Automatically assign pod to correct node after Kubernetes upgrade
After every Kubernetes upgrade, the sites that i host that require persistent volumes always suffer downtime.
I’m scared to update now.
My cluster has 3 node pools with 2 nodes in each. When i click upgrade, when it hits a node with a PV, it detaches the volume and recycles the node. At this point the site(s) are down.
When the node is recycled, the volume isn’t automatically reattached, i have to attach it via the DO Console.
Furthermore, once that is done, i have to add a node label, and in my deployments specify node selectors for that label.
Otherwise, the deployments will stay “Pending” because for whatever reason it cannot launch the pod/deployment on the correct node.
This whole process takes around 20 minutes, in that time my sites are mostly down.
Is there any possible way to make this whole process smoother?