| *** haomaiwang has quit IRC | 00:01 | |
| *** haomaiwang has joined #openstack-nova | 00:01 | |
| *** mrkz has quit IRC | 00:02 | |
| *** su_zhang has quit IRC | 00:02 | |
| *** su_zhang has joined #openstack-nova | 00:03 | |
| *** su_zhang has quit IRC | 00:03 | |
| *** su_zhang has joined #openstack-nova | 00:04 | |
| *** vilobhmm11 has joined #openstack-nova | 00:05 | |
| *** su_zhang has quit IRC | 00:05 | |
| *** darrenc_afk is now known as darrenc | 00:05 | |
| *** su_zhang has joined #openstack-nova | 00:05 | |
| *** su_zhang has quit IRC | 00:06 | |
| *** ivase has joined #openstack-nova | 00:06 | |
| *** su_zhang has joined #openstack-nova | 00:07 | |
| *** EinstCrazy has joined #openstack-nova | 00:10 | |
| *** ivase has quit IRC | 00:11 | |
| *** rk4n has joined #openstack-nova | 00:11 | |
| *** ducttape_ has quit IRC | 00:12 | |
| *** eil397 has quit IRC | 00:13 | |
| *** EinstCrazy has quit IRC | 00:14 | |
| *** markvoelker has quit IRC | 00:15 | |
| *** edtubill has joined #openstack-nova | 00:15 | |
| *** Sukhdev has quit IRC | 00:16 | |
| *** Sukhdev has joined #openstack-nova | 00:17 | |
| *** Sukhdev has quit IRC | 00:18 | |
| *** spandhe has quit IRC | 00:19 | |
| *** vilobhmm11 has quit IRC | 00:19 | |
| *** vilobhmm11 has joined #openstack-nova | 00:20 | |
| *** cdent has joined #openstack-nova | 00:20 | |
| *** vilobhmm11 has quit IRC | 00:21 | |
| *** vilobhmm11 has joined #openstack-nova | 00:21 | |
| *** vilobhmm11 has quit IRC | 00:21 | |
| *** vilobhmm11 has joined #openstack-nova | 00:22 | |
| *** zenoway has joined #openstack-nova | 00:23 | |
| cdent | jaypipes: given copious free time I'd like to add approximate nearest neighbor (in n-dimensional space) and vector comparisons (via pandas series) strategies to you placement bench stuff. But I'll hold off until I actually have the time to know what I'm doing. | 00:24 |
|---|---|---|
| jaypipes | cdent: cool with me :) | 00:26 |
| cdent | I have mental visual model that I think ought to be resolvable by one of those | 00:26 |
| cdent | but it's just over the horizon | 00:26 |
| *** zenoway has quit IRC | 00:27 | |
| *** spandhe has joined #openstack-nova | 00:28 | |
| *** sheel has joined #openstack-nova | 00:38 | |
| *** rfolco has quit IRC | 00:41 | |
| *** rk4n has quit IRC | 00:41 | |
| *** rk4n has joined #openstack-nova | 00:42 | |
| *** diogogmt has joined #openstack-nova | 00:42 | |
| *** cdent has quit IRC | 00:42 | |
| *** rfolco has joined #openstack-nova | 00:43 | |
| *** rfolco has quit IRC | 00:43 | |
| *** markvoelker has joined #openstack-nova | 00:46 | |
| *** zenoway has joined #openstack-nova | 00:47 | |
| *** zenoway has quit IRC | 00:51 | |
| *** hoangcx has joined #openstack-nova | 00:58 | |
| *** ducttape_ has joined #openstack-nova | 00:59 | |
| *** haomaiwang has quit IRC | 01:01 | |
| *** haomaiwang has joined #openstack-nova | 01:01 | |
| *** Bofu2MBP has quit IRC | 01:06 | |
| *** su_zhang has quit IRC | 01:07 | |
| *** su_zhang has joined #openstack-nova | 01:07 | |
| *** su_zhang has quit IRC | 01:08 | |
| *** kfarr has quit IRC | 01:08 | |
| *** gongysh has joined #openstack-nova | 01:08 | |
| *** su_zhang has joined #openstack-nova | 01:08 | |
| *** zenoway has joined #openstack-nova | 01:10 | |
| *** nic has quit IRC | 01:13 | |
| *** Guest75 has joined #openstack-nova | 01:13 | |
| *** Sukhdev has joined #openstack-nova | 01:14 | |
| *** zenoway has quit IRC | 01:15 | |
| *** mylu has quit IRC | 01:17 | |
| *** armax has joined #openstack-nova | 01:21 | |
| *** zhangyufei has joined #openstack-nova | 01:25 | |
| *** gtt116_ has quit IRC | 01:28 | |
| *** penick has quit IRC | 01:28 | |
| *** su_zhang has quit IRC | 01:28 | |
| *** su_zhang has joined #openstack-nova | 01:29 | |
| *** spandhe has quit IRC | 01:31 | |
| *** su_zhang has quit IRC | 01:33 | |
| *** zenoway has joined #openstack-nova | 01:34 | |
| *** lpetrut has quit IRC | 01:36 | |
| *** dmellado has quit IRC | 01:36 | |
| *** sgordon has quit IRC | 01:36 | |
| *** haomaiwang has quit IRC | 01:36 | |
| *** sgordon has joined #openstack-nova | 01:37 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/282136 | 01:38 |
| *** zenoway has quit IRC | 01:38 | |
| *** ducttape_ has quit IRC | 01:42 | |
| *** gyee has quit IRC | 01:46 | |
| *** gyee has joined #openstack-nova | 01:46 | |
| *** erlon has quit IRC | 01:46 | |
| *** jwcroppe has joined #openstack-nova | 01:54 | |
| *** rk4n has quit IRC | 01:55 | |
| *** kairo has joined #openstack-nova | 01:55 | |
| *** rk4n has joined #openstack-nova | 01:56 | |
| *** baoli has joined #openstack-nova | 01:58 | |
| openstackgerrit | Feng Xi Yan proposed openstack/nova: <VMWare> Allocate free bus for new SCSI controller. https://review.openstack.org/272414 | 01:58 |
| *** Sukhdev has quit IRC | 01:59 | |
| *** dims has quit IRC | 01:59 | |
| *** vilobhmm11 has quit IRC | 01:59 | |
| *** yamahata has quit IRC | 02:00 | |
| *** Sukhdev has joined #openstack-nova | 02:01 | |
| *** Sukhdev has quit IRC | 02:01 | |
| *** zhangyufei has quit IRC | 02:02 | |
| *** rk4n has quit IRC | 02:02 | |
| *** vilobhmm11 has joined #openstack-nova | 02:03 | |
| *** vilobhmm11 has quit IRC | 02:03 | |
| *** vilobhmm11 has joined #openstack-nova | 02:03 | |
| *** rk4n has joined #openstack-nova | 02:05 | |
| openstackgerrit | Wenzhi Yu proposed openstack/nova-specs: Modify Nova API to reject invalid operations on physical servers https://review.openstack.org/283904 | 02:06 |
| *** vishwanathj has quit IRC | 02:06 | |
| openstackgerrit | Felix Ma proposed openstack/nova: Catch iscsi VolumeDeviceNotFound when detaching https://review.openstack.org/265681 | 02:06 |
| *** jdurgin1 has quit IRC | 02:07 | |
| *** pcaruana has quit IRC | 02:07 | |
| *** vilobhmm11 has quit IRC | 02:12 | |
| *** haomaiwang has joined #openstack-nova | 02:16 | |
| *** browne has quit IRC | 02:16 | |
| *** apoorvad has quit IRC | 02:19 | |
| *** pcaruana has joined #openstack-nova | 02:19 | |
| *** apoorvad has joined #openstack-nova | 02:22 | |
| *** ljxiash has joined #openstack-nova | 02:24 | |
| *** vishwanathj has joined #openstack-nova | 02:26 | |
| *** vishwana_ has joined #openstack-nova | 02:28 | |
| *** zhangyufei has joined #openstack-nova | 02:30 | |
| *** vishwanathj has quit IRC | 02:31 | |
| gus | johnthetubaguy/hemna: ok, https://review.openstack.org/#/c/277224/ (os-brick/privsep) works. | 02:32 |
| *** mgoddard_ has joined #openstack-nova | 02:32 | |
| gus | Please review (and the depends-on changes in cinder+nova). | 02:33 |
| *** mgoddard has quit IRC | 02:35 | |
| gus | I also need a response to my question on https://review.openstack.org/#/c/277670/: sdague/matt riedemann wanted a release note added somewhere in nova (not os-brick), and I'm asking for some clarification over exactly how they want that done given that there isn't (at this stage) a hard requirement anywhere for nova to update to a privsep version of os-brick. | 02:36 |
| *** houming has joined #openstack-nova | 02:38 | |
| *** Marga_ has quit IRC | 02:39 | |
| *** apoorvad has quit IRC | 02:45 | |
| *** zz_dimtruck is now known as dimtruck | 02:47 | |
| *** mingyu has quit IRC | 02:51 | |
| *** mingyu has joined #openstack-nova | 02:51 | |
| *** dmellado has joined #openstack-nova | 02:54 | |
| *** rk4n has quit IRC | 02:54 | |
| *** vilobhmm11 has joined #openstack-nova | 02:59 | |
| openstackgerrit | Wenzhi Yu proposed openstack/nova: Make sure destination will rollback if live migration fails https://review.openstack.org/270093 | 03:00 |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-nova | 03:01 | |
| *** pcaruana has quit IRC | 03:01 | |
| *** gongysh has quit IRC | 03:02 | |
| *** jyuso1 has joined #openstack-nova | 03:03 | |
| *** hoangcx has quit IRC | 03:06 | |
| *** dmellado has quit IRC | 03:06 | |
| *** zhangyufei has quit IRC | 03:06 | |
| *** zhangyufei has joined #openstack-nova | 03:07 | |
| *** browne has joined #openstack-nova | 03:09 | |
| *** dmellado has joined #openstack-nova | 03:11 | |
| *** dims has joined #openstack-nova | 03:13 | |
| *** pcaruana has joined #openstack-nova | 03:16 | |
| *** gongysh has joined #openstack-nova | 03:19 | |
| *** haomaiwang has quit IRC | 03:20 | |
| *** dmellado has quit IRC | 03:22 | |
| *** salv-orlando has joined #openstack-nova | 03:26 | |
| *** salv-orlando has quit IRC | 03:28 | |
| *** salv-orl_ has quit IRC | 03:29 | |
| *** unicell has quit IRC | 03:33 | |
| *** liusheng has quit IRC | 03:34 | |
| *** liusheng has joined #openstack-nova | 03:34 | |
| *** davideagle has quit IRC | 03:35 | |
| *** mylu has joined #openstack-nova | 03:36 | |
| *** gyee has quit IRC | 03:37 | |
| *** hoangcx has joined #openstack-nova | 03:41 | |
| *** dmellado has joined #openstack-nova | 03:45 | |
| *** links has joined #openstack-nova | 03:46 | |
| *** boris-42 has quit IRC | 03:54 | |
| *** baoli has quit IRC | 03:55 | |
| *** ducttape_ has joined #openstack-nova | 03:58 | |
| *** irenab has quit IRC | 03:59 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/282136 | 03:59 |
| *** irenab has joined #openstack-nova | 04:00 | |
| *** kairo has quit IRC | 04:00 | |
| *** pcaruana has quit IRC | 04:01 | |
| *** dmellado has quit IRC | 04:01 | |
| *** gongysh has quit IRC | 04:02 | |
| *** mylu has quit IRC | 04:02 | |
| *** baoli has joined #openstack-nova | 04:03 | |
| *** liusheng has quit IRC | 04:03 | |
| *** zhangyufei has quit IRC | 04:03 | |
| *** wanghao has quit IRC | 04:04 | |
| *** vilobhmm11 has quit IRC | 04:04 | |
| *** wanghao has joined #openstack-nova | 04:05 | |
| *** mylu has joined #openstack-nova | 04:06 | |
| *** mylu has quit IRC | 04:12 | |
| *** pcaruana has joined #openstack-nova | 04:15 | |
| *** ducttape_ has quit IRC | 04:15 | |
| *** thorst has joined #openstack-nova | 04:16 | |
| *** mylu has joined #openstack-nova | 04:16 | |
| *** thorst has quit IRC | 04:20 | |
| *** thorst has joined #openstack-nova | 04:20 | |
| *** thorst has quit IRC | 04:20 | |
| *** thorst has joined #openstack-nova | 04:21 | |
| *** ljxiash has quit IRC | 04:24 | |
| *** mylu has quit IRC | 04:25 | |
| *** dmellado has joined #openstack-nova | 04:26 | |
| *** sheel has quit IRC | 04:27 | |
| *** baoli has quit IRC | 04:28 | |
| *** baoli has joined #openstack-nova | 04:29 | |
| *** thorst has quit IRC | 04:30 | |
| *** dimtruck is now known as zz_dimtruck | 04:35 | |
| *** hoangcx has quit IRC | 04:35 | |
| *** Marga_ has joined #openstack-nova | 04:36 | |
| *** mylu has joined #openstack-nova | 04:38 | |
| *** dmellado has quit IRC | 04:38 | |
| *** baoli has quit IRC | 04:39 | |
| *** Marga_ has quit IRC | 04:40 | |
| *** Marga_ has joined #openstack-nova | 04:43 | |
| *** mubirru has joined #openstack-nova | 04:45 | |
| jgriffith | johnthetubaguy: hemna FYI I'm working on a patch for Cinder to help with the "stuck in detach" issue that you guys have been discussing on the ML | 04:46 |
| jgriffith | johnthetubaguy: hemna hope to have something up for review tomorrow, I'll give you a shout | 04:47 |
| *** penick has joined #openstack-nova | 04:48 | |
| *** sridharg has joined #openstack-nova | 04:50 | |
| *** penick_ has joined #openstack-nova | 04:50 | |
| *** penick has quit IRC | 04:53 | |
| *** penick_ is now known as penick | 04:53 | |
| *** harlowja_at_home has quit IRC | 04:53 | |
| *** jamielennox is now known as jamielennox|away | 04:54 | |
| *** Marga_ has quit IRC | 04:57 | |
| *** Marga_ has joined #openstack-nova | 04:58 | |
| *** garyk has joined #openstack-nova | 05:00 | |
| *** dmellado has joined #openstack-nova | 05:03 | |
| *** haomaiwang has joined #openstack-nova | 05:04 | |
| *** amotoki has joined #openstack-nova | 05:05 | |
| *** jwcroppe has quit IRC | 05:10 | |
| *** zhangjn has quit IRC | 05:14 | |
| *** mingyu has quit IRC | 05:17 | |
| *** mingyu has joined #openstack-nova | 05:18 | |
| *** unicell has joined #openstack-nova | 05:20 | |
| *** zhangjn has joined #openstack-nova | 05:20 | |
| *** sudipto has joined #openstack-nova | 05:25 | |
| *** hoangcx has joined #openstack-nova | 05:26 | |
| *** liverpooler has quit IRC | 05:27 | |
| *** thorst has joined #openstack-nova | 05:27 | |
| *** zhangyufei has joined #openstack-nova | 05:30 | |
| openstackgerrit | Feng Xi Yan proposed openstack/nova: <VMWARE> Add power state cache for host init and power state sync. https://review.openstack.org/242006 | 05:32 |
| *** thorst has quit IRC | 05:34 | |
| *** zhangyufei has quit IRC | 05:36 | |
| *** salv-orlando has joined #openstack-nova | 05:36 | |
| *** gcb has quit IRC | 05:37 | |
| *** mylu has quit IRC | 05:37 | |
| *** pcaruana has quit IRC | 05:38 | |
| *** mylu has joined #openstack-nova | 05:39 | |
| *** chhavi has joined #openstack-nova | 05:40 | |
| openstackgerrit | Jianghua Wang proposed openstack/nova: xenapi: OVS agent updates the wrong port when using XenServer + Neutron https://review.openstack.org/242846 | 05:41 |
| *** qeelee has joined #openstack-nova | 05:42 | |
| *** Jeffrey4l has joined #openstack-nova | 05:44 | |
| *** mylu has quit IRC | 05:45 | |
| *** ijw has quit IRC | 05:46 | |
| *** dims has quit IRC | 05:48 | |
| *** salv-orlando has quit IRC | 05:49 | |
| *** ljxiash has joined #openstack-nova | 05:49 | |
| *** pcaruana has joined #openstack-nova | 05:53 | |
| *** zhangyufei has joined #openstack-nova | 05:56 | |
| *** liusheng has joined #openstack-nova | 05:56 | |
| *** gcb has joined #openstack-nova | 05:58 | |
| *** haomaiwang has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-nova | 06:01 | |
| *** vilobhmm11 has joined #openstack-nova | 06:02 | |
| *** gongysh has joined #openstack-nova | 06:02 | |
| *** ekuris has joined #openstack-nova | 06:05 | |
| *** gongysh has quit IRC | 06:05 | |
| openstackgerrit | Alex Xu proposed openstack/nova: Generate better validation error message when using name regexes https://review.openstack.org/281143 | 06:08 |
| *** irenab_ has joined #openstack-nova | 06:14 | |
| *** gongysh has joined #openstack-nova | 06:15 | |
| *** fawadkhaliq has joined #openstack-nova | 06:15 | |
| *** gongysh has quit IRC | 06:16 | |
| *** nkrinner has joined #openstack-nova | 06:16 | |
| *** irenab has quit IRC | 06:17 | |
| *** irenab_ is now known as irenab | 06:17 | |
| *** sridharg has quit IRC | 06:17 | |
| *** tyagiprince has joined #openstack-nova | 06:18 | |
| tyagiprince | Hey everyone. In nova.conf, I made this config change. "osapi_compute_unique_server_name_scope = project" | 06:19 |
| tyagiprince | I tried spawning instances with same instance name. In some openstack setup, it works and in others it doesn't work. | 06:20 |
| tyagiprince | Can someone please tell me why is this happening? | 06:20 |
| *** openstack has joined #openstack-nova | 13:21 | |
| sdague | PaulMurray: oh, sorry, wrong tab | 13:21 |
| sdague | yes | 13:21 |
| sdague | you are correct, it's not landed | 13:22 |
| openstackgerrit | Chris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools https://review.openstack.org/281837 | 13:22 |
| PaulMurray | at least I'm in a known state now ! | 13:22 |
| openstackgerrit | Chris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools https://review.openstack.org/281837 | 13:22 |
| sdague | ihrachys: I think that boot image by tag from nova creates a lot more ambiguity and potential randomness | 13:22 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add name to ResourceProvider object https://review.openstack.org/281945 | 13:23 |
| ihrachys | sdague: the assumption would be that there is only one tagged image; if more, fail as you do for multiple hits for the same name. | 13:23 |
| *** subscope has quit IRC | 13:24 | |
| sdague | if glance had a facility that was unique enforced on it's side for a stable label, using that would be fine | 13:24 |
| sdague | but inventing that in a bunch of convention on tags seems really complicated and potentially breaking long term | 13:24 |
| ihrachys | sdague: is your concern just for nova or you think issues for other projects that could take the approach? | 13:25 |
| sdague | basically, I don't want nova getting into this guessing game, and having a junior admin tag a new image on upload then everything fall appart because it's not unique | 13:25 |
| *** Guest65788 has quit IRC | 13:25 | |
| sdague | I think that imageRef is a known, stable, unique identifier | 13:25 |
| sdague | we should boot by something that's known unique | 13:26 |
| *** rods has joined #openstack-nova | 13:26 | |
| ihrachys | sdague: yeah I understand your compexity concern for nova. though for octavia we would still need to handle it somehow, and the glance tag could be a solution. | 13:26 |
| sdague | it could be, however in an ideal world glance would have something else in place that would let you ask for a stable lable | 13:27 |
| sdague | for instance 'RHEL 7.0 base' | 13:27 |
| ihrachys | sdague: yeah, like mark a tag as unique when mapping it, then allowing to atomically switch it | 13:27 |
