rancher2_app_v2 import doesn't import enough data and needs to recreate everything

[MrLuje]

Hello,

• Creates istio app from Apps > Charts
• Once installed, import the resource in terraform

resource "rancher2_app_v2" "monitoring" {
  cluster_id    = var.cluster_id
  name          = "rancher-monitoring-crd"
  namespace     = "cattle-monitoring-system"
  repo_name     = "rancher-charts"
  chart_name    = "rancher-monitoring"
  chart_version = "100.1.3+up19.0.3"
  # values = file("values.yaml")
}

$ terraform import rancher2_app_v2.istio xxxxx.rancher-istio

rancher2_app_v2.istio: Importing from ID "xxx.rancher-istio"...
rancher2_app_v2.istio: Import prepared!
  Prepared rancher2_app_v2 for import
rancher2_app_v2.istio: Refreshing state... [id=xxx.rancher-istio]

Import successful!

The resources that were imported are shown above. These resources are now in
your Terraform state and will henceforth be managed by Terraform.

Current result

• Make a terraform plan (there should be no change)

Terraform used the selected providers to generate the following execution plan. Resource actions are indicated with the following symbols:
-/+ destroy and then create replacement

Terraform will perform the following actions:

  # rancher2_app_v2.istio must be replaced
-/+ resource "rancher2_app_v2" "istio" {
      ~ annotations                 = {} -> (known after apply)
      + chart_name                  = "rancher-istio"
      + chart_version               = "100.4.0+up1.14.1"
      + cleanup_on_fail             = false
      ~ cluster_name                = "flex-dev-rci" -> (known after apply)
      + disable_hooks               = false
      + disable_open_api_validation = false
      + force_upgrade               = false
      ~ id                          = "xxx.rancher-istio" -> (known after apply)
      ~ labels                      = {} -> (known after apply)
      + name                        = "rancher-istio" # forces replacement
      + namespace                   = "istio-system" # forces replacement
      + repo_name                   = "rancher-charts"
      + system_default_registry     = (known after apply)
      + wait                        = true
        # (1 unchanged attribute hidden)

      - timeouts {}
    }

Expected result

There should be almost no change, but it shouldn't try to destroy the resource

[mikekuzak]

I worked around this with ignore changes

resource "rancher2_app_v2" "rancher-istio" {
  count = var.k8s_cluster_istio ? 1 : 0

  # Rancher-istio requires rancher-monitoring
  depends_on = [rancher2_app_v2.applications]

  provider      = rancher2.admin
  cluster_id    = rancher2_cluster_sync.cluster-sync.id
  project_id    = data.rancher2_project.system.id
  name          = "rancher-istio"
  namespace     = "istio-system"
  repo_name     = "rancher-charts"
  chart_name    = "rancher-istio"
  chart_version = "1.7.301"
  values        = file("${path.module}/applications/istio.yaml")

  cleanup_on_fail = true

  lifecycle {
    ignore_changes = [
      cluster_id, project_id
    ]
  }

  timeouts {
    create = "15m"
    update = "15m"
  }
}

[richard-mck]

I've got the same issue. Monitoring was installed via the GUI, I'm now working to set it up on other clusters via Terraform and have imported the existing install.

#terraform/rancher_cluster$> terraform import rancher2_app_v2.rancher_monitoring $CLUSTER_ID.rancher-monitoring
rancher2_app_v2.rancher_monitoring: Importing from ID "$CLUSTER_ID.rancher-monitoring"...
rancher2_app_v2.rancher_monitoring: Import prepared!
  Prepared rancher2_app_v2 for import
rancher2_app_v2.rancher_monitoring: Refreshing state... [id=$CLUSTER_ID.rancher-monitoring]

Import successful!

The resources that were imported are shown above. These resources are now in
your Terraform state and will henceforth be managed by Terraform.

#terraform/rancher_cluster$> terraform plan
... Refresh output removed ...

