it may be easiest to start over and follow the HA install guide exactly as written. You should not need to modify anything with nginx ingress, and you should have 3 nodes, each one with all 3 roles. That cluster is for rancher server only. no other workloads go in there. Your L4 load balancer is outside of this 3 node cluster.
also - there is no “load balancer node”. you bring your own load balancer, external from Rancher.
heiii @superseb
I have one question, I have deployed HA for rancher using the link provided in your comment. but i got the reference from a person of rancher team that another way to deploy the same is just launch 3 masters and 1 worker node from the rancher GUI and that will work as the same.
So what is the need of load-balancer in this case ?
Is rancher is running some nginx pod internally to provide load-balancing ?
Hi Kamlesh,
Would you please tell us, on which node you have installed RKE, Kubectl, helm? Assume 3 etcd,controller node and 4 worker node. Do we need to install in all node?
Regards,
Makesh.
you don;t need to install those tools on any node if you don;t want to. I like to install those on a mgmt host, that is not part of k8s cluster.
Thanks for the detailed explanation.
hi makesh25,
installation of RKE tool, kubectl was done on load balancer node.