| *** salv-orlando has joined #openstack-nova | 00:00 | |
| *** unicell has joined #openstack-nova | 00:00 | |
| *** haomaiwa_ has quit IRC | 00:01 | |
| auggy | edleafe: ha cool, i was a toddler in 1980 ;D | 00:01 |
|---|---|---|
| *** pmathews has quit IRC | 00:01 | |
| *** haomaiwa_ has joined #openstack-nova | 00:01 | |
| *** unicell has quit IRC | 00:01 | |
| *** vilobhmm11 has quit IRC | 00:03 | |
| *** vilobhmm11 has joined #openstack-nova | 00:03 | |
| *** salv-orlando has quit IRC | 00:06 | |
| *** fawadkhaliq has quit IRC | 00:07 | |
| *** fawadkhaliq has joined #openstack-nova | 00:07 | |
| *** gyee has quit IRC | 00:08 | |
| *** vilobhmm11 has quit IRC | 00:08 | |
| *** vilobhmm11 has joined #openstack-nova | 00:09 | |
| *** vilobhmm11 has quit IRC | 00:09 | |
| *** vilobhmm11 has joined #openstack-nova | 00:09 | |
| *** vilobhmm11 has quit IRC | 00:09 | |
| *** ayoung has joined #openstack-nova | 00:11 | |
| *** browne has quit IRC | 00:12 | |
| *** diogogmt has quit IRC | 00:14 | |
| *** mylu has joined #openstack-nova | 00:14 | |
| *** mylu has quit IRC | 00:15 | |
| *** browne has joined #openstack-nova | 00:17 | |
| mriedem | auggy: this probably needs some tweaking since kilo, lots of comments inline https://review.openstack.org/#/c/294337/ | 00:20 |
| mriedem | auggy: might be a decent topic for next weeks' API subteam meeting, if you can make the time | 00:20 |
| mriedem | i think it'd be pretty early for portland | 00:20 |
| *** jaypipes has quit IRC | 00:21 | |
| *** mtanino has quit IRC | 00:21 | |
| *** artom has quit IRC | 00:25 | |
| openstackgerrit | Claudiu Belu proposed openstack/nova: compute: fixes python 3 related unit tests https://review.openstack.org/300262 | 00:27 |
| *** unicell has joined #openstack-nova | 00:29 | |
| * tbachman heads out | 00:30 | |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Add tests for floating_ip private functions https://review.openstack.org/267768 | 00:30 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Remove DictCompat from FixedIP https://review.openstack.org/265951 | 00:30 |
| *** tbachman has quit IRC | 00:31 | |
| gmann | mriedem: did you get chance to see this (swap volume)- https://review.openstack.org/#/c/299830/ | 00:33 |
| gmann | mriedem: still it stuck in detaching/attaching state | 00:33 |
| openstackgerrit | Karen Bradshaw proposed openstack/nova: [WIP] Create parameters.yaml per operation https://review.openstack.org/300264 | 00:35 |
| gmann | mriedem: i see update in your patch, i will update tempest patch accordingly | 00:36 |
| *** yassou has quit IRC | 00:38 | |
| *** mylu has joined #openstack-nova | 00:39 | |
| *** vilobhmm11 has joined #openstack-nova | 00:40 | |
| auggy | mriedem: yeah i keep meaning to join but i end up missing it lol | 00:40 |
| *** READ10 has quit IRC | 00:41 | |
| *** mylu has quit IRC | 00:41 | |
| mriedem | gmann: yeah i don't think tempest can test this w/o devstack being updated to soften the policies in cinder and nova | 00:41 |
| auggy | mriedem: but yeah i'll take a look at your comments and see what i can find out, plus actually set my alarm a little earlier so maybe i'll make that meeting ;D | 00:41 |
| mriedem | auggy: i think it's like 6am for you | 00:41 |
| mriedem | probably just pester sdague about it sometime, and alex_xu | 00:42 |
| auggy | yeah i have it on my calendar | 00:42 |
| *** mylu has joined #openstack-nova | 00:42 | |
| * auggy just needs to take another trip to europe to get her schedule shifted ;) | 00:42 | |
| auggy | it's at 7am for me | 00:42 |
| auggy | normally i wake up around 7 but since this back strain thing sleeping is complicated | 00:43 |
| claudiub | i guess the problems are back? ;) | 00:43 |
| *** mylu has quit IRC | 00:44 | |
| auggy | yeah i can't wait to get BACK into shape ;) | 00:44 |
| *** thorst has joined #openstack-nova | 00:44 | |
| *** mylu has joined #openstack-nova | 00:44 | |
| claudiub | i just hope you won't back down from it. :D | 00:44 |
| auggy | as long as no one talks behind my back :P | 00:45 |
| *** zenoway has joined #openstack-nova | 00:46 | |
| claudiub | well, just make sure to backup a good comeback to make them back off. :D | 00:47 |
| claudiub | ha | 00:47 |
| *** diogogmt has joined #openstack-nova | 00:49 | |
| *** mylu has quit IRC | 00:50 | |
| *** zenoway has quit IRC | 00:51 | |
| *** hoangcx has joined #openstack-nova | 00:51 | |
| *** thorst has quit IRC | 00:51 | |
| *** mylu has joined #openstack-nova | 00:52 | |
| *** markvoelker_ has quit IRC | 00:53 | |
| *** Daisy has joined #openstack-nova | 00:54 | |
| *** edmondsw has joined #openstack-nova | 00:54 | |
| *** Daisy_ has joined #openstack-nova | 00:56 | |
| *** Daisy__ has joined #openstack-nova | 00:58 | |
| gmann | mriedem: yea but should we tests with softening the policy (i do think so), tempest tests everything with default policy and if nova side it make only admin then just negative tests in tempest for non-admin should be enough | 00:59 |
| gmann | (I do *not* think so) | 00:59 |
| *** Jeffrey4l has joined #openstack-nova | 01:00 | |
| *** baoli has joined #openstack-nova | 01:00 | |
| *** Swami has quit IRC | 01:00 | |
| *** Daisy has quit IRC | 01:00 | |
| *** haomaiwa_ has quit IRC | 01:01 | |
| mriedem | gmann: i'm fine with a negative test for a non-admin user trying to do it | 01:01 |
| mriedem | gmann: actually, non-admin would be a negative test, | 01:01 |
| *** efoley has quit IRC | 01:01 | |
| *** haomaiwa_ has joined #openstack-nova | 01:01 | |
| mriedem | gmann: we can test it with an admin | 01:01 |
| mriedem | however, that won't work on stable branches and we're not going to backport the default policy change to stable branches for this, i don't think | 01:01 |
| *** gjayavelu has joined #openstack-nova | 01:01 | |
| mriedem | gmann: so you'd probably need a config option in tempest for enabling the test | 01:02 |
| mriedem | compute-feature-enabled.swap_volume? | 01:02 |
| *** Daisy_ has quit IRC | 01:02 | |
| *** Bofu2MBP has quit IRC | 01:02 | |
| gmann | mriedem: yea, then i feel negative tests in nova and later we can add in tempest if needed. | 01:03 |
| gmann | mriedem: we can tests only admin case ehich should work for all stable | 01:03 |
| *** mylu has quit IRC | 01:03 | |
| gmann | mriedem: because adding feature flag for that might be heavy by considering lot of config option for tempest :) | 01:03 |
| mriedem | true, testing with admin in tempest would pass all branches | 01:04 |
| mriedem | well, maybe not, there was a fix in cinder that might need to be backported | 01:04 |
| mriedem | master and stable/mitaka should be ok though | 01:04 |
| mriedem | we'd have to see about liberty and kilo | 01:04 |
| *** lixiaoy1 has joined #openstack-nova | 01:04 | |
| *** mylu has joined #openstack-nova | 01:04 | |
| gmann | mriedem: nice point, seems like liberty is good(test with admin pass) but yea kilo is issue | 01:05 |
| gmann | mriedem: http://logs.openstack.org/30/299830/2/check/gate-tempest-dsvm-full-kilo/3b281bd/logs/testr_results.html.gz | 01:06 |
| gmann | mriedem: may be we can wait till kilo gets out | 01:06 |
| gmann | which should be soon | 01:06 |
| mriedem | gmann: actually it might be another 6 months | 01:06 |
| mriedem | we'll find out at the summit when we talk about stable branch EOL for kilo | 01:07 |
| gmann | mriedem: hummm, not just 2 stable now (L and M) | 01:07 |
| mriedem | we could backport the cinder change maybe...but it's not really a critical/security fix so not really appropriate for stable/kilo | 01:07 |
| mriedem | gmann: https://etherpad.openstack.org/p/stable-branch-eol-policy-newton | 01:07 |
| *** mylu has quit IRC | 01:09 | |
| *** vilobhmm11 has quit IRC | 01:10 | |
| gmann | mriedem: i see, Thanks nice info | 01:10 |
| *** tbachman has joined #openstack-nova | 01:10 | |
| gmann | mriedem: i too feel now compute-feature-enabled.swap_volume can work | 01:12 |
| *** browne has quit IRC | 01:12 | |
| gmann | mriedem: because i too feel we should not leave swap volume untested in gate in case kilo going till another 6 month. | 01:13 |
| *** angdraug has quit IRC | 01:14 | |
| *** mylu has joined #openstack-nova | 01:15 | |
| auggy | claudiub: lolol | 01:15 |
