Comment 1 for bug 1876769

Hi,

I believe in this case you need to go to the placement database and ensure that a resource provider with same name but different UUID is not present there.

And you have 2 options more or less:

1. either delete record from placement. But this might leave quite some resources without assignment and broken allocations:

# openstack resource provider list
+--------------------------------------+----------------------------------+------------+--------------------------------------+----------------------+
| uuid | name | generation | root_provider_uuid | parent_provider_uuid |
+--------------------------------------+----------------------------------+------------+--------------------------------------+----------------------+
| d5623168-23aa-427c-b1c8-4af1b79cf33b | b2b.blr.example.cloud | 2685 | d5623168-23aa-427c-b1c8-4af1b79cf33b | None |
+--------------------------------------+----------------------------------+------------+--------------------------------------+----------------------+
# openstack resource provider delete d5623168-23aa-427c-b1c8-4af1b79cf33b

After that nova-compute should be able to re-create resource provider and spawn normally.

2. Update MySQL database for placement, by updating UUID with the one you found in the log. for example:

mysql> update placement.resource_providers set uuid = '7e6fb27c-ed5b-4c2c-8373-c99e98da7bcc' where name = 'b2b.blr.example.cloud';

Though I will mark this report as invalid, since it is a generic OpenStack support request, rather then a bug in OpenStack-Ansible.