Upgrade from NGINX Gateway Fabric 2.0.2 to 2.1.1 leaves old CRDs, causing unknown field warnings

[NicolasPires777]

Description:
Hello,

Until a few days ago I was running NGINX Gateway Fabric 2.0.2 on my GKE cluster. I wanted to upgrade to 2.1.1 in order to use workerConnections and autoscaling, but I’m running into issues.

Here’s what I did to upgrade:

# Remove existing gateway
kubectl delete -f gateway.yaml

# Uninstall Helm release
helm uninstall ngf -n nginx-gateway

# Delete CRDs from 2.0.2
kubectl kustomize "https://github.com/nginx/nginx-gateway-fabric/config/crd/gateway-api/standard?ref=v2.0.2" | kubectl delete -f -

Then I reinstalled with 2.1.1:

# Apply CRDs for 2.1.1
kubectl kustomize "https://github.com/nginx/nginx-gateway-fabric/config/crd/gateway-api/standard?ref=v2.1.1" | kubectl apply -f -

# Install Helm release
helm install ngf oci://ghcr.io/nginx/charts/nginx-gateway-fabric \
  --create-namespace -n nginx-gateway \
  -f ngf-values_2-1-1.yaml

When I run this on a fresh dev cluster, everything works fine.
But when I run this on my prod cluster, which previously had 2.0.2 installed, I get the following warnings:

Warning: unknown field "spec.kubernetes.deployment.autoscaling"
Warning: unknown field "spec.workerConnections"

It looks like something from version 2.0.2 is still present, preventing workerConnections and autoscaling from working correctly.

Is there a recommended/official procedure to ensure a clean upgrade from 2.0.2 → 2.1.1, including CRD updates?

Thanks!

[nginx-bot]

Hi @NicolasPires777! Welcome to the project! 🎉

Thanks for opening this issue!
Be sure to check out our Contributing Guidelines and the Issue Lifecycle while you wait for someone on the team to take a look at this.

[bjee19]

Hey @NicolasPires777, thanks for opening this issue.

This is our document on Upgrading NGINX Gateway Fabric, and from the commands you posted above, it seems like you've missed upgrading the NGF specific CRDs. The CRDs you upgraded are the Gateway API CRDs.

Unfortunately uninstalling NGF through helm does not delete the NGF CRDs, and reinstalling it through Helm doesn't upgrade the CRDs.

To upgrade the NGF specific CRDs, you can use this command:

kubectl apply --server-side -f https://raw.githubusercontent.com/nginx/nginx-gateway-fabric/v2.1.1/deploy/crds.yaml

[NicolasPires777]

Thank you very much, that worked! 🙌
I just had to add a flag to force the apply and everything went fine.

By the way, is there a way to disable access logging?
Right now every incoming connection is being logged. Previously, I used to disable it by building a custom image of the data plane based on your image, but now that we already have the worker_connections resolved, I think it would make more sense not to maintain a custom image anymore.

[bjee19]

Hey @NicolasPires777, glad to hear it worked!

You can use the SnippetsFilter CRD to configure/disable access logging. Check out this comment thread on how to do so.

Let me know if you have any other questions!

[NicolasPires777]

The snippets worked!

That’s it — no more questions. Thank you so much again for all your help. Cheers!

[bjee19]

Glad it worked, will close issue for now!