(1) Cluster 1 - MariaDB Master/Slave or Galera Cluster - Server1, Server2, Server3
(2) Cluster 2 - MAriaDB ColumnStore - UM1, UM2 (and PM1, PM2, PM3)
(3) MaxScale configured with
- A service using read-write split router and named server filter.
- In this service servers=Server1,Server2,Server3,UM1,UM2
- In named server filer certain regex pattern go to Server1,Server2,Server3, another regex pattern go to UM1, UM2
- A mariadbmon or galeramon monitor with servers=Server1,Server2,Server3 (note no UM1, UM2 here)
- Manually set UM1 and UM2's server status in MaxScale via MaxCtrl or maxadmin to "slave"
Now send queries from client - verify that regex pattern configured to go to Server1,Server2,Server3 do route to Server1,Server2,Server3 in read-write split manner, and regex pattern configured to go to UM1, and UM2 do route to UM1,UM2
An example config file to be attached later on to this Jira