So installed Rancher on two servers…one on the East Coast and one on the West Coast. Have the main Rancher GUI running on East Coast server (cuz that’s where I am). Got HELM and KUBCTL working fine, able to deploy HELM charts and can do anything I want…Life is pretty good!
Have “ALL” roles servers running on each Coast and as I mentioned everything is working great…what I would like to do…and am sure I am overthinking this…is have a second instance of the Rancher GUI running on a box on the West Coast server just in case I lose connectivity to the East Coast box. I have Portainer running on West Coast as a stop-gap but would really like to have second instance of Rancher GUI that mirrors what is running on the East Coast server.
Run one proper HA installation in one place and register clusters to it instead of trying to register clusters to multiple management servers simultaneously. Even if the server is down for some reason the clusters are still fine, and you can setup access to them directly (authorized cluster endpoint) instead of going through the server for kubectl/helm at all.
Thanks for the feedback…am still working on getting Rancher running as Kubs deployment project (HA style), but to date have not been successful. I really like the whole Rancher experience, it makes things so much easier. Have been using LENS as an alternative but it pales in comparison to Rancher. So many things made more sense (PVC, Service Discovery, etc) when I started using Rancher.