Inline.

J-D

On Thu, May 26, 2011 at 5:12 AM, elton sky <[email protected]> wrote:
> I am new to hbase and I got a few questions:
> 1. tables are divided into regions and allocated to region servers. what is
> the allocation policy?

Ted Yu recently wrote up a blog post about balancing in HBase:
http://zhihongyu.blogspot.com/2011/04/load-balancer-in-hbase-090.html

> 2. From my understanding, if a region server get requests for data on the
> same host, for data locality, all requests will be served by the local host,
> right?

HBase doesn't decide anything, it's a HDFS responsability:
http://hadoop.apache.org/common/docs/r0.20.2/hdfs_design.html#Replica+Selection

>  Unless the max connection for the data node is reached. If this is
> the case, for random read request, can requests access the data
> simultaneously?

I'm not sure I understand this question, could you please clarify?

> 3. When doing a update, e.g. add a new version to a column, how the system
> lock the record?

With a row lock. Regarding HBase's ACID guarantees in general, please
see http://hbase.apache.org/acid-semantics.html

J-D

Reply via email to