| Yingxin | jaypipes: really? what a terrible mistake | 13:27 |
| *** edmondsw has joined #openstack-nova | 13:27 | |
| sdague | ihrachys: well, honestly, your concern is the provider is reving an image right, and it gets a new id every time | 13:28 |
| sdague | right? | 13:28 |
| ihrachys | sdague: yeah. every new version of the image is new id that we need to change in config file and restart to apply | 13:28 |
| jaypipes | Yingxin: it's cool, just letting you know :) | 13:28 |
| sdague | which I think is a very valid concern, and it would be nice if there was a construct of 'label' 'version' | 13:28 |
| sdague | lable + version being forced unique | 13:29 |
| sdague | so you can always ask for newest of label | 13:29 |
| *** qeelee has quit IRC | 13:29 | |
| sdague | which is all a bunch of stuff that doesn't exist, but seems to be what you really want | 13:29 |
| sdague | that you might be able to hack together a 70% solution with tags | 13:30 |
| ihrachys | right, that would be ideal. | 13:30 |
| sdague | if glance did that, I think having nova support it is a no brainer | 13:30 |
| sdague | but the 30% failure space on the tags overload leads you into some really confusing situations for users | 13:31 |
| *** 32NAAD4T7 has quit IRC | 13:31 | |
| sdague | so I wouldn't want that in nova | 13:31 |
| ihrachys | yeah. I may need to work with glance folks on long term and maybe bake something with what I have now in short term. the latter won't probably belong to nova, agreed. | 13:32 |
| ihrachys | and later if we get uniqueness feature, we may want to reconsider it for nova. | 13:32 |
| *** ildikov has quit IRC | 13:33 | |
| *** zhangjn has quit IRC | 13:33 | |
| sdague | ihrachys: ++ | 13:37 |
| *** vladikr has joined #openstack-nova | 13:37 | |
| *** shyama has joined #openstack-nova | 13:38 | |
| sahid | ndipanov, danpb when you have a moment can you ack this change https://review.openstack.org/#/c/277422/ | 13:38 |
| *** mgoddard__ has joined #openstack-nova | 13:38 | |
| *** zhangjn has joined #openstack-nova | 13:38 | |
| *** mubirru has quit IRC | 13:39 | |
| *** dims has joined #openstack-nova | 13:40 | |
| *** mgoddard_ has quit IRC | 13:41 | |
| PaulMurray | andreykurilin, hi - are you working on https://review.openstack.org/#/c/278454/ | 13:42 |
| *** atuvenie__ has joined #openstack-nova | 13:42 | |
| *** chhavi has joined #openstack-nova | 13:43 | |
| andreykurilin | PaulMurray: hi! imo, this patch is finished:) I will talk with mriedem when he will be available. About mentioned bug - I'll post a fix in separate patch in few hours | 13:44 |
| openstackgerrit | Chris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools https://review.openstack.org/281837 | 13:44 |
| *** amotoki has joined #openstack-nova | 13:44 | |
| johnthetubaguy | ihrachys: FWIW, I think we need those features in glance, so we can build well know names on top of that, so you can say "give me the latest Ubuntu 12.04 LTS please", and have that mean the same thing on all certified OpenStack clouds, but I can't say I have got very far with pushing that yet | 13:44 |
| PaulMurray | andreykurilin, we have patches chained up behind you - will your fix affect those ? | 13:45 |
| bauzas | Yingxin: jaypipes: shall I mention my name too ? :D | 13:45 |
| openstackgerrit | Chris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools https://review.openstack.org/281837 | 13:45 |
| *** atuvenie has quit IRC | 13:45 | |
| ihrachys | johnthetubaguy: do we have any patches uploaded? | 13:45 |
| *** ccarmack has joined #openstack-nova | 13:46 | |
| johnthetubaguy | ihrachys: don't even have a spec for that at this point, AFAIK, its just an idea I was talking about on the defcore list, a little bit back, not had chance to move that any further forward sadly. | 13:46 |
| *** shaohe_feng_ has joined #openstack-nova | 13:47 | |
| andreykurilin | PaulMurray: I can put it to start of chain | 13:47 |
| johnthetubaguy | ihrachys: well I presented the approach at the hong kong summit, just just not had chance to move it forward | 13:47 |
| *** EinstCrazy has joined #openstack-nova | 13:47 | |
| *** porunov has joined #openstack-nova | 13:47 | |
| shaohe_feng_ | johnthetubaguy: sdague : ping | 13:48 |
| PaulMurray | andreykurilin, I'll track what's going on with your patch - I need to get the whole chain merged by friday :) | 13:48 |
| jaypipes | bauzas: I had been calling you "Sil-vane" for a long time until I learned it was pronounced "Sil-vaaan". I have been suitably corrected. | 13:48 |
| bauzas | heh | 13:48 |
| bauzas | jaypipes: don't ask alaski how many times I tried to say "Andrew" | 13:49 |
| shaohe_feng_ | johnthetubaguy: sdague : I want to talk about https://review.openstack.org/#/c/258771/27/nova/api/openstack/compute/server_migrations.py | 13:49 |
| andreykurilin | PaulMurray: It would be nice to merge the whole chain as soon as possible:) | 13:50 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add name to ResourceProvider object https://review.openstack.org/281945 | 13:50 |
| shaohe_feng_ | johnthetubaguy: sdague: so I the flavor should be like: "flavor": { "id": "1", "links": [ { "href": "http://openstack.example.com/openstack/flavors/1", "rel": "bookmark" } ] }, | 13:51 |
| sdague | shaohe_feng_: right, that "id" isn't the same id as used by migrations | 13:51 |
| PaulMurray | jaypipes, for max amusement value ask bauzas to say: | 13:51 |
| PaulMurray | Theophilus Thistle, the successful thistle sifter, | 13:51 |
| PaulMurray | In sifting a sieve full of un-sifted thistles, | 13:51 |
| PaulMurray | Thrust three thousand thistles through the thick of his thumb. | 13:51 |
| PaulMurray | If Theophilus Thistle, the successful thistle sifter, | 13:51 |
| PaulMurray | Can thrust three thousand thistles through the thick of his thumb, | 13:51 |
| PaulMurray | See that thou, in sifting a sieve full of un-sifted thistles, | 13:51 |
| PaulMurray | Thrust not three thousand thistles through the thick of thy thumb. | 13:51 |
| bauzas | oh man | 13:51 |
| jaypipes | rofl | 13:52 |
| ihrachys | johnthetubaguy: seems like a feature that would be interesting for multiple projects. I gotta talk to glance folks for sure. :) | 13:52 |
| Bobba | tonyb: You happy with the comments on https://review.openstack.org/#/c/281341/ regarding the CI? I imagine that a -1 might just filter things out for you :) | 13:52 |
| shaohe_feng_ | sdague: should I give the ref link also? | 13:52 |
| sdague | shaohe_feng_: well, lets step back for a minute | 13:52 |
| bauzas | PaulMurray: les chaussettes de l'archiduchesse sont-elles sèches, archi-sèches? | 13:52 |
| sdague | are the old_/new_ flavor ids actually useful in this case? | 13:52 |
| *** zenoway has quit IRC | 13:53 | |
| sdague | because in live migration they are always the same, right? | 13:53 |
| sdague | I kind of wonder if we just delete them entirely from representation for now | 13:53 |
| johnthetubaguy | sdague: good point, thats for resize, dropping them is probably best | 13:53 |
| johnthetubaguy | sdague: I wonder if there are other things in there too, now you mention it | 13:54 |
| sdague | johnthetubaguy: maybe, everything else seems kind of useful | 13:54 |
| johnthetubaguy | yeah, it does | 13:54 |
| johnthetubaguy | just double checked | 13:54 |
| *** sudipto has joined #openstack-nova | 13:55 | |
| sdague | shaohe_feng_: would deleting those fields cause any other issues you see? | 13:55 |
| shaohe_feng_ | sdague: johnthetubaguy: we still need old_/new_ flavor field, but without id? | 13:55 |
| johnthetubaguy | I think just drop it, for now | 13:55 |
| sdague | shaohe_feng_: why do you need to return it to the user? | 13:55 |
| shaohe_feng_ | sdague: we do not need it, user can git it by the showserver \ | 13:56 |
| sdague | shaohe_feng_: right | 13:56 |
| shaohe_feng_ | sdague: IMHO | 13:56 |
| sdague | so how about we delete them from the REST representation here | 13:56 |
| shaohe_feng_ | sdague: agree. | 13:56 |
| sdague | and I think that removes my final concern, then I'm +2 on the patch | 13:56 |
| *** eharney has joined #openstack-nova | 13:57 | |
| sdague | shaohe_feng_: so if you want to respin that, I think we are good | 13:57 |
| shaohe_feng_ | sdague: OK. let remove the flavor field. | 13:58 |
| jaypipes | cdent: doh, should be migrate_engine.connect() :) | 13:58 |
| *** ildikov has joined #openstack-nova | 13:58 | |
| *** atuvenie__ has quit IRC | 13:58 | |
| cdent | Yeah, I fixed that | 13:58 |
| cdent | jaypipes: then that led to the DDL being wrong | 13:58 |
| cdent | but I think I've fixed that too | 13:58 |
| jaypipes | heh | 13:58 |
| shaohe_feng_ | sdague: will ping you after I update it, | 13:59 |
| sdague | shaohe_feng_: sounds great, thank you | 13:59 |
| jaypipes | I think the Intel PCI CI is just dead. | 13:59 |
| sdague | jaypipes: has it ever been really alive? | 14:00 |
| johnthetubaguy | jaypipes: I think mikal was trying to help those folks, not sure where that is at right now | 14:00 |
| *** haomaiwa_ has quit IRC | 14:01 | |
| *** salv-orlando has joined #openstack-nova | 14:01 | |
| *** haomaiwang has joined #openstack-nova | 14:01 | |
| *** cfriesen has joined #openstack-nova | 14:01 | |
| PaulMurray | ndipanov, I know pci passthrough doesn't work with live migration - does it work with cod migration (I've seen a couple of bugs about this) | 14:01 |
| ndipanov | PaulMurray, yes | 14:01 |
| *** rfolco has joined #openstack-nova | 14:01 | |
| ndipanov | er | 14:01 |
| ndipanov | pci passthrough | 14:02 |
| ndipanov | no :( | 14:02 |
| ndipanov | I think not reliably no | 14:02 |
| *** xyang1 has joined #openstack-nova | 14:02 | |
| ndipanov | jaypipes, so there was a bug that was killing it that I quickly tried to fix | 14:02 |
| ndipanov | jaypipes, I think it was the same thing - NFV CI right? | 14:02 |
| ndipanov | jaypipes, never got much love after the patches were posted: | 14:03 |
| ndipanov | https://review.openstack.org/#/c/281483/ | 14:03 |
| ndipanov | (hint: give it love) | 14:03 |
| *** mdrabe has joined #openstack-nova | 14:03 | |
| jaypipes | ndipanov: k, will take a looksie shortly. | 14:03 |
| ndipanov | ah so PCI CI is not the same thing | 14:04 |
| *** efoley__ has joined #openstack-nova | 14:04 | |
| ndipanov | still both are useful | 14:04 |
| *** tyagiprince has quit IRC | 14:04 | |
| *** dimtruck is now known as zz_dimtruck | 14:05 | |
| *** efoley_ has quit IRC | 14:06 | |
| *** edtubill has joined #openstack-nova | 14:06 | |
| *** ctrath has joined #openstack-nova | 14:07 | |
| *** salv-orlando has quit IRC | 14:08 | |
| *** ljxiash has joined #openstack-nova | 14:10 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] Trying latest oslo.* from master https://review.openstack.org/275201 | 14:10 |
| *** rtheis has joined #openstack-nova | 14:13 | |
| *** mriedem has joined #openstack-nova | 14:14 | |
| rtheis | Is there a nova.conf configuration to accept ssh host keys by default for live migration? | 14:14 |
| *** rlrossit_ has joined #openstack-nova | 14:17 | |
| Yingxin | bauzas: :P | 14:18 |
| *** jaosorior has quit IRC | 14:18 | |
| *** zenoway has joined #openstack-nova | 14:18 | |
| *** jaosorior has joined #openstack-nova | 14:19 | |
| *** bklei has joined #openstack-nova | 14:19 | |
| *** BharatK has joined #openstack-nova | 14:20 | |
| *** gmann_ has joined #openstack-nova | 14:21 | |
| dave-mccowan | hello nova devs. i have a question about the use_forwarded_for setting. i want my nova-api.log to show the passed client address. when i set use_forwarded_for = true, it does. are there any other consequences (good or bad) to using this setting? | 14:22 |
| *** ivase has joined #openstack-nova | 14:23 | |
| *** ociuhandu has quit IRC | 14:24 | |
| *** ivase has quit IRC | 14:25 | |
| *** ivase has joined #openstack-nova | 14:25 | |
| *** ociuhandu has joined #openstack-nova | 14:25 | |
| alaski | bauzas: heh. it's mostly that you say it with an accent, not that it's wrong | 14:28 |
| bauzas | WAT? I do have an accent when speaking ? orly ? :D | 14:28 |
| pkoniszewski | rtheis: there isn't one or i'm not aware of that | 14:29 |
| rtheis | thanks | 14:30 |
| sdague | mriedem: https://review.openstack.org/#/c/282012/ respun again. And to help with https://review.openstack.org/#/c/283837 I'm starting to try to unwind the craziness of sending the microversion on every method call explicitly | 14:32 |
| *** smatzek_ has quit IRC | 14:32 | |
| mriedem | sdague: ok, looking | 14:33 |
| mriedem | sdague: so on https://review.openstack.org/#/c/282012/5..6/nova/tests/functional/regressions/test_bug_1522536.py | 14:34 |
| mriedem | are these changes going to be running against v2.1 on v2.1 or v2.latest on v2.1? | 14:34 |
| *** dedery_ has joined #openstack-nova | 14:34 | |
| sdague | v2.1 on v2.1 | 14:34 |
| mriedem | b/c i'd think for the tests to be useful, they should test against v2.latest to sniff out issues before they are merged | 14:34 |
| *** shyama has quit IRC | 14:35 | |
| *** EinstCrazy has quit IRC | 14:35 | |
| sdague | that assumes that there are changes between base and that | 14:35 |
| sdague | I think if we want to do that later, that's fine. I don't think these tests are only useful if they run against v2.latest | 14:37 |
| sdague | because in all the bugs I've been working at so far, they expose on all versions | 14:37 |
| *** dedery has quit IRC | 14:38 | |
| mriedem | i guess my thought process was for this list deleted instances thing, if we had a functional test for that which was testing against 2.1 it would be fine, but 2.16 broke it | 14:39 |
| sdague | sure, which I think we know there | 14:39 |
| mriedem | so i'll add the test for 2.16, and that will handle regressions up to 2.16 | 14:39 |
| sdague | yeh | 14:40 |
| mriedem | but not if like 2.35 breaks something | 14:40 |
| sdague | right, that's true. It's all trade offs | 14:40 |
| *** rtheis has quit IRC | 14:40 | |
| mriedem | what's not a trade off was making my 4 year old watch this with me last night http://www.rottentomatoes.com/m/krull/ | 14:41 |
| *** rtheis has joined #openstack-nova | 14:41 | |
| *** sc68cal has joined #openstack-nova | 14:41 | |
| sdague | because v2.latest would mean if someone broke common code that made v2.11 break, but work in v2.35 it would miss that | 14:41 |
| sdague | nice | 14:41 |
| pkoniszewski | johnthetubaguy: just FYI, i've uploaded review per my AR from yesterdays live migration meeting (fix force_complete according to API guideline) - https://review.openstack.org/#/c/284068/ | 14:42 |
| sdague | mriedem: I'm going to work through getting api_version out of the param list for the client right now | 14:42 |
| *** diogogmt has quit IRC | 14:42 | |
| *** dedery_ has quit IRC | 14:42 | |
| sdague | you can stack on top of that once I get it done | 14:42 |
| *** dedery has joined #openstack-nova | 14:42 | |
| *** rtheis has quit IRC | 14:42 | |
| mriedem | alright | 14:43 |
| *** rtheis has joined #openstack-nova | 14:43 | |
| *** diogogmt has joined #openstack-nova | 14:44 | |
| openstackgerrit | Pawel Koniszewski proposed openstack/nova: Fix return codes in force_complete API https://review.openstack.org/284068 | 14:45 |
| sdague | although now I'm trapped in the massive land of over optimizing test cases through a ton of private methods .... :( | 14:45 |
| openstackgerrit | ShaoHe Feng proposed openstack/nova: Add new APIs and deprecate old API for migrations https://review.openstack.org/258771 | 14:46 |
| *** Daisy_ has joined #openstack-nova | 14:46 | |
| mriedem | sdague: you need the glaive | 14:47 |
| *** armax has quit IRC | 14:47 | |
| mriedem | it's easily accessible in a mountain climb, plus cave with river of fire | 14:47 |
| * mriedem plans on referencing back to krull today | 14:48 | |
| sdague | heh | 14:48 |
| shaohe_feng_ | sdague: johnthetubaguy : https://review.openstack.org/258771 | 14:48 |
| shaohe_feng_ | thanks | 14:48 |
| pkoniszewski | shaohe_feng: why did u remove flavor ids completely? it is part of a response in specification | 14:49 |
| sdague | shaohe_feng_: you have to update the docs and samples | 14:50 |
| *** claudiub|2 has joined #openstack-nova | 14:50 | |
| shaohe_feng_ | sdague: OK, have update the docs and sample | 14:51 |
| shaohe_feng_ | s | 14:51 |
| shaohe_feng_ | pkoniszewski: will update the doc | 14:51 |
| shaohe_feng_ | pkoniszewski: sorry, will update the spec | 14:51 |
| sdague | pkoniszewski: http://eavesdrop.openstack.org/irclogs/%23openstack-nova/%23openstack-nova.2016-02-24.log.html#t2016-02-24T13:52:43 | 14:51 |
| sdague | the fields that were exposed were not useful | 14:52 |
| sdague | and actually problematic because they leak out an index field that is never exposed to the user, and something they can't do anything with | 14:52 |
| pkoniszewski | sdague: it makes sense, just asking because it is still part of a spec, thanks! | 14:52 |
| sdague | pkoniszewski: right, it was caught late in review | 14:52 |
