I am successfully using rancher manager on GCP but I want to build my own home k8s lab with harvester and rancher Specifically rancher manager 2 7 1 on bare metal Ubuntu 22 04 2 LTS installed on k3s v1 24 7+k3s1 integrated with harvester 1 1 1 When trying to provision a new cluster on harvester and trying to set options for the machine pool the UI tries to retrieve k8s clusters v1 harvesterhci io virtualmachineimages and receives a 404 ``` Code Code NotFound Status 404 Message <http clusters management cattle io|clusters management cattle io> v1 not found Cause null FieldName ``` and the UI shows <http clusters management cattle io|clusters management cattle io> v1 not found on my harvester master node a hv01 home rancher kubectl api resources | grep <http clusters management cattle io v1|clusters management cattle io v1> does not produce any output and a curl <https harvester xxx com healthz> produces curl 60 SSL certificate problem self signed certificate in certificate chain which is to be expected as cert management especially an automated process with let s encrypt is not available Would anyone in the community have any ideas of where to look or next steps to begin to troubleshoot

virtualmachineimages vmimage vmimages <http harvesterhci io v1beta1|harvesterhci io v1beta1> true

I am successfully using rancher manager on GCP but I want to rancher-users #general

Title

bright-fireman-42144

02/25/2023, 9:22 PM

I am successfully using rancher manager on GCP but I want to build my own home k8s lab with harvester and rancher. Specifically rancher manager 2.7.1 on bare-metal Ubuntu 22.04.2 LTS installed on k3s v1.24.7+k3s1, integrated with harvester 1.1.1 When trying to provision a new cluster on harvester and trying to set options for the machine pool, the UI tries to retrieve k8s/clusters/v1/harvesterhci.io.virtualmachineimages and receives a 404:

{
  "Code": {
    "Code": "NotFound",
    "Status": 404
  },
  "Message": "<http://clusters.management.cattle.io|clusters.management.cattle.io> \"v1\" not found",
  "Cause": null,
  "FieldName": ""
}

and the UI shows: clusters.management.cattle.io "v1" not found on my harvester master node a hv01:/home/rancher # kubectl api-resources | grep clusters.management.cattle.io/v1 does not produce any output and a curl https://harvester.xxx.com/healthz produces curl: (60) SSL certificate problem: self-signed certificate in certificate chain which is to be expected as cert management, especially an automated process with let's encrypt is not available. Would anyone in the community have any ideas of where to look or next steps to begin to troubleshoot?

had some dinner and digging deeper. Looked for harvesterhci.io.virtualmachineimages using kubectl api-resources and I only see

virtualmachineimages vmimage,vmimages harvesterhci.io/v1beta1 true

how many times am I going to have to destroy everything to make an external rancher deployment work with harvester? I figured I'd follow the suse at home repo but that requires me to use SLES and create an account and get a registration code. The installation of RKE2 requires me to install a specific release cause 2.7.1 won't work on the latest. 'rancher' is used too much as a name/object (rancher labs, rancherd, rancher kubernetes engine, rancher, rancher, rancher...) My apologies. Stress tension breaker.

still can't get the rancher and harvester integration to work, even after using SLES as the base for rke2. This is a support bundle from harvester. Is there any equivalent from Rancher Manager that I can provide to help troubleshoot why it cannot retrieve the information from harvester so a machine pool can be created?

supportbundle_5012811b-0f64-4cab-a386-9443ccdc4717_2023-03-01T18-24-09Z.zip

this looks like a manager issue as it's looking for v1 at itself but can't find it. I am puzzled.

looking at the support matrix, perhaps 2.7.1 is too new for me to be using with harvester 1.1.1 I will try, 2.6.9 and report back.

My lab is a frankenstein mess of cables and a mix of a cisco sb300, a linksys running tomato just to get some more ports closer to the SUSE lab equipment and I think I might be running into https://github.com/coredns/coredns/issues/3305 It at least presents itself in rancher logs when trying to reach https://git.rancher.io/helm3-charts/ So might as well rule out my transport layer.

#general Join Slack