[BUG] rancher2_cluster_v2's rke_config[0].etcd[0].s3_config[0].endpoint doesn't seem to be flexible to a value change during the applied Terraform

[irishgordo]

Rancher Server Setup

• Rancher version: v2.9.1
• Installation option: helm-chart
• Proxy/Cert Details: using cert-manager
• Using Harvester vCluster Addon: https://github.com/irishgordo/rancher-harvester-demo/blob/main/harvester-k8-crds/harvester-vcluster-enablement.yaml

Information about the Cluster

• Kubernetes version: v1.29.8-k3s1
• Downstream RKE2 Version: v1.29.8+rke2r1
• Imported Harvester v1.3.2
• Harvester Node Driver

User Information

• What is the role of the user logged in? Admin

Provider Information

• What is the version of the Rancher v2 Terraform Provider in use? v5.0.0
• What is the version of Terraform in use? v1.9.3

Describe the bug

Let's say you are setting up an S3 Compatible endpoint via Terraform somewhere/of-some-kind/at-some-accessilble-place, in this instance it's using Harvester terraform provider to set up a VM running MinIO that was configured w/ Ansible (Ansible Terraform Provider ) + some cloud-init. It just seems to be complaining that it didn't know the ipv4 address of the rke_config[0].etcd[0].s3_config[0].endpoint for the rancher2_cluster_v2 (even when introducing depends_on props in the resources) with:

╷
│ Error: Provider produced inconsistent final plan
│
│ When expanding the plan for rancher2_cluster_v2.rke2-terraform to include new values learned so far during apply, provider "registry.terraform.io/rancher/rancher2" produced an invalid new value for
│ .rke_config[0].etcd[0].s3_config[0].endpoint: was cty.StringVal(""), but now cty.StringVal("https://192.168.104.251:9000").
│
│ This is a bug in the provider, which should be reported in the provider's own issue tracker.

Workaround

• just re-applying the terraform apply works to overcome the bug

To Reproduce

Please read:

• https://github.com/irishgordo/rancher-harvester-demo/blob/main/harvester-terraform-and-rke2/README.md

This relys heavily on having a single-node Harvester v1.3.2, Rancher v2.9.1 (applied w/ vcluster addon manifest linked in rancher server setup above)

Actual Result

• it just shows off that error, the implementing the workaround overcomes that error

Expected Result

• maybe not show an error, maybe allow for the empty string at first but knowing that when the things it may depends_on finish maybe it will have a value that changes? I guess this all depends... if someone's Terraform is cultivating an S3 Compatible resource on the fly or not... 🤷 🤔 ... but I'm not sure, maybe this is bad practice 😅 - super willing to close this out if what is taking place is totally expected

Screenshots

• no screenshot but there is a giant video ~100MB: https://github.com/irishgordo/rancher-harvester-demo/blob/main/harvester-terraform-and-rke2/media/hybrid-demo-speed-increased-by-ffmpeg.mp4

Additional context

• https://github.com/irishgordo/rancher-harvester-demo/blob/main/harvester-terraform-and-rke2/README.md