Posted	Nick	Remark
#openstack-nova - 2018-01-31
09:44:23	Spazmotic	Going to take the rest of the day off and avoid python.. have a good night all, will be afk :)
09:50:36	openstackgerrit	blkart proposed openstack/nova-specs master: don't show description content in code-block https://review.openstack.org/539478
09:54:11	bauzas	jianghuaw: not sure you saw https://review.openstack.org/#/c/539266/2
10:01:40	jianghuaw	bauzas, yes. I saw it. Just have not got chance to check it due to having been occupied by a kolla deployment problem.
10:01:44	jianghuaw	thanks for asking, surely will look at it before I leave office.
10:01:52	openstackgerrit	Rajesh Tailor proposed openstack/nova master: Allow force-delete even if task_state is not None https://review.openstack.org/530879
10:04:33	bauzas	jianghuaw: ideally, you could make a follow-up with the Xen notes
10:05:14	jianghuaw	Sure.
10:31:09	amotoki	hi, i have a question on 'host' parameters in os-aggregates and os-migrateLive API.
10:31:21	amotoki	I am working on https://review.openstack.org/#/c/538472/ (in horizon).
10:32:02	amotoki	in nova API there are two types of "host". the one is from os-services and the other is from os-hypervisors.
10:33:49	amotoki	I checked the API reference but I am still not clear which is expected for "hosts" in os-aggregates and "host" in migrateLive (and migrate) action.
10:34:06	amotoki	thanks in advance.
10:35:42	bauzas	amotoki: it refers to the os-services hostname
10:36:02	amotoki	bauzas: thanks. is it true for both?
10:36:26	bauzas	amotoki: well, there is a confusion
10:37:09	bauzas	amotoki: long story short, hypervisors provide their own hostnames, so we have a tuple (service hostname, hypervisor hostname) identifying a single node
10:37:28	bauzas	for most of our virt drivers, a compute service only serves a single hypervisor
10:37:44	bauzas	so there is no real problem in adressing the host itself by its service name
10:37:56	bauzas	but for ironic, that paradigm isn't true
10:38:18	bauzas	since a single compute service serves for all the ironic nodes
10:38:24	amotoki	i think the same thing happens for vmware (vcenter) driver.
10:38:28	bauzas	no
10:38:41	bauzas	because nova sees the vCloud as a single node
10:39:00	bauzas	how it's managed within that cloud is beyond its view
10:39:04	amotoki	does nova consider vCloud as a big host?
10:39:08	bauzas	exactly
10:39:19	bauzas	from a scheduling perspective
10:39:34	bauzas	which is what aggregates are for :)
10:41:34	amotoki	aggregates refer to service host name, so how can we distinguish individual hosts? is it determined based on aggregate metadata?
10:42:43	ameeda	Hello, what is the different between "self.volume_api.detach" and " self._detach_volumes" in nova manager ?
10:45:08	bauzas	amotoki: sorry, was at the coffee machine
10:45:24	amotoki	bauzas: no problem. thanks for your support
10:46:55	bauzas	amotoki: so, the answer is "that's not supported"
10:47:17	bauzas	amotoki: all the Aggregate API logic is based on services checks
10:47:53	bauzas	amotoki: that means that if you use ironic and want aggregates, you will only shard your cloud by the compute services, that's it, not by your ironic deployment
10:48:04	bauzas	call it a limitation
10:48:54	bauzas	from my perspective, it's just that ironic introduced a new paradigm (N hypervisors for a single service) without changing the Nova API, which is why it's broken
10:49:14	openstackgerrit	Yikun Jiang (Kero) proposed openstack/nova master: Encode libvirt domain XML in UTF-8 https://review.openstack.org/522161
10:50:29	amotoki	bauzas: thanks for clarification. I see the current logic and it is good to know that :)
10:51:24	openstackgerrit	Stephen Finucane proposed openstack/nova master: trivial: Fix typos in release notes https://review.openstack.org/539490
10:59:26	jianghuaw	bauzas, thanks for creating the documents for vGPU. The patch looks very good. I will add the part for XenAPI.
11:00:13	bauzas	jianghuaw: FWIW, I'm also currently working on providing a functional testcase that will use a fake libvirt driver with mdevs
11:00:56	jianghuaw	bauzas, thanks for the heads-up.
11:01:25	jianghuaw	What's I can help on that>
11:02:01	bauzas	there are a couple of libvirt functional tests already, so I'll land that funct test battery into there
11:02:14	bauzas	you could see if Xen got some
11:02:40	jianghuaw	ok. will check it.
11:02:42	bauzas	later, we could discuss with other folks if we would maybe have a single test battery that would check both drivers, I dunno yet
11:03:33	bauzas	like a base test class that would test what's supported by both drivers, and we would just inherit from it
11:03:45	bauzas	of course, the setup of that test class would be per-driver
11:03:59	jianghuaw	sounds good.
11:04:05	bauzas	jianghuaw: so, nothing from your side yet I guess
11:04:25	jianghuaw	ok.
