Classic clusters: Why does the master status have an ingress IP address for NLB error?

Classic infrastructure

When you run ibmcloud oc cluster get -c <cluster_name_or_ID>, you see the following error message in the Master Status field.

CAE003: Unable to determine the ingress IP address for the network load balancer.

Additionally, when you run ibmcloud oc nlb-dns create to create a subdomain for a network load balancer (NLB), the command might fail with a message that the cluster is not found, the input parameters are incorrect, or you don't have the required roles.

The Konnectivity server could not be configured because load balancer IP address that exposes the default Ingress controller could not be found. The Ingress controller's load balancer service might not have been assigned an IP address because your cluster does not have a subnet with available portable IP addresses, or the load balancer setup did not complete.

Verify that your cluster has available subnets, and that the load balancer setup completed successfully.

Verifying that your cluster has available subnets

Check that your cluster has a Subnet CIDR for public and private subnets. If you set up a private VLAN-only cluster, you might have only a private subnet.

ibmcloud oc cluster get --cluster <cluster_name_or_ID> --show-resources

Example output

NAME:                           <cluster_name>   
...

Subnet VLANs
VLAN ID   Subnet CIDR        Public   User-managed   
2345678   10.xxx.xx.xxx/29   false    false   
2876543   169.xx.xxx.xxx/29  true     false

If the cluster does not have a subnet, create a subnet for the cluster or add an existing subnet from your account to the cluster.
If the cluster does have a subnet, check for available portable IP addresses and if necessary, add more portable IP address by adding a subnet.
Refresh the master to restart the Konnectivity setup so that it uses the available subnet.
```
ibmcloud oc cluster master refresh --cluster <cluster_name_or_ID>
```

Verifying that the load balancer setup completed successfully

Check that the ibm-cloud-provider-ip-* pods for the load balancer are in a Running status.
```
oc get pods -n ibm-system | grep ibm-cloud-provider-ip
```
If a pod is not running, review the Events in the pod details to troubleshoot the issue further.
```
oc describe pod -n kube-system <pod_name>
```
After you resolve the load balancer pod issue, refresh the master to restart the NLB setup.
```
ibmcloud oc cluster master refresh --cluster <cluster_name_or_ID>
```