| *** Daisy__ has quit IRC | 01:16 | |
| *** EinstCrazy has joined #openstack-nova | 01:16 | |
| *** thorst has joined #openstack-nova | 01:17 | |
| *** Jeffrey4l has quit IRC | 01:17 | |
| *** EinstCra_ has joined #openstack-nova | 01:19 | |
| *** tbachman_ has joined #openstack-nova | 01:19 | |
| *** tbachman has quit IRC | 01:19 | |
| *** tbachman_ is now known as tbachman | 01:19 | |
| *** mylu has quit IRC | 01:22 | |
| *** EinstCrazy has quit IRC | 01:22 | |
| *** thorst has quit IRC | 01:23 | |
| *** mylu has joined #openstack-nova | 01:24 | |
| *** knikolla has quit IRC | 01:25 | |
| *** woodster_ has quit IRC | 01:27 | |
| *** mtanino has joined #openstack-nova | 01:28 | |
| *** edmondsw has quit IRC | 01:30 | |
| openstackgerrit | Li Yingjun proposed openstack/nova-specs: Hypervisor notifications spec https://review.openstack.org/299807 | 01:30 |
| *** ChrisBenson has quit IRC | 01:33 | |
| *** krtaylor has joined #openstack-nova | 01:35 | |
| *** Kevin_Zheng has joined #openstack-nova | 01:35 | |
| *** pmathews has joined #openstack-nova | 01:37 | |
| *** fawadkhaliq has quit IRC | 01:40 | |
| *** pmathews has quit IRC | 01:42 | |
| *** liusheng has quit IRC | 01:42 | |
| *** thorst has joined #openstack-nova | 01:42 | |
| *** thorst has quit IRC | 01:43 | |
| *** zz_dimtruck is now known as dimtruck | 01:44 | |
| *** apoorvad has quit IRC | 01:44 | |
| *** mriedem1 has joined #openstack-nova | 01:45 | |
| *** liusheng has joined #openstack-nova | 01:45 | |
| openstackgerrit | Rui Chen proposed openstack/python-novaclient: Fix Resource.__eq__ mismatch semantics of object equal https://review.openstack.org/251020 | 01:46 |
| openstackgerrit | Rui Chen proposed openstack/nova: Create image for suspended instance booted from volume https://review.openstack.org/223382 | 01:46 |
| openstackgerrit | Rui Chen proposed openstack/nova-specs: Task states improvement https://review.openstack.org/216578 | 01:47 |
| *** mriedem has quit IRC | 01:47 | |
| *** pmathews has joined #openstack-nova | 01:48 | |
| *** yamahata has quit IRC | 01:49 | |
| *** mylu has quit IRC | 01:50 | |
| openstackgerrit | Rui Chen proposed openstack/nova: Only return latest fault of each instance in db api https://review.openstack.org/223392 | 01:53 |
| *** fawadkhaliq has joined #openstack-nova | 01:54 | |
| *** dimtruck is now known as zz_dimtruck | 01:55 | |
| *** gjayavelu has quit IRC | 01:58 | |
| *** baoli has quit IRC | 01:58 | |
| *** baoli has joined #openstack-nova | 01:58 | |
| *** zz_dimtruck is now known as dimtruck | 01:59 | |
| *** Sukhdev has quit IRC | 01:59 | |
| *** haomaiwa_ has quit IRC | 02:01 | |
| *** haomaiwang has joined #openstack-nova | 02:01 | |
| *** liunian_ has joined #openstack-nova | 02:01 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: config options: centralize section: "glance" https://review.openstack.org/257153 | 02:03 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [POC] Allow user defined stop https://review.openstack.org/298221 | 02:05 |
| *** pmathews1 has joined #openstack-nova | 02:08 | |
| *** unicell has quit IRC | 02:10 | |
| *** coreywright has quit IRC | 02:10 | |
| *** pmathews has quit IRC | 02:11 | |
| *** achanda has joined #openstack-nova | 02:14 | |
| openstackgerrit | ZhiQiang Fan proposed openstack/nova: Use oslo_config new type PortOpt for port options https://review.openstack.org/238347 | 02:14 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: The 'record' option of the WebSocketProxy should be string https://review.openstack.org/294512 | 02:15 |
| *** haomaiwang has quit IRC | 02:15 | |
| *** mylu has joined #openstack-nova | 02:19 | |
| *** akshai_ has quit IRC | 02:23 | |
| *** baoli has quit IRC | 02:25 | |
| mriedem1 | tonyb: claudiub: to unblock nova on liberty https://review.openstack.org/#/c/300284/ | 02:27 |
| *** coreywright has joined #openstack-nova | 02:27 | |
| tonyb | mriedem1: looking .... | 02:29 |
| *** fawadkhaliq has quit IRC | 02:33 | |
| *** zenoway has joined #openstack-nova | 02:34 | |
| *** efoley has joined #openstack-nova | 02:38 | |
| *** zenoway has quit IRC | 02:38 | |
| *** mriedem1 has quit IRC | 02:38 | |
| *** fawadkhaliq has joined #openstack-nova | 02:42 | |
| *** baoli has joined #openstack-nova | 02:42 | |
| *** fawadkhaliq has quit IRC | 02:43 | |
| *** baoli has quit IRC | 02:45 | |
| *** fawadkhaliq has joined #openstack-nova | 02:45 | |
| *** allen_ga- is now known as allen_gao | 02:46 | |
| *** diogogmt has quit IRC | 02:49 | |
| *** diogogmt has joined #openstack-nova | 02:50 | |
| *** pmathews1 has quit IRC | 02:50 | |
| *** pmathews has joined #openstack-nova | 02:51 | |
| *** gcb has joined #openstack-nova | 02:52 | |
| *** pmathews1 has joined #openstack-nova | 02:53 | |
| *** fawadkhaliq has quit IRC | 02:53 | |
| *** fawadkhaliq has joined #openstack-nova | 02:53 | |
| *** 32NAAPLEF has joined #openstack-nova | 02:54 | |
| *** pmathews has quit IRC | 02:56 | |
| *** tbachman has quit IRC | 02:57 | |
| *** tbachman has joined #openstack-nova | 02:57 | |
| *** 32NAAPLEF has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-nova | 03:01 | |
| *** mylu has quit IRC | 03:08 | |
| *** fawadkhaliq has quit IRC | 03:08 | |
| *** zenoway has joined #openstack-nova | 03:09 | |
| *** mylu has joined #openstack-nova | 03:09 | |
| *** fawadkhaliq has joined #openstack-nova | 03:10 | |
| *** mylu has quit IRC | 03:11 | |
| *** mylu has joined #openstack-nova | 03:11 | |
| *** fawadkhaliq has quit IRC | 03:12 | |
| *** zenoway has quit IRC | 03:14 | |
| *** ljxiash has joined #openstack-nova | 03:14 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Don't use 'updated_at' to check service's status https://review.openstack.org/297085 | 03:18 |
| *** krotscheck is now known as krotscheck_vaca | 03:20 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Enhance value check for option notify_on_state_change https://review.openstack.org/242717 | 03:22 |
| Kevin_Zheng | alex_xu: could you have a quick look on this: https://review.openstack.org/#/c/257153/ ? if you got a sec | 03:24 |
| alex_xu | Kevin_Zheng: yea, checking now | 03:24 |
| Kevin_Zheng | THX | 03:25 |
| openstackgerrit | Diana Clarke proposed openstack/nova: Fix: unable to delete instance when cinder is down https://review.openstack.org/298997 | 03:25 |
| *** diogogmt has quit IRC | 03:25 | |
| alex_xu | Kevin_Zheng: np, done | 03:29 |
| *** fawadkhaliq has joined #openstack-nova | 03:32 | |
| *** vilobhmm11 has joined #openstack-nova | 03:35 | |
| openstackgerrit | Rong Han proposed openstack/nova: Fix resize to same host failed using anti-affinity group https://review.openstack.org/299045 | 03:36 |
| *** links has joined #openstack-nova | 03:41 | |
| allen_gao | alex_xu: would you take a look? thank you. ;) https://review.openstack.org/#/c/299181/ | 03:42 |
| *** zenoway has joined #openstack-nova | 03:45 | |
| *** ChrisBenson has joined #openstack-nova | 03:46 | |
| *** markvoelker has joined #openstack-nova | 03:47 | |
| *** zenoway has quit IRC | 03:50 | |
| *** mylu has quit IRC | 03:54 | |
| *** markvoelker_ has joined #openstack-nova | 03:56 | |
| *** markvoelker has quit IRC | 03:56 | |
| *** efoley has quit IRC | 03:58 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwang has joined #openstack-nova | 04:01 | |
| *** liunian_ has quit IRC | 04:02 | |
| *** liunian_ has joined #openstack-nova | 04:02 | |
| *** liunian_ has quit IRC | 04:02 | |
| *** achanda has quit IRC | 04:04 | |
| *** achanda has joined #openstack-nova | 04:08 | |
| *** mylu has joined #openstack-nova | 04:10 | |
| *** liunian_ has joined #openstack-nova | 04:10 | |
| *** mylu has quit IRC | 04:10 | |
| *** ljxiash has quit IRC | 04:11 | |
| *** ljxiash has joined #openstack-nova | 04:11 | |
| *** mylu has joined #openstack-nova | 04:13 | |
| *** achanda has quit IRC | 04:13 | |
| *** ljxiash has quit IRC | 04:15 | |
| *** harlowja_at_home has joined #openstack-nova | 04:19 | |
| *** zenoway has joined #openstack-nova | 04:21 | |
| *** ChrisBenson has quit IRC | 04:22 | |
| *** Sukhdev has joined #openstack-nova | 04:25 | |
| *** zenoway has quit IRC | 04:25 | |
| *** fawadkhaliq has quit IRC | 04:26 | |
