v95集群节点替换回退操作

(1)需要利用python接口，调用一次gcware.flush_statemachine()
      备注：该刷新命令有时候需要执行2次才重启后见效果
    [gbase@node161 ~]$ python
        Python 2.7.5 (default, Oct 11 2015, 17:47:16) 
        [GCC 4.8.3 20140911 (Red Hat 4.8.3-9)] on linux2
        Type "help", "copyright", "credits" or "license" for more information.
        >>> import gcware
        >>> gcware.flush_statemachine()
        1
        >>> gcware.flush_statemachine()
        1
        >>> exit()

    (2)停止整个集群
    (3)找到最新的snapshot目录，其后数字依次最大的
    (4)如果是coor节点：对应文件为statemachine/COORDINATORSTATE
        cat COORDINATORSTATE
         {
              "nodeid":-1560172352,
              "nodestate":3,
              "gclusterstate":1,
              "gcwarestate":1
            }
        将 "nodestate":3  改为 "nodestate":1