My use case is: I have some servers all over the world, and I need to deploy the same, stateless application on each of them. I’ve been doing this manually for a long time but since Rancher can deploy containers in a “one pod on each node” manner, I’m thinking if it is possible to use Rancher in my setup.
Specific questions I’d like to know:
- Is there a way to make ingress (L4/L7 loadbalancer) only connect to the pod on its own node?
- Do I really need a HA (3-node) etcd and multiple control planes? If I do, should I deploy them as near as possible or it is OK to make them distributed over high-latency WAN?