| *** vladikr_ has quit IRC | 04:26 | |
| *** sdake has joined #openstack-nova | 04:29 | |
| *** ndipanov has quit IRC | 04:29 | |
| *** markvoelker_ has quit IRC | 04:33 | |
| *** markvoelker has joined #openstack-nova | 04:33 | |
| *** dimtruck is now known as zz_dimtruck | 04:34 | |
| *** markvoelker has quit IRC | 04:38 | |
| *** liunian_ has quit IRC | 04:39 | |
| *** mylu has quit IRC | 04:41 | |
| *** vladikr has joined #openstack-nova | 04:42 | |
| *** mtreinish has quit IRC | 04:44 | |
| *** vladikr has quit IRC | 04:44 | |
| *** mylu has joined #openstack-nova | 04:46 | |
| *** mnestratov|2 has quit IRC | 04:47 | |
| *** pradipta has joined #openstack-nova | 04:49 | |
| *** pradipta has quit IRC | 04:50 | |
| *** hoangcx has quit IRC | 04:54 | |
| *** liunian_ has joined #openstack-nova | 04:55 | |
| *** zenoway has joined #openstack-nova | 04:56 | |
| *** mylu has quit IRC | 04:57 | |
| *** rcernin has joined #openstack-nova | 04:57 | |
| *** fawadkhaliq has joined #openstack-nova | 04:57 | |
| *** vladikr has joined #openstack-nova | 04:59 | |
| *** liunian_ has quit IRC | 05:00 | |
| *** haomaiwang has quit IRC | 05:01 | |
| *** zenoway has quit IRC | 05:01 | |
| *** haomaiwa_ has joined #openstack-nova | 05:01 | |
| *** achanda has joined #openstack-nova | 05:02 | |
| *** stevemar_ has quit IRC | 05:02 | |
| *** stevemar has joined #openstack-nova | 05:02 | |
| *** mylu has joined #openstack-nova | 05:05 | |
| *** prateek has joined #openstack-nova | 05:08 | |
| *** prateek has joined #openstack-nova | 05:09 | |
| *** prateek is now known as parora | 05:09 | |
| *** cdelatte has quit IRC | 05:10 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Add validations for volume_size and destination_type https://review.openstack.org/258788 | 05:10 |
| *** achanda has quit IRC | 05:10 | |
| *** liunian_ has joined #openstack-nova | 05:12 | |
| *** liunian_ has quit IRC | 05:17 | |
| *** harlowja_at_home has quit IRC | 05:19 | |
| *** cfriesen has joined #openstack-nova | 05:21 | |
| parora | In the nova api code, I have seen methods using the save_and_reraise_exception() method. But isn't it something that breaks the cli ? it makes the cli throw an exception and | 05:21 |
| parora | i get this on cli when such an exception is thrown [root@localhost devstack]# ERROR (ClientException): Unexpected API Error. Please report this at http://bugs.launchpad.net/nova/ and attach the Nova API log if possible. | 05:21 |
| parora | <class 'nova.exception.UnexpectedDeletingTaskStateError'> | 05:21 |
| *** tpeoples is now known as tpeoples_vacatio | 05:23 | |
| *** cfriesen__ has quit IRC | 05:23 | |
| *** tpeoples_vacatio is now known as tpeoplesvacation | 05:24 | |
| *** davideag_ has joined #openstack-nova | 05:24 | |
| *** jaosorior has joined #openstack-nova | 05:25 | |
| *** davideagnello has quit IRC | 05:26 | |
| alex_xu | allen_gao: done also | 05:27 |
| alex_xu | parora: it will be great if you can file a bug for that | 05:28 |
| openstackgerrit | Jianghua Wang proposed openstack/nova: XenAPI: add unit test for the plugin - glance https://review.openstack.org/299533 | 05:28 |
| *** markvoelker has joined #openstack-nova | 05:29 | |
| *** gjayavelu has joined #openstack-nova | 05:29 | |
| parora | alex_xu i have, what i am trying to do is now that I am trying to resolve that | 05:29 |
| *** toabctl has left #openstack-nova | 05:29 | |
| parora | alex_xu , i filed a bug and am trying to resolve it | 05:30 |
| parora | https://review.openstack.org/#/c/294513/ | 05:30 |
| *** sdake_ has joined #openstack-nova | 05:30 | |
| *** sdake_ has quit IRC | 05:31 | |
| parora | alex_xu for resolving the last set of comments i am thinking of reraising the exception in api class so that the other classes don't try to process the request further | 05:32 |
| *** zenoway has joined #openstack-nova | 05:32 | |
| parora | but the reraise would leave the cli hanged with the message i posted above | 05:32 |
| *** sdake has quit IRC | 05:32 | |
| alex_xu | parora: you should convert the exception to a HTTP*exception in the api layer, like this https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/servers.py#L631 | 05:35 |
| alex_xu | parora: then nova won't consider that exception is an unexpected exception anymore | 05:36 |
| *** zenoway has quit IRC | 05:36 | |
| *** markvoelker has quit IRC | 05:37 | |
| *** pmathews1 has quit IRC | 05:38 | |
| parora | alex_xu that sounds cool, thanks for the help :) | 05:38 |
| alex_xu | parora: np :) | 05:38 |
| *** vilobhmm11 has quit IRC | 05:38 | |
| *** davideag_ has quit IRC | 05:42 | |
| *** mylu has quit IRC | 05:43 | |
| openstackgerrit | Jianghua Wang proposed openstack/nova: XenAPI: add unit test for the plugin - glance https://review.openstack.org/299533 | 05:43 |
| *** liunian_ has joined #openstack-nova | 05:44 | |
| *** vilobhmm11 has joined #openstack-nova | 05:46 | |
| *** liunian_ has quit IRC | 05:48 | |
| *** mylu has joined #openstack-nova | 05:48 | |
| *** liunian_ has joined #openstack-nova | 05:48 | |
| *** mylu has quit IRC | 05:48 | |
| *** mtreinish has joined #openstack-nova | 05:52 | |
| *** nkrinner has joined #openstack-nova | 05:53 | |
| *** esracelik has joined #openstack-nova | 05:53 | |
| *** yamahata has joined #openstack-nova | 05:58 | |
| *** haomaiwa_ has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-nova | 06:01 | |
| *** hoangcx has joined #openstack-nova | 06:03 | |
| *** takashin has joined #openstack-nova | 06:04 | |
| *** ljxiash has joined #openstack-nova | 06:04 | |
| *** ChrisBenson has joined #openstack-nova | 06:06 | |
| *** Marga_ has quit IRC | 06:06 | |
| *** ljxiash has quit IRC | 06:09 | |
| *** ljxiash has joined #openstack-nova | 06:09 | |
| *** haukebruno has joined #openstack-nova | 06:12 | |
| *** ljxiash has quit IRC | 06:14 | |
| *** ChrisBenson has quit IRC | 06:16 | |
| *** liusheng has quit IRC | 06:17 | |
| *** armax has quit IRC | 06:18 | |
| *** liusheng has joined #openstack-nova | 06:19 | |
| *** lpetrut has joined #openstack-nova | 06:22 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [POC] Allow user defined stop https://review.openstack.org/298221 | 06:23 |
| *** Sukhdev has quit IRC | 06:24 | |
| *** Marga_ has joined #openstack-nova | 06:25 | |
| *** browne has joined #openstack-nova | 06:27 | |
| *** Marga_ has quit IRC | 06:29 | |
| *** achanda has joined #openstack-nova | 06:30 | |
| *** browne has quit IRC | 06:32 | |
| *** sdake has joined #openstack-nova | 06:36 | |
| *** mnestratov|2 has joined #openstack-nova | 06:37 | |
| *** Marga_ has joined #openstack-nova | 06:38 | |
| *** fawadkhaliq has quit IRC | 06:38 | |
| *** fawadkhaliq has joined #openstack-nova | 06:38 | |
| *** browne has joined #openstack-nova | 06:39 | |
| *** Marga_ has quit IRC | 06:43 | |
| *** browne has quit IRC | 06:44 | |
| *** browne has joined #openstack-nova | 06:45 | |
| *** jlanoux has joined #openstack-nova | 06:47 | |
| *** mjura has joined #openstack-nova | 06:49 | |
| *** browne has quit IRC | 06:49 | |
| *** ljxiash has joined #openstack-nova | 07:00 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-nova | 07:01 | |
| *** ljxiash has quit IRC | 07:03 | |
| *** ljxiash has joined #openstack-nova | 07:03 | |
| *** salv-orlando has joined #openstack-nova | 07:05 | |
| *** cfriesen has quit IRC | 07:06 | |
| *** e0ne has joined #openstack-nova | 07:07 | |
| *** browne has joined #openstack-nova | 07:08 | |
| *** browne has quit IRC | 07:13 | |
| *** mgoddard has joined #openstack-nova | 07:14 | |
| *** rdopiera has joined #openstack-nova | 07:14 | |
| *** ljxiash has quit IRC | 07:15 | |
| *** matrohon has joined #openstack-nova | 07:16 | |
| *** jamielennox is now known as jamielennox|away | 07:17 | |
| *** markvoelker has joined #openstack-nova | 07:19 | |
| openstackgerrit | yuanzhicao proposed openstack/nova-specs: Add suport for libvirt driver to start lxc from rbd volume. https://review.openstack.org/283957 | 07:19 |
| *** salv-orl_ has joined #openstack-nova | 07:23 | |
| *** sridharg has joined #openstack-nova | 07:24 | |
| *** pcaruana has joined #openstack-nova | 07:25 | |
| *** markvoelker has quit IRC | 07:25 | |
| *** salv-orlando has quit IRC | 07:26 | |
