I’m using Helm + Kubernetes and trying to setup a local cluster to have one PostgreSQL database. However it seems that the giving DNS is not working to connect from any pod in the same namespace.

I’m using this for PostgreSQL: https://artifacthub.io/packages/helm/bitnami/postgresql

Some short findings:

• Pod connecting to the Cluster IP of the postgresql service – do not work
• Pod connecting to the postgresql pod IP – works

These are my findings so far and how I set it up below.

helm upgrade -i postgresql bitnami/postgresql 
                --create-namespace 
                --namespace postgres 
                --values ./postgresql.yaml

My postgresql.yaml (values file):

image:
  registry: docker.io
  repository: bitnami/postgresql
  tag: 16.2.0-debian-12-r15
  digest: ""
  pullPolicy: IfNotPresent
  pullSecrets: []
  debug: false

global:
  postgresql:
    service:
      ports:
        postgresql: "5432"

nameOverride: "postgresql"
fullnameOverride: "postgresql"

service:
  type: ClusterIP
  ports:
    postgresql: 5432

auth:
  enablePostgresUser: true
  postgresPassword: "postgres"
  username: "root"
  password: "root"
  database: "test_db"
  replicationUsername: repl
  replicationPassword: "repl"

  secretKeys:
    adminPasswordKey: postgres-password
    userPasswordKey: password
  usePasswordFiles: false

architecture: standalone

containerPorts:
  postgresql: 5432

audit:
  logHostname: false
  logConnections: false
  logDisconnections: false
  pgAuditLog: ""
  pgAuditLogCatalog: "off"
  clientMinMessages: error
  logLinePrefix: ""
  logTimezone: ""

postgresqlDataDir: /bitnami/postgresql/data

resources:
  limits:
    cpu: 1
    memory: 1000Mi
  requests:
    cpu: 500m
    memory: 500Mi

metrics:
  enabled:

I then installed dnsutils to test it out:

kubectl apply -f dns-test.yaml

With the dns-test.yaml:

apiVersion: v1
kind: Pod
metadata:
  name: dnsutils
  namespace: postgres
spec:
  containers:
    - name: dnsutils
      image: registry.k8s.io/e2e-test-images/jessie-dnsutils:1.3
      command:
        - sleep
        - "infinity"
      imagePullPolicy: IfNotPresent
  restartPolicy: Always

I then tried using the dig command inside the dnsutils pod (within the postgres namespace):

kubectl exec -n postgres -i -t dnsutils — dig postgresql.wikijs.svc.cluster.local

Result:

; <<>> DiG 9.9.5-9+deb8u15-Debian <<>> postgresql.wikijs.svc.cluster.local
;; global options: +cmd
;; connection timed out; no servers could be reached
command terminated with exit code 9

I checked the dnsutils resolve.conf to compare to postgresql:

kubectl exec -ti dnsutils -n postgres -- cat /etc/resolv.conf

Result:

nameserver 10.96.0.10
search postgres.svc.cluster.local svc.cluster.local cluster.local
options ndots:5

Resolve.conf in postgresql pod:

nameserver 10.96.0.10
search postgres.svc.cluster.local svc.cluster.local cluster.local
options ndots:5

I then checked the endpoints for postgresql pod:

NAME              ENDPOINTS         AGE
postgresql        10.1.0.120:5432   107m

So it seems the port is exposed.

So two questions:

• Why can’t I connect to the service using the Cluster IP address to reach the pod(s) that is used for it?
• Why will the DNS not apply to the Cluster IP for the service? Currently nothing is added as far as I understand.