This message was deleted.

I think you may have better luck in the #general channel. With that said, you mention re-using the name is causing the problem but the steps outlined don't seem to reference the new cluster ever working. Did the new cluster ever work with a different name? Any longs on the agent or the rancher instance that shows any errors or verifies the agent connected? It almost sounds like the agent isn't connecting to rancher from the description.

It must make some initial connection - I can see the number and type of nodes listed in cluster manager.

That's fair. Did it ever work with a different name?

Oh, yes. It works fine if I import that cluster as a new name. That's how I've ended up with a four-random-word-name that I want to rename to original-name.

In that case I'm thinking it's a backend bug

Ok, that's a good idea. Do you have some ideas for search requests to see if there were existing reports/issues/PRs?

possibly "stuck pending" or "duplicate name"

Thanks. I'm fighting another fire right now - but I'll get this in.