Lab 6.4 cannot connect to webserver using curl though clusterip is correct.

vsirnapalli

Getting an error "curl: (7) Failed to connect to 10.111.70.140 port 80: Connection refused".
All traffic is open on for EC2 instance

chrispokorni

Hi @vsirnapalli,

In 6.4 most errors are encountered because of mismatching labels and selectors. A Service is able to forward traffic to its expected Endpoints if the selector's key-value pair matches the key-value pair labels on the Pods, and a slight typo could produce a connection refused error.

If it is unrelated to labels and selectors, perhaps more context would help with troubleshooting - the step number, the command, the full output, and possibly other troubleshooting steps attempted with their outputs.

Regards,
-Chris

liorzivi

Hi @chrispokorni ,

This happened the same for me.

1. Validated the network rules for my Azure VMS are opened.

2. The service describe commands returns that the service is defined with:
   Port: 80 80/TCP
   TargetPort: 80/TCP
   NodePort: 80 32000/TCP
   Selector: example=second

3. The pod is configured to run the nginx image and has the corret label:
   Labels: example=second

I was able to reach the nginx service by using the private ip of the pod and the private ip of the service, it only failed while trying the public ip of both my master and worker nodes on Azure.

chrispokorni

Hi @liorzivi,

Azure networking requires additional configuration steps for the CNI network plugin to work, and it is less flexible in terms of supported plugins. For these reasons Azure is not a supported infrastructure for the lab environment, but AWS-EC2 and GCP-GCE should work without any issues.
However, you may be able to find a few discussions with dated solutions for Azure VMs, that students shared in the past on this forum. They could help to get you started with the specific configuration options required by Azure.

Regards,
-Chris

liorzivi

Thanks, got it working