Hi,

I'm having problems setting up replication on HBase 0.94.1. I have
installed HBase on two clusters (the slave is using a pseudo-distributed
installation), and the ids are:

[zk: hbdndl3:2181(CONNECTED) 0] get /hbase/hbaseid
18704@hbnndl1f3853040-98e1-4236-99bf-662196b5aaf4

[zk: hbase-backup:2181(CONNECTED) 3] get /hbase/hbaseid
�9782@hbase-backup55da74bc-47b7-4daf-af65-c37ebb762a3e

As you can see in the second line of the following log, the master is
getting the peer id connecting to the master zookeeper ensemble:

2012-10-04 16:44:18,837 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
regionserver:60020-0x23a261c8591001c Retrieved 35 byte(s) of data from
znode /hbase/replication/peers/1; data=hbase-backup:2181:...
2012-10-04 16:44:18,838 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
connection to cluster: hbase-backup:2181:/hbase opening connection to
ZooKeeper with ensemble
(hbdndl3:2181,hbdndl2:2181,hbdndl1:2181,hbnndl1:2181,hbnndl2:2181)
2012-10-04 16:44:20,496 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
connection to cluster: hbase-backup:2181:/hbase-0x23a261c8591001d Retrieved
36 byte(s) of data from znode /hbase/hbaseid;
data=f3853040-98e1-4236-99bf-66219...
2012-10-04 16:44:20,496 INFO
org.apache.hadoop.hbase.replication.regionserver.ReplicationSource:
Replicating f3853040-98e1-4236-99bf-662196b5aaf4 ->
f3853040-98e1-4236-99bf-662196b5aaf4

Of course the id it gets is the same as the master id and nothing gets
replicated.

Have you any idea of what could be wrong?

Thanks.

Have you set any table with REPLICATION_SCOPE => 1 ?

Can we see more of that log? Put it in a pastebin.com (or similar) please.

J-D

On Thu, Oct 4, 2012 at 8:41 AM, Placido Revilla
<[EMAIL PROTECTED]> wrote:
> Hi,
>
> I'm having problems setting up replication on HBase 0.94.1. I have
> installed HBase on two clusters (the slave is using a pseudo-distributed
> installation), and the ids are:
>
> [zk: hbdndl3:2181(CONNECTED) 0] get /hbase/hbaseid
> 18704@hbnndl1f3853040-98e1-4236-99bf-662196b5aaf4
>
> [zk: hbase-backup:2181(CONNECTED) 3] get /hbase/hbaseid
> �9782@hbase-backup55da74bc-47b7-4daf-af65-c37ebb762a3e
>
> As you can see in the second line of the following log, the master is
> getting the peer id connecting to the master zookeeper ensemble:
>
> 2012-10-04 16:44:18,837 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
> regionserver:60020-0x23a261c8591001c Retrieved 35 byte(s) of data from
> znode /hbase/replication/peers/1; data=hbase-backup:2181:...
> 2012-10-04 16:44:18,838 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
> connection to cluster: hbase-backup:2181:/hbase opening connection to
> ZooKeeper with ensemble
> (hbdndl3:2181,hbdndl2:2181,hbdndl1:2181,hbnndl1:2181,hbnndl2:2181)
> 2012-10-04 16:44:20,496 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
> connection to cluster: hbase-backup:2181:/hbase-0x23a261c8591001d Retrieved
> 36 byte(s) of data from znode /hbase/hbaseid;
> data=f3853040-98e1-4236-99bf-66219...
> 2012-10-04 16:44:20,496 INFO
> org.apache.hadoop.hbase.replication.regionserver.ReplicationSource:
> Replicating f3853040-98e1-4236-99bf-662196b5aaf4 ->
> f3853040-98e1-4236-99bf-662196b5aaf4
>
> Of course the id it gets is the same as the master id and nothing gets
> replicated.
>
> Have you any idea of what could be wrong?
>
> Thanks.

The problem turned out to be that you cannot specify the master zookeeper
ensemble in a zoo.cfg file (very useful if you colocate HBase and
zookeeper) because the replication setup code is borked (the zoo.cfg config
overrides the hack done in the code to create a config object equivalent to
the master one with the remote ensemble).

Once you specify the ensemble in hbase-site.xml everything works as it
should.

Regards.