| *** claudiub has quit IRC | 14:53 | |
| shaohe_feng_ | pkoniszewski: thanks for reminder. | 14:53 |
| sdague | gibi: you about? | 14:53 |
| *** rlrossit_ has quit IRC | 14:53 | |
| *** rlrossit has joined #openstack-nova | 14:53 | |
| sdague | because I'm having a hard time making your test_server_group tests work if we set global microversion per test | 14:53 |
| shaohe_feng_ | sdague: have update the doc and samples. Thanks for remind | 14:55 |
| *** rhagarty_ has quit IRC | 14:55 | |
| smcginnis | mriedem: Krull references? Really? :) | 14:56 |
| bjornar | How can I disable the "rados df" on each compute host? | 14:56 |
| openstackgerrit | Pawel Koniszewski proposed openstack/nova: Fix return codes in force_complete API https://review.openstack.org/284068 | 14:56 |
| rlrossit | smcginnis: don't encourage him | 14:57 |
| smcginnis | rlrossit: haha | 14:58 |
| *** Daisy_ has quit IRC | 14:59 | |
| *** mtanino has joined #openstack-nova | 14:59 | |
| *** mrkz has joined #openstack-nova | 14:59 | |
| *** Daisy has joined #openstack-nova | 15:00 | |
| cfriesen | Anyone here in touch with Phil Day? I've proposed a fix to some changes he made around min_count and network quota constraints: https://review.openstack.org/#/c/283861/1 | 15:00 |
| *** rtheis has quit IRC | 15:00 | |
| *** smatzek_ has joined #openstack-nova | 15:00 | |
| *** haomaiwang has quit IRC | 15:01 | |
| *** ducttape_ has joined #openstack-nova | 15:01 | |
| *** rtheis has joined #openstack-nova | 15:01 | |
| *** armax has joined #openstack-nova | 15:01 | |
| *** haomaiwang has joined #openstack-nova | 15:01 | |
| *** diogogmt has quit IRC | 15:01 | |
| dansmith | alaski: what is your feeling on this? https://review.openstack.org/#/c/282520/5//COMMIT_MSG | 15:02 |
| *** jdillaman has quit IRC | 15:02 | |
| *** smatzek__ has joined #openstack-nova | 15:02 | |
| *** efoley__ has quit IRC | 15:02 | |
| *** krtaylor has quit IRC | 15:02 | |
| *** rtheis has quit IRC | 15:03 | |
| *** rtheis has joined #openstack-nova | 15:03 | |
| *** efoley__ has joined #openstack-nova | 15:03 | |
| shaohe_feng_ | jaypipes: pingt | 15:03 |
| shaohe_feng_ | jaypipes: ping | 15:03 |
| mriedem | is anyone else seeing a bunch of these in tests now? | 15:04 |
| mriedem | ObjectActionError: Object action obj_load_attr failed because: attribute id not lazy-loadable | 15:04 |
| shaohe_feng_ | jaypipes: the unit is bytes. they are come from libvirt. libvirt are byte. | 15:04 |
| sdague | jaypipes: really, you are -1ing for commit message typos? | 15:05 |
| jaypipes | sdague: no. | 15:05 |
| *** Daisy has quit IRC | 15:05 | |
| *** smatzek_ has quit IRC | 15:05 | |
| jaypipes | sdague: I am -1'ing because the REST API does not tell the user what the unit of measurement is. | 15:05 |
| sdague | ok | 15:05 |
| jaypipes | sdague: please read my review comment. | 15:05 |
| *** betherly_ is now known as betherly | 15:06 | |
| sdague | I read the inline bit, gerrit shows that first | 15:06 |
| dansmith | mriedem: link | 15:06 |
| dansmith | ? | 15:06 |
| jaypipes | shaohe_feng: please do not leak libvirt specifics or tribal knowledge out of the public REST API. There is no way a user would know what unit of measurement that is. | 15:06 |
| mriedem | dansmith: this is local, but someone else was complaining about a unit test of theirs failing | 15:07 |
| shaohe_feng_ | jaypipes: any suggestion to improve the patch? | 15:07 |
| mriedem | logstash isn't showing me anything though | 15:07 |
| sdague | jaypipes: so you would be happy with memory_units and disk_units ? | 15:07 |
| dansmith | mriedem: okay, yeah, haven't seen that | 15:08 |
| jaypipes | sdague: no. I would be happy with memory_total_bytes and memory_processed_bytes | 15:08 |
| *** salv-orlando has joined #openstack-nova | 15:08 | |
| sdague | jaypipes: in most cases the important bit is progress, not raw. | 15:08 |
| jaypipes | sdague: same for disk. | 15:08 |
| jaypipes | sdague: I actually thought memory_total was pages, not bytes. there's no way to know unless you have tribal knowledge of libvirt. | 15:08 |
| jaypipes | sdague: thus my request to simply add the unit of measurement to the field names in the output. | 15:08 |
| sdague | ok, that's fair | 15:08 |
| dansmith | which patch is this? I'll -1 for the commit message typos :) | 15:09 |
| jaypipes | dansmith: instigator. | 15:09 |
| dansmith | jaypipes: s/instigator/connoisseur/ | 15:09 |
| mriedem | apparently my functional test is hitting a 500 when deleting the instances | 15:09 |
| mriedem | looks like some rpc stuff isn't stubbed out | 15:09 |
| sdague | mriedem: I can look in a little bit, I'm trying to figure out this other unwind first | 15:10 |
| *** burgerk has joined #openstack-nova | 15:12 | |
| *** krtaylor has joined #openstack-nova | 15:14 | |
| *** sridharg has quit IRC | 15:14 | |
| *** Swami has joined #openstack-nova | 15:14 | |
| sdague | dansmith: when did force down land? | 15:15 |
| dansmith | sdague: which part? | 15:15 |
| bauzas | gibi: around ? | 15:15 |
| sdague | the api | 15:15 |
| dansmith | I want to say liberty | 15:15 |
| sdague | yep, 2.11 | 15:15 |
| sdague | that's why gibi's tests in server_groups don't work | 15:15 |
| *** lyanchih has joined #openstack-nova | 15:15 | |
| sdague | because he's using 2.15 behavior but not accounting for 2.11 changes | 15:16 |
| bauzas | gibi: I wonder if we have any problem with https://github.com/openstack/nova/blob/master/nova/test.py#L143-L159 when a method we want to mock is already decorated | 15:16 |
| *** ivase_ has joined #openstack-nova | 15:16 | |
| bauzas | like with functools.wrap | 15:16 |
| sdague | bauzas: you should delete that code | 15:16 |
| sdague | mock already does that now | 15:17 |
| ansiwen | the metadata caching conflicts with the lazy loading of - for example - flavor of the instance. the object is cached without it, resulting in a database request for each request. | 15:17 |
| sdague | automatically | 15:17 |
| bauzas | sdague: kewl | 15:17 |
| *** ivase has quit IRC | 15:17 | |
| ansiwen | anyway I don't quite understand why it makes sense to store deep copies of python objects in memcached. | 15:18 |
| *** atuvenie__ has joined #openstack-nova | 15:18 | |
| ansiwen | instead of fixing I'd rather remove it | 15:18 |
| *** rhagarty has joined #openstack-nova | 15:19 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: [microversions] fix help msg for versioned args https://review.openstack.org/284186 | 15:20 |
| andreykurilin | mriedem: hi! can we discuss your comment to https://review.openstack.org/#/c/278454 ? | 15:21 |
| alaski | dansmith: on which part of https://review.openstack.org/#/c/282520/5//COMMIT_MSG ? From what I've seen I agree that we won't need a nova-manage command for the migration. I also think it will slow the first scheduler restart but I'm not convinced it will be a problematic slowdown | 15:21 |
| openstackgerrit | Chris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools https://review.openstack.org/281837 | 15:23 |
| cdent | jaypipes: le sigh, still have nitpicky things to get that mysql specific aspect of the migration to go, but maybe nearly there | 15:23 |
| dansmith | alaski: wondering if you think the slow start will be a problem (worth worrying about) | 15:24 |
| johnthetubaguy | I guess its only one per agregate? | 15:25 |
| johnthetubaguy | which at least isn't per host | 15:25 |
| dansmith | johnthetubaguy: correct | 15:25 |
| *** gszasz has quit IRC | 15:25 | |
| *** rtheis has quit IRC | 15:25 | |
| *** BharatK has quit IRC | 15:25 | |
| dansmith | johnthetubaguy: if you had a lot of aggregates, then it would be a thing, but I don't think having a million aggregates is likely | 15:25 |
| mriedem | sdague: well i don't know what i did, but now my test is passing | 15:25 |
| dansmith | johnthetubaguy: or even a thousand | 15:25 |
| mriedem | think i had to start the consoleauth service for some reason | 15:26 |
| *** rtheis has joined #openstack-nova | 15:26 | |
| sdague | mriedem: really? | 15:26 |
| sdague | that seems odd | 15:26 |
| mriedem | the integrated test base does that | 15:26 |
| mriedem | and i was seeing errors in the console service | 15:26 |
| mriedem | messaging timeout (because it wasn't running i guess) | 15:26 |
| sdague | mriedem: sure, but what were you triggering that would need that | 15:26 |
| sdague | that's the thing I'm confused by | 15:26 |
| *** rtheis has quit IRC | 15:26 | |
| mriedem | shrug | 15:26 |
| *** rtheis has joined #openstack-nova | 15:27 | |
| alaski | dansmith: I don't think it's going to be a problem because I agree that it's very unlikely someone will have a large number of aggregates. I think we mention it in a releasenote/doc and move on | 15:28 |
| dansmith | roger | 15:28 |
| johnthetubaguy | alaski: that sounds good | 15:28 |
| *** zz_dimtruck is now known as dimtruck | 15:28 | |
| *** penick has joined #openstack-nova | 15:29 | |
| *** dims has quit IRC | 15:29 | |
| *** salv-orlando has quit IRC | 15:29 | |
| *** gszasz has joined #openstack-nova | 15:29 | |
| dansmith | alaski: bauzas: so a reno in this? | 15:29 |
| dansmith | patch | 15:29 |
| bauzas | dansmith: alaski: yeah hence my +2 | 15:30 |
| bauzas | dansmith: alaski: I don't want to block, just commenting out that we probably would need to document that | 15:30 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: WIP: add functional regression test for bug 1548980 https://review.openstack.org/283837 | 15:30 |
| openstack | bug 1548980 in OpenStack Compute (nova) "nova list --deleted as admin fails with 404" [High,In progress] https://launchpad.net/bugs/1548980 - Assigned to Matt Riedemann (mriedem) | 15:30 |
| mriedem | sdague: ^ that passes now | 15:30 |
| *** dims has joined #openstack-nova | 15:30 | |
| alaski | dansmith: bauzas yep, reno seems appropriate | 15:30 |
| dansmith | bauzas: what reno category? | 15:30 |
| bauzas | dansmith: alaski: like saying in a reno 'upgrade' : if you have a lot of aggregates, the initial scheduler restart will take a bit of time | 15:31 |
| bauzas | dansmith: I think 'upgrades' | 15:31 |
| dansmith | okay | 15:31 |
| bauzas | it's not an issue | 15:31 |
| bauzas | neither a feature | 15:31 |
| *** annasort has quit IRC | 15:31 | |
| sdague | mriedem: did it explode on the delete? | 15:32 |
| bauzas | dansmith: what's cool with that is that you won't need to add a nova-manage | 15:32 |
| dansmith | yeah | 15:32 |
| mriedem | sdague: yeah | 15:32 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add functional regression test for list deleted instances on v2.16 https://review.openstack.org/283837 | 15:32 |
| mriedem | that's what it was failing on | 15:32 |
| bauzas | dansmith: just restarting the scheduler will update all the aggregates, so there will be no need to really have some backwards compat | 15:32 |
| *** atuvenie__ has quit IRC | 15:32 | |
| bauzas | dansmith: and I don't except to see a lof of time for that, because it's just an update for a single column | 15:33 |
| sdague | mriedem: hmmmm... curious | 15:33 |
| mriedem | sdague: https://review.openstack.org/#/c/283837/2..4/nova/tests/functional/regressions/test_bug_1548980.py | 15:33 |
| openstackgerrit | Dan Smith proposed openstack/nova: Allocate uuids for aggregates as they are created or loaded https://review.openstack.org/282520 | 15:34 |
| mriedem | i might have had the api versions screwed up, idk, but only thing i can think of was starting the consoleauth service, | 15:34 |
| dansmith | alaski: bauzas ^ | 15:34 |
| sdague | mriedem: ok, I want to poke at that to understand a bit more once I get this other thing cleaned up | 15:34 |
| *** erno has quit IRC | 15:34 | |
| *** rtheis has quit IRC | 15:34 | |
| *** burt has quit IRC | 15:34 | |
| mriedem | i'll comment that out and see if it starts failing again | 15:34 |
| sdague | just so I can figure out if consoleauth should be in the generic list for an OpenStackCloud fixture pattern | 15:34 |
| *** dims has quit IRC | 15:34 | |
| *** krtaylor has quit IRC | 15:34 | |
| sdague | if we need it for deletes, it seems like yes | 15:35 |
| *** burt has joined #openstack-nova | 15:35 | |
| *** rtheis has joined #openstack-nova | 15:35 | |
| *** burt has quit IRC | 15:35 | |
| *** rhagarty_ has joined #openstack-nova | 15:36 | |
| mriedem | sdague: yup http://paste.openstack.org/show/488037/ | 15:36 |
| mriedem | you have to have the consoleauth service for deletes | 15:36 |
| sdague | ok, good to know. Do we know why? Are we issuing a cleanup for consoles? | 15:37 |
| mriedem | _clean_instance_console_tokens | 15:37 |
| mriedem | yes | 15:37 |
| sdague | oh, yep | 15:37 |
| mriedem | File "nova/compute/manager.py", line 5435, in _clean_instance_console_tokens | 15:37 |
| *** rhagarty has quit IRC | 15:37 | |
| mriedem | honestly i'm tempted to just extend the _IntegratedTestBase which already does all of this setup for me | 15:38 |
| sdague | ok, I'm satisfied with that answer | 15:38 |
| sdague | mriedem: no no no. | 15:38 |
| mriedem | i knew you'd say that | 15:39 |
| sdague | I'm going to build some fixtures here | 15:39 |
| *** yamahata has joined #openstack-nova | 15:39 | |
| sdague | it will get prettier | 15:39 |
| mriedem | SeansFunctionalTestRegressionFixtureFTW | 15:39 |
| sdague | the subclass craziness has to end, because it's so hard to unwind | 15:39 |
| cdent | bless you sdague | 15:40 |
| scottda | ildikov: I've started an etherpad for cinder-nova api changes: https://etherpad.openstack.org/p/cinder-nova-api-changes | 15:40 |
| cdent | inheritance is evil | 15:40 |
| cdent | sometimes | 15:40 |
| ildikov | scottda: cool, thanks much | 15:40 |
| *** efoley__ has quit IRC | 15:40 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: make the opaque network attachment more robust https://review.openstack.org/284121 | 15:40 |
| sdague | cdent: in tests, most of the time | 15:41 |
| mriedem | cdent: yeah that whole DRY idea is crazy.... | 15:41 |
| cdent | mriedem: composition not inheritance | 15:42 |
| cdent | and, actually, DRY in tests is pretty bogus | 15:42 |
| cdent | because the last thing you want tin a test is magic action at a distance | 15:42 |
| mriedem | i guess...until you're like me and banging your head against why your test is failing, but the functional test base already has that handled | 15:43 |
| mriedem | anyway, ignore me | 15:43 |
| *** burt has joined #openstack-nova | 15:43 | |
| dansmith | except when you have a toooon of tests that statically encode the dict representation of some structure, | 15:43 |
| dansmith | which requires a thousand fixes when you tweak that slightly | 15:43 |
| dansmith | which encourages people to not change it to make it better, or do things in the real code so the tests won't fail | 15:43 |
| dansmith | like anything, it's a balance | 15:44 |
| sdague | agree | 15:44 |
| *** nkrinner has quit IRC | 15:44 | |
| sdague | our balance is way too far on the subclass side now | 15:44 |
| dansmith | in some places, for sure | 15:44 |
| shaohe_feng_ | sdague: jaypipes : this is the libvirt doc: https://libvirt.org/html/libvirt-libvirt-domain.html#virDomainJobInfo . | 15:44 |
| shaohe_feng_ | sdague: jaypipes : it is said: Data is measured in bytes unless otherwise specified and is measuring the job as a whole. For VIR_DOMAIN_JOB_UNBOUNDED, dataTotal may be less than the final sum of dataProcessed + dataRemaining in the event that the hypervisor has to repeat some data, such as due to dirtied pages during migration. For VIR_DOMAIN_JOB_BOUNDED, dataTotal shall always equal the sum of dataProcessed + dataRema | 15:45 |
| mriedem | alaski: re: cells v2 + neutron, | 15:45 |
| mriedem | alaski: will neutron api be global for all cells? or will the networks be per-cell? | 15:45 |
| shaohe_feng_ | sdague: jaypipes : will update the patch. thanks for review. | 15:45 |
| sdague | shaohe_feng_: thanks | 15:45 |
| sdague | mriedem: how about I let you fix the test_server_groups tests that are all massively coupled, then we can discusss :) | 15:46 |
| *** pushkaru has joined #openstack-nova | 15:46 | |
| kashyap | mdbooth: When you get a moment, I learnt from Infra folks that which change is causing that "can't merge" errors from Gerrit. Added a comment here with a (non-null) pointer - https://review.openstack.org/#/c/283702/1 | 15:46 |
| alaski | mriedem: neutron should be global for all cells, but networks per cell could be handled with the routed networking effort that Carl brought up in Bristol | 15:46 |
| mriedem | armax: ^ | 15:47 |
| *** krtaylor has joined #openstack-nova | 15:47 | |
| pkoniszewski | jaypipes: hi, still there? | 15:48 |
| mriedem | alaski: was just thinking about the case that you request a specific network, | 15:48 |
| cdent | sdague: I know you've been on this recently, what do I need to create in my local nova test environment to get opportunistc mysql tests happening? | 15:48 |
| mdbooth | kashyap: Thanks! | 15:48 |