| gibi | doffm: thanks for the bug report. Do you plan to fix it or shall I look into it? | 07:29 |
| *** artom has joined #openstack-nova | 07:30 | |
| *** gjayavelu has quit IRC | 07:31 | |
| *** artom has quit IRC | 07:32 | |
| *** e0ne has quit IRC | 07:35 | |
| *** e0ne has joined #openstack-nova | 07:35 | |
| *** zenoway has joined #openstack-nova | 07:37 | |
| *** lpetrut has quit IRC | 07:37 | |
| *** e0ne has quit IRC | 07:40 | |
| *** ljxiash has joined #openstack-nova | 07:41 | |
| *** yamahata has quit IRC | 07:42 | |
| *** ljxiash_ has joined #openstack-nova | 07:42 | |
| *** e0ne has joined #openstack-nova | 07:45 | |
| *** ljxiash has quit IRC | 07:46 | |
| *** jaosorior has quit IRC | 07:46 | |
| *** jaosorior has joined #openstack-nova | 07:46 | |
| *** lixiaoy1 has quit IRC | 07:47 | |
| *** lixiaoy1 has joined #openstack-nova | 07:49 | |
| *** gszasz has joined #openstack-nova | 07:50 | |
| *** ociuhandu has quit IRC | 07:51 | |
| *** amotoki has joined #openstack-nova | 07:52 | |
| *** e0ne has quit IRC | 07:54 | |
| *** tesseract has joined #openstack-nova | 07:56 | |
| *** fawadkhaliq has quit IRC | 07:56 | |
| *** tesseract is now known as Guest90877 | 07:56 | |
| *** e0ne has joined #openstack-nova | 07:57 | |
| *** takashin has left #openstack-nova | 07:59 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-nova | 08:01 | |
| bauzas | good morning Nova | 08:03 |
| *** ndipanov has joined #openstack-nova | 08:06 | |
| *** jistr has joined #openstack-nova | 08:07 | |
| *** e0ne has quit IRC | 08:10 | |
| *** danpb has joined #openstack-nova | 08:10 | |
| *** e0ne has joined #openstack-nova | 08:13 | |
| *** cdent has joined #openstack-nova | 08:16 | |
| *** baoli has joined #openstack-nova | 08:16 | |
| *** e0ne has quit IRC | 08:16 | |
| *** toabctl has joined #openstack-nova | 08:20 | |
| *** e0ne has joined #openstack-nova | 08:20 | |
| openstackgerrit | yuanzhicao proposed openstack/nova-specs: Add suport for libvirt driver to start lxc from rbd volume. https://review.openstack.org/283957 | 08:20 |
| *** baoli has quit IRC | 08:20 | |
| *** sfinucan has joined #openstack-nova | 08:22 | |
| *** achanda has quit IRC | 08:22 | |
| *** cdent has quit IRC | 08:24 | |
| *** cdent has joined #openstack-nova | 08:24 | |
| *** achanda has joined #openstack-nova | 08:25 | |
| *** e0ne has quit IRC | 08:26 | |
| *** efried has quit IRC | 08:26 | |
| cdent | p!spy | 08:28 |
| *** e0ne has joined #openstack-nova | 08:29 | |
| *** vtech has joined #openstack-nova | 08:29 | |
| *** cdent has quit IRC | 08:30 | |
| *** e0ne_ has joined #openstack-nova | 08:31 | |
| *** e0ne has quit IRC | 08:32 | |
| *** achanda has quit IRC | 08:32 | |
| *** cdent has joined #openstack-nova | 08:32 | |
| *** e0ne_ has quit IRC | 08:33 | |
| *** liunian__ has joined #openstack-nova | 08:35 | |
| *** yassine has joined #openstack-nova | 08:38 | |
| *** achanda has joined #openstack-nova | 08:38 | |
| *** liunian_ has quit IRC | 08:39 | |
| *** yassine is now known as Guest81243 | 08:39 | |
| *** Guest81243 is now known as yassou | 08:39 | |
| *** efried has joined #openstack-nova | 08:40 | |
| _gryf | cdent, hi | 08:45 |
| cdent | hi _gryf, what's up? | 08:45 |
| _gryf | cdent, i'm just wondering | 08:45 |
| _gryf | cdent, where could I start if I'd like to add additional resource class | 08:45 |
| _gryf | which would be connected with resources which compute have | 08:46 |
| _gryf | in the current condition of managing resources in nova | 08:46 |
| cdent | do you mean this kind of resource class: https://github.com/openstack/nova/blob/master/nova/objects/fields.py#L305 ? | 08:47 |
| markus_z | tonyb: are you around? | 08:47 |
| _gryf | cdent, yes, basically i'm targeting the new resource objects thing | 08:47 |
| tonyb | markus_z: I am | 08:47 |
| *** e0ne has joined #openstack-nova | 08:48 | |
| markus_z | tonyb: I'm testing the libvirt job stuff and found an issue | 08:48 |
| _gryf | cdent, however, i'd like to play a bit with the thing which require special arrangement before vm can utilize such resource | 08:48 |
| markus_z | tonyb: https://review.openstack.org/#/c/299840/ | 08:48 |
| tonyb | markus_z: looking | 08:48 |
| *** paul-carlton2 has quit IRC | 08:48 | |
| _gryf | and do some cleanup after vm is destroyed | 08:49 |
| cdent | _gryf: All the ResourceClass does at this point is provide a static string representing the class, and an int which is used as an identifier when creating a resource provider, inventory or allocation | 08:49 |
| markus_z | tonyb: The happy path doesn't finish yet | 08:49 |
| *** paul-carlton2 has joined #openstack-nova | 08:49 | |
| cdent | _gryf: so I'm not sure how it fits in with what you're describing | 08:49 |
| cdent | s/sure/clear/ | 08:49 |
| _gryf | cdent, basically i'm talking about using the FPGAs which have to be (re)programmed before the vm can use it | 08:50 |
| *** zhangkepi has joined #openstack-nova | 08:51 | |
| tonyb | markus_z: +2W'd | 08:51 |
| cdent | What resources providers (and classes) give you is some way of tracking how many of something is available and used. Then don't really get involved in the automation of the actual use thereof. | 08:52 |
| markus_z | tonyb: cool, thanks! | 08:52 |
| _gryf | cdent, right. and also claiming I assume? | 08:52 |
| tonyb | markus_z: We probably need to add a cleaup for that | 08:53 |
| *** liunian_ has joined #openstack-nova | 08:53 | |
| cdent | _gryf: eventually, yes | 08:53 |
| *** e0ne has quit IRC | 08:53 | |
| _gryf | cdent, in that case, I have to make sure, that proper programmed fpga is exposed to the vm, isn't it? | 08:53 |
| markus_z | tonyb: Do we? I thought the test node will be destroyed at the end and added things doesn't conflict with anything then. | 08:55 |
| tonyb | markus_z: For nodepool it's probably okay without the clean up but for a developer testing it on a local system it'll be needed | 08:56 |
| tonyb | markus_z: Try doing 2 devstack runs back to back .... | 08:56 |
| markus_z | tonyb: Oh, yes, you're right. I'll push a change for that today. | 08:56 |
| tonyb | markus_z: cool. | 08:56 |
| *** liunian__ has quit IRC | 08:57 | |
| cdent | _gryf: In the context of resource providers, it sounds like you'd want a generic-resource-pool that is in some fashion of class 'FPGA' and it's capacity would need to be managed (by something undefined at this point). An available fpga would only show up in that pool if it were "ready". Have you seen the examples in the generic resource pools spec? | 08:57 |
| *** achanda has quit IRC | 08:57 | |
| _gryf | cdent, I thought that generic resource pool object holds a specific class of resources - hence the question about provideing another class for the fpga | 08:58 |
| *** e0ne has joined #openstack-nova | 08:58 | |
| cdent | Yeah, that's what I'm saying. | 08:59 |
| cdent | _gryf: The question becomes is "FPGA" sufficiently generic and global to warrant a class. Evidently one of the goals of having the classes (I wasn't around during this stage of discussions) was to make sure we didn't keep making more willy nilly. | 09:00 |
| cdent | But if it is, then adding to where I linked above is the place to put the new class. | 09:00 |
| *** haomaiwang has quit IRC | 09:01 | |
| johnthetubaguy | cdent: _gryf: it feels like FPGA needs to be modeled like the PCI devices stuff? | 09:01 |
| cdent | johnthetubaguy: I was under the impression that nothing should ever be modeled on the pci stuff ever again :) | 09:01 |
| *** haomaiwang has joined #openstack-nova | 09:01 | |
| cdent | but yes, it would make some sense for there to be a FPGA_DEVICE alongside PCI_DEVICE | 09:01 |
| *** p01nt3r75 has joined #openstack-nova | 09:02 | |
| cdent | the thing that has me a bit confused is how capacity and allocations are measured for such things | 09:02 |
| johnthetubaguy | cdent: heh, well thats a good point | 09:02 |
| _gryf | johnthetubaguy, actually, it may be - due the fact, that it might be used as a pci card | 09:02 |
| johnthetubaguy | _gryf: ah, interesting, I had those ones integrated into the CPU in my head, but as a PCI card makes sense too | 09:03 |
| _gryf | where some of the fpga parts can be exposed as a functions through sriov, prviding acceleration for a virtual network devices for example | 09:03 |
