我目前正在尝试管理一个openmq集群（使用glassfish 3.0.1），我遇到了一些奇怪的行为。
集群工作了7个月，现在注册了2个经纪人没有任何问题。
我现在需要暂时在群集中添加另外两个代理。以下是我以前的情况：
-------------------------
Host Primary Port
-------------------------
localhost 7676
Cluster ID MyCluster
Cluster is Highly Available true
-------------------------------------------------------------------------------------------------------------
ID of broker Time since last
Broker ID Address State Msgs in store performing takeover status timestamp
-------------------------------------------------------------------------------------------------------------
Broker1 192.168.0.1:7676 OPERATING 5 6 seconds
Broker2 192.168.0.2:7676 OPERATING 8 6 seconds
然后，我在另外两台服务器上成功启动了另外两个代理，我得到了：
-------------------------
Host Primary Port
-------------------------
localhost 7676
Cluster ID MyCluster
Cluster is Highly Available true
-------------------------------------------------------------------------------------------------------------
ID of broker Time since last
Broker ID Address State Msgs in store performing takeover status timestamp
-------------------------------------------------------------------------------------------------------------
Broker1 192.168.0.1:7676 OPERATING 5 6 seconds
Broker2 192.168.0.2:7676 OPERATING 8 6 seconds
Broker3 192.168.0.3:7676 OPERATING 5 6 seconds
Broker4 192.168.0.4:7676 OPERATING 8 6 seconds
应用程序运行良好，配置自动使用两个新的代理。使用以下命令从集群停止代理时会发生此问题：
./imqcmd shutdown bkr
在其中一台服务器上。 ./imqcmd list bkr的结果如下：
-------------------------
Host Primary Port
-------------------------
localhost 7676
Cluster ID MyCluster
Cluster is Highly Available true
-------------------------------------------------------------------------------------------------------------
ID of broker Time since last
Broker ID Address State Msgs in store performing takeover status timestamp
-------------------------------------------------------------------------------------------------------------
Broker1 192.168.0.1:7676 OPERATING 5 6 seconds
Broker2 192.168.0.2:7676 OPERATING 8 6 seconds
Broker3 192.168.0.3:7676 TAKEOVER_COMPLETE 0 Broker1 6 seconds
Broker4 192.168.0.4:7676 OPERATING 8 6 seconds
一切似乎都没问题，接管是由Broker1执行的，但是当我查看glassfish的server.log时，我发现了以下行：
[C4003]: Error occurred on connection creation [192.168.0.3:7676]. - cause: java.net.ConnectException: Connection refused|#]
就像玻璃鱼试图连接到倒下的经纪人一样。
我错过了什么吗？
感谢您的帮助。
答案 0 :(得分：0)
缺少的命令行是：
imqdbmgr remove bkr -n Broker3
then a list will ouput:
-------------------------------------------------------------------------------------------------------------
ID of broker Time since last
Broker ID Address State Msgs in store performing takeover status timestamp
-------------------------------------------------------------------------------------------------------------
Broker1 192.168.0.1:7676 OPERATING 5 6 seconds
Broker2 192.168.0.2:7676 OPERATING 8 6 seconds
Broker4 192.168.0.4:7676 OPERATING 8 6 seconds
Broker3不再在代理HA集群中注册。