| sdague | cdent: as mysql admin user.... | 15:48 |
| mriedem | belmoreira: how many IP address per cell at CERN? | 15:48 |
| mriedem | *addresses | 15:49 |
| mriedem | rough idea | 15:49 |
| mriedem | how many compute nodes | 15:49 |
| alaski | mriedem: it's dependent on works in progress, but the idea is that network segments will be tracked as resource pools in Nova so when a network is passed in Nova can determine which pool to schedule to | 15:49 |
| sdague | create user 'openstack_citest'@'localhost' identified by 'openstack_citest'; grant all on *.* to 'openstack_citest'@'localhost'; | 15:49 |
| openstackgerrit | ShaoHe Feng proposed openstack/nova: Add new APIs and deprecate old API for migrations https://review.openstack.org/258771 | 15:49 |
| mriedem | alaski: ok, i assumed that was the answer | 15:49 |
| mriedem | sdague: i'm just arguing to argue | 15:50 |
| *** efoley__ has joined #openstack-nova | 15:51 | |
| openstackgerrit | Swaminathan Vasudevan proposed openstack/nova: Implement setup_networks_on_host for Neutron networks https://review.openstack.org/275073 | 15:51 |
| belmoreira | mriedem number of nodes per cell is ~200 | 15:51 |
| *** andrearosa_web has joined #openstack-nova | 15:51 | |
| mriedem | dougwig: ^ | 15:51 |
| *** haplo37 has joined #openstack-nova | 15:52 | |
| belmoreira | mriedem number of IPs varies a lot depending in the cell use case | 15:52 |
| *** Sukhdev has joined #openstack-nova | 15:52 | |
| shaohe_feng_ | sdague: jaypipes :done. https://review.openstack.org/258771 waiting for gate test. thanks. | 15:52 |
| belmoreira | mriedem however we have ~5 segments per cell | 15:53 |
| cdent | thanks sdague, I am suitably empowered now | 15:53 |
| openstackgerrit | Sean M. Collins proposed openstack/os-vif: [WIP] Add mtu attribute to Network object https://review.openstack.org/284209 | 15:53 |
| doffm | belmoreira: Thats alot of total segments. | 15:53 |
| *** mdrabe has quit IRC | 15:53 | |
| sdague | cdent: that should go in the docs somewhere... just haven't gotten around to that yet | 15:54 |
| kylek3h | dansmith: I had some questions for you regarding migrate_data.py...specifically the LiveMigrateData objects. | 15:54 |
| belmoreira | doffm yes... a segment per ~40 nodes | 15:54 |
| dansmith | kylek3h: okay | 15:54 |
| shaohe_feng_ | sdague: what is WG? API WG | 15:55 |
| kylek3h | dansmith: With the changeset that introduced sending LiveMigrateData objects around between the services we're no longer able to do live migration for the PowerVM compute driver. | 15:55 |
| *** jaosorior has quit IRC | 15:55 | |
| kylek3h | dansmith: For PowerVM we need to pass more data between the hosts than we can fit in the libvirt or xen objects. | 15:55 |
| dansmith | kylek3h: that's good, because those objects aren't for you :) | 15:55 |
| sc68cal | jaypipes: https://review.openstack.org/284209 - first baby steps | 15:55 |
| kylek3h | dansmith: I've looked and I don't see a way to properly register an object for PowerVM through the driver since conductor needs it. | 15:55 |
| kylek3h | dansmith: Anything I'm missing? | 15:56 |
| sdague | shaohe_feng_: ?? | 15:56 |
| kylek3h | dansmith: For PowerVM it seems we need to propose a change set for a specific PowerVMLiveMigrateData object. | 15:56 |
| dansmith | kylek3h: yeah, drivers don't register objects | 15:56 |
| *** matrohon has quit IRC | 15:56 | |
| dansmith | kylek3h: yeah, but we're not going to take that into master for an out of tree driver | 15:56 |
| danpb | kylek3h: you'd have to install the python module that defines the objects in your conductor nodes | 15:56 |
| kylek3h | dansmith: What would you advise? | 15:56 |
| jaypipes | sc68cal: coolio. | 15:57 |
| danpb | ie, you'd need a separate nova-powervm-objects python dist | 15:57 |
| *** sridharg has joined #openstack-nova | 15:57 | |
| danpb | and your nova-powervm-driver python dist would depend on that | 15:57 |
| *** daemontool has quit IRC | 15:57 | |
| dansmith | kylek3h: something like danpb says, but it's going to have to require changes to core code to import that | 15:57 |
| dansmith | so they get registerd | 15:57 |
| danpb | that way you can install the object definitions on conductor + compute nodes | 15:57 |
| kylek3h | dansmith: right...is there a mechanism for that today? | 15:57 |
| dansmith | kylek3h: no | 15:58 |
| shaohe_feng_ | sdague: I have seen this comments on "From the API WG guidance" , do not know what stand for. | 15:58 |
| *** penick has quit IRC | 15:58 | |
| dansmith | kylek3h: because we don't support out of tree stuff | 15:58 |
| *** baoli has quit IRC | 15:58 | |
| kylek3h | dansmith: understand but we're working hard to get in tree, but facing these types of hurdles. | 15:58 |
| *** baoli has joined #openstack-nova | 15:59 | |
| sdague | shaohe_feng_: right api working group | 15:59 |
| shaohe_feng_ | sdague: got it. thank you. | 15:59 |
| sdague | shaohe_feng_: https://wiki.openstack.org/wiki/API_Working_Group | 15:59 |
| danpb | kylek3h: well to be honest you don't need to have 100% coverage of all apis to submit it in tree | 15:59 |
| danpb | kylek3h: in fact when we took ironic in tree, we made them split it up so their initial driver contained just minimal featurs | 16:00 |
| pkoniszewski | sdague: can you have a quick look at https://review.openstack.org/#/c/284068/3 ? Im trying to fix force_complete API according to the API guideline | 16:00 |
| danpb | IOW you don't need to solve the live migration objects issue out of tree - its not a blocker for submitting & merging your driver in tree | 16:00 |
| kylek3h | danpb: okay, understood. | 16:00 |
| danpb | live migration can be added later once its is in tree | 16:00 |
| *** haomaiwang has quit IRC | 16:01 | |
| dansmith | presumably they just want to fix the regression | 16:01 |
| dansmith | which isn't going to be easy | 16:01 |
| *** haomaiwang has joined #openstack-nova | 16:01 | |
| kylek3h | danpb: well, we're kind of going backwards since we had live migration work well... | 16:01 |
| *** zenoway has quit IRC | 16:01 | |
| *** erlon has joined #openstack-nova | 16:02 | |
| pkoniszewski | jaypipes: should we include units (bytes) in python-novaclient too? Im talking about this change https://review.openstack.org/#/c/281335/ | 16:02 |
| *** daemontool has joined #openstack-nova | 16:02 | |
| danpb | right but you're out of tree, so we don't particularly care about regressions in out of tree, so unfortunately that's life | 16:02 |
| dansmith | this ^ | 16:02 |
| kylek3h | dansmith: understood. I have a prototype that works .... it's about 20 LOC in migrate_data.py :-) | 16:02 |
| danpb | if you want to avoid this, focus should be on doing everything possible & neccessary to get in tree asap | 16:02 |
| jaypipes | pkoniszewski: yes, sir, thank you :) | 16:03 |
| *** dims has joined #openstack-nova | 16:03 | |
| pkoniszewski | jaypipes: thanks! | 16:03 |
| *** lyanchih has quit IRC | 16:03 | |
| dansmith | so actually | 16:03 |
| dansmith | kylek3h: conductor never sees this object, right? | 16:03 |
| dansmith | or maybe it does because of the pre call | 16:03 |
| *** tomoiaga has quit IRC | 16:03 | |
| kylek3h | it sends it so it needs the definition | 16:03 |
| kylek3h | right | 16:03 |
| danpb | kylek3h: as a partially sucky hack, you could create your own daemon startup program ie instead of using /usr/bin/nova-conductor, create a /usr/bin/nova-conductor-powervm | 16:04 |
| mriedem | danpb: actually, | 16:04 |
| mriedem | conductor manager is pluggable | 16:04 |
| kylek3h | dansmith: I tried putting it in the driver itself and some of the orchestration worked but eventually conductor blew up | 16:04 |
| dansmith | kylek3h: yep | 16:04 |
| openstackgerrit | Dmitry Stepanenko proposed openstack/nova: Go away from using PciDeviceStats consume_requests https://review.openstack.org/243131 | 16:04 |
| openstackgerrit | Dmitry Stepanenko proposed openstack/nova: Deprecate ComputeNode pci_device_pools field https://review.openstack.org/247692 | 16:04 |
| kylek3h | mriedem: oh.... | 16:04 |
| openstackgerrit | Dmitry Stepanenko proposed openstack/nova: Add PciDevicePoolList.get_all method https://review.openstack.org/250757 | 16:04 |
| openstackgerrit | Dmitry Stepanenko proposed openstack/nova: Added support_requests method to PciDevicePoolList https://review.openstack.org/242087 | 16:04 |
| *** efoley__ has quit IRC | 16:04 | |
| mriedem | CONF.conductor.manager | 16:04 |
| dansmith | mriedem: we should remove that | 16:04 |
| danpb | mriedem: oh, so they can subclass that and register their object that way | 16:04 |
| mriedem | danpb: yeah maybe | 16:05 |
| danpb | we should totally delete that config parameter | 16:05 |
| kylek3h | ok, let me look at that. It may be the piece I was missing! | 16:05 |
| sdague | dansmith: push a deprecation patch then | 16:05 |
| dansmith | sdague: yeah, I should | 16:05 |
| openstackgerrit | Sean Dague proposed openstack/nova: make microversion a client level construct for tests https://review.openstack.org/284220 | 16:05 |
| johnthetubaguy | dansmith: there are a few of those we need to deprecate... dang it | 16:05 |
| sdague | I'll +2 the heck out of a deprecation patch for that | 16:05 |
| * danpb wonders how many config params we've got left which let people load arbitrary code still | 16:05 | |
| dansmith | johnthetubaguy: yeah | 16:05 |
| danpb | i know there's the firewall_driver config parameter we should kill | 16:05 |
| johnthetubaguy | danpb: way more than you would hope | 16:05 |
| mriedem | compute and network manager are also pluggable | 16:06 |
| mriedem | lots of the managers | 16:06 |
| mriedem | if not all | 16:06 |
| mriedem | identity backend, all the things | 16:06 |
| johnthetubaguy | yeah, that one is a move to stevdore things, some are just pain kill them | 16:06 |
| *** efoley__ has joined #openstack-nova | 16:06 | |
| sdague | mriedem: https://review.openstack.org/284220 will make your patch simpler | 16:06 |
| dansmith | yeah, mostly kill | 16:06 |
| johnthetubaguy | dansmith: +1 | 16:06 |
| mriedem | btw, | 16:06 |
| mriedem | are there specific goals for the powervm driver to hit that are stated somewhere? | 16:07 |
| mriedem | i know (1) ci and (2) users | 16:07 |
| johnthetubaguy | hmm, probably only in an either mad | 16:07 |
| johnthetubaguy | oops, etherpad | 16:07 |
| johnthetubaguy | the other was diverse contributor base, would be ideal | 16:08 |
| kylek3h | mriedem: I know thorst has been working on that. | 16:08 |
| *** browne has joined #openstack-nova | 16:08 | |
| mriedem | it's just kind of harsh to say, sorry we broke you but that's life so suck it up, work on getting in tree, but oh btw we do'nt have anything stating what you have to do to get in tree | 16:08 |
| *** salv-orlando has joined #openstack-nova | 16:08 | |
| *** mdrabe has joined #openstack-nova | 16:08 | |
| mriedem | johnthetubaguy: the hyperv, vmware and xenapi drivers arguably don't have a diverse contributor base | 16:09 |
| mriedem | libvirt does b/c it's the reference | 16:09 |
| dansmith | mriedem: by that logic, we'd never be able to do anything that broke something out of tree that doesn't belong in nova | 16:09 |
| mriedem | dansmith: i'm not saying we can't | 16:09 |
| dansmith | mriedem: I don't think libvirt being reference has much to do with the contributor base | 16:09 |
| dansmith | I think it has to do with the user base | 16:09 |
| danpb | yeah, IMHO diverse contributor base is a non-requirement | 16:10 |
| *** mjura has quit IRC | 16:10 | |
| danpb | what we actually need is a /responsiveness/ contributor base committed to ongoing work | 16:10 |
| mriedem | i'm not saying libvirt is the reference b/c of it's contributor base | 16:10 |
| sdague | danpb: agree, I'm not sure where that popped up in conversation. xenserver driver definitely isn't | 16:10 |
| mriedem | i agree it's the user base b/c it's free | 16:10 |
| *** yamahata has quit IRC | 16:10 | |
| johnthetubaguy | that was in the summit, maybe my memory playing tricks with me | 16:10 |
| johnthetubaguy | I think it was diverse user base, more than anything | 16:11 |
| johnthetubaguy | well, really just a user base | 16:11 |
| mriedem | anyway, i will shut up again | 16:11 |
| sdague | it seemed like the conversation was: ci, because we expect that for everyone | 16:11 |
| danpb | IMHO requirements for a new driver should be pretty much CI testing + committed group of maintainers + real world users | 16:11 |
| sdague | users, because we totally go powned by that in the past | 16:11 |
| danpb | ie, we don't want to accept a toy academic project with no real users | 16:12 |
| johnthetubaguy | we have lxd on its way as well, I think, so we need to get this straight somewhere | 16:12 |
| dansmith | johnthetubaguy: how about "has a mission in line with nova's mission" ? | 16:12 |
| dansmith | that would be an easy way to answer the lxd crowd, IMHO | 16:12 |
| johnthetubaguy | danpb: yeah, I thats a bit part of it, those toy projects are best not being upstream | 16:12 |
| sdague | danpb: well, and also in the case we had a driver for this hypevisor before that burned a ton of review time, and then ended up with "oh forgot to tell you all, no one has ever run this code outside our test lab" | 16:13 |
| johnthetubaguy | sdague: yeah, thats where the higher bar comes from | 16:13 |
| danpb | yeah, that's why i said real world users, as opposed to toy research project users | 16:13 |
| johnthetubaguy | dansmith: true, thats got to be in there | 16:13 |
| *** Kennan has quit IRC | 16:13 | |
| mriedem | well, lxd vs lxc? | 16:13 |
| sdague | danpb: sure, basically in agreement | 16:13 |
| mriedem | ci for lxc is busted | 16:13 |
| dansmith | mriedem: lxc is basically free | 16:13 |
| mriedem | it's in tree b/c it's convenient | 16:13 |
| mriedem | yeah, that | 16:14 |
| danpb | mriedem: LXD is basically more akin to Docker these days | 16:14 |
| dansmith | so it gets a pass, IMHO, but I'd be the same way if we were pressed to add it today | 16:14 |
| *** diogogmt has joined #openstack-nova | 16:14 | |
| *** annasort has joined #openstack-nova | 16:14 | |
| *** Kennan has joined #openstack-nova | 16:14 | |
| mriedem | danpb: ok | 16:14 |
| *** mgoddard__ has quit IRC | 16:14 | |
| sdague | I think we've told a lot of people in the past that if you show up as a libvirt driver, you get a pass | 16:14 |
| danpb | it tries todo the whole container lifecycle & image mgmt thing now | 16:14 |
| *** mgoddard has joined #openstack-nova | 16:14 | |
| sdague | like the power kvm and z kvm folks | 16:14 |
| dansmith | aye | 16:14 |
| dansmith | because that addresses most of the problem of adding a new driver | 16:15 |
| johnthetubaguy | danpb: oh, I missed that happening, good to know | 16:15 |
| johnthetubaguy | sdague: yeah, +1 on the libvirt thing, well, its a lower bar at least | 16:15 |
| *** annasort_ has joined #openstack-nova | 16:15 | |
| cdent | jaypipes: so the sticking point on this migration 318 is the use of server_default=text('0'). When the migration is compared with the model the two 'text' objects are different and it blows up. We need to 'text' otherwise 0 is added to the DDL as a string. I can avoid all this just by not using server_default at all, and just sticking with just default. What will that break if I do? | 16:15 |
| johnthetubaguy | well, parallels is probably the recent example | 16:15 |
| sdague | we explicitly tell most groups 'go be a libvirt driver, you'll largely get a pass' | 16:15 |
| *** apearson has joined #openstack-nova | 16:15 | |
| danpb | for libvirt we'd still at minimum like to see a committed set of maintainers for the port | 16:15 |
| openstackgerrit | Sean M. Collins proposed openstack/os-vif: [WIP] Add mtu attribute to Network object https://review.openstack.org/284209 | 16:15 |
| sdague | danpb: agree | 16:16 |
| danpb | CI testing would be pretty desirable too | 16:16 |
| jaypipes | cdent: just ditch the server_default. let's just move on. | 16:16 |
| cdent | jaypipes: or I can go the long way round and change the sync test, but.... blech | 16:16 |
| cdent | word | 16:16 |
| cdent | thanks | 16:16 |
| danpb | since although the nova code should be largely unchanged, this doesn't mean the hypervisor supports the same features as existing libvirt hypervisors, so we like see where we stand on feature parity | 16:16 |
| danpb | in general a libvirt driver has very little ongoing dev burden, compared to a completely new nova driver though | 16:17 |
| openstackgerrit | Chris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools https://review.openstack.org/281837 | 16:17 |
| cdent | jaypipes: I think that ^ gets it then. bauzas, alaski, dansmith, if you please, at your leisure, etc | 16:18 |
| *** gszasz has quit IRC | 16:18 | |
| jaypipes | cdent: will do as soon as I see the gate check succeed. | 16:19 |
| *** annasort has quit IRC | 16:19 | |
