K
Q

How to access pod in k8 cluster via url

4/14/2020

I have a service running in a cluster in a namespace:

NAME                                    TYPE           CLUSTER-IP       EXTERNAL-IP                                                                     PORT(S)                                                       AGE
amundsen-frontend                       LoadBalancer   10.100.59.220    a563823867e6f11ea82a90a9c116adac-124ae00284b50400.elb.us-west-2.amazonaws.com   80:31866/TCP                                                  70m

And when I run pods:

kubectl get pods
NAME                                                  READY   STATUS    RESTARTS   AGE
amundsen-frontend-595b49d856-mkbjj                    1/1     Running   0          74m
amundsen-metadata-5df6c6c8d8-nrk9f                    1/1     Running   0          74m
amundsen-search-c8b7cd9f6-mspzr                       1/1     Running   0          74m
dsci-amundsen-elasticsearch-client-65f858c656-znjfd   1/1     Running   0          74m
dsci-amundsen-elasticsearch-data-0                    1/1     Running   0          74m
dsci-amundsen-elasticsearch-master-0                  1/1     Running   0          74m

I'm not really sure what to do here. How do I access the url? Can I port forward in development? What do I do in production? The front-end pod is one I want to access, so is the search pod.

This is what's in my charts.yaml for helm:

frontEnd:
  ##
  ## frontEnd.serviceName -- The frontend service name.
  ##
  serviceName: frontend
  ##
  ## frontEnd.imageVersion -- The frontend version of the metadata container.
  ##
  imageVersion: 2.0.0
  ##
  ## frontEnd.servicePort -- The port the frontend service will be exposed on via the loadbalancer.
  ##
  servicePort: 80
-- Jwan622
kubernetes-helm

Similar Questions

Kubernetes service account - token signature validation
How to run multiple insurance of MySQL in kubernetes
Inter pods communication with dns name not working in kubernetes
Prometheus storage persistent understanding when installing from [stable/prometheus] on kubernets/eks with Grafana

1 Answer

4/16/2020

With so little information I don't know if I can solve your problem, but will try to help you find it.

To start with it will be helpful if we can see your service and pod config?

kubectl get sa amundsen-frontend -o yaml
kubectl get pod amundsen-frontend-595b49d856-mkbjj -o yaml

You can try to reach the fronted from another pod, this will help figure out if the problem is in the pod or ingress layer. To gain shell access inside search pod container run:

kubectl exec -it amundsen-search-c8b7cd9f6-mspzr  --container <<name of container>>  -- sh

If you have only one container in the pod you can omit the container part from the command above

Once inside check if your are able to reach amundsen-frontend-595b49d856-mkbjj with curl

curl amundsen-frontend-595b49d856-mkbjj
curl amundsen-frontend-595b49d856-mkbjj:31866

If you are able to establish communication, then look for the problem in the ingress layer. You may want to look at your ingress logs to see why it's timing out.

Network security groups in AWS as also worth exploring.

Is your ingress configured properly?

-- Filip Nikolov
Source: StackOverflow