Read-Only Attributes of SKE lead to re-apply for resource compliance #1078
Description
Description
When I create a SKE cluster using opentofu during a tofu apply and later will run another tofu apply, I will be informed that there are attributes (in this case the read-only attributes of the Master and Worker Node about the version) will be known after apply, which leads to unnecessary opentofu noise and unnecessary drifts + re-applys.
Steps to reproduce
- Create a SKE Cluster in OpenTofu
- Apply
- Apply again
resource "stackit_ske_cluster" "this" {
project_id = stackit_resourcemanager_project.this.project_id
name = "skepoc"
kubernetes_version_min = "1.33.5"
node_pools = [
{
name = "worker-1"
machine_type = "g1a.2d"
minimum = 1
maximum = 1
availability_zones = ["eu01-2"]
os_name = "flatcar"
os_version_min = "4230.2.4"
volume_size = 32
volume_type = "storage_premium_perf2"
},
{
name = "worker-2"
machine_type = "g1a.2d"
minimum = 1
maximum = 1
availability_zones = ["eu01-2"]
os_name = "flatcar"
os_version_min = "4230.2.4"
volume_size = 32
volume_type = "storage_premium_perf2"
}
]
network = {
id = stackit_network.this.network_id #Required for SKE Clusters in a SNA
}
}Actual behavior
The SKE Read-Only Attributes for the version are set to known after apply for every new apply after cluster creation.
Expected behavior
There shouldnt be chaning attribute noise.
Please describe the behavior which you would expect from the STACKIT Terraform provider in that case.
Environment
OpenTofu v1.9.0
on linux_amd64
- provider registry.opentofu.org/gitlabhq/gitlab v18.3.0
- provider registry.opentofu.org/stackitcloud/stackit v0.71.0
Additional information
Workaround in my case, but the issue should be resolved anyway.
lifecycle {
ignore_changes = [
node_pools[0].os_version_used,
node_pools[1].os_version_used,
kubernetes_version_used
]
}