EMR on EKS: managed endpoint terminates without any useful clue

Question

I am trying to get an EMR Studio running on an existing EKS cluster by following the instructions here: https://docs.aws.amazon.com/emr/latest/ManagementGuide/emr-studio-create-eks-cluster.html
Most of those steps went smoothly, but the issue occurred when I tried to create a managed endpoint with the following command:
```
aws emr-containers create-managed-endpoint --name [ENDPOINT_NAME] --virtual-cluster-id [VIRTUAL_CLUSTER_ID] --type JUPYTER_ENTERPRISE_GATEWAY --release-label emr-6.2.0-latest --execution-role-arn [EXECUTION_ROLE_ARN] --certificate-arn [CERTIFICATE_ARN]
``` 
The endpoint has been created, it stayed in CREATING status for a few minutes, then it terminated, without giving any clues what might be the problem.
The result of listing the endpoints is the following:
```
{
            "id": "2w5v7j307hut1",
            "name": "ENDPOINT_NAME",
            "arn": "ENDPOINT_ARN",
            "virtualClusterId": "VIRTUAL_CLUSTER_ID",
            "type": "JUPYTER_ENTERPRISE_GATEWAY",
            "state": "TERMINATED_WITH_ERRORS",
            "releaseLabel": "emr-6.2.0-latest",
            "executionRoleArn": "EXECUTION_ROLE_ARN",
            "certificateArn": "CERTIFICATE_ARN",
            "createdAt": "2022-05-09T19:58:21+00:00",
            "stateDetails": "Unknown",
            "failureReason": "INTERNAL_ERROR",
            "tags": {}
        },
```
Could anyone give some suggestions how to troubleshoot this? Is there a way to access logs to figure out what could be wrong?

Thanks a lot!
Simon

Answer

hello, this can happen when all the steps for creating the managed endpoint were not followed properly. You can follow below docs :

https://docs.aws.amazon.com/emr/latest/EMR-on-EKS-DevelopmentGuide/connect-emr-studio.html
https://catalog.us-east-1.prod.workshops.aws/workshops/1f91e1d4-5587-40ff-8d5d-54fc86e0ddc1/en-US/advanced/emr-studio

In my experience, this usually happens when the loadbalancer and ingress controller has some issue while creation. You should consider checking them, as below :

```
$ kubectl get deployment -n kube-system aws-load-balancer-controller
NAME                           READY   UP-TO-DATE   AVAILABLE   AGE
aws-load-balancer-controller   1/1     1            1           2m33s

# you might see some output above, note, it should be ready and available then check the status of ingress controller

$ kubectl get ingress -A
```

EMR on EKS: managed endpoint terminates without any useful clue

Relevanter Inhalt