| johnthetubaguy | _gryf: the vGPU stuff is an interesting twist, I wondering how far we are from vFPGAs? | 09:03 |
| johnthetubaguy | oh... | 09:03 |
| _gryf | while the other slots can be used exclusively for other purposes | 09:03 |
| _gryf | so it goes a bit complicated there… | 09:04 |
| * johnthetubaguy it must be summer, a bee just flew into my home office | 09:04 | |
| johnthetubaguy | _gryf: so crazy idea, is the FPGA more like a local read-only cinder volume... stay with me... you need to apply an "image" to it during the boot process? | 09:06 |
| *** kbyrne has quit IRC | 09:06 | |
| *** EinstCrazy has joined #openstack-nova | 09:06 | |
| _gryf | johnthetubaguy, for a specific acceleration - yes, it is kind of image, which have to be pass to the software which program the fpga | 09:06 |
| _gryf | in my stupid poc I've been using glance as storage, though | 09:07 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: neutron: Update the port with a MAC address for PFs https://review.openstack.org/262586 | 09:07 |
| johnthetubaguy | _gryf: I think that actually makes sense, its just a specific "image" type | 09:07 |
| *** EinstCra_ has quit IRC | 09:07 | |
| _gryf | johnthetubaguy, yup | 09:07 |
| _gryf | the only difference is, that it doesn;t go to the vm directly | 09:08 |
| johnthetubaguy | hmm, it feels like it might have to be a new concept we expose via the API, but scheduler wise it sounds very like SR-IOV | 09:09 |
| _gryf | I really like to discuss how it might fit into nova during the summit. | 09:10 |
| *** kbyrne has joined #openstack-nova | 09:11 | |
| * _gryf is looking for the right etherpad | 09:11 | |
| *** markvoelker has joined #openstack-nova | 09:11 | |
| vtech | Hello all, just out of curiosity, is there any reason why cold VM migration doesn't support destination hypervisor selection? This feature would be very useful in OpenStack Watcher and and the current adhoc implementation (https://github.com/openstack/watcher/blob/master/watcher/common/nova_helper.py#L57) in Watcher directly doesn't make much sense as I think it should be in nova. | 09:12 |
| *** PaulMurray has joined #openstack-nova | 09:13 | |
| *** sdake has quit IRC | 09:13 | |
| *** markvoelker has quit IRC | 09:16 | |
| PaulMurray | good morning | 09:16 |
| _gryf | johnthetubaguy, is meetup and unconference section in https://etherpad.openstack.org/p/newton-nova-summit-ideas alright for the topic, or i get it completely wrong? | 09:19 |
| bauzas | vtech: the current destination option for live-mig is really bad now, I would want to fix that first before proposing a new way for cold-mig | 09:20 |
| johnthetubaguy | _gryf: yeah, I think that would work | 09:21 |
| _gryf | johnthetubaguy, thanks | 09:21 |
| johnthetubaguy | _gryf: its good to get it on the etherpad, for sure | 09:21 |
| *** jaosorior has quit IRC | 09:21 | |
| johnthetubaguy | bauzas: +1 that | 09:21 |
| johnthetubaguy | vtech: is there a reason watcher doesn't live-migrate the VM? | 09:22 |
| vtech | johnthetubaguy, well I can imagine cases where some of the VMs are no in active state, but you want to move them anyway. | 09:22 |
| johnthetubaguy | vtech: ah, yep, thats a good reason to want it | 09:23 |
| vtech | bauzas, does it mean you plan to add an option to choose destination hypervisor for cold migration? | 09:23 |
| bauzas | vtech: well, what I actually don't like is to see Watcher proposing a destination without asking the scheduler | 09:23 |
| bauzas | vtech: I know Watcher a little and I think something is perhaps wrong with that logic | 09:24 |
| *** zhangkepi has quit IRC | 09:24 | |
| bauzas | vtech: given Nova will expose a new scheduler REST API for proposing resources, I wonder if Watcher should still select the destination itself rather than asking the scheduler | 09:24 |
| johnthetubaguy | bauzas: it would be better if it "trusted" the nova scheduler to pick a better place | 09:25 |
| bauzas | johnthetubaguy: that's my point, I can understand Watcher not being sure that the scheduler would give them the best destination | 09:25 |
| bauzas | johnthetubaguy: but in that case, I'd rather discuss on the use-cases to see how we can help them | 09:25 |
| vtech | e.g. we've been working on watcher consolidation strategy, which basically moves VM within the cluster to achieve more energy efficient distribution turning off unused hypervisors. | 09:25 |
| johnthetubaguy | but if it doesn't, then all new instances will be going in the wrong place | 09:25 |
| bauzas | vtech: I did that in the past with Climate/Blazar by using aggregates, you don't need to provide a destination | 09:26 |
| johnthetubaguy | well, we might need scheduler hints I guess | 09:26 |
| vtech | so therefore I would like to be able to move VM ideally no matter what state they are in. | 09:26 |
| bauzas | johnthetubaguy: yeah that's what I did 3 years ago | 09:26 |
| bauzas | vtech: so, see, I totally get your use-case of energy-efficiency | 09:27 |
| *** amotoki has quit IRC | 09:27 | |
| bauzas | vtech: I just feel that you shouldn't try to boil the ocean and calculate all the resource usage | 09:27 |
| johnthetubaguy | maybe here is a better way at looking at things, if the new instances are not landing in the right place, you will never get finished | 09:28 |
| bauzas | I actually tried to boil this in the past, and I finally got convinced that no 3rd party engine can play with resource placement unless Nova dedicates a specific set of hosts to that 3rd party engine | 09:29 |
| johnthetubaguy | identifying candidate servers to shutdown, and disabling them, then asking the scheduler to move the instances is all good stuff, but yeah, I think letting the scheduler pick where (most of the time) would be the best approach | 09:29 |
| johnthetubaguy | bauzas: I think you are correct | 09:30 |
| johnthetubaguy | the slight irony being, we might be close to making that possible, but lets ignore that | 09:30 |
| bauzas | it's possible | 09:30 |
| bauzas | and it was possible in Grizzly | 09:30 |
| bauzas | using aggregates and a specific out-of-tree filter | 09:31 |
| bauzas | but I don't want to solve that problem | 09:31 |
| bauzas | here, the point is, I got Watcher's idea to try to rearrange things | 09:32 |
| bauzas | I just feel Watcher should trigger on metrics and action on Nova | 09:32 |
| bauzas | but leave Nova itself do the selection | 09:32 |
| vtech | I still think watcher makes a lot of sense, as you can change your cluster workload distribution dynamically using different strategies. | 09:33 |
| bauzas | vtech: and I do | 09:33 |
| bauzas | vtech: I'm not saying Watcher is useless | 09:33 |
| bauzas | vtech: I certainly support it even | 09:33 |
| bauzas | vtech: I just think that selecting the host is not a Watcher thing | 09:34 |
| bauzas | (or you shard your cloud between what Nova manages and what Watcher manages) | 09:34 |
| *** e0ne has quit IRC | 09:36 | |
| vtech | yep, I understand your point. But for now, given the status and directions of both, having a possibility to choose destination hypervisor at cold migration would be very helpful. | 09:36 |
| vtech | The problem is | 09:37 |
| vtech | when watcher produces a list of actions, lets say few migrations followed by turning off unused host | 09:38 |
| vtech | it could happened VM doesn't end up at the destination which is expected and "predicted" by the strategy | 09:39 |
| vtech | if it relies on nova shceduler to do the placement. | 09:39 |
| bauzas | vtech: well, Watcher could still produce a list of outputs to Nova asking to move those instances | 09:39 |
| bauzas | without specifying a destination | 09:39 |
| bauzas | oh, I see | 09:40 |
| bauzas | interesting | 09:40 |
| bauzas | what Watcher could do is to ask the scheduler to move those instances, but giving hints | 09:41 |
| johnthetubaguy | bauzas: +1 thats what I was saying really, its about influencing the scheduler to make a good enough decision | 09:41 |
| bauzas | mmmm | 09:41 |
| bauzas | that's weights | 09:42 |
| vtech | the problem is different use cases will have different "good" decicisions | 09:42 |