| *** annasort_ is now known as annasort | 16:19 | |
| *** dedery_ has joined #openstack-nova | 16:19 | |
| cdent | :some slack emoji in jaypipes' direction: | 16:19 |
| cdent | bah, pep8 | 16:20 |
| *** zenoway has joined #openstack-nova | 16:21 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools https://review.openstack.org/281837 | 16:22 |
| *** liverpooler has quit IRC | 16:22 | |
| *** dedery has quit IRC | 16:22 | |
| *** ekuris has quit IRC | 16:24 | |
| *** dedery_ has quit IRC | 16:24 | |
| openstackgerrit | Dmitry Stepanenko proposed openstack/nova: Switch HostState to use pci device pools https://review.openstack.org/258417 | 16:25 |
| thorst | mriedmen johnthetubaguy danpb dansmith sdague: Sorry for the delay, was afk. I agree with the stated goals (contributor base, user base, CI). But at the same time one of the stated things in Tokyo was that nova doesn’t want out-of-tree drivers doing ‘weird’ things. I just want to make sure that this won’t be considered ‘weird’ later on. We just don’t want to do the wrong thing here... | 16:25 |
| danpb | thorst: you mean the migration object hack ? | 16:25 |
| dansmith | thorst: you mean hacking things up to get your object registered? | 16:25 |
| thorst | yeah - a separate objects project. | 16:26 |
| dansmith | you don't need a separate project | 16:26 |
| openstackgerrit | Dmitry Stepanenko proposed openstack/nova: Change HostManager to load pci_dev_pools from db https://review.openstack.org/260023 | 16:26 |
| danpb | that's not something you'd submit when you merged it in-tree, so its not a issue | 16:26 |
| *** apearson has quit IRC | 16:27 | |
| thorst | danpb: OK. That's the main thing. We'd just add that to the existing object when merged in-tree | 16:27 |
| *** apearson has joined #openstack-nova | 16:27 | |
| dansmith | of course | 16:28 |
| thorst | cool...I was just worried when the review came it would be a 'what is this?' moment. But sounds like its well understood and won't cause concern later | 16:28 |
| *** belmoreira has quit IRC | 16:31 | |
| *** harlowja_at_home has joined #openstack-nova | 16:31 | |
| *** bjornar__ has joined #openstack-nova | 16:32 | |
| *** gszasz has joined #openstack-nova | 16:32 | |
| PaulMurray | hi mriedem - have you seen responses to your comments here: https://review.openstack.org/#/c/278454/11 | 16:32 |
| PaulMurray | mriedem, that patch is blocking several live migration ones | 16:32 |
| *** ihrachys has quit IRC | 16:33 | |
| PaulMurray | mriedem, don't know if andreykurilin has had a chance to talk to you about it - he said he would | 16:33 |
| mriedem | PaulMurray: looking | 16:33 |
| mriedem | he pinged earlier | 16:33 |
| PaulMurray | mriedem, thanks - good - sorry for being eager - got three patches queued by others | 16:35 |
| mriedem | sheesh | 16:35 |
| mriedem | replied | 16:35 |
| *** dstepanenko_ has quit IRC | 16:36 | |
| *** dstepanenko has joined #openstack-nova | 16:37 | |
| *** gszasz has quit IRC | 16:38 | |
| *** atuvenie__ has joined #openstack-nova | 16:38 | |
| *** fawadkhaliq has joined #openstack-nova | 16:38 | |
| *** ljxiash has quit IRC | 16:39 | |
| Bobba | Vek: Regarding https://review.openstack.org/#/c/283962/, what do you mean by a separate set of tests? Do you mean independent of nova/tests/unit? | 16:40 |
| *** 32NAAC6RR has quit IRC | 16:42 | |
| sdague | cdent: I see we both o_O at the same time on that | 16:42 |
| cdent | yeah | 16:42 |
| cdent | um | 16:42 |
| cdent | wat | 16:42 |
| *** ducttape_ has quit IRC | 16:42 | |
| *** ducttape_ has joined #openstack-nova | 16:43 | |
| *** gyee has joined #openstack-nova | 16:43 | |
| cdent | sdague: I for reasons not entirely clear to me took it a bit personal...curse of REST and all that. feh! | 16:43 |
| sdague | heh | 16:44 |
| *** apoorvad has joined #openstack-nova | 16:44 | |
| sdague | I mostly want to know why people keep saying it | 16:44 |
| sdague | like I feel like there is some post somewhere people read that makes them think it's true | 16:44 |
| *** mgoddard_ has joined #openstack-nova | 16:44 | |
| *** ducttape_ has quit IRC | 16:45 | |
| *** jistr has quit IRC | 16:45 | |
| *** ducttape_ has joined #openstack-nova | 16:45 | |
| *** oomichi has joined #openstack-nova | 16:46 | |
| *** baoli_ has joined #openstack-nova | 16:46 | |
| *** vilobhmm11 has joined #openstack-nova | 16:46 | |
| *** apearson has quit IRC | 16:47 | |
| auggy | sdague: thanks for posting that, i honestly thought maybe i was just missing something | 16:47 |
| * auggy usually blames myself | 16:47 | |
| *** mgoddard has quit IRC | 16:47 | |
| *** moshele has quit IRC | 16:48 | |
| *** baoli has quit IRC | 16:48 | |
| sdague | mriedem: hmm... this is interesting | 16:52 |
| sdague | in trying to fix the server_groups tests | 16:52 |
| sdague | evacuate > 2.11 fails | 16:52 |
| mriedem | sdague: http://docs.openstack.org/developer/nova/api_microversion_history.html#id12 | 16:54 |
| mriedem | ? | 16:54 |
| *** zenoway has quit IRC | 16:54 | |
| sdague | mriedem: yep, that's is | 16:54 |
| sdague | but, the way it explodes is wonky | 16:54 |
| *** zenoway has joined #openstack-nova | 16:55 | |
| sdague | oh, wait, maybe the stack trace is all fail | 16:55 |
| sdague | mriedem: http://paste.openstack.org/show/488048/ | 16:55 |
| sdague | is that the thing you were chasing | 16:55 |
| claudiub|2 | hello folks. If you have a minute, can anyone take a look at a bugfix? It's 13 lines long and it already has a +2. https://review.openstack.org/#/c/258614/ | 16:55 |
| openstackgerrit | Sean Dague proposed openstack/nova: WIP: update test_server_groups to client microversion https://review.openstack.org/284257 | 16:56 |
| sheel | cdent: hi there | 16:56 |
| mriedem | sdague: i was seeing that too yeah | 16:56 |
| *** Guest51435 is now known as mgagne | 16:56 | |
| *** mgagne has quit IRC | 16:56 | |
| *** mgagne has joined #openstack-nova | 16:56 | |
| mriedem | sdague: kylek3h might have a patch for htat | 16:57 |
| mriedem | *that | 16:57 |
| *** Sukhdev has quit IRC | 16:57 | |
| mriedem | the cancel events NoneType thing | 16:57 |
| sheel | cdent: its regarding https://review.openstack.org/#/c/282862/ ..... review comments are incorporated... | 16:57 |
| sheel | cdent: kindly go through it.. | 16:57 |
| bauzas | johnthetubaguy: you know what ? I'm getting mad trying to play with test_conductor for asserting the calls that we discussed this morning | 16:57 |
| *** lbragstad has quit IRC | 16:58 | |
| cdent | thanks sheel, will do soon | 16:58 |
| sheel | cdent: thank you.. | 16:58 |
| *** zenoway has quit IRC | 16:58 | |
| sdague | mriedem: is that posted anywhere? | 16:59 |
| *** dims has quit IRC | 17:00 | |
| mriedem | sdague: oh i was thinking of this https://review.openstack.org/#/c/281553/ | 17:00 |
| *** amotoki has quit IRC | 17:00 | |
| *** haomaiwang has quit IRC | 17:01 | |
| *** haomaiwang has joined #openstack-nova | 17:01 | |
| sdague | right, that seems different | 17:01 |
| *** lbragstad has joined #openstack-nova | 17:02 | |
| *** thorst has quit IRC | 17:02 | |
| johnthetubaguy | bauzas: oh, thats no fun, it sounded like it should be an easy one, in theory | 17:02 |
| bauzas | johnthetubaguy: the real problem is that test_conductor is messy | 17:03 |
| bauzas | johnthetubaguy: all the checks are only for the manager | 17:03 |
| *** thorst has joined #openstack-nova | 17:03 | |
| *** thorst has quit IRC | 17:03 | |
| bauzas | johnthetubaguy: but I think I'm about to provide a good catch | 17:03 |
| *** atuvenie__ has quit IRC | 17:03 | |
| bauzas | just telling you FYI | 17:03 |
| *** dims has joined #openstack-nova | 17:03 | |
| sdague | mriedem: I wonder if we now have an instance where cancel_events gets called twice | 17:04 |
| johnthetubaguy | bauzas: cool | 17:04 |
| bauzas | johnthetubaguy: tbc, we need to be very sure about the UTs we accept for test_conductor | 17:04 |
| mriedem | sdague: was just going to say that | 17:04 |
| bauzas | because the coverage is very small | 17:04 |
| bauzas | (for the RPC versioning I mean) | 17:04 |
| sdague | mriedem: it seems like a simple safety check | 17:05 |
| mriedem | do the tests kill the compute service twice? | 17:05 |
| sdague | but I'm going to try to get through this other patch first | 17:05 |
| sdague | maybe.... | 17:05 |
| *** yamahata has joined #openstack-nova | 17:06 | |
| *** belmoreira has joined #openstack-nova | 17:06 | |
| mriedem | sdague: i see it | 17:07 |
| mriedem | the base class uses ServiceFixture | 17:07 |
| mriedem | which has self.addCleanup(self.service.kill) | 17:07 |
| mriedem | so the functional tests don't need to expliclity call it | 17:07 |
| *** apearson has joined #openstack-nova | 17:07 | |
| bauzas | last fun stuff, when using the ServiceFixture, it seems that assert_called_with doesn't work with a mock.ANY arg :) | 17:08 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add functional regression test for list deleted instances on v2.16 https://review.openstack.org/283837 | 17:09 |
| bauzas | it literally tells me that my expected mock.ANY is different from a dict :) | 17:09 |
| bauzas | meh | 17:09 |
| sdague | bauzas: review? | 17:11 |
| bauzas | not yet | 17:11 |
| bauzas | locally | 17:11 |
| bauzas | http://paste.openstack.org/show/488054/ | 17:11 |
| *** pbacterio has quit IRC | 17:12 | |
| bauzas | with http://paste.openstack.org/show/488057/ | 17:12 |
| *** zenoway has joined #openstack-nova | 17:14 | |
| sdague | interesting | 17:14 |
| *** achanda has joined #openstack-nova | 17:15 | |
| *** achanda has quit IRC | 17:15 | |
| *** sheel has quit IRC | 17:15 | |
| *** achanda has joined #openstack-nova | 17:15 | |
| *** sfinucan has quit IRC | 17:16 | |
| *** sheel has joined #openstack-nova | 17:16 | |
| *** omnipresent has joined #openstack-nova | 17:17 | |
| bauzas | sdague: I'll workaround that by mocking the image api, but see, it's a bit weirdo | 17:19 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Cleanup service.kill calls in functional tests https://review.openstack.org/284268 | 17:19 |
| mriedem | sdague: cleans up those redundant service.kill calls ^ | 17:19 |
| *** jlanoux has quit IRC | 17:20 | |
| mriedem | wondering if i should add a check in the cancel_all_events method too | 17:20 |
| *** zenoway has quit IRC | 17:20 | |
| *** mvk has quit IRC | 17:22 | |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: libvirt: delete the last file link in _supports_direct_io() https://review.openstack.org/224764 | 17:23 |
| *** edmondsw has quit IRC | 17:23 | |
| sdague | mriedem: cool, the check in cancel_all_events wouldn't hurt | 17:23 |
| *** mpavone has quit IRC | 17:25 | |
| *** gyee has quit IRC | 17:25 | |
| mriedem | ok i'll do that | 17:25 |
| *** ihrachys has joined #openstack-nova | 17:26 | |
| *** zenoway has joined #openstack-nova | 17:27 | |
| *** ihrachys has quit IRC | 17:29 | |
| *** omnipresent has quit IRC | 17:30 | |
| auggy | sdague: so i was having trouble duplicating the behavior in the mysql cli because i wasn't including the delete column in my insert, but when i did, it triggered the duplicate key error at the cli | 17:31 |
| auggy | because i can read unique constraints!! | 17:31 |
| sdague | ah, nice | 17:32 |
| *** rtheis has quit IRC | 17:32 | |
| *** rtheis has joined #openstack-nova | 17:33 | |
| *** josecastroleon1 has quit IRC | 17:34 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: remove deprecation warnings from oslo_versionedobjects https://review.openstack.org/284286 | 17:35 |
| auggy | the other odd thing, why is deleted NULL as default? i mean logically, if you're inserting wouldn't deleted be 0?? | 17:35 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Cleanup service.kill calls in functional tests https://review.openstack.org/284268 | 17:36 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Gracefully handle cancelling all events more than once https://review.openstack.org/284287 | 17:36 |
| *** sudipto has quit IRC | 17:36 | |
| *** sahid has quit IRC | 17:37 | |
| *** zenoway has quit IRC | 17:37 | |
| andreykurilin | mriedem: just replied to your comment:) If you still disagree, I'll post new patch-set | 17:37 |
| *** zenoway has joined #openstack-nova | 17:38 | |
| *** rtheis has quit IRC | 17:39 | |
| *** mrkz has quit IRC | 17:40 | |
| sdague | auggy: I forget, possibly an oversight | 17:40 |
| auggy | it sets it to an id on delete | 17:40 |
| *** mrkz has joined #openstack-nova | 17:41 | |
| openstackgerrit | Andrea Rosa proposed openstack/python-novaclient: Support for abort an ongoing live migration https://review.openstack.org/284296 | 17:41 |
| mriedem | andreykurilin: so i guess i'll pull down https://review.openstack.org/#/c/284186/ first and then https://review.openstack.org/#/c/278454/ and see if the help message thing is fixed | 17:41 |
| auggy | it's ok, i was just curious, i assumed it was a binary field | 17:41 |
| andreykurilin | mriedem: thanks | 17:43 |
| mriedem | andreykurilin: can you write a test for the help message fix? | 17:44 |
| *** nic has joined #openstack-nova | 17:45 | |
| *** sridharg has quit IRC | 17:46 | |
| *** ociuhandu has quit IRC | 17:46 | |
| *** zenoway has quit IRC | 17:47 | |
| *** edmondsw has joined #openstack-nova | 17:47 | |
| *** zenoway has joined #openstack-nova | 17:48 | |
| *** moshele has joined #openstack-nova | 17:48 | |
| andreykurilin | mriedem: I didn't write it yet, since I wanted to work on more shell improvements(hint message about --os-compute-api-version after command description), but don't finished it yet. | 17:49 |
| johnthetubaguy | sdague: jaypipes: for this live-migrate status, just thinking this negative test is a bit risky, feels like we should be picking out the known good states here? https://review.openstack.org/#/c/258771/29/nova/db/sqlalchemy/api.py | 17:49 |
| mriedem | andreykurilin: it doesn't seem to really work like i expected | 17:50 |
| andreykurilin | mriedem: will write it in 10 minutes | 17:50 |
| *** rtheis has joined #openstack-nova | 17:50 | |
| mriedem | andreykurilin: i pulled both changes down and now the help i get is this: http://paste.openstack.org/show/488066/ | 17:50 |
| sdague | johnthetubaguy: I'm fine either way. I do see your concern about the lack of formal states and what that does later | 17:51 |
| *** lucasagomes is now known as lucas-hungry | 17:51 | |
| sdague | I also think that could be fixed later | 17:51 |
| sdague | as it doesn't really change the semantics of the API | 17:51 |
| andreykurilin | mriedem: It works like I expected(print the proper argument for chosen version) | 17:51 |
| *** knikolla has joined #openstack-nova | 17:52 | |
| mriedem | andreykurilin: for chosen version.... | 17:52 |
| mriedem | how is the version chosen? | 17:52 |
| andreykurilin | --os-compute-api-version | 17:52 |
| mriedem | i tried using --os-compute-api-version 2.20 with the help but it failed | 17:52 |
| mriedem | ohhhhh | 17:52 |
| mriedem | http://paste.openstack.org/show/488067/ | 17:53 |
| mriedem | i had the option in the wrong place | 17:53 |
| andreykurilin | :) | 17:54 |
| *** rdopiera has quit IRC | 17:54 | |
| mriedem | andreykurilin: ok, so i'm good with both if we can add a test for the shell fix, and order the changes such that we land the shell fix first | 17:55 |
| *** Marga_ has quit IRC | 17:55 | |
| andreykurilin | mriedem: ok. I'll write(and fix) unit tests for patch for help message and rebase v2.21 on it | 17:56 |
| *** rtheis has quit IRC | 17:56 | |
| *** rtheis has joined #openstack-nova | 17:56 | |
| PaulMurray | mriedem, andreykurilin - hi - oh, you're already sorting it out. I was about to talk about that | 17:57 |
| andreykurilin | :) | 17:57 |
| *** moshele has quit IRC | 17:57 | |
| *** zenoway has quit IRC | 17:57 | |
| *** jdillaman has joined #openstack-nova | 17:58 | |
| mriedem | the task master | 17:58 |
| *** zenoway has joined #openstack-nova | 17:59 | |
| mriedem | PaulMurray: are you a certified PMP? | 17:59 |
| openstackgerrit | Kevin Benton proposed openstack/nova: Use MTU value from Neutron in OVS/LB VIF wiring https://review.openstack.org/283847 | 17:59 |
| *** scheuran has quit IRC | 17:59 | |
| *** claudiub|2 has quit IRC | 18:00 | |
| *** haomaiwang has quit IRC | 18:01 | |
| *** rtheis has quit IRC | 18:01 | |
| *** haomaiwang has joined #openstack-nova | 18:01 | |
| *** derekh has quit IRC | 18:01 | |
| *** eil397 has joined #openstack-nova | 18:02 | |
| *** mgoddard_ has quit IRC | 18:02 | |
| *** ociuhandu has joined #openstack-nova | 18:02 | |
| *** mgoddard has joined #openstack-nova | 18:03 | |
| mriedem | kevinbenton: heh, looking at your mtu patch | 18:05 |
| mriedem | https://github.com/openstack/nova/commit/44670135be0640622bf6ab74552aa4a9efd21b0c | 18:05 |
| mriedem | back in juno we deprecated the config option | 18:05 |
| mriedem | we should probably actually use the deprecated_for_removal=True in the option so it warns when it's used | 18:05 |
