Bug: stackit_postgresflex_instance node_type unknown

[mitterle-sit]

Description

Tofu apply fails when using provider v0.105.0

Steps to reproduce

resource "stackit_postgresflex_instance" "example" {
  project_id = "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
  name       = "example-instance"
  network = {
    acl = ["XXX.XXX.XXX.X/XX", "XX.XXX.XX.X/XX"]
  }
  flavor = {
    cpu       = 4
    ram       = 8
  }
  storage = {
    class = "premium-perf2-stackit"
    size  = 5
  }
 replicas = 3
  version  = "17"
}

1. Create an instance with old provider version before v0.105.0 e.g. v0.104.0
2. Upgrade your provider to v0.105.0

Actual behavior

Error: Provider returned invalid result object after apply
│
│ After the apply operation, the provider still indicated an unknown value
│ for
│ stackit_postgresflex_instance.example.flavor.node_type.
│ All values must be known after apply, so this is always a bug in the
│ provider and should be reported in the provider's own repository. OpenTofu
│ will still save the other known object values in the state.

Expected behavior

Apply works

Environment

• OS: macOS
• Terraform version (see terraform --version): OpenTofu v1.12.3
• Version of the STACKIT Terraform provider: v0.105.0

[SerseusWasTaken]

Hello,
thank you for reporting this issue. Unfortunately, i am unable to reproduce the issue.
The configuration you provided is only compatible with provider version v0.105.0 and higher. So i have two questions:

1. Can you please provide the exact configuration you used with provider version v0.104.0?
2. Did you change anything in the configuration while upgrading the provider versions?

Thank you.

[mitterle-sit]

Hi @SerseusWasTaken sorry for the example typo.
Here a full example from our module

resource "stackit_postgresflex_instance" "postgres_database_instance" {
  project_id      = var.stackit_project_id
  name            = var.database_instance_name
  acl             = local.database_acl
  backup_schedule = var.database_backup_schedule
  flavor = {
    cpu = var.database_cpu
    ram = var.database_memory_in_gb
  }
  replicas = var.database_replicas
  storage = {
    class = var.database_storage_class
    size  = var.database_storage_size_in_gb
  }
  version = var.postgres_version

  lifecycle {
    prevent_destroy = true
  }
}

~ update in-place (current -> planned)
OpenTofu will perform the following actions:
  # module.postgres_database.stackit_postgresflex_instance.postgres_database_instance will be updated in-place
  ~ resource "stackit_postgresflex_instance" "postgres_database_instance" {
      ~ backup_schedule = "00 00 * * *" -> "0 0 * * *"
      ~ flavor          = {
            id          = "2.4"
          + node_type   = (known after apply)
            # (3 unchanged attributes hidden)
        }
        id              = "***,***,****"
        name            = "xxx"
      ~ network         = {
          ~ access_scope     = "PUBLIC" -> (known after apply)
          ~ acl              = [
              - "xxx.xxx.xxx/32",
            ] -> (known after apply)
          + instance_address = (known after apply)
          + router_address   = (known after apply)
        } -> (known after apply)
        # (10 unchanged attributes hidden)
    }

Plan: 0 to add, 1 to change, 0 to destroy.
module.postgres_database.stackit_postgresflex_instance.postgres_database_instance: Modifying... [id=***,***,****]
╷
╷
│ Error: Provider returned invalid result object after apply
│ 
│ After the apply operation, the provider still indicated an unknown value
│ for
│ module.postgres_database.stackit_postgresflex_instance.postgres_database_instance.flavor.node_type.
│ All values must be known after apply, so this is always a bug in the
│ provider and should be reported in the provider's own repository. OpenTofu
│ will still save the other known object values in the state.
╵

[mitterle-sit]

If apply a second time the issue is gone seems to be a state set problem at some point in conjunction maybe with older instances

[mitterle-sit]

Next interesting issue also gone after second apply maybe also related to #1628 ?

╷
│ Error: Provider produced inconsistent result after apply
│ 
│ When applying changes to
│ module.postgres_database.stackit_postgresflex_instance.postgres_database_instance,
│ provider "provider[\"registry.opentofu.org/stackitcloud/stackit\"]"
│ produced an unexpected new value: .flavor.description: was
│ cty.StringVal("Medium, Compute optimized"), but now
│ cty.StringVal("PostgreSQL-Flex-4.8-Single-EU01").
│ 
│ This is a bug in the provider, which should be reported in the provider's
│ own issue tracker.
╵
Error: Process completed with exit code 1.

And on another instance:

  ╷
  │ Error: Provider produced inconsistent result after apply
  │ 
  │ When applying changes to
  │ module.postgres_database.stackit_postgresflex_instance.postgres_database_instance,
  │ provider "provider[\"registry.opentofu.org/stackitcloud/stackit\"]"
  │ produced an unexpected new value: .flavor.id: was cty.StringVal("4.8"), but
  │ now cty.StringVal("4.8-replica").
  │ 
  │ This is a bug in the provider, which should be reported in the provider's
  │ own issue tracker.

[SerseusWasTaken]

I'm still not able to reproduce the issue using OpenTofu v1.12.5 and the following config:

resource "stackit_postgresflex_instance" "example" {
  project_id = local.project_id
  name       = "example-instance"
  acl        = ["0.0.0.0/24"]
  flavor = {
    cpu = 2
    ram = 4
  }
  storage = {
    class = "premium-perf2-stackit"
    size  = 5
  }
  backup_schedule = "0 2 * * *"
  replicas        = 3
  version         = "17"
  lifecycle {
    prevent_destroy = true
  }
}

I'm using v0.104.0 to create, and v0.105.0 afterwards to run the second apply and it runs without an error on my machine.

[SerseusWasTaken]

I was able to reproduce the second issue you mentioned now. We will have a look at this.