| johnthetubaguy | if you fixate on making sure new instances go in the correct place, based on what watcher knows, I think thats interesting | 09:42 |
| *** amotoki has joined #openstack-nova | 09:42 | |
| vtech | and that's exactly what watcher decision engine is solving | 09:42 |
| johnthetubaguy | vtech: right, and the scheduler needs configuring for different use cases, and has hints for different use cases | 09:42 |
| bauzas | vtech: well, that's a NP-complex problem you know | 09:43 |
| bauzas | vtech: so, we can try to solve it by doing multiple iterations | 09:43 |
| vtech | yep, but we did a first fit based heuristic which works reasonably well | 09:43 |
| vtech | yep, we did it. | 09:44 |
| bauzas | vtech: how do you know that the metrics your heuristics are based upon are correct? | 09:44 |
| johnthetubaguy | so lets go into specifics for a second... turning off some servers | 09:44 |
| johnthetubaguy | watcher could pick the best servers to turn off | 09:44 |
| johnthetubaguy | then disable those hosts | 09:44 |
| *** amotoki has quit IRC | 09:44 | |
| johnthetubaguy | then tell the scheduler to move the VMs on those hosts to a "new" place, knowing the disabled hosts will not be picked | 09:44 |
| johnthetubaguy | then you turn off the hosts | 09:44 |
| vtech | https://github.com/openstack/watcher/blob/master/watcher/decision_engine/strategy/strategies/vm_workload_consolidation.py | 09:44 |
| *** lpetrut has joined #openstack-nova | 09:45 | |
| bauzas | johnthetubaguy: we approached that in the past | 09:45 |
| *** mvk has joined #openstack-nova | 09:46 | |
| bauzas | johnthetubaguy: that's a bit suboptimal | 09:46 |
| *** amotoki has joined #openstack-nova | 09:46 | |
| bauzas | johnthetubaguy: hence our use of aggregates | 09:46 |
| bauzas | which is faster | 09:46 |
| bauzas | coupled with the fact that we were not having forced_down in Grizzly :p | 09:47 |
| *** e0ne has joined #openstack-nova | 09:47 | |
| johnthetubaguy | disabled should be enough here | 09:47 |
| vtech | bauzas: well it's based on telemetry + nova utillzation data and approximate the values as the workloads moves across the cluster | 09:47 |
| bauzas | johnthetubaguy: no, not really | 09:47 |
| bauzas | johnthetubaguy: because if you disable an host, it won't accept new instances | 09:48 |
| bauzas | johnthetubaguy: so you reduce your capacity | 09:48 |
| johnthetubaguy | agreed its suboptimal here, but my question is really, what cases do we really need to worry about, and how do we improve them | 09:48 |
| * PaulMurray is listening in johnthetubaguy bauzas vtech | 09:48 | |
| johnthetubaguy | bauzas: right, I am talking about trying to turn off a hypervisor to save power, in this case | 09:48 |
| johnthetubaguy | bauzas: so disable is fine, other cases we could probably enable after we move the VMs | 09:49 |
| johnthetubaguy | hmm, I didn't mean "disable is fine" | 09:49 |
| vtech | heh, didn't expect such a strong feedback for a simple question :D | 09:49 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: config options: Centralise 'image_file_url' options https://review.openstack.org/298799 | 09:50 |
| *** amotoki has quit IRC | 09:51 | |
| *** artom has joined #openstack-nova | 09:51 | |
| bauzas | johnthetubaguy: well, that can work on a lease concept where you know how long the instance will stay on the host | 09:51 |
| johnthetubaguy | vtech: its a problem we call care about getting fixed, and have been thinking about it for a while, I guess we haven't shared too much | 09:51 |
| johnthetubaguy | bauzas: I am confused, I am thinking about the one use cases of turning off unused hypervisors, just as a specific example | 09:51 |
| bauzas | johnthetubaguy: sorry if I got you confused | 09:52 |
| PaulMurray | johnthetubaguy, bauzas vtech all this is picking around the edges of the problem - there shouldn't be two different decision processes (scheduelr and external optimizer) | 09:52 |
| bauzas | johnthetubaguy: I approached that problem in the past so I have a bit of background that I hardly explain | 09:52 |
| bauzas | johnthetubaguy: in particular energy efficiency | 09:52 |
| johnthetubaguy | bauzas: I mean I get reservations for some of the use cases, just not this one | 09:52 |
| vtech | the think is the load of VMs is dynamic and my change, then you want to do the optimization not only once | 09:52 |
| bauzas | johnthetubaguy: I'm not talking of asking Watcher to do reservations, I just feel it's hard to do optimization on an non-finite resource usage | 09:53 |
| bauzas | PaulMurray: I agree with you, that's what I said at the beginning of the convo | 09:54 |
| vtech | so scheduler can do a very good placement of the VM at the beggining but if the load changes there can be better placement. | 09:54 |
| bauzas | PaulMurray: either Watches takes all credit for managing pools of hosts, or it leaves the scheduler doing that | 09:54 |
| bauzas | vtech: sure, and that's why I like Watcher | 09:54 |
| johnthetubaguy | PaulMurray: so the way I thought about this before (which might be totally wrong), is that external optimizer need to spot when things are "out of balance" and ask the scheduler to pick a better slot, but the scheduler picks the slot for both initial boot and when replacing, based on information its given | 09:55 |
| *** amotoki has joined #openstack-nova | 09:56 | |
| bauzas | johnthetubaguy: a POC could be just some hints passed to the scheduler that a weigher could consume | 09:56 |
| *** amotoki has quit IRC | 09:56 | |
| bauzas | but we don't accept hints for move operations | 09:56 |
| PaulMurray | I guess the scheduelr can be set up to favour groups of hosts and watcher can provide that group as input | 09:56 |
| vtech | bauzas, and this part is very unlikely to be managed by nova-scheduler. | 09:56 |
| johnthetubaguy | bauzas: yeah, I am thinking we should accept hints for the move, and the destination choice is just one such hint (like boot) | 09:57 |
| bauzas | PaulMurray: yeah, aggregates is one way of doing this | 09:57 |
| *** daemontool has joined #openstack-nova | 09:57 | |
| vtech | ok guys, I have to go, thanks for the feedback. | 09:57 |
| bauzas | and Watcher can play with the host placement within aggregates to force the destination | 09:57 |
| *** achanda has joined #openstack-nova | 09:58 | |
| bauzas | that's one way of doing affinity without explicitly passing hints to the move operation | 09:58 |
| PaulMurray | bauzas, the case I'm thinking of is cooling - there is no need to turn hosts off for that - but will want to favor parts of the data centre | 09:58 |
| *** amotoki has joined #openstack-nova | 09:59 | |
| bauzas | PaulMurray: yeah, I know | 09:59 |
| johnthetubaguy | vtech: to answer your question, I think migrate should take a host parameter, I guess we are just not sure you should use it | 09:59 |
| johnthetubaguy | for cooling, you can get the metrics system in the scheduler know whats happening, I think | 09:59 |
| johnthetubaguy | then the external system can find the "hot" vms in the hot area, and ask for a new location for them, or something like that | 10:00 |
| johnthetubaguy | historically hot vm I guess | 10:00 |
| bauzas | I think that's what Watcher does | 10:00 |
| bauzas | they use the notification bus to get metrics | 10:00 |
| *** haomaiwang has quit IRC | 10:01 | |
| bauzas | and detect the hot spots | 10:01 |
| vtech | right, but this is more a question of dividing scheduling between watcher and nova which I cannot answer. For now in the way how watcher and nova currently works, using VM migration for further cluster optimization is a good enough solution for us. | 10:01 |
| bauzas | that detection pattern provides a list of destinations AFAICU | 10:01 |
| *** haomaiwang has joined #openstack-nova | 10:01 | |
| PaulMurray | bauzas, johnthetubaguy I think the question will be how to influence the scheduler - I'm not sure it would be on a per request basis | 10:01 |
| PaulMurray | vtech, ^ | 10:01 |
| johnthetubaguy | if scheduler is making sure new VMs are going in the right place, it should be able to find a better place for the existing one | 10:01 |
| openstackgerrit | Rui Chen proposed openstack/python-novaclient: Fix ServerGroup.NAME_ATTR https://review.openstack.org/299093 | 10:02 |
