解决supervisor节点启动失败apache.zookeeper.KeeperException$NodeExistsException: KeeperErrorCode = NodeExists

问题描述：

Ambari UI 启动Storm

过一会即Supervisors 挂掉一个
查看对应节点的supervisor日志 /var/log/storm/supervisor.log

tail -n 5000 supervisor.log | grep ERROR -A 10

其中一个条错误

2020-04-28 08:56:36.675 o.a.s.d.s.Supervisor main [ERROR] Failed to start supervisor

java.lang.RuntimeException: org.apache.storm.shade.org.apache.zookeeper.KeeperException$NodeExistsException: KeeperErrorCode = NodeExists for /supervisors /8210c2a5-f8da-4ef6-910b-7fe9f95f1acb
at org.apache.storm.utils.Utils.wrapInRuntime(Utils.java:1532) ~[storm-core-1.1.0.2.6.2.0-205.jar:1.1.0.2.6.2.0-205]
at org.apache.storm.zookeeper.Zookeeper.createNode(Zookeeper.java:149) ~[storm-core-1.1.0.2.6.2.0-205.jar:1.1.0.2.6.2.0-205]
at org.apache.storm.cluster.ZKStateStorage.set_ephemeral_node(ZKStateStorage.java:150) ~[storm-core-1.1.0.2.6.2.0-205.jar:1.1.0.2.6.2.0-205]
at org.apache.storm.cluster.StormClusterStateImpl.supervisorHeartbeat(StormClusterStateImpl.java:419) ~[storm-core-1.1.0.2.6.2.0-205.jar:1.1.0.2.6.2.0-205]
at org.apache.storm.daemon.supervisor.timer.SupervisorHeartbeat.run(SupervisorHeartbeat.java:85) ~[storm-core-1.1.0.2.6.2.0-205.jar:1.1.0.2.6.2.0-205]
at org.apache.storm.daemon.supervisor.Supervisor.launch(Supervisor.java:202) ~[storm-core-1.1.0.2.6.2.0-205.jar:1.1.0.2.6.2.0-205]
at org.apache.storm.daemon.supervisor.Supervisor.launchDaemon(Supervisor.java:255) ~[storm-core-1.1.0.2.6.2.0-205.jar:1.1.0.2.6.2.0-205]
at org.apache.storm.daemon.supervisor.Supervisor.main(Supervisor.java:374) ~[storm-core-1.1.0.2.6.2.0-205.jar:1.1.0.2.6.2.0-205]

解决办法：

删除zookeeper中 /storm/supervisors/8210c2a5-f8da-4ef6-910b-7fe9f95f1acb 目录

/usr/hdp/2.6.2.0-205/zookeeper/bin/zkCli.sh
rmr /storm/supervisors/8210c2a5-f8da-4ef6-910b-7fe9f95f1acb

重启storm 即可