I have latest Rancher v2.6.5 and try to install RKE2 and I can not bypass new credential screen

[rr4444]

I have latest Rancher v2.6.5 and try to install RKE2 and I can not bypass new credential screen

Add new token

And there is no token to select

Originally posted by @olexiyb in #125 (comment)

[mehran-prs]

I have the same issue too.

[simonostendorf]

I have got the same issue.

[TitusKirch]

Me too

[mehdisadeghi]

fwiw, the option to create a cloud credential does not appear for Hetzner under Cluster Management -> Croud Credentials -> Create menu entry.

Moreover, according to the Rancher Docs:

The values of this dropdown is based on the active node drivers in Rancher.

I don't know whether this is a Rancher bug or ui-driver-heztner bug. I checked Rancher issues and found at least one similar issue. However, it might not really be related to this issue at all.

I'm also interested in fixing this if I know where the issue lies. For the time being, I'll stick to RKE which is fully functional in my experience.

Edit: minor formatting & typo

[simonostendorf]

Cloud Credential could be specified with annotations inside node-driver (view other node drivers in API to see examples).
But driver does not work after creating credential. There is no communication with the Hetzner API i could recognize.

[simonostendorf]

Hello @mehdisadeghi,
I found more information about the error i am (we are) getting with the Hetzner node driver inside rancher:

Error will be that the cluster creation is stuck on "waiting for viable init node".
Interesting error message is:

[ERROR] error syncing 'fleet-default/test-pool1-f7bcb304-l8jcz': handler machine-provision: nodedrivers.management.cattle.io "hetzner" not found, requeuing

Messages before (multiple times):

[ERROR] defaultSvcAccountHandler: Sync: error handling default ServiceAccount of namespace key=p-rhgk6, err=Operation cannot be fulfilled on namespaces "p-rhgk6": the object has been modified; please apply your changes to the latest version and try again

I recorded the network traffic and there is no outgoing connection to the Hetzner API so its an internal rancher error interacting with the driver.

Hope anyone has an idea to fix this because i dont want to use custom clusters on preprovisioned Hetzner nodes because I want to scale up via Rancher GUI.

Notes: I am using Rancher 2.7.0 inside a docker container, Hetzner node driver 3.10.0 and no UI driver (because its only to look fancy and outdated)

[simonostendorf]

Here is a solution:
rancher/rancher#34782 (comment)

[mehdisadeghi]

Thanks for the follow up and the solution and happy new year @simonostendorf!

P.S.: I'll update as soon as I apply the fix.