unable to get output from nginx-ingress external ip

Question

unable to get output from nginx-ingress external ip

Prem Jha 6

Hi,

i am trying to install ingress-nginx into the AKS cluster using the kubernetes helm repository (https://kubernetes.github.io/ingress-nginx) and it gets successfully deployed.

and i also got the external ip of ingress but whenever i am trying to hit the external ip over the internet it is giving me gateway timeout error. even i have also checked my NSG rules.

what i have tried :

i thought the problem is in the cluster bt when i installed the nginx from the NGINX official helm repository (https://helm.nginx.com/stable) i got the external ip of nginx.

and when hit that external ip i got the outpit (404 nginx ) which is fine.

but the NGINX official helm repository doesnt comes with the admission controllers.

can you please tell me how to resolve this i want to use kubernetes nginx helm repository(https://kubernetes.github.io/ingress-nginx) but it seems not to be working

srbhatta-MSFT 8,586 Reputation points Microsoft Employee

2022-12-26T11:56:43.483+00:00
Hello @Prem Jha ,
Welcome to Microsoft QnA.
Could you please check the logs of the Ingress controller pods. You can do this by running the following command:

kubectl logs -n ingress-nginx -l app=ingress-nginx

1 answer

Your answer

srbhatta-MSFT 8,586 Reputation points Microsoft Employee

2022-12-26T11:56:43.483+00:00

Hello @Prem Jha ,
Welcome to Microsoft QnA.
Could you please check the logs of the Ingress controller pods. You can do this by running the following command:

kubectl logs -n ingress-nginx -l app=ingress-nginx

Answer 1

Amjad Aljunaidi 6 Microsoft Employee

actually it is know issue with http/s and TCP
so you should add this annotation to make sure it is configure correctly
"service.beta.kubernetes.io/azure-load-balancer-health-probe-request-path: /healthz"

already explained here https://github.com/Azure/AKS/releases/tag/2022-09-11

"
Behavioral Changes
For Kubernetes 1.24+ the services of type LoadBalancer with appProtocol HTTP/HTTPS will switch to use HTTP/HTTPS as health probe protocol (while before v1.24.0 it uses TCP). And / will be used as the default health probe request path. If your service doesn’t respond 200 for /, please ensure you're setting the service annotation service.beta.kubernetes.io/port_{port}_health-probe_request-path or service.beta.kubernetes.io/azure-load-balancer-health-probe-request-path (applies to all ports) with the correct request path to avoid service breakage.
"

Rudhisundar Beura 0 Reputation points

2023-06-18T18:50:26.9233333+00:00

Yes, this seems to be the case.
Daniel Förberg 0 Reputation points

2023-08-10T15:11:44.49+00:00

Hi, but have the same issue as the requester here. But the public ip for nginx do not give 404 it timeouts after a while. Seems something is blocking the traffic. Followed instructions to have the subnet and agentpool NSG the same. But no difference. AKS CNI and NGinx. Kube Version 1.26.6. Thanks in advance. Installed Nginx with recommended version for Azure v1.81 with the Kube Manifest.
Prem Jha 6 Reputation points

2023-08-11T02:02:23.0233333+00:00

Hi i was facing this issue earlier when i was trying to install the ingress controller from the helm in this case it didnt worked.

after that i tried installing it through manifest file of ingress controller https://github.com/kubernetes/ingress-nginx/blob/main/deploy/static/provider/cloud/deploy.yaml and it started working.

reason i am thinking is that : ingress controller dont have proper role to control the ingress controller svc (means that public ip) [i verified it by installing only the "kind: Role" after helm and it also worked]

straight forward way is to use manifest file of ingress controller and modify it according to your needs

Thanks
Daniel Förberg 0 Reputation points

2023-08-11T07:07:35.7833333+00:00

@Prem Jha Hi thanks but i have used the manifests to install the Ingress Controller from start the recommended version v1.8.1, from the link you are providing here. What exactly do you mean by install only Kind: "Role", which role cause in Azure have set the roles on the MSI user identity. Or do mean some specific role in Kubernetes. Cause the symtom is not only for Nginx Ingress Controller. Also for Garafana, Prometheus and ArgoCD. Each have LoadBalancer kind. The default network policies and some for ArgoCD default ones also. Use ArgoCD with port forwarding. Must be something Azure/AKS specific. The annotations mentioned in this thread could that be the reason ?
Daniel Förberg 0 Reputation points

2023-08-11T07:19:42.7766667+00:00
Also followed this guide: https://learn.microsoft.com/en-us/troubleshoot/azure/azure-kubernetes/custom-nsg-blocks-traffic . The conclusion of that guide the pods are accessible and services can access the, but the ingress with external IP cannot reach it service or such. It times out after a while.
The message is as follows:
curl -v http://EIP:port

Trying EIP:port

connect to EIP port port failed: Operation timed out

Failed to connect to EIP port after 75007 ms: Couldn't connect to server

Closing connection 0

curl: (28) Failed to connect to EIP port after 75007 ms: Couldn't connect to server

The NSG are in place for the inbounds and equally in subnet NSG. As recommended. Both Nodes and Pods runs in same subnet. The services have their own CIDR as default 10.0.0.0/16. Everything is setup with terraform providers. Thanks in Advance.
Daniel Förberg 0 Reputation points

2023-08-11T07:43:03.9566667+00:00

Also found now that this should be a workaround also,

Another on the topic the annotation provided does work someone else had pointed that out it works for the ingress controller but not the ingresses it uses

After some more research, we have found a workaround which is to enable the default backend (https://kubernetes.github.io/ingress-nginx/user-guide/default-backend/) on the nginx controller by setting the following defaultBackend.enabled: true, but it doesn't sound right since it's not reaching the azure health endpoint anymore but just the default backend that nginx deploys now.

Default backend - Ingress-Nginx Controller

https://github.com/Azure/AKS/issues/3621
Daniel Förberg 0 Reputation points

2023-08-11T08:50:35.7066667+00:00

Tried to apply the according to latest post but i says no change, not applied cause that param did not exist. Used:
kubectl -n ingress-nginx patch deployment ingress-nginx-controller -p '{"spec": {"defaultBackend": {"enabled": true}}}'

Share via

unable to get output from nginx-ingress external ip

1 answer

Your answer