We upgraded our cluster last night and we ran into the problem of not having docker running in /var/lib/docker, so we redeployed all of our hosts and we no we can no longer log into the rancher server.
It simply timesout and forces us to auth again with Github. It appears that the /identities endpoint is timing out upon login.
We’ve tried disabling auth in the DB and that gets us into the UI, but very, very, very slowly and it appears that many of our commands aren’t accepted. The Processes list seems saturated with tasks that appear to be the server trying to execute tasks on dead hosts.