how to set up high-availability kubernetes masters
We need to deploy redundant haproxy load-balancers(VRRP) and 2~3 apiservers to build high-availability clusters on bare-metal
But, we have 4 bare-metal servers.(the number of nodes will increase)
So we are thinking of installing load-balancers in VMs assigned flat IPs on Master nodes, as you see the picture.
Is there any issues in this structure? or please let me know better one..
Similar Questions
1 Answer
Seems like an overkill to build HA control plane and have one worker node, but I guess you plan to add more in future. Other then that it looks solid. Although I assume you also have some network equipment in the environment. This could open a possibility of ditching dedicated haproxy vrrp for apiserver in favor of having loadbalancer for it solved on ie. your router/gateway level.