11. Ingress¶

Ingress Controller¶

We have dozen configure manager tools for kubernetes.

But some of them is more popular, such as HashiCorp Terraform, RedHat Ansible and Kubernetes native configure manager Kustomize.

helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx

update repository

helm repo update

check to added successfully

helm repo list

install chart

helm install my-ingress-nginx ingress-nginx/ingress-nginx --namespace ingress-nginx --create-namespace

check helm releases

helm list -A

Layers of problems locations

curl http://web-service-ip:node-port

Describe the service and check it's parameters especially Endpoints, Service and Pods

kubectl describe svc <service-name>

kubectl get pods

Pay attention to 'eady' and 'Status'

3.1. Get describe and logs of resources

kubectl describe pods <pod-name>

Check the pod events

3.2. Get pod logs

kubectl logs <pod-name>
kubectl logs <pod-name> -f # follows pod logs
kubectl logs <pod-name> -f --previous # logs of previous container

kubectl get nodes

kubectl -n kube-system get pods

NOTE: It is work if cluster set up via Kubeadm.

If not, you should check via systemd

systemctl status kube-apiserver
systemctl status kube-control-plane
systemctl status kube-schedular

Then get logs via journalctl or kubectl logs

kubectl get nodes

kubectl describe nodes <worker-node-name>

2.1. If your worker node is on 'NOT RAEDY' status, check your 'Kubelet' first

2.2. Get more focus on 'Conditions' column

2.3. If your node status is 'UNKNOWN' that means the 'Kubelet' could not connect to 'API-Server'

check the kubelet via ssh

NOTE: the 'Best Practice' for this purpose is use monitoring tools, like 'Prometheus' via 'Node Exporter'

systemctl status kubelet
journalctl -u kubelet -f

openssl x509 -in /var/lib/kubelet/pki/kubelet.crt -text -noout