I have set up a 3 node percona cluster in amazon ec2. As our client required, i also had to set up an asynchronous slaves in another datacenter as well. Each async slave pointed to single percona cluster node with log-slave-updates and bin-log enabled.
However the async slaves failed to maintain the replication process for long. At least once a week i have to build the slaves from scratch. These async slaves are used only for backup purpose.
Is there any alternative method(best way) of doing this?
Percona Cluster ---> async slaves ?