node deploy fails with noop network_interface reporting no VIFs found

[Julia Kreger]

Attempting to spin up a configuration consisting of noop + neutron network interfaces, where the dhcp_provider (which will begin to see deprecation removals in Pike), a node configured with the "noop" network_interface, returns an error indicating no VIFs found:

Installed via bifrost, utilizing some in-development patches for neutron deployment support, with ironic git from master as of April 1st, 2017. Port details are naturally minimal for a noop network_interface based deploy.

+------------------------+--------------------------------------------------------------------------+
| Property | Value |
+------------------------+--------------------------------------------------------------------------+
| boot_interface | |
| chassis_uuid | None |
| clean_step | {} |
| console_enabled | False |
| console_interface | |
| created_at | 2017-04-01T20:04:44+00:00 |
| deploy_interface | |
| driver | agent_ipmitool |
| driver_info | {u'ipmi_port': u'623', u'ipmi_username': u'admin', u'deploy_kernel': |
| | u'http://192.168.122.1:8080/ipa.vmlinuz', u'ipmi_address': |
| | u'192.168.122.1', u'deploy_ramdisk': |
| | u'http://192.168.122.1:8080/ipa.initramfs', u'ipmi_password': u'******'} |
| driver_internal_info | {u'is_whole_disk_image': True} |
| extra | {} |
| inspect_interface | |
| inspection_finished_at | 2017-04-02T04:11:33+00:00 |
| inspection_started_at | None |
| instance_info | {u'root_gb': 10, u'image_source': |
| | u'http://192.168.122.1:8080/deployment_image.qcow2', u'image_type': u |
| | 'whole-disk-image', u'configdrive': u'******', u'image_url': u'******', |
| | u'image_disk_format': u'qcow2', u'image_checksum': |
| | u'ee1eca47dc88f4879d8a229cc70a07c6'} |
| instance_uuid | None |
| last_error | Failed to prepare to deploy. Error: No VIFs found for node |
| | 4e41df61-84b1-5856-bfb6-6b5f2cd3dd11 when attempting to update DHCP BOOT |
| | options. |
| maintenance | False |
| maintenance_reason | None |
| management_interface | |
| name | testvm1 |
| network_interface | |
| power_interface | |
| power_state | power off |
| properties | {u'memory_mb': u'3072', u'cpu_arch': u'x86_64', u'local_gb': u'9', |
| | u'cpus': u'1', u'capabilities': u'cpu_vt:true,cpu_hugepages:true'} |
| provision_state | deploy failed |
| provision_updated_at | 2017-04-02T04:16:31+00:00 |
| raid_config | |
| raid_interface | |
| reservation | None |
| resource_class | |
| target_power_state | None |
| target_provision_state | active |
| target_raid_config | |
| updated_at | 2017-04-02T04:16:31+00:00 |
| uuid | 4e41df61-84b1-5856-bfb6-6b5f2cd3dd11 |
| vendor_interface | |
+------------------------+--------------------------------------------------------------------------+

[Vladyslav Drok (vdrok)]

So you mean the case where you have dhcp provider = neutron and some nodes having noop net interface, and some having neutron? Yeah, neutron net interface requires dhcp provider to be neutron too, and I'm not sure how to deal with this, apart from deprecation of dhcp providers. So let's do this as a fix :)