Posted	Nick	Remark
#openstack-nova - 2017-07-03
09:44:06	openstackgerrit	Alex Xu proposed openstack/nova master: placement: return traits in the ProviderSummary object https://review.openstack.org/479766
09:44:09	openstackgerrit	Béla Vancsics proposed openstack/nova master: Reduce (notification) test duplication https://review.openstack.org/391428
09:47:38	openstackgerrit	NidhiMittalHada proposed openstack/nova master: Nova treats hostname as pattern https://review.openstack.org/474949
10:27:57	openstackgerrit	Stephen Finucane proposed openstack/nova master: Update policy descriptions for base. https://review.openstack.org/472622
10:27:57	openstackgerrit	Stephen Finucane proposed openstack/nova master: Use oslo.polcy DocumentedRuleDefault https://review.openstack.org/472634
10:27:58	openstackgerrit	Stephen Finucane proposed openstack/nova master: Consistent policies https://review.openstack.org/479774
10:27:58	openstackgerrit	Stephen Finucane proposed openstack/nova master: Remove 'create_rule_default' https://review.openstack.org/479773
10:38:51	openstackgerrit	Alex Xu proposed openstack/nova master: placement: return traits in the ProviderSummary object https://review.openstack.org/479766
10:38:52	openstackgerrit	Alex Xu proposed openstack/nova master: placement: add traits to the 'GET /allocation_candidates' endpoint https://review.openstack.org/479776
10:42:41	openstackgerrit	Andrey Kurilin proposed openstack/python-novaclient master: [ci] Use pseudo-random names for new resources https://review.openstack.org/479777
11:54:50	openstackgerrit	Alex Szarka proposed openstack/nova master: Transform instance-live_migration_abort notification https://review.openstack.org/420453
11:56:50	openstackgerrit	Andrey Kurilin proposed openstack/python-novaclient master: [ci] Use pseudo-random names for new resources https://review.openstack.org/479777
12:00:46	openstackgerrit	Luis Pigueiras proposed openstack/nova master: Handle keypair not found from metadata server using cells https://review.openstack.org/476122
12:01:40	openstackgerrit	Andrey Kurilin proposed openstack/python-novaclient master: [ci] Use pseudo-random names for new resources https://review.openstack.org/479777
12:09:44	jaypipes	morning folks.
12:14:47	bauzas	jaypipes: morning
12:25:35	bhagyashri_s	johnthetubaguy: Hi Could you please review patch : https://review.openstack.org/#/c/409644/ ? Addressed all review comments.
12:35:32	openstackgerrit	Matthew Booth proposed openstack/nova master: Allow wrapping of closures https://review.openstack.org/479801
12:35:32	openstackgerrit	Matthew Booth proposed openstack/nova master: Automatically revert resize which fails on destination https://review.openstack.org/462521
12:35:33	openstackgerrit	Matthew Booth proposed openstack/nova master: Ensure errors_out_migration errors out migration https://review.openstack.org/479802
12:40:12	jaypipes	bauzas: https://review.openstack.org/#/q/topic:bug/1686116
12:40:28	jaypipes	bauzas: that's a bug we've now run into a few times in our support team
12:40:36	bauzas	jaypipes: okay, looking it
12:40:52	jaypipes	bauzas: the code in general is tough to follow, but not because of sahid's stuff. just general ickiness with the BDM layer.
12:41:06	bauzas	FWIW, just had an internal mail server migration, now has 700 emails to review :(
12:41:16	jaypipes	bauzas: my support folks have applied it and said it fixes issues they've seen with virtio-scsi drives.
12:41:24	bauzas	jaypipes: okay, I trust you
12:41:35	jaypipes	bauzas: yeah, I hear ya :) good luck with your emails :)
12:46:36	sfinucan	jaypipes: I know it's not the top priority for the cycle, but is there anything more I can do on nested-resource-providers stuff?
12:47:01	jaypipes	sfinucan: nope, not right now. reviews on https://review.openstack.org/#/q/topic:bug/1686116 would be great though
12:47:46	sfinucan	jaypipes: yup, sahid pinged me earlier. On the to-do list for today
12:52:23	jaypipes	sfinucan: cool, cheers :)
12:54:52	openstackgerrit	Stephen Finucane proposed openstack/nova master: Update policy descriptions for base. https://review.openstack.org/472622
12:54:53	openstackgerrit	Stephen Finucane proposed openstack/nova master: Remove 'create_rule_default' https://review.openstack.org/479773
12:54:53	openstackgerrit	Stephen Finucane proposed openstack/nova master: Consistent policies https://review.openstack.org/479774
13:05:35	openstackgerrit	Ildiko Vancsa proposed openstack/nova master: Add Cinder v3 detach to swap_volume https://review.openstack.org/456971
13:05:36	openstackgerrit	Ildiko Vancsa proposed openstack/nova master: Remove check_detach https://review.openstack.org/446671
13:05:37	openstackgerrit	Ildiko Vancsa proposed openstack/nova master: WIP/POC Implement new attach Cinder flow https://review.openstack.org/330285
13:14:06	openstackgerrit	Kaitlin Farr proposed openstack/nova master: Remove deprecated keymgr code https://review.openstack.org/439855
13:22:39	openstackgerrit	Kaitlin Farr proposed openstack/nova master: Remove deprecated keymgr code https://review.openstack.org/439855
13:27:34	openstackgerrit	Andrey Kurilin proposed openstack/python-novaclient master: DO NOT MERGE: debug failed test https://review.openstack.org/479818
13:31:51	sfinucan	jaypipes: Comments left on that but you might be able to answer of the top of your head: can you hotplug SCSI devices?
13:32:19	jaypipes	sfinucan: you mean virtual SCSI devices?
