Hitting on an issue upgrading from 1.4.1 to 1.4.2. Applied the upgrade with

kubectl apply -f <https://releases.rancher.com/harvester/v1.4.2/version.yaml>

I see the Upgrade button on the dashboard, click it, select 1.4.2, check the boxes, but when I click Upgrade, I get

admission webhook "<http://validator.harvesterhci.io|validator.harvesterhci.io>" denied the request: machine fleet-local/custom-87ecea6cbf6f is not running

Potentially pertinent info: I had deployed a single node Rancher via Docker to manage this cluster, but then removed the cluster from Rancher, which seems to have cleaned up successfully. Logging is enabled, so I can happily provide the logs (though I may need to be told where to grab them).

This is also one of those painful things lately. I also happened to try rancher long time ago (and removed it according to instructions). Well last two upgrades have shown that if you try rancher there is no going back.

Honestly, my issues wasn't related to that

It was probably some of those n issues I also faced.

removed the Harvester cluster from Rancher and it was all fine - root cause was the Longhorn pre-drained issue I mentioned below