Kong (with DB) installation using helm on EKS

Kubernetes
1

Hi, I am trying to install Kong using helm on the EKS cluster. I am using Postgres RDS as DB.

Below is the content of values.yaml file

env:
  database: "postgres"
  pg_host: aws-kong.123456789.a-aws-region.rds.amazonaws.com
  pg_port: 5432
  pg_user: dummyadmin
  pg_password: dummypassword
  pg_database: kongdb
  nginx_worker_processes: "1"
  proxy_access_log: /dev/stdout
  admin_access_log: /dev/stdout
  admin_gui_access_log: /dev/stdout
  portal_api_access_log: /dev/stdout
  proxy_error_log: /dev/stderr
  admin_error_log: /dev/stderr
  admin_gui_error_log: /dev/stderr
  portal_api_error_log: /dev/stderr
  prefix: /kong_prefix/
admin:
  annotations: {}
  #  service.beta.kubernetes.io/aws-load-balancer-proxy-protocol: "*"
  http:
    enabled: true
    servicePort: 8001
    containerPort: 8001
    parameters: []
  tls:
    enabled: true
    servicePort: 8444
    containerPort: 8444

Using below helm command to install Kong

helm upgrade -i kong -n kong -f values.yaml kong/kong --set ingressController.installCRDs=false;

Once the above helm command is executed, the pods are not coming to running state, below is status, it is stuck to initialize.

NAME                                           READY   STATUS     RESTARTS   AGE
kong-kong-6447659bbf-9jl9d        0/2     Init:0/1           0          7m59s
kong-kong-init-migrations-87r74   0/1     Init:0/1          0          7m59s

@traines @hbagdi Please help, what can be the issue?

2

These two services require your DB to run first, you can describe your pod to see the specific pod information.

3

The init containers do indeed check for database availability. You can check

kubectl logs kong-kong-init-migrations-87r74 -c wait-for-postgres
kubectl logs kong-kong-6447659bbf-9jl9d -c wait-for-db

Those should roughly indicate what’s failing when they try to connect. The first should generally fail with either a DNS failure or some sort of TCP connection failure. The latter seems more likely; I’d first check to confirm that you have the necessary AWS network ACLs and security restrictions to allow traffic between your EKS workers and the database instance:
https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_VPC.Scenarios.html
https://docs.aws.amazon.com/eks/latest/userguide/sec-group-reqs.html

4

Thanks for pointing out @traines.

It was an issue with a TCP connection between Pods and AWS RDS. After fixing the TCP connection, I can see kong migration in Completed state and kong in Running state.

$ kubectl get pods -n kong
**NAME                                                READY   STATUS      RESTARTS   AGE**
kong-kong-6447659bbf-9jl9d            2/2        Running          1                   2h
kong-kong-init-migrations-87r74       0/1       Completed      0                  2h