[
https://issues.apache.org/jira/browse/HBASE-2822?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12922750#action_12922750
]
Jean-Daniel Cryans commented on HBASE-2822:
-------------------------------------------
You got it right Jeff, so currently you can do master-master as long as tables
that are scoped for replication on one cluster aren't on the other. Also now
you can chain replication (the code I quoted in the description of this jira is
what prevented you from doing it before). For example, here at SU we replicate
from the web-serving prod to a "backup" cluster, which also contains some
tables that are scoped for replication to a "dev" cluster that our front-end
engineers use for obvious reasons.
> Enable Master-Master replication
> --------------------------------
>
> Key: HBASE-2822
> URL: https://issues.apache.org/jira/browse/HBASE-2822
> Project: HBase
> Issue Type: Sub-task
> Reporter: Jean-Daniel Cryans
> Assignee: Jean-Daniel Cryans
> Fix For: 0.92.0
>
>
> There's nothing inherently wrong in the replication design that prevents
> master-master, but it was disabled in the code in order to reduce the scope
> of HBASE-2223. Specifically, this code from Replication just needs a simple 2
> lines modification to always instantiate ReplicationSink if this.replication
> is true:
> {code}
> public void startReplicationServices() throws IOException {
> if (this.replication) {
> if (this.replicationMaster) {
> this.replicationManager.init();
> } else {
> this.replicationSink =
> new ReplicationSink(this.conf, this.stopRequested);
> }
> }
> }
> {code}
> Enable and test.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
