Any update on this? We encountered this on a production cluster. Did the
zookeeper bug ever get filed or fixed?

On Fri, Nov 16, 2012 at 11:23 AM, Scott Alcaide <[EMAIL PROTECTED]> wrote:
> Any update on this? We encountered this on a production cluster. Did the
> zookeeper bug ever get filed or fixed?
>

If hbase is not managing ZK, and it gets restarted, the client will
ride over the zk restart fine.

If cluster is restarted under a client, I've not tried that in a
while.  Anyone else know if client will just ride over it?

St.Ack

On Sat, Nov 17, 2012 at 11:11 AM, Stack <[EMAIL PROTECTED]> wrote:
> On Fri, Nov 16, 2012 at 11:23 AM, Scott Alcaide <[EMAIL PROTECTED]> wrote:
>> Any update on this? We encountered this on a production cluster. Did the
>> zookeeper bug ever get filed or fixed?
>>
>
> If hbase is not managing ZK, and it gets restarted, the client will
> ride over the zk restart fine.
>
> If cluster is restarted under a client, I've not tried that in a
> while.  Anyone else know if client will just ride over it?
>

Should be easy to test.  Open shell, connect to cluster and do some
operations, restart cluster and see if shell keeps going.
St.Ack