Bug 1653788

Failed_QA:

Environemnt:
(undercloud) [stack@undercloud-0 ~]$ yum info openstack-ironic-conductor
Loaded plugins: search-disabled-repos
Available Packages
Name        : openstack-ironic-conductor
Arch        : noarch
Epoch       : 1
Version     : 11.1.2
Release     : 0.20190112034459.4279258.el7ost
Size        : 5.4 k
Repo        : rhelosp-14.0-puddle/x86_64
Summary     : The Ironic Conductor
URL         : http://www.openstack.org
License     : ASL 2.0
Description : Ironic Conductor for management and provisioning of physical machines

In virtual deployment removed the cpu_arch for both controller/compute

Example:

| properties                 | capabilities:boot_option='local', capabilities:profile='controller', resources:CUSTOM_CONTROLLER='1', resources:DISK_GB='0', resources:MEMORY_MB='0', resources:VCPU='0' |


openstack baremetal node show controller-1 -f value -c properties
{u'memory_mb': u'32768', u'local_gb': u'39', u'cpus': u'8', u'capabilities': u'profile:controller,boot_option:local'}


See the following error in /var/log/containters/nova/nova-compute.log

2019-03-05 14:39:21.899 1 WARNING nova.virt.ironic.driver [req-dca133d4-5358-4801-9018-885613f3f6ad - - - - -] cpu_arch not defined for node '8bf495e2-d763-4ef2-9b96-8e186e33f0│·············

this looks like it may require a patch on nova

(In reply to Riccardo Pittau from comment #5)
> this looks like it may require a patch on nova

What needs to happen in nova?

Hey Tony, apologies, I should have elaborated more.
I'm not an expert on how nova works under the hood, but having a look at the code in the nova ironic driver, my understanding is that this is the expected behavior when the cpu_arch is not set, so it looks like this is working as intended.
If the cpu_arch is not set, nova will return an empty list of supported instances for a node, and the deploy will not happen.

Mike, did the deployment actually fail? Also this issue is about ironic validation, not about nova behavior.

Dmitry,
Yes, the deployment failed when we removed the cpu_arch.  Let me know if we need to re-create an environment?

Then could you paste the actual failure please? It will be interesting to see the cause.

openstack stack failures list overcloud
overcloud.Controller.1.Controller:
  resource_type: OS::TripleO::ControllerServer
  physical_resource_id: 65456b08-d180-4cdd-8aa8-e7bac455a44a
  status: CREATE_FAILED
  status_reason: |
    ResourceInError: resources.Controller: Went to status ERROR due to "Message: No valid host was found. There are not enough hosts available., Code: 500"
overcloud.Controller.0.Controller:
  resource_type: OS::TripleO::ControllerServer
  physical_resource_id: f7a0d60c-cb48-4783-87c8-b53222eeb66d
  status: CREATE_FAILED
  status_reason: |
    ResourceInError: resources.Controller: Went to status ERROR due to "Message: No valid host was found. There are not enough hosts available., Code: 500"
overcloud.Controller.2.Controller:
  resource_type: OS::TripleO::ControllerServer
  physical_resource_id: db25708e-d848-444d-87d1-ec0fd9c2e328
  status: CREATE_FAILED
  status_reason: |
    ResourceInError: resources.Controller: Went to status ERROR due to "Message: No valid host was found. There are not enough hosts available., Code: 500"
overcloud.Compute.0.NovaCompute:
  resource_type: OS::TripleO::ComputeServer
  physical_resource_id: 85607a65-e299-42f1-afb6-7a2cea8fa4ce
  status: CREATE_FAILED
  status_reason: |
    ResourceInError: resources.NovaCompute: Went to status ERROR due to "Message: No valid host was found. There are not enough hosts available., Code: 500"

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2019:0587