Error: string is required when updating from 1.10.0 to 1.10.1

[avthart]

Error: string is required with terraform plan when we update from 1.10.0 to 1.10.1

We are still using Terraform v0.12.x

Currently, trying to find which string is required.

[abhi1693]

I have the same issue with terraform v0.13

[rawmind0]

This seems to be a terraform issue, hashicorp/terraform#25752. It should be addressed on tf v0.13.1. Could you please try to upgrade to this tf version?? If not working with tf v0.13.1, could you please add more detailed data to be able to reproduce the issue??

[abhi1693]

I upgraded to v0.13.1 and didn't work. I just deleted the state files and re-did everything after which it started working.

[nickvth]

Tomorrow I will add some tf debug log, deleting state file is not an option.

[nickvth]

@rawmind0 Debug log

2020/09/01 14:11:50 [DEBUG] ReferenceTransformer: "rancher2_cluster.cluster" references: []
2020-09-01T14:11:50.252Z [DEBUG] plugin.terraform-provider-rancher2_v1.10.1: 2020/09/01 14:11:50 [WARN] unexpected type cty.String for map in json state
2020-09-01T14:11:50.252Z [DEBUG] plugin.terraform-provider-rancher2_v1.10.1: 2020/09/01 14:11:50 [WARN] unexpected type cty.String for map in json state
2020/09/01 14:11:50 [ERROR] <root>: eval: *terraform.EvalReadState, err: string is required
2020/09/01 14:11:50 [ERROR] <root>: eval: *terraform.EvalSequence, err: string is required

[rawmind0]

@nickvth , thanks to provide debug but i need to reproduce the issue. I've made tf upgrades with cluster data and resource defined but working fine to me. Could you please provide rancher2_cluster.cluster tf definition to test with it??

[nickvth]

@nickvth , thanks to provide debug but i need to reproduce the issue. I've made tf upgrades with cluster data and resource defined but working fine to me. Could you please provide rancher2_cluster.cluster tf definition to test with it??

Did you first create a cluster with < 1.10.1 and then upgrade plugin and then rerun terraform plan?

[rawmind0]

Did you first create a cluster with < 1.10.1 and then upgrade plugin and then rerun terraform plan?

Obviously, on tf 0.12 and 0.13 but not getting any issue.

[fad3t]

Hi, I'm having the same issue when upgrading the provider from 1.10.0 to 1.10.1 or 1.10.2. I tried multiple Terraform 0.13.x releases, same problem. Here's the corresponding cluster definition:

resource "rancher2_cluster" "rke-backend-01" {
  name = "rke-backend-01"
  cluster_auth_endpoint {
    fqdn = "api.rke-backend-01.domain.com:6443"
  }
  rke_config {
    authentication {
      strategy = "x509|webhook"
      sans = [
        "api.rke-backend-01.domain.com",
        "10.10.10.10"
      ]
    }
    private_registries {
      url        = "registry.domain.com"
      user       = "docker"
      password   = var.registry_password
      is_default = true
    }
    network {
      plugin = "calico"
    }
    services {
      kube_api {
        service_cluster_ip_range = "10.43.0.0/16"
        pod_security_policy      = true
        secrets_encryption_config {
          enabled = true
        }
      }
      kube_controller {
        cluster_cidr             = "10.42.0.0/16"
        service_cluster_ip_range = "10.43.0.0/16"
      }
      kubelet {
        cluster_domain     = "cluster.local"
        cluster_dns_server = "10.43.0.10"
      }
      etcd {
        backup_config {
          enabled = true
          s3_backup_config {
            endpoint    = "s3.domain.com"
            bucket_name = "bucket_rancher"
            access_key  = var.s3_access_key
            secret_key  = var.s3_secret_key
            custom_ca   = filebase64("ca.pem")
          }
        }
      }
    }
    ingress {
      extra_args = {
        "enable-ssl-passthrough" = ""
      }
    }
  }
  default_pod_security_policy_template_id = data.rancher2_pod_security_policy_template.unrestricted.id
  enable_cluster_monitoring               = true
  cluster_monitoring_input {
    answers = {
      "grafana.persistence.enabled"             = true
      "grafana.persistence.size"                = "10Gi"
      "grafana.persistence.storageClass"        = "csi-rbd-sc"
      "prometheus.persistence.enabled"          = "true"
      "prometheus.persistence.size"             = "30Gi"
      "prometheus.persistence.storageClass"     = "csi-rbd-sc"
      "prometheus.resources.core.limits.cpu"    = "1000m"
      "prometheus.resources.core.limits.memory" = "1500Mi"
    }
  }
}

[rawmind0]

@fad3t , thanks for your example. Finally i could reproduce the issue.

Submitted PR #456 to fix it. Once merged, cutting new 1.10.3 release.

[fad3t]

Great, thanks a lot @rawmind0 !