The way we've moved data from cluster A to cluster B is:
1. Turn on replication from A -> B
2. Run map reduce export on A and map reduce import on B
The time stamp functionality of hbase makes it so the import to B wont overwrite any of the newer
replicated data. Once the import is complete we simply point the clients at cluster B and we're done.
However I'm not sure if the replication functionality is in 0.20.6.
~Jeff
On 2/11/2011 3:48 PM, Ted Yu wrote:
Are you referring to https://issues.apache.org/jira/browse/HBASE-3451 ?
Our clusters are in same DC but not on same hdfs.
Our issue is that the table to be transferred doesn't exist on destination
cluster.
Thanks
On Fri, Feb 11, 2011 at 2:41 PM, Stack<[email protected]> wrote:
On Fri, Feb 11, 2011 at 2:32 PM, Ted Yu<[email protected]> wrote:
QA can use export/import tool
I wonder if there is a faster way of transferring whole table across
clusters.
You saw the recent mozilla posting by xavier about wrangling
x-datacenter copying? If not, check it out.
Are your clusters in same DC or even on same HDFS?
St.Ack
--
Jeff Whiting
Qualtrics Senior Software Engineer
[email protected]
