having trouble enabling ingress nginx addon

[mainlyetcetera]

What Happened?

been attempting to use ingress nginx addon, and it's been failing unpredictably in ways that don't make sense. all pods, deployments, and services look like their ports should line up, but the nginx service is convinced there's no endpoints. so back-tracking, and trying to create a cleaner poc, and now i can't successfully enable the addon at all.

listing os as other - it's PopOs from System76, and is pretty close to Ubuntu.

Attach the log file

log.txt

minikube_addons.log

Operating System

Other

Driver

None (Baremetal)

[mainlyetcetera]

i have tried troubleshooting from this and this, to no avail. would really appreciate some help with this.

[mainlyetcetera]

haha managed to reach a solution, it seems. there were a couple moving pieces. first:

• i had minikube addons enable ingress going, which tried to add the controller pod, controller svc, and admission svc
  • the controller pod was failing per the above errors
• after wiping the cluster, i tried installing ingress with helm install ingress-nginx oci://ghcr.io/nginx/charts/nginx-ingress --version 2.1.0 -n ingress-nginx --create-namespace
  • this gave me what looked like a working controller pod, but nothing in terms of an admission svc, and connecting any ingress resources error'd as such: "no admission svc found"
  • enabling the ingress addon after installing this controller pod did not rectify the issue - i'm assuming that's because the admission svc the addon spun up wasn't aware of helm's controller pod
• after wiping the cluster again, i found these docs, which gave me a different helm command to deploy ingress. stressing i needed to wipe the cluster before this, as the other installations left behind a whole host of other resources (i was on deleting clusternodes before i gave up) that were blocking the given command. but, after wiping the cluster, this command spun up a working controller and admission svc, and it actually works:

helm upgrade --install ingress-nginx ingress-nginx \
  --repo https://kubernetes.github.io/ingress-nginx \
  --namespace ingress-nginx --create-namespace

i haven't taken the time to understand the difference between the github and gcr repos, but:

• the mismatch is somewhere around: gcr only sets up the controller on its face (maybe there was a way to follow that up with the admission svc i just didn't find)
• it was paramount to wipe the cluster in-between each try, as the previous tries were leaving behind several background resources that would then conflict with the current attempt