[
https://issues.apache.org/jira/browse/HBASE-2195?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13095047#comment-13095047
]
Lars Hofhansl commented on HBASE-2195:
--------------------------------------
I have a patch working.
Part of this is splicing the cluster id into the Puts and Deletes generated at
the Slave (ReplicationSink) and then writing it to the WAL on the Slave.
Currently I am using {get|set}Attribute(...) on both Puts and Deletes. Then in
HRegion I check for these attributes and write the cluster id to the log.
What the general thought here? Do this? Or increment the versions of the Put
and Delete classes and add a new byte member for the cluster id? Or something
entirely different that does not involve core (HRegion) needing to know about
cluster ids now?
> Support cyclic replication
> --------------------------
>
> Key: HBASE-2195
> URL: https://issues.apache.org/jira/browse/HBASE-2195
> Project: HBase
> Issue Type: Sub-task
> Components: replication
> Reporter: Jean-Daniel Cryans
>
> We need to support cyclic replication by using the cluster id of each HlogKey
> and stop replicating when it goes back to the original cluster.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
