Our current setup is Mysql Server 5.5 as MASTER → asynchronous replication → 3 Node PXC. We are using the PXC as read only replicas right now. The PXC is receiving all updates via asynchronous replication from the existing MASTER.
I need to create a new Asynchronous slave in our office, which we use for reporting but I want it’s Master to be one of the cluster members and not the existing master as that master will be shutdown once we migrate writes to the PXC.
I haven’t seen any documentation on this setup or how to create a asynchronous replica from an existing PXC. Of note, I have an existing replica from the old master already in the office. I would love to essentially use a change master statement if possible and not recreate it from scratch as it is quite large.
If you are migrating an existing slave of MASTER to become a slave of a PXC node without re-clone, you’d need to:
a) stop writes on MASTER, wait for PXC to catch up
b) CHANGE MASTER on slave to new PXC node’s current master binlog position
c) restart writes on MASTER
If you can’t stop writes on the MASTER, then you’ll probably have to reclone your slave.
Pretty much the same way, except you setup one one in the cluster as a slave to your master with CHANGE MASTER TO. The slave in question needs log-slave-updates to replicate changes it sees from the master to the rest of the cluster.
Jay, do you happen to know if there there an existing write up on how to accomplish this with PXC? I have one stack set by Marco, but looking to set up a second set of stacks as well.
No jay i don’t want like this i need suppose i have one existing mysql server that working fine now i configure the PXC and its also working fine …but i need to configure the this that is… i need from EXISTING Mysql server data----to------PXC…thats means Existing mysql server is working in front end and pxc will work on back end…it means all entries come in existing mysql and its will diliver to pxc…is is possiable…
Fine, you just backup the existing mysql server and restore that on your first cluster node (the node you bootstrap). After the cluster is setup, one node in the cluster must replicate from the existing master (standard async mysql replication) as I mentioned above.
This is very similar to any HOWTO you may find on the internet about setting up a mysql slave. I cannot write you a full set of steps here.