13:32:24	sfinucan	yup
13:32:31	sfinucan	maybe hotplug is the wrong term
13:32:45	jaypipes	sure, you can attach/detach scsi block devices.
13:33:19	sfinucan	OK, so my comment on that is whether always incrementing this unit parameter when attaching a scsi device is a good idea
13:33:46	sfinucan	I'd imagine if we attach/detach a few times on a long-running VM, we'd hit that 256 device limit?
13:34:02	sfinucan	...but maybe it's a YAGNI kind of thing
13:34:23	jaypipes	sfinucan: I think the whole max(unit) logic prevents that, no?
13:35:41	sfinucan	jaypipes: not really. If every device gets max(unit) + 1, then we could eventually hit 257, right?
13:36:08	sfinucan	once we've executed that many attach operations, that is
13:36:16	jaypipes	sfinucan: oh, I see what you mean... meh, not sure I'm too concerned about that.
13:36:51	sfinucan	jaypipes: Think it's a small enough edge case to ignore so?
13:37:08	jaypipes	sfinucan: yeah. considering rebooting the VM would "clear that out", right?
13:38:08	sfinucan	jaypipes: I don't think so? The XML isn't rebuilt on a reboot, right? Only a resize/migrate?
13:38:35	jaypipes	sfinucan: is it? I thought it was rebuilt on reboot...
13:42:51	sfinucan	jaypipes: Looks like the complete rebuild only happens if libvirt doesn't support 'reset' https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L2422-L2424
13:43:30	sfinucan	However, it is an edge case. I'll wait an see what sahid says and if he's OK with it, it can slide
13:44:28	jaypipes	https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L2428
13:44:31	jaypipes	sfinucan: ^
13:45:20	openstackgerrit	Ildiko Vancsa proposed openstack/nova master: WIP/POC Implement new attach Cinder flow https://review.openstack.org/330285
13:45:28	sfinucan	jaypipes: The comments lie. Grr
13:47:28	sfinucan	That wouldn't be the case for soft_reboot though, I guess - I've yet to see any undefine calls yet https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L2384-L2385
13:48:26	jaypipes	sfinucan: no, I'm referring to a hard reboot.
13:53:15	sfinucan	Folks with +2 permissions: would appreciate reviews on this series today --> https://review.openstack.org/#/q/topic:doc-migration+project:openstack/nova
13:53:20	sfinucan	I want to get started on moving the guides from openstack-manuals in here this week
14:02:27	alex_xu	looks like we don't have scheduler meeting today
14:02:37	jaypipes	sfinucan: ++
14:02:54	jaypipes	alex_xu: I'm always happy to chat about scheduler :)
14:03:02	jaypipes	alex_xu: but yeah, a number of US folks are out.
14:03:19	alex_xu	jaypipes: ah, holiday?
14:04:34	jaypipes	alex_xu: July 4th tomorrow so lots of folks taking today off too.
14:04:48	alex_xu	jaypipes: cool, got it
14:18:03	sfinucan	moshele: You about. I've a question about SR-IOV devices and PCI affinity
14:19:29	sfinucan	moshele: Namely, if I boot a VM with an SR-IOV device, then will that VM be forced to use CPUs from the same NUMA node as the actual PCI device. I thought this was the case
14:20:23	openstackgerrit	Jay Pipes proposed openstack/nova master: placement: scheduler uses allocation candidates https://review.openstack.org/476631
14:20:24	openstackgerrit	Jay Pipes proposed openstack/nova master: WIP - placement: claim resources in scheduler https://review.openstack.org/476632
14:20:24	openstackgerrit	Jay Pipes proposed openstack/nova master: placement: support GET /allocation_candidates https://review.openstack.org/475448
14:21:41	openstackgerrit	dane-fichter proposed openstack/nova master: Mask password when logging connection_info during swap volume https://review.openstack.org/466533
14:27:32	moshele	sfinucan: this is what was implemented https://review.openstack.org/#/c/140290/
14:29:30	mriedem	wubba wubba wubba
14:30:13	moshele	sfinucan: is that answer you question?
14:31:10	mriedem	jaypipes: so you fixed the pep8 issue in https://review.openstack.org/#/c/475448/ but did you see the functional tests failing listing traits?
14:31:18	mriedem	something going on with the lock or loading traits from the library it seems
14:31:38	mriedem	http://logs.openstack.org/48/475448/21/check/gate-nova-tox-functional-ubuntu-xenial/59c59c0/console.html#_2017-07-01_01_12_45_285206
14:31:45	mriedem	i could recreate it locally, but couldn't figure out what was causing it
14:32:46	jaypipes	mriedem: I thought edleafe- had fixed that in a patch last week... lemme check.
14:33:26	openstackgerrit	Jay Pipes proposed openstack/nova master: placement: scheduler uses allocation candidates https://review.openstack.org/476631
14:33:27	openstackgerrit	Jay Pipes proposed openstack/nova master: placement: support GET /allocation_candidates https://review.openstack.org/475448
14:33:28	openstackgerrit	Jay Pipes proposed openstack/nova master: WIP - placement: claim resources in scheduler https://review.openstack.org/476632
14:33:32	jaypipes	mriedem: pulled in master ^^
14:33:45	sfinucan	moshele: Kind of. That's for PCI devices, but would that include VF devices?
14:33:45	jaypipes	mriedem: I think that func test failure is fixed.
14:33:47	sfinucan	I guess
14:33:51	jaypipes	mriedem: will see in about an hour ;)
14:33:58	sfinucan	* I assume it would, but I'm not sure
14:34:00	moshele	moshele
14:34:00	moshele	sfinucan
14:34:06	mriedem	jaypipes: ok i can hunt for a recent change related to that

Earlier   Later