Debugging failed node creation in vSphere: how to disable auto-recreate

I asked this before on the Slack channel, but didn’t get much response there.

I’m trying to deploy clusters on vSphere using Rancher Server. There’s an error installing docker on the machine. Now, to debug I want to enter the machine through SSH to see what’s going wrong. The problem is Rancher deletes the machine and re-creates it if the deployment fails, even if I have auto-replace set to 0 in the cluster configuration.

It seems that the auto-replace setting gets ignored for new nodes. Is there a way to tell rancher not to re-create new nodes if their deployment fails?