Hey all, quick question. Rancher 2.10 on-prem in VMs, SLE Micro on RKE2, all fresh build. I’m wondering why the load balancer needs to be external to the Rancher Server nodes. Seems like something in in-cluster load balancer could do pretty easily, can we not use something inside the rancher RKE2 cluster for this?
The article above reference refers to the below article guidance on setting up a nginx for a load balancer, but using a single VM for load balancing doesn’t work for a solution that needs to be HA.
There are a few other discussions on here on this topic, but I really don’t want to build or recommend snowflake configurations. So my recommendation will be to either build a 2 node HA LB external to the cluster using HAProxy, Traefik or NGINX, or just use DNS to point to one node and move it when performing maintenance (yuck). Curious if anyone else has a cleaner way to handle this problem.
I did try to use MetalLB in the RKE2 cluster and that didn’t seem to work based on my attempt.
Curious for your thoughts here.