vSphere provider: sparse VMDK is not completely cleaned up with vsphere_file

[hashibot]

This issue was originally opened by @sputnik13 as hashicorp/terraform#8058. It was migrated here as part of the provider split. The original body of the issue is below.

---

When a sparse VMDK is uploaded with vsphere_file and used with a vsphere_virtual_machine, additional files get generated due to the VMDK being a sparse VMDK. If I have keep_on_remove = "true" on the VM so that terraform destroy can delete the vsphere_file, the additional files generated for the sparse VMDK is not cleaned up properly. This results in subsequent runs of terrform apply failing.

Sample terraform template below...

resource "vsphere_file" "k8s-master-bootdisk" {
    datacenter = "${var.datacenter}"
    datastore = "${var.datastore}"
    source_file = "${var.linux_vmdk_source}"
    destination_file = "${var.prefix}/master-bootdisk.vmdk"
}

resource "vsphere_virtual_machine" "k8s-master" {
    name = "${var.prefix}-master"
    #folder = "${vsphere_folder.k8s_folder.path}"
    datacenter = "${var.datacenter}"
    cluster = "${var.cluster}"

    connection {
        user = "ubuntu"
        key_file = "${var.key_file}"
    }

    network_interface {
        label = "${var.network}"
    }

    cdrom {
        datastore = "${var.datastore}"
        path = "${vsphere_file.cidata.destination_file}"
    }

    disk {
        controller_type = "ide"
        datastore = "${var.datastore}"
        vmdk = "${vsphere_file.k8s-master-bootdisk.destination_file}"
        bootable = "true"
        keep_on_remove = "true"
    }
}

[vancluever]

Hey all, I'm changing this one to an enhancement - I'm not too sure if this will be possible as VMDKs are possibly the only place that this magic functionality occurs and depending on if there is a way we can handle this in the API, there may not be too much we can do here.

In the meantime I would recommend using different methods to upload VMDKs.

Thanks!

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. If you feel I made an error 🤖 🙉 , please reach out to my human friends 👉 hashibot-feedback@hashicorp.com. Thanks!