Greetings all. I’m new to percona and have setup some test clusters.
Current test involves 2 - 3 node clusters in different datacenters.
I have then setup master master SQL replication cross cluster on each node 3 of each cluster. Some thing like this.
east-node-1 west-node-1
| |
east-node-2 west-node-2
| |
east-node-3 <-------> west-node-3
The idea is to have fast replication on each cluster, but eventual compliance on the other datacenter.
Here is what is interesting, and I think I know why, but want to see what the community says.
If I create or change a table on either node3, all 6 nodes have the replicated change very quickly.
If I create or change a table on nodes 1 or 2 of a cluster, ONLY the clusters 3 nodes where the change was made replicates. So the other cluster does not see the changes.
I pretty much stopped there, as if schema changes don’t replicate I’d be concerned with the data replications.
I’m guessing it has something to do with the replication in the percona layer not being seen by the mySQL replication on the master node-3.
Is there a better way to accomplish this goal?
Much thanks.
Tim