problem with creating vapp_vm with official plugin

[ghost]

Sorry for getting back to you that late. But I could not test earlier. Currently I am having the problem that my VM cannot be created anymore.

Terraform v0.11.10

• provider.vcd v2.0.0

Terraform Version

Terraform v0.11.10

• provider.vcd v2.0.0

Affected Resource(s)

Please list the resources as a list, for example: vcd_vapp_vm.my_vm00

Terraform Configuration Files

resource "vcd_network_routed" "net00" {
        name = "${var.vcdnet_name}"
        edge_gateway = "${var.vcdnet_edge_gateway}"
        gateway = "${var.vcdnet_gateway}"
                static_ip_pool {
                start_address = "${var.vcdnet_static_ip_pool_start}"
                end_address = "${var.vcdnet_static_ip_pool_end}"
        }
}

resource "vcd_catalog" "MMIG-Test-Catalog" {
name = "${var.mmigcatalogname}"
description = "OS Templates"
delete_force = true
delete_recursive = true
}

resource "vcd_vapp" "mmigvapp00" {
        depends_on = ["vcd_network_routed.net00"]
        name = "${var.vcd_vapp_name}"
        power_on = false
        network_name = "${var.vcdnet_name}"
}

resource "vcd_vapp_vm" "my_vm00" {
        depends_on = ["vcd_vapp.mmigvapp00"]
        vapp_name = "${var.vcd_vapp_name}"
        network_name = "${var.vcdnet_name}"
        power_on = false
        name = "${var.vmnamelinux}"
        catalog_name = "${var.catalognamelinux}"
        template_name = "${var.templatenamelinux}"
        memory = 4096
        cpus = 2
        initscript = <<EOF
#!/bin/bash
touch /root/terraformtest.txt
EOF
}

resource "vcd_catalog_media" "myNewMedia" {
  depends_on = ["vcd_catalog.MMIG-Test-Catalog"]
  catalog = "${var.mmigcatalogname}"

  name                 = "testvm iso"
  description          = "cloud init test"
  media_path           = "/root/iso/seed.iso"
  upload_piece_size    = 10
  show_upload_progress = true
}

resource "vcd_inserted_media" "myInsertedMedia" {
  depends_on  = ["vcd_vapp_vm.my_vm00"]
  catalog = "${var.mmigcatalogname}"
  name = "testvm iso"

  vapp_name = "mmig-vcloud-providers-testspace"
  #vm_name = "cust1_mgmt_vm_lx_1"
  vm_name = "${var.vmnamelinux}"
}

Debug Output

https://gist.github.com/ybaumy/a1fcab843ab3901b2fde871bc9a1a1de

Panic Output

Expected Behavior

VM should have been created but it isn't completely. Then the uploaded ISO file should be attached to the VM

Actual Behavior

That is what i want to know. Am I missing something.

Steps to Reproduce

terraform plan + apply

Important Factoids

I dont know the vCloud Version. Would have to ask.

References

[lvirbalas]

@ybaumy the error says that "error instantiating a new VM: API Error: 400: [ ... ] There is already a VM named \"cust1_mgmt_vm_lx_1\"."}

Are you sure your environment doesn't have any remnants from previous setup (of which the state file has been lost)? Could you try removing everything and starting from empty?

[ghost]

actually that is what i did. rm -f statefile + backup terraform plan terraform init

and I made sure none of the resources were still in the vOrg created.

So it should have been a clean state. That is why I am wondering why this happens.

[ghost]

@lvirbalas i found on the vcd gui an error that is strange.. at least for me.

Recomposed Virtual Application 'mmig-vcloud-providers-testspace' (9ee97503-6089-43a6-930c-72ba801d0ade)

[ b78c21aa-0414-473b-8206-af2e4cb724fd ] Unable to perform this action. Contact your cloud administrator.

This is the error probably why it fails. It just does not say so in the terraform trace output.

[vbauzys]

I think I found the issue, fix on the way.

[vbauzys]

Fix in https://github.com/vmware/go-vcloud-director/pull/150 . When this one will be merged, terraform can be updated too.

[ghost]

@vbauzysvmware Just talked to the vCloud Engineer. He provided me another error message.

`[ b78c21aa-0414-473b-8206-af2e4cb724fd ] The operation is not supported on the object.

• The operation is not supported on the object. `

So it seems that some API call to the VM is wrong?

[vbauzys]

Yes it is,constructing API call happens one issue, which is solved in PR.

[vbauzys]

Additional note from my investigation: with current configuration vapp creation won't use network name - it will create raw vApp with only name, all other configuration details ignored. Current implementation expects template_name and catalog_name to be provided, and then all other vApp properties will be used.

[ghost]

@vbauzysvmware Ok. But I am using template_name + catalog_name in the resource defintion. So sorry but I do not understand what should be wrong here.

resource "vcd_vapp_vm" "my_vm00" { depends_on = ["vcd_vapp.mmigvapp00"] vapp_name = "${var.vcd_vapp_name}" network_name = "${var.vcdnet_name}" power_on = false name = "${var.vmnamelinux}" catalog_name = "${var.catalognamelinux}" template_name = "${var.templatenamelinux}"

But I can tell you what was wrong in the code. The "ip = x.x.x.x" attribute was missing, since we have to use static_ip_pool and not dhcp.

I rerun the terraform apply and provided an IP address in the vcd_vapp_vm statement block but I am getting the same error.

[vbauzys]

I was talking about vcd_vapp resource. Did you use code from PR I mentioned? Binary update will happen with new release only. In reality it's just one line change in vapp.go

[vbauzys]

Fix come around and in terraform-providers:vmware-stash-while-master-pr-open until merged to master

[vbauzys]

2.1.0 version released with the fix. Let me us know it it's good now :)