so it fails reaching out the GKE cluster, but it finished creating the cluster?

For the dns-domain, I use the comanyname.com . Not sure if it needed,

If you don't have a domain around, you could just use local DNS resolution or nip.io

I ran the flyte deploy on a newly created Google Cloud project and a new bucket

Could my problem be solved?

Here is the output - it is very long: # kubectl_manifest.cert-manager-issuer will be created + resource "kubectl_manifest" "cert-manager-issuer" { + api_version = "cert-manager.io/v1" + apply_only = false + field_manager = "kubectl" + force_conflicts = false + force_new = false + id = (known after apply) + kind = "Issuer" + live_manifest_incluster = (sensitive value) + live_uid = (known after apply) + name = "letsencrypt-production" + namespace = "flyte" + server_side_apply = false + uid = (known after apply) + validate_schema = true + wait_for_rollout = true + yaml_body = (sensitive value) + yaml_body_parsed = <<-EOT apiVersion: cert-manager.io/v1 kind: Issuer metadata: name: letsencrypt-production namespace: flyte spec: acme: email: roman@vinci4d.ai privateKeySecretRef: name: letsencrypt-production server: https://acme-v02.api.letsencrypt.org/directory solvers: - http01: ingress: ingressClassName: nginx EOT + yaml_incluster = (sensitive value) } Plan: 9 to add, 0 to change, 0 to destroy. ────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── Note: You didn't use the -out option to save this plan, so Terraform can't guarantee to take exactly these actions if you run "terraform apply" now.

hey Roman So, I reproduced this behaviour and while there's a workaround, this is something the module should handle better. Specifically, with most of the resources in the

ingress

module (like cert-manager) we're basically trying to create resources or detect resource state on a GKE cluster that doesn't exist yet. So if you could create an Issue to capture this problem it'd be great, I'll take point on separating GKE to a different module and state, so the resources go after the cluster (module dependencies are not enough for this). While it works, it makes apply/destroy operations unstable. I saw this issue when trying to reuse an existing project and resources so I removed state for the resources that were complaining, like

terraform state rm helm_release.flyte-core

for example, and after that plan and apply work

so I removed state for the resources that were complaining, like

terraform state rm helm_release.flyte-core

so, would you recommend me to remove the resources and re-run 'terraform apply'?