[
https://issues.apache.org/jira/browse/HBASE-7158?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13532863#comment-13532863
]
Lars Hofhansl commented on HBASE-7158:
--------------------------------------
Awesome. So for CopyTable it would tag the KVs with the id of the source
cluster. For import/importTSV it would tag them with the local cluster id.
For the import case we might sometimes want to set a specific cluster ID (for
example if we did an export in cluster X and moved that data to cluster Y, we
might specify cluster X as the cluster id when we import to cluster Y). But
that can be on an other issue.
+1
> Allow CopyTable to identify the source cluster (for replication scenarios)
> --------------------------------------------------------------------------
>
> Key: HBASE-7158
> URL: https://issues.apache.org/jira/browse/HBASE-7158
> Project: HBase
> Issue Type: Bug
> Reporter: Lars Hofhansl
> Attachments: HBASE-7158-0.94-v1.patch
>
>
> When I worked on HBASE-2195 I added a mechanism for an edit to identify its
> source cluster, so that replication would not bounce it back to the source.
> See: {{this.clusterId =
> zkHelper.getUUIDForCluster(zkHelper.getZookeeperWatcher());}} in
> ReplicationSource, and {{put.setClusterId(entry.getKey().getClusterId());}}
> in ReplicationSink.
> In master-master replication scenarios, it would very useful if CopyTable
> would identify the source cluster (by tagging each Put/Delete with the source
> clusterId before applying it).
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