| bauzas | PaulMurray: well, Watcher could influence a resource API to modify the weights | 10:02 |
| *** achanda has quit IRC | 10:02 | |
| bauzas | PaulMurray: that would necessarly benefit for new instances | 10:03 |
| bauzas | PaulMurray: you know, that new resource API we're committed to deliver :) | 10:03 |
| * bauzas that Neutron thread on naming on making me ROFL | 10:04 | |
| PaulMurray | bauzas, yes, I'm following | 10:04 |
| PaulMurray | vtech, I'll check out watcher - do you have a pointer for more info (appart from github) | 10:04 |
| bauzas | like I said in the past, Watcher doesn't really need a destination field for cold-mig, they can use the aggregates API to manipulate the destinations | 10:05 |
| bauzas | https://wiki.openstack.org/wiki/Watcher | 10:05 |
| *** allen_gao has quit IRC | 10:05 | |
| PaulMurray | bauzas, thanks | 10:05 |
| *** markvoelker has joined #openstack-nova | 10:06 | |
| *** e0ne has quit IRC | 10:08 | |
| *** allen_gao has joined #openstack-nova | 10:10 | |
| *** EinstCrazy has quit IRC | 10:11 | |
| *** vilobhmm11 has quit IRC | 10:11 | |
| *** ihrachys has joined #openstack-nova | 10:12 | |
| *** markvoelker has quit IRC | 10:12 | |
| *** e0ne has joined #openstack-nova | 10:12 | |
| *** sdague has joined #openstack-nova | 10:14 | |
| *** amotoki has quit IRC | 10:16 | |
| *** hoangcx has quit IRC | 10:16 | |
| *** cdent has quit IRC | 10:17 | |
| PaulMurray | bauzas, johnthetubaguy I have a patch to remove ERT - should relnotes go in upgrade and features ? | 10:21 |
| bauzas | PaulMurray: technically, it's no longer a deprecation, right? | 10:22 |
| johnthetubaguy | PaulMurray: a remove sounds like just an upgrade thing to me | 10:22 |
| *** ljxiash_ has quit IRC | 10:22 | |
| bauzas | PaulMurray: maybe even if that, putting it in "deprecations" section as a reminder + some upgrade section could be nice | 10:22 |
| PaulMurray | bauzas, it was deprecated begining last cycle | 10:22 |
| *** ljxiash has joined #openstack-nova | 10:22 | |
| bauzas | PaulMurray: yeah I know | 10:22 |
| bauzas | PaulMurray: I just wonder if just putting an upgrade section is okay | 10:23 |
| PaulMurray | bauzas, does depreications include things that are already gone ? | 10:23 |
| bauzas | PaulMurray: technically, no :( | 10:23 |
| bauzas | "Deprecations should not be used for something that is removed in the release, use upgrade section instead. " | 10:24 |
| bauzas | http://docs.openstack.org/developer/reno/usage.html#creating-new-release-notes | 10:24 |
| bauzas | PaulMurray: okay, I guess you have your answer... | 10:24 |
| *** daemontool_ has joined #openstack-nova | 10:24 | |
| PaulMurray | thanks - I could have found that but to be honest I never find what I'm looking for | 10:24 |
| bauzas | yeah, I'm usually lazy too :p | 10:26 |
| bauzas | but I'm French, so I guess it's something people already know :p | 10:26 |
| *** ljxiash has quit IRC | 10:27 | |
| *** daemontool has quit IRC | 10:27 | |
| bauzas | oh oops, my manners | 10:27 |
| *** bauzas is now known as bauwser | 10:27 | |
| *** smatzek has joined #openstack-nova | 10:29 | |
| *** ociuhandu has joined #openstack-nova | 10:31 | |
| *** zhangkepi has joined #openstack-nova | 10:33 | |
| *** gcb has quit IRC | 10:37 | |
| *** gcb has joined #openstack-nova | 10:38 | |
| *** alaski is now known as lascii | 10:38 | |
| *** zhangkepi has quit IRC | 10:39 | |
| openstackgerrit | Artom Lifshitz proposed openstack/python-novaclient: Virtual device tagging client support https://review.openstack.org/299222 | 10:40 |
| *** cdent has joined #openstack-nova | 10:43 | |
| *** amotoki has joined #openstack-nova | 10:45 | |
| openstackgerrit | Sean Dague proposed openstack/nova: WIP: demonstrate using one of the new yaml files https://review.openstack.org/300403 | 10:46 |
| *** e0ne has quit IRC | 10:47 | |
| *** e0ne has joined #openstack-nova | 10:52 | |
| vtech | PaulMurray I would start with the official Watcher documentation, or join #openstack-watcher channel | 10:59 |
| *** haomaiwang has quit IRC | 11:01 | |
| *** haomaiwang has joined #openstack-nova | 11:01 | |
| openstackgerrit | John Garbutt proposed openstack/nova: Config options: Centralize compute options https://review.openstack.org/295938 | 11:02 |
| *** markvoelker has joined #openstack-nova | 11:02 | |
| *** e0ne has quit IRC | 11:04 | |
| *** markvoelker has quit IRC | 11:06 | |
| *** Marga_ has joined #openstack-nova | 11:07 | |
| *** liunian_ has quit IRC | 11:08 | |
| *** e0ne has joined #openstack-nova | 11:08 | |
| *** haomaiwang has quit IRC | 11:10 | |
| *** e0ne has quit IRC | 11:11 | |
| *** Marga_ has quit IRC | 11:11 | |
| *** GheRivero has joined #openstack-nova | 11:11 | |
| *** baoli has joined #openstack-nova | 11:12 | |
| openstackgerrit | Paul Murray proposed openstack/nova: Remove extensible resource tracking https://review.openstack.org/300420 | 11:13 |
| *** liunian_ has joined #openstack-nova | 11:13 | |
| PaulMurray | johnthetubaguy, bauwser FYI ^^ | 11:13 |
| *** baoli has quit IRC | 11:16 | |
| *** liunian_ has quit IRC | 11:17 | |
| PaulMurray | vtech, thanks | 11:18 |
| * artom is assuming dansmith isn't awake yet... | 11:18 | |
| sdague | artom: that's a good assumption. He's an early riser, but not this early | 11:19 |
| *** tbachman_ has joined #openstack-nova | 11:21 | |
| *** tbachman has quit IRC | 11:22 | |
| *** tbachman_ is now known as tbachman | 11:22 | |
| PaulMurray | artom when he does wake up he will probably be called superdan (its a friday) | 11:28 |
| *** achanda has joined #openstack-nova | 11:30 | |
| artom | PaulMurray, thanks :) I'm aware of the Friday nick craziness :) | 11:31 |
| artom | Hawaiian shirts for IRC :D | 11:31 |
| *** e0ne has joined #openstack-nova | 11:33 | |
| *** amotoki has quit IRC | 11:34 | |
| *** thorst has joined #openstack-nova | 11:35 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Enhanced error handling for rest_parameters parser https://review.openstack.org/300427 | 11:35 |
| *** achanda has quit IRC | 11:35 | |
| *** daemontool_ is now known as daemontool | 11:37 | |
| * tbachman notes its an extra special Friday — Apirl Fool’s Friday NICKs anyone? | 11:38 | |
| *** rk4n has joined #openstack-nova | 11:38 | |
| openstackgerrit | Sean Dague proposed openstack/nova: WIP: demonstrate using one of the new yaml files https://review.openstack.org/300403 | 11:39 |
| *** rtheis has joined #openstack-nova | 11:40 | |
| *** baoli has joined #openstack-nova | 11:42 | |
| openstackgerrit | Prateek Arora proposed openstack/nova: Delete traces of in-progress snapshot on VM being deleted https://review.openstack.org/294513 | 11:42 |
| *** baoli_ has joined #openstack-nova | 11:45 | |
| *** rook-lappy has joined #openstack-nova | 11:46 | |
| *** baoli has quit IRC | 11:48 | |
| *** e0ne has quit IRC | 11:50 | |
| *** Goneri has quit IRC | 11:50 | |
| snikitin | johnthetubaguy: Hi! Can I ask you to reapprove instance tags spec for newton? https://review.openstack.org/#/c/299160/ | 11:52 |
| *** rfolco has joined #openstack-nova | 11:53 | |
| *** e0ne has joined #openstack-nova | 11:54 | |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP resource-pools/placement api https://review.openstack.org/293104 | 11:56 |
| *** e0ne has quit IRC | 11:57 | |
| openstackgerrit | Yuiko Takada proposed openstack/nova: Fix not supported error message https://review.openstack.org/300435 | 12:02 |
| *** mvk_ has joined #openstack-nova | 12:03 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Enhanced error handling for rest_parameters parser https://review.openstack.org/300427 | 12:06 |
| *** mvk has quit IRC | 12:07 | |
| *** FL1SK has quit IRC | 12:08 | |
| openstackgerrit | Yuiko Takada proposed openstack/nova: Fix error message of nova baremetal-node-delete https://review.openstack.org/300437 | 12:08 |
| *** e0ne has joined #openstack-nova | 12:09 | |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: [WIP] Add console log offset https://review.openstack.org/292473 | 12:11 |
| *** gcb has quit IRC | 12:11 | |
| *** ducttape_ has joined #openstack-nova | 12:12 | |
| *** haomaiwang has joined #openstack-nova | 12:16 | |
