ZooKeeper集群环境介绍：

bigdata131 192.168.126.131
bigdata132 192.168.126.132
bigdata133 192.168.126.133

1.启动ZooKeeper集群并查看状态

[root@bigdata131 ~]# zkServer.sh start
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
[root@bigdata131 ~]# jps
1186 QuorumPeerMain
1213 Jps
[root@bigdata131 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: follower

[root@bigdata132 ~]# zkServer.sh start
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
[root@bigdata132 ~]# jps
1187 QuorumPeerMain
1220 Jps
[root@bigdata132 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: leader

[root@bigdata133 ~]# zkServer.sh start
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
[root@bigdata133 ~]# jps
1184 QuorumPeerMain
1212 Jps
[root@bigdata133 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: follower

可以看到：初始化时bigdata132上启动的是leader，bigdata131和bigdata133上启动的是follower。

2.杀死leader进程后再查看状态

[root@bigdata132 ~]# kill -9 1187
[root@bigdata132 ~]# jps
1263 Jps
[root@bigdata132 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Error contacting service. It is probably not running.

[root@bigdata131 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: follower

[root@bigdata133 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: leader

可以看到：杀死bigdata132上的leader后，ZooKeeper通过内部的选举算法将bigdata133上的follower选举成为新的leader。

3.再次杀死leader进程后再查看状态

[root@bigdata133 ~]# kill -9 1184
[root@bigdata133 ~]# jps
1292 Jps
[root@bigdata133 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Error contacting service. It is probably not running.

[root@bigdata131 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Error contacting service. It is probably not running.
[root@bigdata131 ~]# jps
1186 QuorumPeerMain
1322 Jps

可以看到：杀死bigdata133上的leader后，bigdata131上的follower并没有变成leader，而是出错了，但是进程还在。

4.重启一台ZooKeeper服务器

[root@bigdata132 ~]# zkServer.sh start
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
[root@bigdata132 ~]# jps
1317 QuorumPeerMain
1342 Jps
[root@bigdata132 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: follower

[root@bigdata131 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: leader

可以看到：重新启动bigdata132上的ZooKeeper后，bigdata131上的ZooKeeper又可以用了，变成了新的leader，bigdata132上的成了follower。

5.再重启一台ZooKeeper服务器

[root@bigdata133 ~]# zkServer.sh start
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
[root@bigdata133 ~]# jps
1386 Jps
1359 QuorumPeerMain
[root@bigdata133 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: follower

[root@bigdata131 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: leader

[root@bigdata132 ~]# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /root/trainings/zookeeper-3.4.10/bin/../conf/zoo.cfg
Mode: follower

可以看到：bigdata133上新启动的的ZooKeeper直接作为follower存在，bigdata131还是leader，bigdata132还是follower。

总结：

• ZooKeeper集群的节点个数不一定非得是2n+1，但是最好是2n+1，好处是可以有效减少选举的投票次数；
• ZooKeeper集群能够工作的节点数不能少于2，否则将出现没有leader的情况，比如step3的情形。注意和ZooKeeper StandAlone模式的区别：StandAlone能够工作的节点数是1。
• ZooKeeper集群恢复的节点直接作为follower存在，不能影响恢复前的集群状态，因为恢复前的leader可能正在对外提供服务。
• ZooKeeper集群内部选举leader使用端口3888，ZooKeeper集群内部各个节点间的数据同步使用端口2888，ZooKeeper对外提供服务使用端口2181。