This repository was archived by the owner on Nov 17, 2022. It is now read-only.
k8s-spot-rescheduler doesn't seem to iterate through on-demand nodes and keeps trying to drain undrainable nodes #55
Description
We're using kops 1.10.0 and k8s 1.10.11. We're using two separate instance groups (IG), nodes (on-demand) and spots (spot). As mentioned in #54, the rescheduler thinks it can move all pods, fails due to a PDB, and leaves the node underutilized and tainted. And here's where my issue comes in, it keeps trying to drain that node over and over again. If it can't drain the node due to whatever reason, e.g. PDBs, availability zone conflicts (#53), etc., it should move on to the next on-demand node and try that one.