On Mon, Oct 8, 2012 at 6:59 PM, Jean-Daniel Cryans <[EMAIL PROTECTED]>wrote:

> Have you set any table with REPLICATION_SCOPE => 1 ?
>
> Can we see more of that log? Put it in a pastebin.com (or similar) please.
>
> J-D
>
> On Thu, Oct 4, 2012 at 8:41 AM, Placido Revilla
> <[EMAIL PROTECTED]> wrote:
> > Hi,
> >
> > I'm having problems setting up replication on HBase 0.94.1. I have
> > installed HBase on two clusters (the slave is using a pseudo-distributed
> > installation), and the ids are:
> >
> > [zk: hbdndl3:2181(CONNECTED) 0] get /hbase/hbaseid
> > 18704@hbnndl1f3853040-98e1-4236-99bf-662196b5aaf4
> >
> > [zk: hbase-backup:2181(CONNECTED) 3] get /hbase/hbaseid
> > �9782@hbase-backup55da74bc-47b7-4daf-af65-c37ebb762a3e
> >
> > As you can see in the second line of the following log, the master is
> > getting the peer id connecting to the master zookeeper ensemble:
> >
> > 2012-10-04 16:44:18,837 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
> > regionserver:60020-0x23a261c8591001c Retrieved 35 byte(s) of data from
> > znode /hbase/replication/peers/1; data=hbase-backup:2181:...
> > 2012-10-04 16:44:18,838 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
> > connection to cluster: hbase-backup:2181:/hbase opening connection to
> > ZooKeeper with ensemble
> > (hbdndl3:2181,hbdndl2:2181,hbdndl1:2181,hbnndl1:2181,hbnndl2:2181)
> > 2012-10-04 16:44:20,496 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
> > connection to cluster: hbase-backup:2181:/hbase-0x23a261c8591001d
> Retrieved
> > 36 byte(s) of data from znode /hbase/hbaseid;
> > data=f3853040-98e1-4236-99bf-66219...
> > 2012-10-04 16:44:20,496 INFO
> > org.apache.hadoop.hbase.replication.regionserver.ReplicationSource:
> > Replicating f3853040-98e1-4236-99bf-662196b5aaf4 ->
> > f3853040-98e1-4236-99bf-662196b5aaf4
> >
> > Of course the id it gets is the same as the master id and nothing gets
> > replicated.
> >
> > Have you any idea of what could be wrong?
> >
> > Thanks.
>

Ah it's not only replication that does that, if you use CopyTable
you'll have the same issue. See also
https://issues.apache.org/jira/browse/HBASE-4614

J-D

On Tue, Oct 9, 2012 at 1:05 AM, Placido Revilla
<[EMAIL PROTECTED]> wrote:
> The problem turned out to be that you cannot specify the master zookeeper
> ensemble in a zoo.cfg file (very useful if you colocate HBase and
> zookeeper) because the replication setup code is borked (the zoo.cfg config
> overrides the hack done in the code to create a config object equivalent to
> the master one with the remote ensemble).
>
> Once you specify the ensemble in hbase-site.xml everything works as it
> should.
>
> Regards.
>
> On Mon, Oct 8, 2012 at 6:59 PM, Jean-Daniel Cryans <[EMAIL PROTECTED]>wrote:
>
>> Have you set any table with REPLICATION_SCOPE => 1 ?
>>
>> Can we see more of that log? Put it in a pastebin.com (or similar) please.
>>
>> J-D
>>
>> On Thu, Oct 4, 2012 at 8:41 AM, Placido Revilla
>> <[EMAIL PROTECTED]> wrote:
>> > Hi,
>> >
>> > I'm having problems setting up replication on HBase 0.94.1. I have
>> > installed HBase on two clusters (the slave is using a pseudo-distributed
>> > installation), and the ids are:
>> >
>> > [zk: hbdndl3:2181(CONNECTED) 0] get /hbase/hbaseid
>> > 18704@hbnndl1f3853040-98e1-4236-99bf-662196b5aaf4
>> >
>> > [zk: hbase-backup:2181(CONNECTED) 3] get /hbase/hbaseid
>> > �9782@hbase-backup55da74bc-47b7-4daf-af65-c37ebb762a3e
>> >
>> > As you can see in the second line of the following log, the master is
>> > getting the peer id connecting to the master zookeeper ensemble:
>> >
>> > 2012-10-04 16:44:18,837 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
>> > regionserver:60020-0x23a261c8591001c Retrieved 35 byte(s) of data from
>> > znode /hbase/replication/peers/1; data=hbase-backup:2181:...
>> > 2012-10-04 16:44:18,838 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
>> > connection to cluster: hbase-backup:2181:/hbase opening connection to
>> > ZooKeeper with ensemble
>> > (hbdndl3:2181,hbdndl2:2181,hbdndl1:2181,hbnndl1:2181,hbnndl2:2181)
>> > 2012-10-04 16:44:20,496 DEBUG org.apache.hadoop.hbase.zookeeper.ZKUtil:
>> > connection to cluster: hbase-backup:2181:/hbase-0x23a261c8591001d
>> Retrieved
>> > 36 byte(s) of data from znode /hbase/hbaseid;
>> > data=f3853040-98e1-4236-99bf-66219...
>> > 2012-10-04 16:44:20,496 INFO
>> > org.apache.hadoop.hbase.replication.regionserver.ReplicationSource:
>> > Replicating f3853040-98e1-4236-99bf-662196b5aaf4 ->
>> > f3853040-98e1-4236-99bf-662196b5aaf4
>> >
>> > Of course the id it gets is the same as the master id and nothing gets
>> > replicated.
>> >
>> > Have you any idea of what could be wrong?
>> >
>> > Thanks.
>>