Acorn won't work with Civo's default node port based traefik ingress

[cjellick]

When using acorn on a civo cluster with the default node port based traefik ingress controller, any acorns apps that are deployed will have their endpoints sit in pending forever.

This is because the ingress controller isn't updating the status of the ingress with loadbalancer information.

To get trafefik to update ingress status, you have to add one of the following fields to its configuraiton:

- --providers.kubernetesingress.ingressendpoint.ip=<host ip>
- --providers.kubernetesingress.ingressendpoint.hostnaem=<host name>
- --providers.kubernetesingress.ingressendpoint.publishedservice=<service name>

Note that Civo has a "LoadBalancer" traefik ingress which adds this to the configuration:

- --providers.kubernetesingress.ingressendpoint.publishedservice=kube-system/traefik

and everything works as expected.

Here are some related upstream issues in traefik:

• Publish IPs from ClusterIP and NodePort services on Ingresses traefik/traefik#7972
• Traefik v2 Kubernetes Ingress Controller doesn't update Ingress resource status traefik/traefik#6303

We can't address this directly on the acorn side as having the status of the ingress properly updated is a requirement. However, we'll be adding some pre-flight checks to ensure ingress is fully functional as part of acorn install

[cjellick]

Found while testing obot-platform/obot#452

[cjellick]

The chart in civo should be fixed. I'll test later today and close this out if it is.

[cjellick]

Woohoo! @saiyam1814 fixed the ingress in civo. This is working now