On Thu, Feb 5, 2009 at 1:35 PM, Yair Even-Zohar <[email protected]>wrote:

> I try to backup hbase is a brute force, yet a faster method by copying
> everything under the adoop.tmp.dir which includes the namenode, data
> etc.


This will work while your dataset is small.  A variant when your data gets
bigger is to shutdown hbase and then do a MR copy of your hbase rootdir.
See the 'distcp' option on ./bin/hadoop..

St.Ack

Reply via email to