Cross-posting from #general, since I just realized this channel exists.

Are you passing the

Provider

instance in as the

provider

to all of the Kubernetes objects you're creating?

Yes, at least I’ve triple-checked that I am. Is there a good general way to debug this kind of problem? Like a way to view the providers for all of the resources that I’m creating?

For debugging, I’ve set a garbage

KUBECONFIG

env var and watched for which resources fail.

Yeah, that’s the state I’m in right now pretty much, except I think I’m passing in the correct provider rather than a garbage one. This is what I have in the

__init__

of my `ComponentResource`:

provider = pulumi_kubernetes.Provider('kubernetes-provider', kubeconfig=args.kubeconfig)
        self._opts = opts.merge(pulumi.ResourceOptions(provider=provider)).merge(pulumi.ResourceOptions(parent=self))

And this is the code that is creating the object that is failing:

self._configmap_license = pulumi_kubernetes.core.v1.ConfigMap(
                f'{self._name}-license',
                metadata=meta_v1.ObjectMetaArgs(
                    name=f'{self._name}-license',
                    namespace=self._args.namespace_name
                ),
                data={
                    'seeq_server_license': license_file.read()
                },
                opts=self._opts
            )

Could it be that

merge

is retaining a bad provider somehow? It looks like providers can be a list or collection, and

merge

is documented as retaining all items in all collections.

That’s an interesting thought. That might be exactly what’s happening.

np!