[Openstack Provider] Instance with volumes are not idempotent

[Rob Coward]

Hi,

I'm having a problem using the openstack provider to create an instance using volumes. Our openstack deploy defaults to using ephemeral disk, so to setup say a database server on persistent ceph-backed volume storage I'm trying to use the following:

resource "openstack_blockstorage_volume_v1" "db_system" {
  name = "db_system"
  size = 30
  image_id = "${var.IMAGE_ID}"
  description = "System volume for MySQL"
}


resource "openstack_blockstorage_volume_v1" "db_data" {
  region = ""
  name = "db_data"
  description = "data volume for MySQL"
  size = 50
}


resource "openstack_compute_instance_v2" "database" {
  name            = "db01"
  flavor_name     = "${var.INSTANCE_TYPE}"
  key_pair        = "${openstack_compute_keypair_v2.ssh-keypair.name}"
  security_groups = ["${openstack_networking_secgroup_v2.local_ssh.name}",
                     "${openstack_networking_secgroup_v2.local_mysql.name}"]


  block_device {
    uuid = "${openstack_blockstorage_volume_v1.db_system.id}"
    source_type = "volume"
    boot_index = 0
    volume_size = "${openstack_blockstorage_volume_v1.db_system.size}"
    destination_type = "volume"
    delete_on_termination = true
  }


  volume {
    volume_id = "${openstack_blockstorage_volume_v1.db_data.id}"
  }


  network {
    name = "${openstack_networking_network_v2.dmz.name}"
  }
}

This works when the terraforrm apply is first run, however subsequent plans show:

~ openstack_compute_instance_v2.database
    volume.#:                   "2" => "1"
    volume.481488857.volume_id: "1e242a36-be4e-4afa-b9ad-26066f12fa31" => "1e242a36-be4e-4afa-b9ad-26066f12fa31"




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

and if I try another apply, i get the following error:

Error applying plan:


1 error(s) occurred:


* openstack_compute_instance_v2.database: Expected HTTP response code [202 204] when accessing [DELETE https://beta.openstack.skyscapecloud.com:13774/v2.1/51c7ec94d0a1420f9f55113dadbf2ec4/servers/02cee6da-298a-4f7e-968a-3ce81a0ebc4e/os-volume_attachments/1292ef21-3fa7-491c-ae93-c56131950cf8], but got 403 instead
{"forbidden": {"message": "Can't detach root device volume", "code": 403}}

Has anyone else come across this ? Is there a different approach I can use ?

Thanks in advance,

Rob Coward

[Rob Coward]

After some more digging, it seems that if I remove the 2nd volume and just boot from the db_system volume, everything works as expected - it's something to do with having the additional volume attached to the VM that triggers the issue.

[Joe Topjian]

Hi Rob,

This definitely looks like a bug -- something about booting from a volume + attaching additional volumes is not being reported correctly.

Can you open a bug report / issue on Github? Please tag me (@jtopjian) and I'll take a look.

Thanks,

Joe

On Wed, Jun 22, 2016 at 7:09 AM, Rob Coward <r...@jive-videos.net> wrote:

After some more digging, it seems that if I remove the 2nd volume and just boot from the db_system volume, everything works as expected - it's something to do with having the additional volume attached to the VM that triggers the issue.

--
This mailing list is governed under the HashiCorp Community Guidelines - https://www.hashicorp.com/community-guidelines.html. Behavior in violation of those guidelines may result in your removal from this mailing list.
 
GitHub Issues: https://github.com/hashicorp/terraform/issues
IRC: #terraform-tool on Freenode
---
You received this message because you are subscribed to the Google Groups "Terraform" group.
To unsubscribe from this group and stop receiving emails from it, send an email to terraform-too...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/terraform-tool/66c8ebdf-5360-4762-9252-5eff47bbfd9c%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Rob Coward]

Hi Joe,

I've raised https://github.com/hashicorp/terraform/issues/7297 for the bug.

Thanks

Rob