RKE ignores cluster_cidr and service

1

When I bring up my cluster, it is ignoring the network ranges I’ve specified in my cluster.yml. It is using the default values for the network ranges: 10.43.0.0/16 for services and 10.42.0.0/16 for pods, even though I’ve specified completely different ranges in my cluster.yml (see below). I’ve read through the documentation for the cluster.yml several times and cannot see what I’m doing wrong here. I was hoping someone could point me in the right direction on this. Thanks.

From cluster.yml

kube-api:
  # IP range for any services created on Kubernetes
  #  This must match the service_cluster_ip_range in kube-controller
  service_cluster_ip_range: 192.168.128.0/17
  # Expose a different port range for NodePort services
  service_node_port_range: 30000-32767
  pod_security_policy: false

kube-controller:
  # CIDR pool used to assign IP addresses to pods in the cluster
  cluster_cidr: 192.168.0.0/17
  # IP range for any services created on Kubernetes
  # # This must match the service_cluster_ip_range in kube-api
  service_cluster_ip_range: 192.168.128.0/17

From cluster.rkestate:

 "kubeApi": {
          "image": "rancher/hyperkube:v1.20.11-rancher1",
          "serviceClusterIpRange": "10.43.0.0/16",
...
"kubeController": {
          "image": "rancher/hyperkube:v1.20.11-rancher1",
          "clusterCidr": "10.42.0.0/16",

Versons:

$ rke  --version
rke version v1.3.1
$ grep PRETTY /etc/os-release
PRETTY_NAME="Red Hat Enterprise Linux 8.4 (Ootpa)"
2

Please share the full cluster.yml, the services need to be under the services key, as shown at Rancher Docs: Example Cluster.ymls

3

I knew the problem had to be something somewhat obvious - I had messed up the indentation of my cluster.yml so that services were not properly under the “services:” key. After correcting the file structure, the cluster came up correctly. Thank you!!

1 Like