Within each DC, M/S - M/S replication with Keepalived for each Icinga2 Cluster.
I’m looking for some simple and reliable options to have DB replication across DCs to have Geo redundancy for Icinga2 DBs.
With some research, I think Galera cluster seems a bit complicated. The circular replication is not suitable and should be avoided.
Thanks
Based on personal experiences, I would recommend a traditional master-replica setup. You could set up one master MariaDB node in DC A and the replica node in DC B. If you want more high availability, just add another replica node to each DC.
This requires a good-enough link between the DCs, but since there is already an Icinga 2 HA setup across this DCs, a good connection is required anyways.
Personally, I would argue against a master-master MariaDB setup or setting up Galera, as it usually makes recovery much harder.
Thank for your inputs. in each DC I prefer 2 nodes M/S - M/S for HA. Maybe a slave in each DC and replicate it with the other slave in another DC would work for DB Geo redundancy.