' ignored for proxy 'rabbitmq_cluster' as it requires HTTP mode.
[ALERT] 077/133130 (20245) : Starting proxy rabbitmq_admin: cannot bind socket [10.5.9.168:15672]
解决报错
[root@localhost haproxy]# vim /etc/sysctl.conf
net.ipv4.ip_nonlocal_bind = 1
[root@localhost haproxy]# sysctl -p
[root@localhost haproxy]# haproxy -f /etc/haproxy/haproxy.cfg
[WARNING] 077/133427 (20252) : parsing [/etc/haproxy/haproxy.cfg:45] : 'option httplog' not usable with proxy 'rabbitmq_cluster' (needs 'mode http'). Falling back to 'option tcplog'.
[WARNING] 077/133427 (20252) : config : 'option forwardfor' ignored for proxy 'rabbitmq_cluster' as it requires HTTP mode.
[root@localhost haproxy]#  ps -ef|grep haproxy
haproxy  20253     1  0 13:34 ?        00:00:00 haproxy -f /etc/haproxy/haproxy.cfg
root     20255 20145  0 13:34 pts/0    00:00:00 grep haproxy
#重新载入配置文件
[root@localhost haproxy]# service haproxy reload
[root@localhost keepalived]# more /etc/keepalived/haproxy_check.sh
#!/bin/bash
A=`ps -C haproxy --no-header |wc -l`
if [ $A -eq 0 ];then
/etc/init.d/keepalived stop
fi
[root@localhost keepalived]# /etc/init.d/keepalived restart
#查看keepalived日志
[root@localhost keepalived]# tail -f -n 30 /var/log/messages
Mar 18 13:42:30 localhost Keepalived_vrrp[20281]: VRRP sockpool: [ifindex(2), proto(112), unicast(1), fd(10,11)]
Mar 18 13:42:30 localhost Keepalived_healthcheckers[20280]: Using LinkWatch kernel netlink reflector...
Mar 18 13:42:31 localhost Keepalived_vrrp[20281]: VRRP_Instance(VI_1) Transition to MASTER STATE
Mar 18 13:42:31 localhost Keepalived_vrrp[20281]: VRRP_Instance(VI_1) Received lower prio advert, forcing new election
Mar 18 13:42:32 localhost Keepalived_vrrp[20281]: VRRP_Instance(VI_1) Entering MASTER STATE
Mar 18 13:42:32 localhost Keepalived_vrrp[20281]: VRRP_Instance(VI_1) setting protocol VIPs.
Mar 18 13:42:32 localhost Keepalived_vrrp[20281]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 10.5.9.188
Mar 18 13:42:32 localhost Keepalived_healthcheckers[20280]: Netlink reflector reports IP 10.5.9.188 added
Mar 18 13:42:37 localhost Keepalived_vrrp[20281]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 10.5.9.188