Using 6 nodes for ZK is a bit odd. Actually, it is a bit even.
If all of the nodes are involved in the quorum, you will get lower write
throughput than with 5 nodes and slightly higher chance of failure since it
is more likely to get 3/6 node failures versus 3/5 failures.
What motivated your choice of 6 nodes?
On Tue, Oct 16, 2012 at 11:06 PM, yang.li <[EMAIL PROTECTED]> wrote:
> Hi, all:
> I'm in charge of a zookeeper cluster including six nodes. It worked well
> for the last six months,
> but yesterday when I want to list the children of a specific path
> something is just wrong. Here is the dump info as shown below:
> [zk: zk-6:2181(CONNECTED) 1] ls /dp/monitor_root/child/CDSkafkaSensor/msg
> 2012-10-17 13:47:28,719 [myid:] - WARN
> [main-SendThread(m32p118.bfdabc.com:2181):ClientCnxn$SendThread@1057] -
> Session 0x63a6d4272590001 for server
> m32p118.bfdabc.com/192.168.32.118:2181, unexpected error, closing socket
> connection and attempting reconnect
> java.io.IOException: Packet len5362775 is out of range!
> WatchedEvent state:Disconnected type:None path:null
> Exception in thread "main"
> KeeperErrorCode = ConnectionLoss for
> at org.apache.zookeeper.ZooKeeper.getChildren(ZooKeeper.java:1448)
> at org.apache.zookeeper.ZooKeeper.getChildren(ZooKeeper.java:1476)
> at org.apache.zookeeper.ZooKeeperMain.run(ZooKeeperMain.java:323)
> at org.apache.zookeeper.ZooKeeperMain.main(ZooKeeperMain.java:282)
> I have tried many ways to solve this problem, but none of those works,
> even the "super user" method.
> Now neither can Ｉ list the node, nor can I delete it . So I really really
> need a help. Thank you!