launch a instance which name include spot failed

[yan tan]

can not launch a instance which name include spot and number like 14.4.

nova-conductor.log find this error.

2017-01-11 16:27:46.928 23 ERROR nova.scheduler.utils [req-ed6f17d5-ef49-412e-8e7e-703b7ed5738b eef7aab2a2bf4067a4baf27cdbe62cc6 90e82be1b4a34c62a27d4fde986e00c2 - - -] [instance: 46fe3b71-0c90-458c-8be6-cd9dcd8733a3] Error from last host: compute1 (node compute1): [u'Traceback (most recent call last):\n', u' File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 1926, in _do_build_and_run_instance\n filter_properties)\n', u' File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 2116, in _build_and_run_instance\n instance_uuid=instance.uuid, reason=six.text_type(e))\n', u"RescheduledException: Build of instance 46fe3b71-0c90-458c-8be6-cd9dcd8733a3 was re-scheduled: Invalid input for dns_name. Reason: '14.4' not a valid PQDN or FQDN. Reason: TLD '4' must not be all numeric.\nNeutron server returns request_ids: ['req-9a2fa7cb-35cd-462f-973d-5c618b72b18d']\n"]
2017-01-11 16:27:47.113 23 WARNING nova.scheduler.utils [req-ed6f17d5-ef49-412e-8e7e-703b7ed5738b eef7aab2a2bf4067a4baf27cdbe62cc6 90e82be1b4a34c62a27d4fde986e00c2 - - -] Failed to compute_task_build_instances: No valid host was found. There are not enough hosts available.

[Akihiro Motoki (amotoki)]

It happens even with CLI. http://paste.openstack.org/show/628012/
Looks line a nova bug.

[melanie witt (melwitt)]

It looks like this is because the instance 'name' is used as the instance hostname, which is then passed to update the neutron port's dns_name and then that call fails because it is not a valid FQDN. The top-level domain name cannot be all numeric.

When we set instance.hostname, we run it through utils.sanitize_hostname() first to ensure the resulting hostname conforms to the RFC-952 and RFC-1123 specs. So it looks like utils.sanitize_hostname() needs to be enhanced to handle this *or* we need to check the resulting hostname for RFC conformance and fail earlier, before the RPC cast to conductor. That way the user will receive a 400 from the API instead of getting an instance in ERROR state.

[Choi-Sung-Hoon (knu-cse)]

Hi, i assigned this bug to me to contribute for the first time.

I just installed devstack and tried the CLI command in the http://paste.openstack.org/show/628012/ and it works fine.

Am i doing something wrong?

here's the output.