Terraform used the selected providers to generate the following execution plan. Resource actions are indicated with the following symbols:
-/+ destroy and then create replacement

Terraform will perform the following actions:

  # rancher2_app_v2.rancher_monitoring must be replaced
-/+ resource "rancher2_app_v2" "rancher_monitoring" {
      ~ annotations                 = {} -> (known after apply)
      + chart_name                  = "rancher-monitoring"
      + chart_version               = (known after apply)
      + cleanup_on_fail             = false
      ~ cluster_name                = "$CLUSTER_NAME" -> (known after apply)
      + disable_hooks               = false
      + disable_open_api_validation = false
      + force_upgrade               = false
      ~ id                          = "$CLUSTER_ID.rancher-monitoring" -> (known after apply)
      ~ labels                      = {} -> (known after apply)
      + name                        = "rancher-monitoring" # forces replacement
      + namespace                   = "cattle-monitoring-system" # forces replacement
      + repo_name                   = "rancher-charts"
      + system_default_registry     = (known after apply)
      + wait                        = true
        # (1 unchanged attribute hidden)

      - timeouts {}
    }

Plan: 1 to add, 0 to change, 1 to destroy.

In my case, I haven't done any significant config as yet so replacing the install isn't a huge issue but I can see situations where this might be a problem.

It also appears that attempts replace the resource may fail:

#terraform/rancher_cluster$> terraform apply
 ... Refresh output removed ...

Terraform used the selected providers to generate the following execution plan. Resource actions are indicated with the following symbols:
-/+ destroy and then create replacement

Terraform will perform the following actions:

  # rancher2_app_v2.rancher_monitoring must be replaced
-/+ resource "rancher2_app_v2" "rancher_monitoring" {
      ~ annotations                 = {} -> (known after apply)
      + chart_name                  = "rancher-monitoring"
      + chart_version               = (known after apply)
      + cleanup_on_fail             = false
      ~ cluster_name                = "$CLUSTER_NAME" -> (known after apply)
      + disable_hooks               = false
      + disable_open_api_validation = false
      + force_upgrade               = false
      ~ id                          = "$CLUSTER_ID.rancher-monitoring" -> (known after apply)
      ~ labels                      = {} -> (known after apply)
      + name                        = "rancher-monitoring" # forces replacement
      + namespace                   = "cattle-monitoring-system" # forces replacement
      + repo_name                   = "rancher-charts"
      + system_default_registry     = (known after apply)
      + wait                        = true
        # (1 unchanged attribute hidden)

      - timeouts {}
    }

Plan: 1 to add, 0 to change, 1 to destroy.

Do you want to perform these actions?
  Terraform will perform the actions described above.
  Only 'yes' will be accepted to approve.

  Enter a value: yes

rancher2_app_v2.rancher_monitoring: Destroying... [id=$CLUSTER_ID.rancher-monitoring]
╷
│ Error: Error removing App V2 : action [uninstall] not available on [&{ collection map[self:https://rancher.$DOMAIN_NAME/k8s/clusters/$CLUSTER_ID/v1/catalog.cattle.io.apps/rancher-monitoring] map[]}]

My actual app config is minimal:

resource "rancher2_app_v2" "rancher_monitoring" {
  chart_name = "rancher-monitoring"
  cluster_id = rancher2_cluster.main.id
  name       = "rancher-monitoring"
  namespace  = "cattle-monitoring-system"
  repo_name  = "rancher-charts"
}

[abinet]

We tried the app2 import as well and had same issue. Apparently the format of the resource id is wrong. It should be namespaced terraform import rancher2_app_v2.istio <cluster-id>.<namespace>/<app-name>

[snasovich]

This issue is for Apps & Marketplace related functionality so changed team Mapps. FYI @gunamata - wondering if you can fit it into your team's backlog and how it ranks among other things on your team's plate.

[gunamata]

@snasovich , Sure. We will review this during our triage meeting, however, this will be considered mostly for a release after the Q4 release.