Nginx fails to reload due to long server name

[kate-osborn]

Describe the bug
Nginx fails to reload with the following error message:

nginx: [emerg] could not build server_names_hash, you should increase server_names_hash_bucket_size: 64

To Reproduce
Steps to reproduce the behavior:

1. Deploy NKG
2. Create the cafe app:
3. Apply the following manifest:

apiVersion: gateway.networking.k8s.io/v1beta1
kind: Gateway
metadata:
  name: gateway
  labels:
    domain: k8s-gateway.nginx.org
spec:
  gatewayClassName: nginx
  listeners:
  - name: http
    port: 80
    protocol: HTTP
---
apiVersion: gateway.networking.k8s.io/v1beta1
kind: HTTPRoute
metadata:
  name: coffee
spec:
  parentRefs:
  - name: gateway
    sectionName: http
  hostnames:
  - "abcdefghijklmn.too.long.server.name.example.org" // 47 characters seems to be the limit
  rules:
  - matches:
    - path:
        type: PathPrefix
        value: /coffee
    backendRefs:
    - name: coffee
      port: 80

5. Get the logs of the nginx container: kubectl logs -n nginx-gateway <nginx-gateway-pod> -c nginx -- nginx -T
   Look for the following error message:

2023/05/09 15:22:56 [emerg] 41#41: could not build server_names_hash, you should increase server_names_hash_bucket_size: 64

Expected behavior
Nginx should reload successfully and the route should be configured.

Your environment

• Version of the NGINX Kubernetes Gateway: main, commit: d62cf56
• Version of Kubernetes: 1.26
• Kubernetes platform (e.g. Mini-kube or GCP): kind

[saiyam1814]

thanks for creating this :)

[kate-osborn]

Discussed as a team today and decided to fix this by increasing the default server_names_hash_bucket_size to 256 and server_names_hash_max_size to 1024. NGINX Ingress Controller uses these defaults, which should work in most cases.

In a later release, we can make these values configurable or even calculate them based on the length of the server name.