Hi,
I have hbase setup in pseudo distributed atop HDFS ( again pseudo
distributed mode ) in production.
This Hbase version is 0.94.7 and Hadoop version is 1.1.2.

I was planning to have a fully distributed 2 node cluster on 2 new machines
with latest stable Hbase and Hadoop versions and then copy data from
existing cluster to this new cluster using copyTable.

But now i have to add a new machine in existing psuedo distributed mode
cluster to make it in distributed mode.

My question is how  this can be achieved cleanly.

My another concern is "dfs.replication" property which i set to one in both
Hadoop and Hbase configuration.So when i upgrade existing pseudo
distributed mode to fully distributed , how will data be replicated to
newly added machine ?

Also will regions automatically move to new machine or i will have to run
some shell command  to manually move it (balance command ) ?

Please help.

-- 
Thanks and Regards,
Vimal Jain

Reply via email to