You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hbase.apache.org by Ken Weiner <ke...@gumgum.com> on 2009/07/22 02:57:49 UTC

zookeeper problem starting hbase 0.20.0 trunk

I am trying to start HBase in distributed mode on EC2.  HBase is not
starting properly and it seems to be caused by Zookeeper not being able to
elect a leader.
We are using an HBase-managed Zookeeper.  There is one ZK on each of 3
regionservers.
I didn't change any default ports.
The following error keeps repeating in the logs:

2009-07-22 00:41:53,940 INFO
org.apache.zookeeper.server.quorum.LeaderElection: Server address:
domU-43-31-49-02-E8-L2.compute-1.internal/10.244.229.1:2888
2009-07-22 00:41:54,139 WARN
org.apache.zookeeper.server.quorum.LeaderElection: Ignoring exception while
looking for leader
java.net.SocketTimeoutException: Receive timed out
        at java.net.PlainDatagramSocketImpl.receive0(Native Method)
        at
java.net.PlainDatagramSocketImpl.receive(PlainDatagramSocketImpl.java:136)
        at java.net.DatagramSocket.receive(DatagramSocket.java:712)
        at
org.apache.zookeeper.server.quorum.LeaderElection.lookForLeader(LeaderElection.java:170)
        at
org.apache.zookeeper.server.quorum.QuorumPeer.run(QuorumPeer.java:484)

I did a netstat -a -n | grep LISTEN and there is nothing listening on port
2888 on any server which probably explains why the QuorumPeer can't
connect.  Is there supposed to be something listnening on 2888?  Any ideas
on how to fix this?  Thanks.

-Ken

Re: zookeeper problem starting hbase 0.20.0 trunk

Posted by Ken Weiner <ke...@gumgum.com>.
Got the answer to my problem.  I had to enable UDP traffic on port 2888 for
the Amazon security group for which I'm running my HBase instances.
Everything started working once I did that.  Thanks to Jon Gray and others
in the IRC channel for helping.

On Tue, Jul 21, 2009 at 5:57 PM, Ken Weiner <ke...@gumgum.com> wrote:

> I am trying to start HBase in distributed mode on EC2.  HBase is not
> starting properly and it seems to be caused by Zookeeper not being able to
> elect a leader.
> We are using an HBase-managed Zookeeper.  There is one ZK on each of 3
> regionservers.
> I didn't change any default ports.
> The following error keeps repeating in the logs:
>
> 2009-07-22 00:41:53,940 INFO
> org.apache.zookeeper.server.quorum.LeaderElection: Server address:
> domU-43-31-49-02-E8-L2.compute-1.internal/10.244.229.1:2888
> 2009-07-22 00:41:54,139 WARN
> org.apache.zookeeper.server.quorum.LeaderElection: Ignoring exception while
> looking for leader
> java.net.SocketTimeoutException: Receive timed out
>         at java.net.PlainDatagramSocketImpl.receive0(Native Method)
>         at
> java.net.PlainDatagramSocketImpl.receive(PlainDatagramSocketImpl.java:136)
>         at java.net.DatagramSocket.receive(DatagramSocket.java:712)
>         at
> org.apache.zookeeper.server.quorum.LeaderElection.lookForLeader(LeaderElection.java:170)
>         at
> org.apache.zookeeper.server.quorum.QuorumPeer.run(QuorumPeer.java:484)
>
> I did a netstat -a -n | grep LISTEN and there is nothing listening on port
> 2888 on any server which probably explains why the QuorumPeer can't
> connect.  Is there supposed to be something listnening on 2888?  Any ideas
> on how to fix this?  Thanks.
>
> -Ken
>