yaml change causes destroy -> create instead of in place.

[yctn]

every time i change something in the yaml it tries to destroy the yaml and recreate it. it never actualy in-place replace it. this causes operators to do a full restart of clusters sometimes.

i suspect that is because the full yaml is in the terraform state.
is this expected?
or am i simply doing something wrong here?
my tf code is as following.

`data "kubectl_path_documents" "kafka-yaml" {
pattern = "./manifest/kafka/*.yaml.tpl"
}

resource "kubectl_manifest" "kafka-GRA7" {
provider = kubectl.GRA7
for_each = toset(data.kubectl_path_documents.kafka-yaml.documents)
yaml_body = each.value
wait = false
wait_for_rollout = false
}`

[alekc]

As discussed in alekc/terraform-provider-kubectl#50, a proper way to do it is to use manifests and not documents.
I.e.

data "kubectl_path_documents" "manifests-directory-yaml" {
  pattern = "./manifests/*.yaml"
}
resource "kubectl_manifest" "directory-yaml" {
  for_each  = data.kubectl_path_documents.manifests-directory-yaml.manifests
  yaml_body = each.value
}

Documentation has been updated on the fork