Hi, Running EKS in AWS where everything is setup with Pulumi. Upgraded EKS version and also Pulumi code from v1beta to v1. Also creating an ALB in Pulumi code with ingress. Now I am facing this error and have no idea how to resolve. Can anyone help pointing out what options I have to resolve it? Regards,

kubernetes:elbv2.k8s.aws/v1beta1:IngressClassParams (kube-system/alb):

error: resource kube-system/alb was not successfully created by the Kubernetes API server : ingressclassparams.elbv2.k8s.aws "alb" already exists

does this issue roughly match what you’re seeing? https://github.com/pulumi/pulumi-kubernetes/issues/2035

Yes, I think that is the identical issue. I also use

k8s.helm.v3.Chart()

. Seems it was resolved for some by switching

k8s.helm.v3.Chart() to k8s.helm.v3.Release()

but unsure if I can do that on an existing setup? Can I just change the Pulumi code? Any other options?

if you can tolerate some downtime you can apply DeleteBeforeReplace to the chart. switching to a release should also work but you’ll probably need to delete the chart first. this is also worth understanding first https://www.pulumi.com/registry/packages/kubernetes/how-to-guides/choosing-the-right-helm-resource-for-your-use-case/

If you're trying to apply DeleteBeforeReplace to the entire Helm chart, that is currently not directly supported by Pulumi's Helm integration.

Downtime would be OK so can try that route. Do DeleteBeforeReplace really work on the entire chart or how do you mean?

we should probably update those docs - nowadays you can use a v4 chart with a transform to add the resource option to child resources https://www.pulumi.com/docs/concepts/options/transforms/

Here is the Pulumi code creating the chart. We are using V3. NOTE! We have not updated this code for years but now upgrading Pulumi nuget packages we get this new problem where "alb" already exists. Not sure how to use transforms in this case? Does the code example help to better understand the problem? Do not know how to proceed being completely stuck now and cannot deploy the ingress changes (new domains/services etc).

what were the packages and versions you upgraded from/to?

Left -> from, right -> to. Also changed all "V1Beta1" to "V1" because newer Kubernetes version required that.

what does your pulumi preview look like?

image.png

i’m still not clear on why it’s trying to create the new ingress class (was the chart version also updated?). from the issue it sounds like a bug in v3.x of the provider — not sure if it also exists in v4.x, but one thing that changes between the two is server side apply, so conflicts like this aren’t raised (levi’s comment on the issue explains how you can turn this on for v3) there’s always the blunt hammer of deleting the existing ingressclassparams and then updating to let pulumi re-create it. not pretty but it should unblock you i think.