[
https://issues.apache.org/jira/browse/HBASE-2195?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13096195#comment-13096195
]
stack commented on HBASE-2195:
------------------------------
So, we'd go with using the attribute and the attribute would have the uuid for
the cluster. It'd be set on every edit rather than per every HLogEdit (which
could have one or a batch of edits). Chatting w/ J-D, he reminds me that there
is not yet a means of correlating the byte-level cluster identifier with the
uuid cluster id we have for CPs (there is no cluster registry where the byte
gets mapped to uuid).
Downsides:
+ Going this route would make the byte id redundant; might as well clear it
out if we are not going to use it
+ Fat UUID repeated per individual edit (in a Map so fatter still, rather than
as data member bits in byte array)
Upsides:
+ 'safe' -- no need of cluster registry
+ Minimal changes required to core types getting cyclic replication working
Whats your thought Lars... get this attribute based version working then
optimize space-used, etc., later?
> 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
> Attachments: 2195.txt
>
>
> 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