| mriedem | could be a separate change that builds on the neutron change you have | 18:05 |
| jaypipes | johnthetubaguy: yeah, that's a very good point. | 18:06 |
| dansmith | ansiwen: we probably just need to pre-load flavor data when we pull it for caching | 18:06 |
| dansmith | ansiwen: (expected_attrs=['flavor']) | 18:06 |
| dansmith | ansiwen: I think the flavor info is probably always going to be hit from metadata in almost every case, so just always pre-querying it is probably best | 18:06 |
| *** porunov has quit IRC | 18:07 | |
| ansiwen | dansmith: I was doing this with a dummy get_flavor(), and indeed it fixes the problem | 18:07 |
| dansmith | yeah, so just add it to expected_attrs on your initial query and it'll be even better.. pulled in the first go | 18:07 |
| dansmith | although I thought I already did that for this metadata one.. have you checked in master? | 18:08 |
| ansiwen | so this is really the right way to fix it? shouldn't lazy loads trigger an update of the cache? | 18:08 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add name to ResourceProvider object https://review.openstack.org/281945 | 18:09 |
| dansmith | ansiwen: I think flavor should probably always be queried | 18:10 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: [microversions] Enable 2.21 https://review.openstack.org/278454 | 18:11 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: [microversions] fix help msg for versioned args https://review.openstack.org/284186 | 18:11 |
| dansmith | ansiwen: the problem with updating the cache on lazy load would be that you're refreshing the image, but without fresh data, right? so you'd have to wait longer for the cache to flush so you got some other instance attribute just because you happened to touch flavor, right? | 18:11 |
| ansiwen | dansmith: I didn't check master. I'm investigating this on a kilo installation. | 18:11 |
| dansmith | ansiwen: 3a761270581d1ac61a3b4669c130d211f1ad5a17 | 18:11 |
| andreykurilin | mriedem: I added unit tests and rebased a change on it | 18:12 |
| *** thorst has joined #openstack-nova | 18:13 | |
| openstackgerrit | Sean Dague proposed openstack/nova: complete the removal of api_version from rest client parameters https://review.openstack.org/284257 | 18:13 |
| *** dedery has joined #openstack-nova | 18:14 | |
| *** andreykurilin__ has joined #openstack-nova | 18:14 | |
| cdent | jaypipes, dansmith, bauzas, alaski: gate finally says go: https://review.openstack.org/#/c/281837/ | 18:14 |
| openstackgerrit | Sean Dague proposed openstack/nova: add a place for functional test to block specific regressions https://review.openstack.org/282012 | 18:14 |
| ansiwen | dansmith: not sure what you mean with "fresh data". isn't the current instance object authoritative? | 18:15 |
| dansmith | ansiwen: 1. Query metadata, cache instance | 18:16 |
| dansmith | ansiwen: 2. Update display_name from the API externally | 18:16 |
| dansmith | ansiwen: 3. Lazy-load flavor, which renews the instance TTL | 18:16 |
| *** zenoway has quit IRC | 18:16 | |
| dansmith | ansiwen: 4. Wait longer to notice the display_name change from the metadata client side | 18:16 |
| *** ctrath has quit IRC | 18:17 | |
| *** zenoway has joined #openstack-nova | 18:17 | |
| *** dedery_ has joined #openstack-nova | 18:18 | |
| ansiwen | dansmith: so, when lazy-load of flavor triggers a cache update in 3, why isn't the display_name updated as well? | 18:19 |
| *** dedery has quit IRC | 18:19 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Deprecate conductor manager option https://review.openstack.org/284332 | 18:20 |
| dansmith | ansiwen: because that's not how we do lazy-loads | 18:20 |
| sdague | doffm: so you realized you -1ed a patch because there wasn't a completely developer reference for a model in which we've landed 0 patches | 18:20 |
| ansiwen | dansmith: IIUC the memcached cache object is just an deep object dump of meta_data object, right? | 18:21 |
| dansmith | ansiwen: we could, but we'd need to be super careful not to overwrite any dirty fields, and potentially we'd be changing the consistent state of the instance for the base attributes | 18:21 |
| *** zenoway has quit IRC | 18:21 | |
| sdague | mriedem: I think we've got about 8 patches in flight that build upon - https://review.openstack.org/#/c/282012/ | 18:21 |
| dansmith | ansiwen: I don't know how it gets serialized into memcache, TBH.. a deepcopy isn't really the right thing to do, TBH, but that may be what it's doing | 18:21 |
| sdague | so we should start landing this stuff, or abandon it all and rage quit :) | 18:22 |
| ansiwen | dansmith: I think I'm missing some important information about the architecture | 18:22 |
| *** belmoreira has quit IRC | 18:23 | |
| ansiwen | dansmith: it's just my observation. if getting the metadata object from the cache "overwrites" meta_data.instance.flavor state (because it fetches the flavor again and again), it seems to be a deep copy | 18:24 |
| dansmith | ansiwen: I was talking about lazy loads in the instance object, not anything metadata specific | 18:25 |
| dansmith | ansiwen: also, I don't know what meta_data.instance.flavor is | 18:25 |
| *** jwcroppe has quit IRC | 18:26 | |
| dansmith | jaypipes: did you do that db patch we talked about yesterday yet? | 18:26 |
| cdent | jaypipes: yes, that double alter really frustrated me | 18:27 |
| ansiwen | dansmith: the metadata object has an instance attribute, which is a nova object, which again has a flavor attribute | 18:28 |
| jaypipes | dansmith: currently working on it, sir. | 18:28 |
| PaulMurray | johnthetubaguy, just saw your comment on https://review.openstack.org/#/c/258771 - I wondered if that really matters now - the fucntion is the same as the existing one above | 18:28 |
| dansmith | jaypipes: okay | 18:28 |
| kevinbenton | mriedem: yeah, i laughed when i looked at the commit for that and saw it was in juno-eol | 18:28 |
| PaulMurray | johnthetubaguy, oops - I'm about to turn into a pumpkin if I don't run....... | 18:28 |
| * PaulMurray runs | 18:29 | |
| dansmith | ansiwen: what is "the metadata object" ? | 18:29 |
| *** Jeffrey4l has quit IRC | 18:29 | |
| *** lpetrut has quit IRC | 18:29 | |
| ansiwen | dansmith: class InstanceMetadata | 18:29 |
| dansmith | oh, that python object, I see | 18:29 |
| ansiwen | from nova/api/metadata/base.py | 18:30 |
| ansiwen | dansmith: that's exactly the object that get's cached in memcached | 18:30 |
| dansmith | ansiwen: yeah, I don't see where that caching happens | 18:31 |
| dansmith | because you have to serialize it in some way, and just json()ing it or something wouldn't retain the lazy-loading behavior | 18:31 |
| ansiwen | dansmith: https://github.com/openstack/nova/blob/master/nova/api/metadata/handler.py#L94 | 18:31 |
| dansmith | ansiwen: that's caching the metadata result, right? | 18:32 |
| dansmith | none of the internals, just the response we would have sent to the client, I thought | 18:32 |
| ansiwen | dansmith: exactly | 18:32 |
| *** PaulMurray has quit IRC | 18:32 | |
| dansmith | I dunno what you're asking then :) | 18:33 |
| ansiwen | dansmith: well, it uses the memcache library, which stores the whole object transparantly. so, I suppose it MUST do a deep copy, else you would have dangling references | 18:33 |
| dansmith | dims: is that right? are we like pickling things into memcache? | 18:33 |
| ansiwen | it just stores a python object. and when I resurrect it, the embedded instance object has an "old" state | 18:34 |
| * dims catches up | 18:34 | |
| *** Marga_ has joined #openstack-nova | 18:34 | |
| dansmith | ansiwen: it can't just "store a python object" -- it has to serialize it in some way | 18:34 |
| dansmith | and using something like pickle to do that would be ... terrible | 18:34 |
| dansmith | and non-portable | 18:34 |
| dansmith | and a security risk | 18:34 |
| ansiwen | dansmith: of course it has to | 18:35 |
| *** zhangjn has quit IRC | 18:35 | |
| *** zhangjn has joined #openstack-nova | 18:36 | |
| *** READ10 has joined #openstack-nova | 18:36 | |
| ansiwen | dansmith: it's just my observation. that's also why I was wondering, why you would store a mashalled object in memcached, since it is not useful for any other nova-api process | 18:37 |
| ansiwen | dansmith: but there is still a high likelyhood, that I got it completely wrong | 18:37 |
| ansiwen | dansmith: I have to run and move home. I'll be back later. I'm very interested in understanding this. | 18:38 |
| mriedem | andreykurilin: +W on https://review.openstack.org/#/c/284186/ | 18:39 |
| *** josecastroleon has joined #openstack-nova | 18:39 | |
| dansmith | ansiwen: okay well, clearly I don't have the full picture of the backend of this, so not sure how much more I can help, but interested in not pickling if that's what we're doing :) | 18:39 |
| dims | dansmith : what we store in memcached has not changed with the move to oslo.cache, it was what it was before :) | 18:40 |
| dansmith | dims: that's fine, I just figure you'd know | 18:41 |
| *** liverpooler has joined #openstack-nova | 18:41 | |
| dims | snikitin : pkholkin_ : do you know exactly if pickling is what happens under the covers? | 18:41 |
| *** irenab_ has joined #openstack-nova | 18:42 | |
| *** irenab has quit IRC | 18:44 | |
| *** irenab_ is now known as irenab | 18:44 | |
| dims | dansmith : ansiwen : i believe we have not added support for other formats (Encoding/Decoding data into another format - http://dogpilecache.readthedocs.org/en/latest/usage.html) and the default is still pickling | 18:44 |
| *** mylu has joined #openstack-nova | 18:44 | |
| dansmith | dims: doesn't that seem kinda scary to you? | 18:45 |
| dansmith | like, I'm surprised that it even works for nova-objects and leaves them working enough to lazy-load things | 18:45 |
| *** chhavi has quit IRC | 18:45 | |
| dansmith | scary in the security sense | 18:45 |
| *** ijw has joined #openstack-nova | 18:46 | |
| mriedem | doffm: i think you can drop your -1 on this https://review.openstack.org/#/c/278454/ | 18:46 |
| mriedem | andreykurilin: you might be missing a test in https://review.openstack.org/#/c/278454/ | 18:46 |
| *** irenab_ has joined #openstack-nova | 18:48 | |
| dims | dansmith : yep it does | 18:48 |
| *** danpb has quit IRC | 18:48 | |
| *** irenab has quit IRC | 18:50 | |
| *** irenab_ is now known as irenab | 18:50 | |
| *** Marga_ has quit IRC | 18:50 | |
| *** ijw has quit IRC | 18:50 | |
| *** ljxiash has joined #openstack-nova | 18:51 | |
| *** paul-carlton1 has quit IRC | 18:51 | |
| *** Marga_ has joined #openstack-nova | 18:51 | |
| ansiwen | dansmith: that was exactly my thought, how can that work? But I concluded that I just don't understand it. | 18:51 |
| doffm | mriedem: Yeahhhh. | 18:51 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add name to ResourceProvider object https://review.openstack.org/281945 | 18:52 |
| *** zenoway has joined #openstack-nova | 18:52 | |
| dims | ansiwen : dansmith : threw in a but, but did not make it scary - https://bugs.launchpad.net/oslo.cache/+bug/1549438 | 18:53 |
| openstack | Launchpad bug 1549438 in oslo.cache "Support option to specify Encoding/Decoding data" [Undecided,New] | 18:53 |
| *** ctrath has joined #openstack-nova | 18:53 | |
| *** rtheis has joined #openstack-nova | 18:53 | |
| dansmith | dims: okay | 18:53 |
| *** efoley__ has quit IRC | 18:54 | |
| ansiwen | So, do you agree, that it makes no sense to use memcached in the first pace, because it can't be shared anyway? | 18:54 |
| *** daemontool has quit IRC | 18:55 | |
| *** ljxiash has quit IRC | 18:55 | |
| *** zenoway has quit IRC | 18:57 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: [microversions] Enable 2.21 https://review.openstack.org/278454 | 18:57 |
| andreykurilin__ | mriedem: fixed:) | 18:58 |
| ansiwen | dims: I don't think that encoding would help here. The problem is, we try to cache a whole object tree here | 18:58 |
| dims | ansiwen : then we should fix what we store/retrieve :) | 18:58 |
| ansiwen | dims: exactly | 18:59 |
| cdent | dansmith, dims: dogpiles interface of encouraging the caching of methods is bad news as it makes people think they can just magically cache objects. Much better to explicitly serialize to at least a native dict | 18:59 |
| *** ijw has joined #openstack-nova | 18:59 | |
| cdent | bbl | 18:59 |
| *** cdent has quit IRC | 18:59 | |
| dansmith | agree | 19:00 |
| dansmith | agree | 19:00 |
| dansmith | oops :) | 19:00 |
| dansmith | but .. really agree :) | 19:00 |
| sdague | mriedem: ok, I'm going to step away from the keyboard for a bit. Review on the stack of test patches would be nice. Once those actually get in I can dump context and start working on the metadata server testing stack. | 19:00 |
| dims | +1 to do better :) | 19:01 |
| *** haomaiwang has quit IRC | 19:01 | |
| *** 16WAAEDST has joined #openstack-nova | 19:01 | |
| ansiwen | I think a first improvement would be to always use the internal cache in this case | 19:02 |
| mriedem | andreykurilin__: +2 | 19:03 |
| mriedem | now you have to summon Vek | 19:03 |
| *** ijw has quit IRC | 19:04 | |
| andreykurilin__ | mriedem: wow! thanks:) I'm happy | 19:04 |
| andreykurilin__ | Vek: hi! I need your good mark for https://review.openstack.org/278454 :) | 19:05 |
| *** ijw has joined #openstack-nova | 19:05 | |
| openstackgerrit | Sean M. Collins proposed openstack/os-vif: Fix calls to create_ovs_vif_port https://review.openstack.org/284353 | 19:06 |
| openstackgerrit | Sean M. Collins proposed openstack/os-vif: [WIP] Add mtu attribute to Network object https://review.openstack.org/284209 | 19:07 |
| *** lpetrut has joined #openstack-nova | 19:08 | |
| *** josecastroleon has quit IRC | 19:09 | |
| mriedem | melwitt: i think you're probably super happy with this now https://review.openstack.org/#/c/282012/ | 19:09 |
| *** Sukhdev has joined #openstack-nova | 19:10 | |
| *** browne has quit IRC | 19:11 | |
| mriedem | sdague: if you had pep8 passing that would be easier :) | 19:11 |
| mriedem | oh hi-o! | 19:11 |
| dims | mriedem : i keep getting bitten by pep8 too :) | 19:12 |
| *** jwcroppe has joined #openstack-nova | 19:13 | |
| *** jbernard_ is now known as jbernard | 19:13 | |
| *** zenoway has joined #openstack-nova | 19:14 | |
| openstackgerrit | Kevin Benton proposed openstack/nova: Use MTU value from Neutron in OVS/LB VIF wiring https://review.openstack.org/283847 | 19:14 |
| *** rodrigods has quit IRC | 19:16 | |
| *** rodrigods has joined #openstack-nova | 19:16 | |
| *** zenoway has quit IRC | 19:16 | |
| *** zenoway has joined #openstack-nova | 19:17 | |
| *** READ10 has quit IRC | 19:19 | |
| *** belmoreira has joined #openstack-nova | 19:20 | |
| openstackgerrit | Sean M. Collins proposed openstack/os-vif: [WIP] Add mtu attribute to Network object https://review.openstack.org/284209 | 19:21 |
| *** irenab_ has joined #openstack-nova | 19:23 | |
| *** irenab has quit IRC | 19:25 | |
| *** jwcroppe has quit IRC | 19:25 | |
| *** irenab has joined #openstack-nova | 19:26 | |
| *** irenab_ has quit IRC | 19:28 | |
| *** claudiub|2 has joined #openstack-nova | 19:28 | |
| *** eil397 has quit IRC | 19:29 | |
| *** ctrath has quit IRC | 19:30 | |
| doffm | andreykurilin__: Yeah thanks for that patch. I suppose I should have done that after getting the instance actions server stuff in. | 19:33 |
| *** browne has joined #openstack-nova | 19:37 | |
| *** ducttape_ has quit IRC | 19:38 | |
| *** irenab has quit IRC | 19:39 | |
| *** fawadkhaliq has quit IRC | 19:40 | |
| *** irenab has joined #openstack-nova | 19:41 | |
| *** yamahata has quit IRC | 19:41 | |
| *** rtheis has quit IRC | 19:42 | |
| *** cfcp1 has joined #openstack-nova | 19:42 | |
| *** rtheis has joined #openstack-nova | 19:43 | |
| *** mylu has quit IRC | 19:44 | |
| *** mylu has joined #openstack-nova | 19:44 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Test to validate obj_make_compatible() strips old attributes https://review.openstack.org/280795 | 19:45 |
| dansmith | can we please land that ASAP? It's rebase-bait :( | 19:45 |
| dansmith | alaski: jaypipes: mriedem ^ | 19:45 |
| andreykurilin__ | doffm: my pleasure. I'm happy to do novaclient better :) | 19:45 |
| *** mylu has quit IRC | 19:46 | |
| *** rtheis has quit IRC | 19:47 | |
| *** mylu has joined #openstack-nova | 19:48 | |
| *** omnipresent has joined #openstack-nova | 19:49 | |
| *** ducttape_ has joined #openstack-nova | 19:49 | |
| sdague | mriedem: doh! | 19:49 |
| mriedem | sdague: on the base change, i had one thing about the version in the header | 19:50 |
| mriedem | checking for that | 19:50 |
| mriedem | then i'm +2 | 19:50 |
| openstackgerrit | Sean Dague proposed openstack/nova: complete the removal of api_version from rest client parameters https://review.openstack.org/284257 | 19:50 |
| *** rtheis has joined #openstack-nova | 19:51 | |
