error: can't use Helm Release with unreachable cluster. Reason: "unable to load schema information from the API server: Get \"https://10.0.1.20:6443/openapi/v2?timeout=32s\": dial tcp 10.0.1.20:6443: connect: no route to host"

[mitchellmaler]

Issue details

I am using the new Release resource but I am unable to use it in the same up prior to the kubernetes control-plane and provider being created. It tries to connect to a cluster that hasn't been built yet and errors on the preview.

     Type                                                               Name                                                                 Plan       Info
     pulumi:pulumi:Stack                                                homelab-iac-main
 +   ├─ pkg:index:k3sNode                                               k3s-cp-01                                                            create
 +   │  ├─ pulumi-python:dynamic:Resource                               k3s-cp-01-ignition                                                   create
 +   │  ├─ pulumi-python:dynamic:Resource                               k3s-cp-01-before-node                                                create
 +   │  ├─ vsphere:index:VirtualMachine                                 k3s-cp-01                                                            create
 +   │  └─ pulumi-python:dynamic:Resource                               k3s-cp-01-after-node                                                 create
 +   ├─ pulumi:providers:kubernetes                                     k8s-cluster-provider                                                 create
     └─ kubernetes:helm.sh/v3:Release                                   truenas-iscsi-csi                                                               1 error

Diagnostics:
  kubernetes:helm.sh/v3:Release (truenas-iscsi-csi):
    error: can't use Helm Release with unreachable cluster. Reason: "unable to load schema information from the API server: Get \"https://10.0.1.20:6443/openapi/v2?timeout=32s\": dial tcp 10.0.1.20:6443: connect: no route to host"

I have my provider and the release setup to depend on the node resources but that doesn't seem to make a difference

k3s_cp_01 = K3sNode(
    name="k3s-cp-01",
    args=K3sNodeArgs(
        token=k3s_cluster_token,
        cluster_api_ip=k3s_cluster_ip,
        cluster_api_fqdn=k3s_cluster_fqdn,
        node_folder=k3s_cluster_folder,
        admin_kube_config=admin_kube_config,
        node_ip=k3s_cluster_ip,
        server=True,
        initial_server=True,
        server_ca_cert=k3s_cluster_server_ca,
        server_ca_key=k3s_cluster_server_ca_key,
        client_ca_cert=k3s_cluster_client_ca,
        client_ca_key=k3s_cluster_client_ca_key,
    ),
)


kubernetes_provider = pulumi_kubernetes.Provider(
    resource_name="k8s-cluster-provider",
    args=pulumi_kubernetes.ProviderArgs(
        kubeconfig=admin_kube_config.kube_config, enable_dry_run=True
    ),
    opts=ResourceOptions(depends_on=[k3s_cp_01.vm, k3s_cp_01.node_manager_after_node]),
)

truenas_iscsi_csi = Release(
    resource_name="truenas-iscsi-csi",
    args=ReleaseArgs(
        repository_opts=RepositoryOptsArgs(
            repo="https://democratic-csi.github.io/charts/",
        ),
        values=truenas_iscsi_csi_values,
        chart="democratic-csi",
        name="democratic-csi",
        namespace="kube-system",
    ),
    opts=ResourceOptions(
        delete_before_replace=True,
        provider=kubernetes_provider,
        depends_on=snapshotter_resources + [k3s_cp_01.node_manager_after_node],
    ),
)

Steps to reproduce

1. Setup code to create a control-plane, provider, and deploy a release
2. run up

Expected: The Release preview to ignore that the cluster doesn't exist yet since the control-plane node and provider is being created in the same run
Actual: tries to connect to a none existing control-plane api

[emirozmen07]

I'm still having the same issue.