Are you reading the right instructions, the ones I pointed you at in
previous mail?  Its 60000 that the master listens on.  It puts up a UI on
60010 by default.  0 region servers doesn't sound right if you started
some.

On Mon, Nov 9, 2009 at 5:42 PM, Chris Bates <
[email protected]> wrote

..

>
>
> Also, another thing to note is that when I try running the shut down
> script,
> the master node takes infinitely long to shut down and I always have to
> manually kill the process.
>
>
Whats it doing?  If no regionservers, there is nothing for it to shutdown.


>
> [had...@crunch hbase-0.20.1]$ bin/start-hbase.sh
>
> crunch2: Warning: Permanently added 'crunch2' (RSA) to the list of known
> hosts.
>

This happens when you ssh between hosts?   Usually its when a host is seen
for the first time.  Can you ssh w/o a password to all regionservers (as was
asked previous)?


crunch2: zookeeper running as process 16233. Stop it first.
>
> chanel: zookeeper running as process 29181. Stop it first.
>
> crunch3: zookeeper running as process 28820. Stop it first.
>
>

Who is running zk?  HBase?  Or did you start it yourself.  Would seem the
latter given above messages.


chanel: regionserver running as process 7254. Stop it first.

>
> crunch2: regionserver running as process 13502. Stop it first.
>
> chanel2: regionserver running as process 21510. Stop it first.
>
> crunch3: regionserver running as process 32497. Stop it first.
>
> chris: regionserver running as process 6480. Stop it first.
>
>
The regionservers are up and running?  But they are not reporting into the
master?


20.1/bin/../lib/native/Linux-i386-32]2009-11-09

> 17:44:15,803 INFO org.apache.hadoop.hbase.master.HMaster: My address is
> 172.16.1.220:600002009-11-09 17:44:16,474 INFO
>

Looks like you have the right address here.


> 2009-11-09 17:44:16,787 INFO org.apache.zookeeper.ClientCnxn: Server
> connection successful
>
> Looks like it can connect to zk fine.




2009-11-09 17:51:16,800 INFO org.apache.hadoop.hbase.master.ServerManager: 0
> region servers, 0 dead, average load NaN
>
>
The regionservers are not coming in and connecting.  Did you change the port
they are expecting to find master on?

Whats in their logs?
St.Ack




> On Mon, Nov 9, 2009 at 5:26 PM, stack <[email protected]> wrote:
>
> > On Mon, Nov 9, 2009 at 12:59 PM, Chris Bates <
> > [email protected]> wrote:
> >
> > > Hi all,
> > >
> > >   I'm new to HBase and trying to configure it to run with a cluster of
> 6
> > > machines.  I'm logged into the Master Node and ran the start scripts as
> > > directed from the Wiki,
> >
> >
> > You are reading
> >
> >
> http://hadoop.apache.org/hbase/docs/r0.20.1/api/overview-summary.html#overview_description
> > ?
> >
> >
> > </configuration>
> > >
> > > here's the ugly error:
> > >
> > > hbase(main):001:0> create 'stations', {NAME=>'info',VERSIONS=>1}
> > > 09/11/09 14:54:09 INFO client.HConnectionManager$TableServers:
> getMaster
> > > attempt 0 of 5 failed; retrying after sleep of 2000
> > > java.net.ConnectException: Connection refused
> > >
> >
> >
> >
> > Is HBASE up?  Check the master logs.  Check the UI.
> >
> > Sorry, the ruby shell can dump ugly stack traces.  There's an issue to
> > tighten them up.
> >
> > St.Ack
> >
>

Reply via email to