Hi everyone, i'm trying to have rancher 2.6 deploy a kubernetes cluster on openstack and use openstack as the cloud provider. So far i can get rancher to deploy the cluster and if i don't specify any cloud provider the cluster runs fine. However if i specify openstack as the cloud provider the cluster fails and i see the cloud following error for cattle-cluster-agent pod:

Events:
  Type     Reason            Age    From               Message
  ----     ------            ----   ----               -------
  Warning  FailedScheduling  13m    default-scheduler  0/1 nodes are available: 1 node(s) had untolerated taint {<http://node.cloudprovider.kubernetes.io/uninitialized|node.cloudprovider.kubernetes.io/uninitialized>: true}. preemption: 0/1 nodes are available: 1 Preemption is not helpful for scheduling.
  Warning  FailedScheduling  8m40s  default-scheduler  0/1 nodes are available: 1 node(s) had untolerated taint {<http://node.cloudprovider.kubernetes.io/uninitialized|node.cloudprovider.kubernetes.io/uninitialized>: true}. preemption: 0/1 nodes are available: 1 Preemption is not helpful for scheduling.

did you configure the cloud provider? Its still tainted because the cloud-provider hasn’t initialized the node yet, which usually means the cloud provider isn’t configured properly.

yes i did or at least this is the config i applied based on the rancher docs:

cloud_provider:
    name: openstack
    openstackCloudProvider:
      global:
        username: "xxx"
        password: "****"
        auth-url: "<https://xxxx:5000/v3>"
        tenant-id: "xxxxxxxxxxxxxxxxxx"
        domain-id: "default"
        region: "RegionOne"
        tls-insecure: true
      load_balancer:
        lb-version: "v2"
        subnet-id: "xxxxxxxxx"
        floating-subnet-id: "xxxxx"
        use-octavia: true
        floating-network-id: "xxxxxxx"
        create-monitor: false
        manage-security-groups: true
      block_storage:
        ignore-volume-az: true
        trust-device-path: false
        bs-version: "v2"
      metadata:
        request-timeout: 0

checked the logs on the openstack cluster?

you mean check the logs on openstack itself or the kubernetes cluster deployed by rancher ?

on the cluster. there will probably be some errors from the openstack cloud controller

looks like the cloud controller never got installed in the first place:

oh, is this RKE2? We don’t have a packaged cloud provider for openstack on RKE2…

oh ok so i have to install openstack-ccm manually ?

correct

what about if i use rke does that have the openstack-ccm included ?

I am not sure, I’m an rke2/k3s dev so I don’t use rke much… but its more likely.

that ok i want to test RKE2 anyway because of it's use of containerd

yeah, I might be biased but I would go with rke2 or k3s 😉

looks like installing openstack-ccm failed because rke2 is using the port used by ccm:

error: failed to create listener: failed to listen on 127.0.0.1:10258: listen tcp 127.0.0.1:10258: bind: address already in use
root@testk8s-pool1-ad60a11d-cd48t:~# lsof -i:10258
COMMAND    PID USER   FD   TYPE DEVICE SIZE/OFF NODE NAME
cloud-con 2283 root    7u  IPv4  24904      0t0  TCP localhost:10258 (LISTEN)
root@testk8s-pool1-ad60a11d-cd48t:~# ps aux | grep 2283
root        2283  0.3  0.6 751948 24444 ?        Ssl  21:06   0:02 cloud-controller-manager --allocate-node-cidrs=true --authentication-kubeconfig=/var/lib/rancher/rke2/server/cred/cloud-controller.kubeconfig --authorization-kubeconfig=/var/lib/rancher/rke2/server/cred/cloud-controller.kubeconfig --bind-address=127.0.0.1 --cloud-provider=rke2 --cluster-cidr=10.42.0.0/16 --configure-cloud-routes=false --kubeconfig=/var/lib/rancher/rke2/server/cred/cloud-controller.kubeconfig --node-status-update-frequency=1m0s --profiling=false
root       12485  0.0  0.0   7004  2212 pts/2    S+   21:15   0:00 grep --color=auto 2283
root@testk8s-pool1-ad60a11d-cd48t:~#

yeah, you should disable the built-in ccm

ok let me try again