| oomichi | sdague: hi, about https://review.openstack.org/#/c/282012/ | 19:52 |
| *** zenoway has quit IRC | 19:53 | |
| oomichi | what is a difference from Tempest tests? is that more lightwieght tests? | 19:53 |
| jaypipes | sdague: may I get your help please? http://paste.openstack.org/show/488087/ been a couple months since I ran tox on nova locally and yet again I'm bombing on some pip versioning thing :( | 19:53 |
| *** bklei has quit IRC | 19:54 | |
| mriedem | dansmith: so to test this, i add or remove an attribute on an object and if i didn't add a compat check it should fail, right | 19:54 |
| sdague | mriedem: so, it feels like that level of extra checking on headers to ensure that no one can use it wrong is about 2 orders of magnitude more locked down than any other convenience interface we have in the test stack | 19:55 |
| sdague | as such, it seems overkill now, and could have unintended fallout / unwind later | 19:55 |
| rlrossit | dansmith: I was hoping we wouldn't have to add another static dict that we have to keep updating like the fingerprint dict... | 19:55 |
| *** rtheis has quit IRC | 19:56 | |
| sdague | oomichi: for anything that doesn't require an actual kvm guest, testing in the functional tree is at least 10 times if not 100 or 1000 times more efficent | 19:56 |
| *** rtheis has joined #openstack-nova | 19:56 | |
| sdague | jaypipes: looking | 19:56 |
| mriedem | sdague: i thought you wanted things done in a very specific way? | 19:56 |
| mriedem | because part of the problem is random craziness in the tests | 19:57 |
| sdague | mriedem: can we not solve this in parts? | 19:57 |
| mriedem | blerg, sec | 19:57 |
| mriedem | i have to finish up dan's review first | 19:58 |
| sdague | mriedem: I'm feeling like the answer to cleaning up parts of this is that unless there is a fully formed perfect architecture for 20,000 existing tests we can't do anything | 19:58 |
| sdague | jaypipes: virtualenv --version? | 19:58 |
| *** irenab_ has joined #openstack-nova | 19:58 | |
| *** irenab has quit IRC | 19:59 | |
| *** irenab_ is now known as irenab | 19:59 | |
| oomichi | sdague: I see, thanks | 19:59 |
| *** omnipresent has quit IRC | 20:00 | |
| edleafe | jaypipes: yeah, you need to update virtualenv | 20:00 |
| edleafe | jaypipes: a few people got burned by that recently | 20:00 |
| *** zenoway has joined #openstack-nova | 20:00 | |
| sdague | oomichi: for instance, there is no way we could land a tempest test for these kinds of regressions. We'd end up with 4 hour tempest run. | 20:00 |
| *** 16WAAEDST has quit IRC | 20:01 | |
| *** penick has joined #openstack-nova | 20:01 | |
| sdague | many of them run in 2 or 3 seconds in the functional tree, and work on any laptop | 20:01 |
| *** rtheis has quit IRC | 20:01 | |
| *** haomaiwang has joined #openstack-nova | 20:01 | |
| jaypipes | sdague, edleafe: should I install virtualenv in a particular way? via pip? | 20:01 |
| sdague | jaypipes: via pip | 20:02 |
| jaypipes | k | 20:02 |
| *** omnipresent has joined #openstack-nova | 20:02 | |
| sdague | however, I found that 14 is dicey on ubuntu | 20:02 |
| dansmith | rlrossit: well, my magic 8 ball can't tell me when a thing was added | 20:02 |
| sdague | pip install -U "virtualenv<14" | 20:02 |
| sdague | is probably the most reliable | 20:02 |
| *** irenab has quit IRC | 20:02 | |
| jaypipes | ffs,,, | 20:03 |
| jaypipes | sdague: installed virtualenv 13.1.2. same error after clearing .tox directory. tried virtualenv 14.0.2. same error. :( | 20:03 |
| sdague | jaypipes: https://github.com/pypa/virtualenv/issues/869 | 20:03 |
| sdague | jaypipes: hmmm.... | 20:03 |
| sdague | tox --version? | 20:03 |
| jaypipes | lemme check that | 20:04 |
| jaypipes | jaypipes@uberbox:~/repos/nova$ tox --version | 20:04 |
| jaypipes | 2.1.1 imported from /usr/lib/python3/dist-packages/tox/__init__.py | 20:04 |
| oomichi | sdague: yeah, I got the merit so well. | 20:04 |
| *** ociuhandu has quit IRC | 20:04 | |
| sdague | jaypipes: python3 you say? | 20:04 |
| *** irenab has joined #openstack-nova | 20:04 | |
| *** zenoway has quit IRC | 20:05 | |
| jaypipes | sdague: yes... is there something bad about that? | 20:05 |
| *** dedery_ has quit IRC | 20:05 | |
| sdague | well, what virtualenv / pip is python 3 using | 20:06 |
| sdague | because if you are executing tox as python3 | 20:06 |
| sdague | jaypipes@uberbox:~/repos/nova$ pip --version | 20:06 |
| sdague | pip 8.0.2 from /usr/local/lib/python2.7/dist-packages (python 2.7) | 20:06 |
| sdague | isn't the pip you are using | 20:06 |
| jaypipes | le sigh. | 20:07 |
| *** zenoway has joined #openstack-nova | 20:07 | |
| jaypipes | python packaging is becoming almost as bad as ruby. | 20:07 |
| mriedem | dansmith: should dropping attributes be caught by that test, or only adding? | 20:07 |
| mriedem | b/c https://gist.github.com/mriedem/c10fb017ee10c7cd2d33 wasn't a problem | 20:07 |
| dansmith | well, you can't drop things | 20:08 |
| dansmith | without a major bump | 20:08 |
| dansmith | so it's looking for the opposite | 20:08 |
| jaypipes | sdague: so, if I run tox -epy27, it creates the virtualenv but then attempts to do: /home/jaypipes/repos/nova/.tox/py27/bin/pip -chttps... | 20:08 |
| mriedem | ok, | 20:09 |
| mriedem | testing adding a thing w/o compat | 20:09 |
| sdague | jaypipes: /home/jaypipes/repos/nova/.tox/py27/bin/pip --version | 20:09 |
| jaypipes | sdague: | 20:09 |
| jaypipes | jaypipes@uberbox:~/repos/nova$ /home/jaypipes/repos/nova/.tox/py27/bin/pip --version | 20:09 |
| jaypipes | pip 1.5.6 from /home/jaypipes/repos/nova/.tox/py27/local/lib/python2.7/site-packages (python 2.7) | 20:09 |
| sdague | yep so that's the problem | 20:09 |
| sdague | because tox is in your python 3 toolchain, and `which pip` is in your python 2 toolchain, pip install -U virutualenv only updated python2 | 20:10 |
| sdague | your python 3 virtualenv is ancient | 20:10 |
| *** omnipresent has quit IRC | 20:10 | |
| sdague | pip install -U tox probably fixes it | 20:10 |
| sdague | because that will stomp the system tox to python2 | 20:11 |
| mriedem | dansmith: https://gist.github.com/mriedem/c4f31e9849a260cb7a0d doesn't fail | 20:11 |
| jaypipes | sdague: http://paste.openstack.org/show/488091/ | 20:11 |
| *** omnipresent has joined #openstack-nova | 20:12 | |
| dansmith | mriedem: that's because instance isn't in the table | 20:12 |
| sdague | jaypipes: tox --version? | 20:12 |
| *** yassine__ has quit IRC | 20:12 | |
| mriedem | dansmith: ok, that's what i thought | 20:12 |
| jaypipes | sdague: jaypipes@uberbox:~/repos/nova$ tox --version | 20:12 |
| jaypipes | 2.1.1 imported from /usr/lib/python3/dist-packages/tox/__init__.py | 20:12 |
| sdague | do you have a pip3 ? | 20:13 |
| jaypipes | nope. or at least, which pip3 doesn't return anything. | 20:13 |
| mriedem | dansmith: fixed ips is in the table though and this doesn't fail either https://gist.github.com/mriedem/c6e869b8a91e1f98ea04 | 20:14 |
| sdague | jaypipes: sudo apt-get remove python-tox ? | 20:15 |
| dansmith | mriedem: I just got done adding all the fields that we've added over the last two weeks because this was failing, so it should be working, | 20:15 |
| dansmith | mriedem: but instance should also be in there | 20:15 |
| dansmith | so I dunno | 20:15 |
| dansmith | I feel like maybe I don't care enough to fix this | 20:15 |
| *** zenoway has quit IRC | 20:16 | |
| *** rhagarty_ has quit IRC | 20:16 | |
| *** rhagarty_ has joined #openstack-nova | 20:16 | |
| jaypipes | sdague: lol. wtf... http://paste.openstack.org/show/488093/ | 20:16 |
| sdague | jaypipes: honestly, I've tried very hard to not let python3 creep into my toolchain for doing stuff to openstack for reasons like this, because it kind of all falls appart | 20:16 |
| mriedem | dansmith: yeah... | 20:16 |
| *** zenoway has joined #openstack-nova | 20:17 | |
| sdague | jaypipes: dpkg -S /usr/lib/python3/dist-packages/tox/__init__.py | 20:17 |
| sdague | figure out where the damn thing is coming from | 20:17 |
| sdague | or, alternatively `which tox` | 20:17 |
| jaypipes | jaypipes@uberbox:~/repos/nova$ dpkg -S /usr/lib/python3/dist-packages/tox/__init__.py | 20:17 |
| jaypipes | tox: /usr/lib/python3/dist-packages/tox/__init__.py | 20:17 |
| sdague | because you just need a python2 tox, and then I think you'll be good | 20:18 |
| jaypipes | jaypipes@uberbox:~/repos/nova$ which tox | 20:18 |
| jaypipes | /usr/local/bin/tox | 20:18 |
| sdague | yeh, so apt-get remove that | 20:18 |
| sdague | and does the file look like it will run python2 | 20:18 |
| _gryf | jaypipes, I think that tox may just trying to use wrong version of interpreter to run the tests | 20:18 |
| sdague | _gryf: no, it's not getting that far | 20:19 |
| sdague | _gryf: tox will use the right version to run the tests once it can build environments, it's the bootstrapping thats the issue | 20:19 |
| jaypipes | sdague: ah.... so I needed to apt-get remove *both* python-tox and tox. | 20:19 |
| sdague | jaypipes: honestly, I would apt-get autoremove | 20:20 |
| sdague | there is a bunch of stuff in that list that might get you in trouble | 20:20 |
| sdague | python3-virtualenv tox virtualenv | 20:20 |
| *** omnipresent has quit IRC | 20:20 | |
| *** mylu has quit IRC | 20:20 | |
| _gryf | sdague, well, i've stepped on something opposite - running the py3 tests makes for some reason using py2 version of python to actually build the package | 20:20 |
| sdague | _gryf: in which circumstance? | 20:20 |
| _gryf | sdague, my pet project :) wel, the situation was as follows | 20:21 |
| *** cleong has joined #openstack-nova | 20:21 | |
| sdague | system python happens at the top level | 20:21 |
| _gryf | i have the py2 env as a default one | 20:21 |
| dansmith | mriedem: any chance your test is failing to run? | 20:21 |
| _gryf | and wrote py3 little project | 20:21 |
| dansmith | mriedem: I just noticed that the new ServiceStatusPayload object is broken :( | 20:21 |
| sdague | if the environment isn't one that tox implicitly understands, then you have to set the python version field | 20:21 |
| *** mylu has joined #openstack-nova | 20:22 | |
| _gryf | made the tox conf for building and running the tests | 20:22 |
| *** omnipresent has joined #openstack-nova | 20:22 | |
| sdague | _gryf: https://github.com/openstack/nova/blob/master/tox.ini#L29 | 20:22 |
| _gryf | where it turns out that setup.py, which was prepared for py3 was invoked with… py2 :/ | 20:22 |
| *** zenoway_ has joined #openstack-nova | 20:24 | |
| *** irenab_ has joined #openstack-nova | 20:24 | |
| *** irenab has quit IRC | 20:25 | |
| *** irenab_ is now known as irenab | 20:25 | |
| *** zenoway has quit IRC | 20:25 | |
| sdague | _gryf: what did you name your environment? | 20:25 |
| _gryf | sdague, does it matter in any way? | 20:26 |
| sdague | _gryf: yes | 20:26 |
| *** rk4n has quit IRC | 20:26 | |
| sdague | because tox has some very magical ones that it does the right thing for | 20:26 |
| sdague | and everything else, you are on your own | 20:27 |
| sdague | https://bitbucket.org/hpk42/tox/src/4dbfb1f1eb86b35cbd2c83461e1e294651e89cf6/tox/config.py?at=default&fileviewer=file-view-default#config.py-24 | 20:27 |
| *** Bjoern_ has joined #openstack-nova | 20:27 | |
| _gryf | sdague, http://paste.openstack.org/show/488096/ it was something like this | 20:27 |
| _gryf | nothing fancy | 20:27 |
| sdague | the envlist was actually py34, and not py3 | 20:28 |
| *** cdent has joined #openstack-nova | 20:28 | |
| jaypipes | sdague: k, I apt-get remove'd tox, python-tox, python3-tox, virtualenv, python-virtualenv, python3-virtualenv, then did this: http://paste.openstack.org/show/488097/ | 20:28 |
| jaypipes | sdague: amazing. | 20:28 |
| *** omnipresent has quit IRC | 20:28 | |
| _gryf | yet, because i'd like to have the package to be built, setup.py is invoked with py2 | 20:28 |
| _gryf | sdague, yep, that's what i had installed that day | 20:29 |
| *** zenoway_ has quit IRC | 20:29 | |
| sdague | _gryf: well if you show me the actual file I might be able to debug it for you | 20:29 |
| sdague | jaypipes: yeh, so just re source your .bashrc right? | 20:29 |
| *** zenoway has joined #openstack-nova | 20:30 | |
| jaypipes | sdague: gah. doh. /me weeps. | 20:30 |
| jaypipes | sdague: ok, running now. | 20:30 |
| *** irenab_ has joined #openstack-nova | 20:30 | |
| _gryf | sdague, i didn;t upload it anywehre right now, because of that weirdness | 20:30 |
| _gryf | sdague, but i will, and i'll appreciate your help :D | 20:31 |
| *** irenab has quit IRC | 20:32 | |
| *** irenab_ is now known as irenab | 20:32 | |
| jaypipes | sdague: thx much for your help, Sean | 20:32 |
| cdent | sdague, jaypipes was that bashrc src to get a 'hash -r' | 20:33 |
| *** omnipresent has joined #openstack-nova | 20:34 | |
| sdague | jaypipes: no problem. In return might I suggest - https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:functional_mv | 20:36 |
| *** zenoway_ has joined #openstack-nova | 20:37 | |
| *** zenoway has quit IRC | 20:37 | |
| *** omnipresent has quit IRC | 20:40 | |
| cfcp1 | hi, I'm wondering if someone might have a minute to give me some advice, I'm trying to setup an admin user so that I can start running VMs, but running into trouble on the keystone-install section, while installing python-keystoneclient | 20:43 |
| cfcp1 | would anyone be able to advise? | 20:43 |
| cfcp1 | the error im getting is: | 20:43 |
| cfcp1 | There are no enabled repos. | 20:43 |
| cfcp1 | Run "yum repolist all" to see the repos you have. | 20:43 |
| cfcp1 | You can enable repos with yum-config-manager --enable <repo> | 20:43 |
| *** rook-rook has quit IRC | 20:46 | |
| *** ihrachys has joined #openstack-nova | 20:47 | |
| *** Marga_ has quit IRC | 20:48 | |
| *** Sukhdev has quit IRC | 20:48 | |
| *** Marga_ has joined #openstack-nova | 20:48 | |
| *** boris-42 has joined #openstack-nova | 20:53 | |
| *** belmoreira has quit IRC | 20:53 | |
| *** jwcroppe has joined #openstack-nova | 20:54 | |
| mriedem | cfcp1: this is not the channel to ask that | 20:56 |
| mriedem | try #openstack | 20:56 |
| cfcp1 | alright thanks | 20:57 |
| mriedem | dansmith: which test of mine? | 20:57 |
| dansmith | mriedem: my test you were running | 20:57 |
| dansmith | but tbh, I'm done with it | 20:57 |
| auggy | jaypipes: i ended up just using pyenv and i created an openstack virtualenv, pyenv lets you do it by directory | 20:57 |
| dansmith | the instance bug is because I wasn't running 1.1->1.0 on anything | 20:57 |
| dansmith | and after doing that, all kinds of other stupid behaviors from other objects popped out | 20:57 |
| auggy | in perl we have this neat tool called perlbrew that lets you run different perls all over the place ;) | 20:57 |
| dansmith | and so I just don't care anymore | 20:57 |
| mriedem | dansmith: it ran {0} nova.tests.unit.objects.test_objects.TestObjectAttributes.test_obj_make_compatible [0.213682s] ... ok | 20:57 |
| *** cfcp1 has left #openstack-nova | 20:58 | |
| mriedem | dansmith: you have to care! | 20:58 |
| dansmith | mriedem: okay, well, when I run it in tox/testr, I get that but when I run with unittest, I get a can-never-be-true failure | 20:58 |
| *** rlrossit has quit IRC | 20:59 | |
| dansmith | and rlrossit hurt my feelings | 20:59 |
| mriedem | he'll do that | 20:59 |
| *** penick has quit IRC | 21:00 | |
| * melwitt cues Phil Collins song | 21:00 | |
| mriedem | which one? | 21:00 |
| *** haomaiwang has quit IRC | 21:01 | |
| dansmith | heh | 21:01 |
| melwitt | "I don't care anymore" | 21:01 |
| melwitt | of course | 21:01 |
| mriedem | oh | 21:01 |
| mriedem | i thought sisudio | 21:01 |
| *** haomaiwang has joined #openstack-nova | 21:01 | |
| auggy | or the land of confusion? | 21:01 |
| mriedem | *sussudio | 21:01 |
| *** achanda has quit IRC | 21:01 | |
| *** xyang1 has quit IRC | 21:02 | |
| * cdent realizes that mriedem is the dood in american psycho | 21:02 | |
| mriedem | loved the movie | 21:02 |
| mriedem | the book was rough | 21:02 |
| mriedem | it's about 1000x more graphic | 21:02 |
| cdent | we should exchange business cards | 21:03 |
| mriedem | mine is bone | 21:03 |
| * cdent lets dansmith follow up on that one | 21:03 | |
| *** aysyd has quit IRC | 21:03 | |
| *** ducttape_ has quit IRC | 21:04 | |
| *** ducttape_ has joined #openstack-nova | 21:04 | |
| auggy | i keep thinking of jim carrey as the lead dude in the american psycho movie, even though it was actually batman | 21:05 |
| *** andrearosa_web has quit IRC | 21:06 | |
| * cdent tries to imagine that | 21:06 | |
| *** rtheis has joined #openstack-nova | 21:06 | |
| mriedem | jim carey makes everything infinitely worse | 21:07 |
| *** Kennan has quit IRC | 21:08 | |
| *** Kennan has joined #openstack-nova | 21:08 | |
| *** annasort has quit IRC | 21:08 | |
| *** ctrath has joined #openstack-nova | 21:08 | |
| *** mylu has quit IRC | 21:09 | |
