Apache storm Supervisor routinely shutting down worker - apache-storm

Question

I made topology in Apache Storm(0.9.6) with kafka-storm, zookeeper(3.4.6)
(3 zookeeper each node, and 3 supervisor each node. operate 3 topology)
I add 2 storm&zookeeper nodes and change topology.worker configuration 3 to 5.
But after 2 nodes, storm supervisor routinely shutting down worker. Checked with iostat command, read and write throughput is under 1mb.
In supervisor log, show like below.
2016-10-19T15:07:38.904+0900 b.s.d.supervisor [INFO] Shutting down and clearing state for id ee13ada9-641e-463a-9be5-f3ed66fdb8f3. Current supervisor time: 1476857258. State: :timed-out, Heartbeat: #backtype.storm.daemon.common.WorkerHeartbeat{:time-secs 1476857226, :storm-id "top3-17-1476839721", :executors #{[36 36] [6 6] [11 11] [16 16] [21 21] [26 26] [31 31] [-1 -1] [1 1]}, :port 6701}
2016-10-19T15:07:38.905+0900 b.s.d.supervisor [INFO] Shutting down b278933f-f9c7-4189-b615-1d70c7988f17:ee13ada9-641e-463a-9be5-f3ed66fdb8f3
2016-10-19T15:07:38.907+0900 b.s.util [INFO] Error when trying to kill 9306. Process is probably already dead.
2016-10-19T15:07:44.948+0900 b.s.d.supervisor [INFO] Shutting down and clearing state for id d6df820a-7c29-4bff-a606-9e8e36fafab2. Current supervisor time: 1476857264. State: :disallowed, Heartbeat: #backtype.storm.daemon.common.WorkerHeartbeat{:time-secs 1476857264, :storm-id "top3-17-1476839721", :executors #{[-1 -1]}, :port 6701}
2016-10-19T15:07:44.949+0900 b.s.d.supervisor [INFO] Shutting down b278933f-f9c7-4189-b615-1d70c7988f17:d6df820a-7c29-4bff-a606-9e8e36fafab2
2016-10-19T15:07:45.954+0900 b.s.util [INFO] Error when trying to kill 11171. Process is probably already dead.
2016-10-19T15:07:45.954+0900 b.s.d.supervisor [INFO] Shut down b278933f-f9c7-4189-b615-1d70c7988f17:d6df820a-7c29-4bff-a606-9e8e36fafab2
And in zookeeper.out log... show like below(xxx ip address is another storm zookeeper address)
2016-09-20 02:31:06,031 [myid:5] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn#1007] - Closed socket connection for client /xxx.xxx.xxx.xxx:39426 which had sessionid 0x5574372bbf00004
2016-09-20 02:31:08,116 [myid:5] - WARN [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn#357] - caught end of stream exception
EndOfStreamException: Unable to read additional data from client sessionid 0x5574372bbf0000a, likely client has closed socket
at org.apache.zookeeper.server.NIOServerCnxn.doIO(NIOServerCnxn.java:228)
at org.apache.zookeeper.server.NIOServerCnxnFactory.run(NIOServerCnxnFactory.java:208)
at java.lang.Thread.run(Thread.java:745)
I don't know why worker is down routinely. How can i fix it? Is this something wrong?
Oh, my zookeeper and storm configuration is like below
zoo.cfg(same all nodes)
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/log/zkdata/1
clientPort=2181
server.1=storm01:2888:3888
server.2=storm02:2888:3888
server.3=storm03:2888:3888
server.4=storm04:2888:3888
server.5=storm05:2888:3888
autopurge.purgeInterval=1
storm.yaml
storm.zookeeper.servers:
- "storm01"
- "storm02"
- "storm03"
- "storm04"
- "storm05"
storm.zookeeper.port: 2181
zookeeper.multiple.setup:
follower.port:2888
election.port:3888
nimbus.host: "storm01"
storm.supervisor.hosts:
- "storm01"
- "storm02"
- "storm03"
- "storm04"
- "storm05"
supervisor.slots.ports:
- 6700
- 6701
- 6702
- 6703
- 6704
storm.local.dir: /log/storm-data
worker.childopts: "-Xmx5120m -Djava.net.preferIPv4Stack=true"
topology.workers: 5
storm.log.dir: /log/storm-log