cloudnative-pg-operator

PoC for deploying and operating the CloudNativePG operator

Installation

→ # install operator in `cnpg-system` namespace
→ kubectl apply -f \
    https://raw.githubusercontent.com/cloudnative-pg/cloudnative-pg/release-1.20/releases/cnpg-1.20.2.yaml

→ kubectl get deployments \
    -n cnpg-system cnpg-controller-manager

→ kubectl describe deployments \
    -n cnpg-system cnpg-controller-manager

Deployment

Secrets

We create two secrets, one for the admin and one for the app user, where in both cases the password equals the username.

Cluster

→ kubectl apply -f \
    https://raw.githubusercontent.com/m99coder/cloudnative-pg-operator/main/cluster.yaml

→ # all pods are labeled with the cluster name
→ # kubectl get pods -n cnpg -l cnpg.io/cluster=cnpg-pg15-cluster
→ kubectl get pods -n cnpg
NAME                  READY   STATUS    RESTARTS   AGE
cnpg-pg15-cluster-1   1/1     Running   0          7m13s
cnpg-pg15-cluster-2   1/1     Running   0          6m23s
cnpg-pg15-cluster-3   1/1     Running   0          5m29s

While spinning up the cluster, one can see different sidecars in action. The first is initdb for setting up the initializing the database server. As soon as the primary is ready, the two secondaries run join sidecars before coming up on their own.

→ kubectl get svc
NAME                   TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)    AGE
cnpg-pg15-cluster-r    ClusterIP   10.100.63.101   <none>        5432/TCP   85m
cnpg-pg15-cluster-ro   ClusterIP   10.100.204.36   <none>        5432/TCP   85m
cnpg-pg15-cluster-rw   ClusterIP   10.100.8.124    <none>        5432/TCP   85m

→ kubectl get pods/cnpg-pg15-cluster-1 -o json | jq '.spec.containers[].ports'
[
  {
    "containerPort": 5432,
    "name": "postgresql",
    "protocol": "TCP"
  },
  {
    "containerPort": 9187,
    "name": "metrics",
    "protocol": "TCP"
  },
  {
    "containerPort": 8000,
    "name": "status",
    "protocol": "TCP"
  }
]

→ kubectl expose pods/cnpg-pg15-cluster-1 \
    --name=cnpg-pg15-cluster-1-metrics \
    --port=9187 \
    -n cnpg

→ kubectl get ep
NAME                          ENDPOINTS                                                    AGE
cnpg-pg15-cluster-1-metrics   172.24.220.162:9187                                          12s
cnpg-pg15-cluster-r           172.24.220.162:5432,172.24.220.178:5432,172.24.247.51:5432   4h22m
cnpg-pg15-cluster-ro          172.24.220.178:5432,172.24.247.51:5432                       4h22m
cnpg-pg15-cluster-rw          172.24.220.162:5432                                          4h22m

→ kubectl port-forward
    -n cnpg \
    svc/cnpg-pg15-cluster-1-metrics 9187

→ open http://localhost:9187/metrics

Monitoring

→ kubectl get crds -n monitoring

→ # get endpoints for:
→ #   - alert manager (port 9093)
→ #   - prometheus (port 9090)
→ kubectl get endpoints -n monitoring

→ kubectl get podmonitors.monitoring.coreos.com -n cnpg
NAME                AGE
cnpg-pg15-cluster   152m

→ kubectl port-forward \
    -n monitoring \
    svc/prometheus-app 9090

Seems like the PodMonitor is not properly configured to monitor the cluster pods.

Cleanup

→ # operator and cluster
→ kubectl delete -f \
    https://raw.githubusercontent.com/m99coder/cloudnative-pg-operator/main/cluster.yaml
→ kubectl delete -f \
    https://raw.githubusercontent.com/cloudnative-pg/cloudnative-pg/release-1.20/releases/cnpg-1.20.2.yaml

m99coder/cloudnative-pg-operator