You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hbase.apache.org by yanivG <ya...@gmail.com> on 2014/06/18 09:56:18 UTC
HBase leader node - Is there such an animal?
Hi,
I have noticed that from time to time, one hbase node (out of 6) is more
active.
More active = its replication queue is higher, disk IO is higher and so on.
Is there a reason for that? Will it be the same one all the time, or it will
randomly move between nodes?
--
View this message in context: http://apache-hbase.679495.n3.nabble.com/HBase-leader-node-Is-there-such-an-animal-tp4060544.html
Sent from the HBase User mailing list archive at Nabble.com.
Re: HBase leader node - Is there such an animal?
Posted by Ted Yu <yu...@gmail.com>.
bq. by hashing the first 2 bytes.
Does this achieve good distribution ? You may want to include more bytes.
On Wed, Jun 18, 2014 at 11:23 PM, yanivG <ya...@gmail.com> wrote:
> Yes, I thought it as well, although I distribute the keys by hashing the
> first 2 bytes.
>
>
>
> --
> View this message in context:
> http://apache-hbase.679495.n3.nabble.com/HBase-leader-node-Is-there-such-an-animal-tp4060544p4060581.html
> Sent from the HBase User mailing list archive at Nabble.com.
>
Re: HBase leader node - Is there such an animal?
Posted by yanivG <ya...@gmail.com>.
Yes, I thought it as well, although I distribute the keys by hashing the
first 2 bytes.
--
View this message in context: http://apache-hbase.679495.n3.nabble.com/HBase-leader-node-Is-there-such-an-animal-tp4060544p4060581.html
Sent from the HBase User mailing list archive at Nabble.com.
Re: HBase leader node - Is there such an animal?
Posted by Ted Yu <yu...@gmail.com>.
What you described sounds like a hot spot.
I have never heard of 'leader node' in hbase.
One cause for hotspot could be due to (suboptimal) schema design.
Cheers
On Wed, Jun 18, 2014 at 12:56 AM, yanivG <ya...@gmail.com> wrote:
> Hi,
> I have noticed that from time to time, one hbase node (out of 6) is more
> active.
> More active = its replication queue is higher, disk IO is higher and so on.
>
> Is there a reason for that? Will it be the same one all the time, or it
> will
> randomly move between nodes?
>
>
>
> --
> View this message in context:
> http://apache-hbase.679495.n3.nabble.com/HBase-leader-node-Is-there-such-an-animal-tp4060544.html
> Sent from the HBase User mailing list archive at Nabble.com.
>