| mriedem | doffm: can you update https://review.openstack.org/#/c/201606/ after the cells meeting and then i'll +2 | 21:09 |
| *** mylu has joined #openstack-nova | 21:09 | |
| *** thorst is now known as thorst_afk | 21:10 | |
| *** pcaruana has quit IRC | 21:10 | |
| *** ihrachys has quit IRC | 21:17 | |
| *** mingyu has quit IRC | 21:19 | |
| *** rlrossit_ has joined #openstack-nova | 21:19 | |
| *** lpetrut has quit IRC | 21:19 | |
| *** mingyu has joined #openstack-nova | 21:19 | |
| mriedem | kevinbenton: tsk tsk https://review.openstack.org/#/c/283847/ | 21:21 |
| *** mylu has quit IRC | 21:24 | |
| *** smatzek__ has quit IRC | 21:24 | |
| *** salv-orl_ has joined #openstack-nova | 21:27 | |
| sdague | on the live migration front is self.driver.live_migration_abort(instance) async? | 21:27 |
| jaypipes | sdague: should be, yes. | 21:27 |
| sdague | so, sending start and end notifications seems weird | 21:27 |
| *** kylek3h is now known as kylek3h__away | 21:28 | |
| sdague | https://review.openstack.org/#/c/277971/15/nova/compute/manager.py | 21:28 |
| rlrossit_ | dansmith: maybe I'm too strung out, but from what I can see, your magic 8 ball has no premonition of history in the objects, so there's no way we can enforce removing fields right? | 21:28 |
| rlrossit_ | so if someone adds a field, we don't know it wasn't there before, so we can't assert there should be something in object_attributes for it | 21:28 |
| jaypipes | sdague: oh, sorry, I thought you were referring to whether the REST API call is a 202 or not. | 21:28 |
| jaypipes | sdague: it should be a 202... | 21:28 |
| rlrossit_ | if they do add it to the static dict, then we can start asserting on it | 21:29 |
| jaypipes | sdague: I wasn't referring to whether the driver call was an RPC cast. | 21:29 |
| *** salv-orlando has quit IRC | 21:29 | |
| sdague | jaypipes: we're in the compute manager, this isn't rpc | 21:29 |
| jaypipes | sdague: so those two notifications essentially wrap the dom.abortJob() call. I don't see a problem with that in order to have external systems understand how long the abort took? | 21:30 |
| sdague | self.driver.live_migration_abort(instance) -> dom.abortJob() | 21:30 |
| sdague | well, only if that's synchronous, right | 21:30 |
| sdague | if that's just a signal then the answer is 0 | 21:31 |
| sdague | that was the question | 21:31 |
| jaypipes | I see. | 21:31 |
| jaypipes | ok, sorry to interrupt you. I don't know the answer to that question. | 21:31 |
| sdague | oh, no, it looks like it's a sync call - https://libvirt.org/sources/java/javadoc/org/libvirt/Domain.html#abortJob() | 21:32 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add test for unshelve in the conductor API https://review.openstack.org/284420 | 21:33 |
| openstackgerrit | Chris Friesen proposed openstack/nova: check max_net_count against min_count when booting https://review.openstack.org/283861 | 21:33 |
| bauzas | https://review.openstack.org/284420 : the longest UT ever | 21:34 |
| *** ekarlso- has quit IRC | 21:34 | |
| *** ekarlso- has joined #openstack-nova | 21:34 | |
| jaypipes | bauzas: you know my opinion on the shelve API :) | 21:34 |
| cfriesen | ^ Minor fixup when booting multiple instances at once. | 21:35 |
| bauzas | jaypipes: heh, yeah :) | 21:35 |
| bauzas | jaypipes: but we have to support it :D | 21:35 |
| sdague | bauzas: can you actually sleep 0? | 21:35 |
| bauzas | sdague: not really | 21:35 |
| bauzas | sdague: I tried | 21:35 |
| sdague | how about 0.01 | 21:36 |
| bauzas | maybe, lemme try locally | 21:36 |
| bauzas | sdague: btw. I saw my problem about mock.ANY | 21:36 |
| sdague | adding real sleep delays into the tests ends up covering up other async issues | 21:36 |
| openstackgerrit | Brian Haley proposed openstack/nova: Implement setup_networks_on_host for Neutron networks https://review.openstack.org/275073 | 21:36 |
| sdague | bauzas: what was it? | 21:36 |
| bauzas | it wasn't due to mock.ANY but rather the instance object | 21:36 |
| sdague | ah | 21:36 |
| bauzas | which is dehydrated over the wirtre | 21:36 |
| bauzas | so the ref changes | 21:37 |
| sdague | ok, that makes me feel better about the universe | 21:37 |
| bauzas | and since 2 objects aren't __eq__ so... | 21:37 |
| jaypipes | sdague: omg test_server_groups makes me vomit. | 21:39 |
| dansmith | rlrossit_: it cat with a 1.0 element that removes everything, which I was trying to avoid for the mess | 21:40 |
| dansmith | rlrossit_: but I'm giving up on it, so just don't worry about it | 21:40 |
| dansmith | *can | 21:40 |
| rlrossit_ | dansmith: :( | 21:41 |
| openstackgerrit | Kevin Benton proposed openstack/nova: Use MTU value from Neutron in OVS/LB VIF wiring https://review.openstack.org/283847 | 21:41 |
| *** krtaylor has quit IRC | 21:41 | |
| *** mtanino__ has joined #openstack-nova | 21:41 | |
| sdague | jaypipes: yeh, well one thing at a time | 21:43 |
| *** achanda has joined #openstack-nova | 21:43 | |
| *** achanda has quit IRC | 21:45 | |
| bauzas | sdague: so, sleep(0.01) sucks, but sleep(.1) does | 21:45 |
| *** ndipanov has quit IRC | 21:45 | |
| jaypipes | sdague: yeah, I know... just slogging through the test code :) | 21:45 |
| bauzas | sdague: ideally, I would signal that the mock is ready, because either time makes me a bit worried | 21:47 |
| bauzas | it's racy a lot | 21:47 |
| *** cleong has quit IRC | 21:48 | |
| bauzas | sdague: hence me initially proposing 0.5 as a safeguard - to make sure that a slow node wouldn't -1 the job | 21:48 |
| bauzas | sdague: so, open to ideas for seeing how we could mock a ServiceFixture related method | 21:48 |
| sdague | bauzas: is this addressed if yuo do castascall? | 21:49 |
| bauzas | sdague: oh | 21:49 |
| *** _ducttape_ has joined #openstack-nova | 21:50 | |
| sdague | self.useFixture(cast_as_call.CastAsCall(self.stubs)) | 21:50 |
| sdague | it's used in all the full stack tests, as well as a bunch of conductor tests | 21:50 |
| bauzas | yeah | 21:50 |
| bauzas | I forgot that | 21:50 |
| bauzas | that would help for sure | 21:51 |
| bauzas | because it should wait for the return | 21:51 |
| sdague | yep | 21:51 |
| bauzas | great catch, thanks | 21:51 |
| sdague | no prob | 21:51 |
| bauzas | I was knowing the trick, but I forgot :( | 21:51 |
| bauzas | testing now | 21:51 |
| *** mylu has joined #openstack-nova | 21:53 | |
| bauzas | awesomeness | 21:53 |
| *** ducttape_ has quit IRC | 21:54 | |
| bauzas | sdague: <3 | 21:54 |
| *** Sukhdev has joined #openstack-nova | 21:54 | |
| *** krtaylor has joined #openstack-nova | 21:55 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] Trying latest oslo.* from master https://review.openstack.org/275201 | 21:56 |
| *** jwcroppe has quit IRC | 21:57 | |
| *** gmann_ has quit IRC | 21:58 | |
| *** oomichi has quit IRC | 22:00 | |
| *** oomichi_ has joined #openstack-nova | 22:00 | |
| *** oomichi_ is now known as oomichi | 22:00 | |
| *** sc68cal has quit IRC | 22:00 | |
| *** haomaiwang has quit IRC | 22:01 | |
| *** haomaiwang has joined #openstack-nova | 22:01 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add test for unshelve in the conductor API https://review.openstack.org/284420 | 22:03 |
| *** Sukhdev has quit IRC | 22:03 | |
| *** knikolla has quit IRC | 22:05 | |
| *** ljxiash has joined #openstack-nova | 22:07 | |
| openstackgerrit | Mark Doffman proposed openstack/nova: Creates flavor* tables in API database https://review.openstack.org/201606 | 22:08 |
| mriedem | sdague: your thing is +W | 22:08 |
| sdague | mriedem: thanks | 22:08 |
| sdague | ok, have a good evening folks | 22:10 |
| *** ljxiash has quit IRC | 22:10 | |
| *** daemontool has joined #openstack-nova | 22:11 | |
| openstackgerrit | Mark Doffman proposed openstack/nova: Creates flavor* tables in API database https://review.openstack.org/201606 | 22:13 |
| *** Sukhdev has joined #openstack-nova | 22:14 | |
| *** dims has quit IRC | 22:14 | |
| *** _ducttape_ has quit IRC | 22:14 | |
| *** ducttape_ has joined #openstack-nova | 22:15 | |
| *** Sukhdev has quit IRC | 22:17 | |
| *** Sukhdev has joined #openstack-nova | 22:17 | |
| *** raildo is now known as raildo-afk | 22:19 | |
| *** cody-somerville has joined #openstack-nova | 22:19 | |
| *** smatzek__ has joined #openstack-nova | 22:21 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Create Instance from BuildRequest if not in a cell https://review.openstack.org/263927 | 22:21 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Populate instance_mappings during boot https://review.openstack.org/263925 | 22:21 |
| *** daemontool has quit IRC | 22:23 | |
| *** _ducttape_ has joined #openstack-nova | 22:25 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add functional regression test for list deleted instances on v2.16 https://review.openstack.org/283837 | 22:25 |
| *** penick has joined #openstack-nova | 22:27 | |
| *** ducttape_ has quit IRC | 22:28 | |
| *** mrsoul has quit IRC | 22:28 | |
| *** Sukhdev has quit IRC | 22:29 | |
| *** haplo37 has quit IRC | 22:29 | |
| *** oomichi_ has joined #openstack-nova | 22:30 | |
| *** oomichi_ has quit IRC | 22:30 | |
| *** oomichi_ has joined #openstack-nova | 22:30 | |
| *** oomichi_ has quit IRC | 22:31 | |
| mriedem | oomichi: do you know takashin? | 22:31 |
| mriedem | or if he is ever on irc | 22:31 |
| mriedem | ? | 22:31 |
| *** mrsoul has joined #openstack-nova | 22:32 | |
| *** delattec has joined #openstack-nova | 22:34 | |
| mriedem | oomichi: if so, please tell him to reply or remove his -1 from this :) https://review.openstack.org/#/c/279839/ | 22:34 |
| oomichi | mriedem: I don't know takashin, which patch owner? | 22:35 |
| * oomichi is setting up ZNC, not familer with it yet | 22:35 | |
| *** mylu has quit IRC | 22:36 | |
| oomichi | mriedem: ah, he is the other company guy and I met him once | 22:36 |
| oomichi | will catch him via his coworker | 22:37 |
| *** cdelatte has quit IRC | 22:37 | |
| *** smatzek__ has quit IRC | 22:37 | |
| mriedem | thanks, i assumed maybe he was on irc when i wasn't due to timezone or something | 22:37 |
| *** mylu has joined #openstack-nova | 22:38 | |
| *** vishwana_ is now known as vishwanathj | 22:39 | |
| oomichi | mriedem: yeah, now he just wakes up his timezone | 22:40 |
| oomichi | 7:40am | 22:40 |
| mriedem | oomichi: obviously he's not hard-core such as yourself working already :) | 22:41 |
| mriedem | or do you work from home? | 22:41 |
| oomichi | mriedem: I have migrated to USA now 2 week ago, we are on the same tz ;) | 22:42 |
| mriedem | oh wow | 22:42 |
| mriedem | where are you living? | 22:42 |
| *** rtheis has quit IRC | 22:42 | |
| oomichi | San Jose, CA | 22:42 |
| mriedem | ah, 2 hours different from me, but much closer than japan | 22:43 |
| oomichi | mriedem: yeah, nice to meet you | 22:43 |
| oomichi | but now I am in Boston for QA code sprint with mtreinish :) | 22:44 |
| *** aorourke has quit IRC | 22:44 | |
| *** burgerk has quit IRC | 22:45 | |
| mriedem | oh right | 22:45 |
| *** rtheis has joined #openstack-nova | 22:45 | |
| *** aorourke has joined #openstack-nova | 22:45 | |
| *** rfolco has quit IRC | 22:46 | |
| *** apearson has quit IRC | 22:46 | |
| doffm | alaski, jaypipes: I'm going through resource pools. I don't see anything cells specific in there. Both of you are happy that you have a solution? | 22:47 |
| *** jdillaman has quit IRC | 22:48 | |
| *** zenoway_ has quit IRC | 22:48 | |
| mriedem | the first rule of cells + resource pools, | 22:49 |
| mriedem | is not to ask about cells + resource pools | 22:49 |
| jaypipes | doffm: yeah. the agreement was that we would house the resource_providers schema tables in the cell child database and the nova resource-pool-create would require an --aggregate-uuid argument. | 22:50 |
| doffm | jaypipes: Ok, that makes sense, but you were opposed to having the cell_id in the api. If the resource pool goes in a cell won't we still need a cell_id to know where to create it? | 22:52 |
| jaypipes | doffm: the aggregate_uuid -> cell_id will be mapped in the API database. | 22:52 |
| doffm | Ahhhh. | 22:52 |
| doffm | I was missing that bit. | 22:52 |
| mriedem | bauzas: alaski: +2 on https://review.openstack.org/#/c/201606/ now that doffm fixed the flavors.root_gb nullable thing | 22:53 |
| jaypipes | doffm: :) | 22:54 |
| *** ctrath has quit IRC | 22:54 | |
| *** sheel has quit IRC | 22:57 | |
| alaski | doffm: yeah, I think we have it worked out with the above. but definitely poke holes if you see them | 22:59 |
| *** mriedem has quit IRC | 23:00 | |
| *** jdillaman has joined #openstack-nova | 23:00 | |
| *** zenoway has joined #openstack-nova | 23:00 | |
| *** haomaiwang has quit IRC | 23:01 | |
| *** thorst_afk is now known as thorst | 23:01 | |
| *** haomaiwang has joined #openstack-nova | 23:01 | |
| *** burt has quit IRC | 23:01 | |
| *** claudiub has joined #openstack-nova | 23:02 | |
| doffm | alaski: Still thinking about it. Is this described in the spec anywhere? If not I'm fine with that, will review as-is. But we should get down the cells plan soonish. | 23:02 |
| doffm | I see the aggregates stuff there, but not the mapping. | 23:02 |
| alaski | just the aggregate stuff, not the cells specifics | 23:03 |
| *** mdrabe has quit IRC | 23:03 | |
| alaski | until we move aggregates up to the api db we can't really use them to map | 23:04 |
| *** zenoway has quit IRC | 23:04 | |
| *** claudiub|2 has quit IRC | 23:05 | |
| doffm | Yeah. | 23:05 |
| doffm | I'd better get the spec up for moving aggregates to the api. | 23:06 |
| *** dims has joined #openstack-nova | 23:08 | |
| *** rlrossit_ has quit IRC | 23:10 | |
| openstackgerrit | Angus Lees proposed openstack/nova: Add os-brick rootwrap filter for privsep https://review.openstack.org/277670 | 23:10 |
| openstackgerrit | Angus Lees proposed openstack/nova: Require privsep-based os-brick https://review.openstack.org/277671 | 23:10 |
| *** achanda has joined #openstack-nova | 23:10 | |
| openstackgerrit | Bartosz Górski proposed openstack/python-novaclient: Adds missing internationalization for help message https://review.openstack.org/284458 | 23:13 |
| *** ccarmack has quit IRC | 23:13 | |
| openstackgerrit | Angus Lees proposed openstack/nova: Add os-brick rootwrap filter for privsep https://review.openstack.org/277670 | 23:14 |
| openstackgerrit | Angus Lees proposed openstack/nova: Require privsep-based os-brick https://review.openstack.org/277671 | 23:14 |
| *** _ducttape_ has quit IRC | 23:15 | |
| *** Sukhdev has joined #openstack-nova | 23:15 | |
| *** rtheis has quit IRC | 23:16 | |
| *** rtheis has joined #openstack-nova | 23:16 | |
| *** edtubill has quit IRC | 23:16 | |
| *** rtheis has quit IRC | 23:17 | |
| *** rtheis has joined #openstack-nova | 23:17 | |
| *** rtheis has quit IRC | 23:17 | |
| *** pushkaru has quit IRC | 23:18 | |
| *** rk4n has joined #openstack-nova | 23:18 | |
| *** abitha has joined #openstack-nova | 23:20 | |
| jaypipes | dansmith: just wrapping up the compute_node_get_all patch. | 23:21 |
| *** eharney has quit IRC | 23:23 | |
| *** Swami has quit IRC | 23:24 | |
| *** zenoway has joined #openstack-nova | 23:25 | |
| *** vilobhmm11 has quit IRC | 23:25 | |
| *** vilobhmm11 has joined #openstack-nova | 23:25 | |
| dansmith | jaypipes: okay, I'll start on the one on top in the morning | 23:26 |
| jaypipes | dansmith: sorry for the delay. | 23:26 |
| dansmith | s'ok | 23:26 |
| jaypipes | dansmith: just adding some tests to double-check my SQL-work. | 23:26 |
| dansmith | it'll be what it'll be | 23:27 |
| *** armax has quit IRC | 23:27 | |
| *** armax has joined #openstack-nova | 23:28 | |
| *** armax has quit IRC | 23:28 | |
| *** dimtruck is now known as zz_dimtruck | 23:28 | |
| *** zenoway has quit IRC | 23:29 | |
| *** cody-somerville has quit IRC | 23:29 | |
| *** Guest75 has joined #openstack-nova | 23:35 | |
| *** Guest75 has quit IRC | 23:37 | |
| shaohe_feng_ | johnthetubaguy: ping. I's like to talk about the in progress migration | 23:40 |
| shaohe_feng_ | sdague: ping. | 23:41 |
| *** hemna is now known as hemnafk | 23:41 | |
| *** mylu has quit IRC | 23:44 | |
| *** pushkaru has joined #openstack-nova | 23:45 | |
| *** Bjoern_ has quit IRC | 23:46 | |
| jaypipes | shaohe_feng: it's nearly midnight for johnthetubaguy, FYI. | 23:46 |
| *** cdent has quit IRC | 23:47 | |
| *** mylu has joined #openstack-nova | 23:48 | |
| shaohe_feng_ | jaypipes: got it. thank you. | 23:48 |
| *** thorst has quit IRC | 23:48 | |
| *** Guest75 has joined #openstack-nova | 23:49 | |
| *** mylu has quit IRC | 23:50 | |
| *** pushkaru has quit IRC | 23:52 | |
| shaohe_feng_ | jaypipes: would you like to look at my patch? https://review.openstack.org/#/c/258771/29 | 23:57 |
| jaypipes | shaohe_feng: yup, will do shortly. | 23:57 |
| shaohe_feng_ | jaypipes: Thanks. | 23:58 |
| *** mrkz has quit IRC | 23:58 | |
| *** y_sawai has joined #openstack-nova | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!