I'm facing for the first time to a real scaling problem with a db-heavy webapp.
We are studying Galera and other cluster systems.
I ask you what is the ideal way to use Galera.
I ipothize having 3 nodes
- write on only one node and read from other two
- write and read from all of 3
- write in roundrobin using a balancer
NOTE: this is not a opinion-related question. I am here to ask you to clarify what the goal of Galera, what is the intended, suggested, use
NOTE 2: we need scaling on reading operations than on writing