This message was deleted.

I've had several issues lately where a cluster gets stuck in provisioning when nodes don't join at the right time. It simply stops trying to re-provision at all, even though nodes are there in a registering state. (Specifically with ETCD node queries) I've also seen (in the past) where nodes don't register properly if the hostnames aren't what it expects (Like if an aws provider expects one hostname and the local dns assigns another) - I've wanted a button for a long time which would force rancher to execute/restart the provisioning loop check