11:04:31	bauzas	just maybe if you have time, try to see how to mock your Xen API the less invasive way
11:05:17	jianghuaw	k
11:05:22	bauzas	ideally, I wouldn't want to mock every single libvirt call like we do in the unittests, rather just pass a libvirt fixture that would setup my host like I want
11:05:34	bauzas	hence the fakelibvirt module reading
11:44:00	openstackgerrit	Merged openstack/nova master: Fix documentation nits in set_and_clear_allocations https://review.openstack.org/531001
11:44:54	openstackgerrit	Surya Seetharaman proposed openstack/nova master: Marker reset option for nova-manage map_instances https://review.openstack.org/539501
12:08:45	openstackgerrit	Mark Goddard proposed openstack/nova master: Update docstring for get_traits virt driver method https://review.openstack.org/539505
12:08:46	openstackgerrit	Mark Goddard proposed openstack/nova master: Remove unused method _parse_node_instance_info https://review.openstack.org/539507
12:08:46	openstackgerrit	Mark Goddard proposed openstack/nova master: Add resource_class to fields in ironic node cache https://review.openstack.org/539506
12:08:47	openstackgerrit	Mark Goddard proposed openstack/nova master: Request only instance_uuid in ironic node list https://review.openstack.org/539509
12:08:47	openstackgerrit	Mark Goddard proposed openstack/nova master: Include only required fields in ironic node cache https://review.openstack.org/539508
12:37:54	openstackgerrit	Merged openstack/nova master: SchedulerReportClient.set_aggregates_for_provider https://review.openstack.org/533808
12:45:49	sauloaugusto	Hi , How can I deploy on openstack-nova for a error in database ?
12:51:32	openstackgerrit	Merged openstack/nova stable/ocata: Handle exception on adding secgroup https://review.openstack.org/515421
12:51:40	openstackgerrit	Merged openstack/nova stable/pike: Don't persist could-be-stale InstanceGroup fields in RequestSpec https://review.openstack.org/528225
12:51:48	openstackgerrit	Merged openstack/nova stable/ocata: Don't persist could-be-stale InstanceGroup fields in RequestSpec https://review.openstack.org/528226
13:02:22	openstackgerrit	Lajos Katona proposed openstack/nova master: WIP: Add nested resources to server moving tests https://review.openstack.org/527728
13:03:18	openstackgerrit	Chris Dent proposed openstack/nova master: Remove microversion fallback code from report client https://review.openstack.org/528794
13:07:01	openstackgerrit	Ameed Ashour proposed openstack/nova master: detach instance volumes when VM creation fails https://review.openstack.org/528385
13:30:37	openstackgerrit	Merged openstack/nova stable/pike: Bumping functional test job timeouts https://review.openstack.org/539319
13:30:45	openstackgerrit	Merged openstack/nova master: Don't launch guestfs in a thread pool if guestfs.debug is enabled https://review.openstack.org/526748
13:30:56	openstackgerrit	Merged openstack/nova master: Don't log a warning for InstanceNotFound in detach_interface https://review.openstack.org/530969
13:31:07	openstackgerrit	Merged openstack/nova stable/ocata: Make sure that functional test triggered on sample changes https://review.openstack.org/536936
13:41:16	openstackgerrit	Merged openstack/nova stable/ocata: Bumping functional test job timeouts https://review.openstack.org/539320
13:42:36	openstackgerrit	Dan Smith proposed openstack/nova master: Pass limit to /allocation_requests https://review.openstack.org/531517
14:01:08	zigo	I got issue with Pike's console-log:
14:01:11	zigo	ERROR (UnicodeEncodeError): 'ascii' codec can't encode characters in position 22096-22098: ordinal not in range(128)
14:01:11	zigo	# nova console-log 92847293-3b4d-44cc-8cdd-6cb57e23b0cf
14:01:14	zigo	WTF ? :)
14:01:18	zigo	bauzas: ^
14:04:49	efried	zigo: Is there a stack trace?
14:04:58	zigo	edleafe: Let me check.
14:06:02	zigo	edleafe: No stack trace in the logs.
14:07:09	lyarwood	zigo: can you run the nova command again with --debug and pastebin the output?
14:07:27	openstackgerrit	Matt Riedemann proposed openstack/nova stable/pike: Don't launch guestfs in a thread pool if guestfs.debug is enabled https://review.openstack.org/539541
14:07:58	openstackgerrit	Merged openstack/nova master: Fix the order of target host checks https://review.openstack.org/526225
14:08:18	zigo	lyarwood: http://paste.openstack.org/show/658151/
14:08:19	openstackgerrit	Merged openstack/nova master: Fix missing 'if_notifications_enabled' decorator https://review.openstack.org/537745
14:09:19	jroll	zigo: definitely a client issue, py2 or py3?
14:09:35	zigo	jroll: py2
14:09:45	zigo	I got the same issue with openstackclient
14:10:01	jroll	probably just need a .decode('utf8') or something there
14:10:24	zigo	Note that this is Debian Unstable ...
14:10:31	zigo	Because I can ... (tm)
14:10:32	zigo	:)
14:10:52	zigo	So, something bleeding edge may be the cause.

Earlier   Later