We have 2 docker swarms in 2 different geographical locations. We opted for 2 separate docker swarms to avoid latency issues by one swarm forwarding to containers in another datacenter.
Let's call it Swarm-US and Swarm-EU.
We have Percona setup in Swarm-US in multiple nodes so we can serve an application, and we would like to replicate this to Swarm-EU.
On top of that we will have a dns load balancer that will forward EU users to Swarm-EU and other users to Swarm-US, and failover from one to the other if a datacenter blows up :)
With this, we are aiming to have a fully replicated system where web applications can be served if one of the datacenters goes down.
We are having some trouble setting Percona up for this system, since we would need the keys from the percona nodes in the 2 swarms to be the same, otherwise sync won't work.
Are we approaching this the wrong way, or what are we missing?