| *** gszasz has quit IRC | 12:16 | |
| *** mvk_ has quit IRC | 12:17 | |
| *** ljxiash has joined #openstack-nova | 12:20 | |
| *** liusheng has quit IRC | 12:23 | |
| *** liusheng has joined #openstack-nova | 12:23 | |
| *** ljxiash has quit IRC | 12:24 | |
| *** ljxiash has joined #openstack-nova | 12:24 | |
| snikitin | johnthetubaguy: John, thank you for spec! And could you remove your -2 from tags patches, please? https://review.openstack.org/#/c/128940/ https://review.openstack.org/#/c/248805/ https://review.openstack.org/#/c/268932/ | 12:26 |
| *** aysyd has joined #openstack-nova | 12:27 | |
| *** ducttape_ has quit IRC | 12:27 | |
| *** cdelatte has joined #openstack-nova | 12:29 | |
| *** wuhg has quit IRC | 12:29 | |
| *** erlon has joined #openstack-nova | 12:32 | |
| *** dave-mccowan has joined #openstack-nova | 12:32 | |
| *** achanda has joined #openstack-nova | 12:32 | |
| *** zehua has quit IRC | 12:34 | |
| *** zehua has joined #openstack-nova | 12:36 | |
| *** achanda has quit IRC | 12:37 | |
| johnthetubaguy | snikitin: cool, its merged, I think we have everything in place now | 12:39 |
| snikitin | johnthetubaguy: huge thank! | 12:39 |
| *** pradk has joined #openstack-nova | 12:40 | |
| johnthetubaguy | snikitin: no problem, thanks for reaching out | 12:40 |
| sdague | johnthetubaguy: if you could reset -2s on all of these - https://review.openstack.org/#/q/topic:bp/use-glance-v2-api+status:open that would be great as well | 12:40 |
| johnthetubaguy | sdague: that one doesn't have the spec approved yet, how did you think we should deal with that one? | 12:41 |
| *** devananda has quit IRC | 12:43 | |
| sdague | we know we are going to move forward with glance v2 support in some way, right? | 12:43 |
| openstackgerrit | Srushti Gadadare proposed openstack/nova: Fix invalid uuid warnings in compute api testcases https://review.openstack.org/300453 | 12:43 |
| *** devananda has joined #openstack-nova | 12:44 | |
| sdague | I get that all the details aren't sorted, but unless you think we aren't going to do it at all, -2ing all the patches seems odd | 12:44 |
| sdague | because we could land some infrastructure that's uncontroversial I think which narrows the problem | 12:44 |
| *** mpavone has joined #openstack-nova | 12:46 | |
| *** dims_ has quit IRC | 12:49 | |
| johnthetubaguy | sdague: its just we normally don't remove -2 until a blueprint is approved | 12:49 |
| *** dims has joined #openstack-nova | 12:49 | |
| *** mvk_ has joined #openstack-nova | 12:50 | |
| openstackgerrit | Paul Murray proposed openstack/nova: Remove extensible resource tracking https://review.openstack.org/300420 | 12:50 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: config options: Centralize 'nova.rpc' options https://review.openstack.org/281942 | 12:50 |
| openstackgerrit | Cédric LECOMTE proposed openstack/nova: Archive instance_actions and instance_actions_event https://review.openstack.org/299474 | 12:50 |
| johnthetubaguy | sdague: agreed we should merge stuff that we can make progress on earlier rather than later, just unsure how to track that with the tooling we have | 12:50 |
| *** swamireddy has quit IRC | 12:51 | |
| sdague | ok, I guess I misunderstood and didn't realize we leave everything in a locked up state that long | 12:51 |
| johnthetubaguy | normally any patch that references an unapproved bp gets a -2, at any point in the cycle really | 12:52 |
| johnthetubaguy | the idea was really just so its easier to focus reviews on the approved stuff | 12:52 |
| *** sahid has joined #openstack-nova | 12:53 | |
| johnthetubaguy | but I am not sure that process is totally working for us, for stuff like this where there is stuff we should just land | 12:53 |
| *** rlrossit has joined #openstack-nova | 12:53 | |
| *** edmondsw has joined #openstack-nova | 12:53 | |
| *** zengine has joined #openstack-nova | 12:54 | |
| sdague | ok, it seems there is no reproposed spec at this point | 12:54 |
| sdague | I can't even find the last one in gerrit | 12:55 |
| sdague | only the liberty one | 12:55 |
| *** pmathews has joined #openstack-nova | 12:57 | |
| openstackgerrit | Srushti Gadadare proposed openstack/nova: Fix invalid uuid warnings https://review.openstack.org/300458 | 12:57 |
| *** e0ne has quit IRC | 12:57 | |
| *** pmathews has quit IRC | 12:57 | |
| *** vladikr has quit IRC | 12:57 | |
| sdague | ok, I guess I'll just start reproposing it | 12:57 |
| *** vladikr has joined #openstack-nova | 12:57 | |
| sdague | though, honestly, it would probably be better if someone doing the work is going to propose that | 12:58 |
| *** akshai has joined #openstack-nova | 13:00 | |
| *** mtanino has quit IRC | 13:01 | |
| *** haomaiwang has quit IRC | 13:01 | |
| *** edleafe is now known as figleaf | 13:01 | |
| *** haomaiwang has joined #openstack-nova | 13:01 | |
| *** akshai_ has joined #openstack-nova | 13:02 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: handle reserved pages size https://review.openstack.org/292500 | 13:02 |
| openstackgerrit | sahid proposed openstack/nova: virt: reserved number of mempages on compute host https://review.openstack.org/292499 | 13:02 |
| *** eharney has joined #openstack-nova | 13:03 | |
| *** swamireddy has joined #openstack-nova | 13:05 | |
| *** akshai has quit IRC | 13:05 | |
| *** amotoki has joined #openstack-nova | 13:06 | |
| *** cdelatte has quit IRC | 13:08 | |
| *** EinstCrazy has joined #openstack-nova | 13:09 | |
| *** cleong has joined #openstack-nova | 13:11 | |
| *** aysyd has quit IRC | 13:11 | |
| *** links has quit IRC | 13:12 | |
| openstackgerrit | Srushti Gadadare proposed openstack/nova: Fix invalid uuid warnings in api testcases https://review.openstack.org/300477 | 13:12 |
| *** READ10 has joined #openstack-nova | 13:13 | |
| *** ccarmack has joined #openstack-nova | 13:14 | |
| *** ctrath has joined #openstack-nova | 13:14 | |
| *** Goneri has joined #openstack-nova | 13:14 | |
| *** e0ne has joined #openstack-nova | 13:15 | |
| *** aysyd has joined #openstack-nova | 13:15 | |
| *** cdelatte has joined #openstack-nova | 13:17 | |
| *** mjura has quit IRC | 13:18 | |
| *** mjura has joined #openstack-nova | 13:20 | |
| *** ivase has joined #openstack-nova | 13:22 | |
| *** salv-orlando has joined #openstack-nova | 13:23 | |
| *** dansmith is now known as bizarroman | 13:25 | |
| *** mnestratov|2 has quit IRC | 13:25 | |
| openstackgerrit | Srushti Gadadare proposed openstack/nova: Fix invalid uuid warnings in cell api testcases https://review.openstack.org/300482 | 13:25 |
| gibi | sahid: hi! I'm looking at your last question in https://review.openstack.org/#/c/292499/ | 13:25 |
| *** bizarroman is now known as bizarrodan | 13:25 | |
| bauwser | heh | 13:26 |
| *** burt has joined #openstack-nova | 13:26 | |
| bauwser | April Fools Day nick ? | 13:26 |
| gibi | sahid: If I remember correctly I did restarted all nova services when tested patchset 11 alone. | 13:26 |
| bizarrodan | https://en.wikipedia.org/wiki/Bizarro | 13:26 |
| *** salv-orl_ has quit IRC | 13:26 | |
| * bauwser lawn_bauwer | 13:26 | |
| *** bauwser is now known as lawn_bauwer | 13:27 | |
| *** clenimar has joined #openstack-nova | 13:27 | |
| gibi | sahid: but then I tested the two pach together and it worked fine. So now that you squashed the two patch together I have no problem with it. | 13:27 |
| *** dims_ has joined #openstack-nova | 13:27 | |
| *** e0ne has quit IRC | 13:29 | |
| *** pmathews has joined #openstack-nova | 13:29 | |
| *** EinstCrazy has quit IRC | 13:30 | |
| *** xyang1 has joined #openstack-nova | 13:32 | |
| *** jistr has quit IRC | 13:32 | |
| *** openstackgerrit has quit IRC | 13:33 | |
| *** openstackgerrit has joined #openstack-nova | 13:34 | |
| *** achanda has joined #openstack-nova | 13:34 | |
| *** krtaylor has quit IRC | 13:35 | |
| *** salv-orlando has quit IRC | 13:35 | |
| *** vtech has quit IRC | 13:36 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add Allocation and AllocationList objects https://review.openstack.org/282442 | 13:36 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add ResourcePool and ResourcePoolList objects https://review.openstack.org/284963 | 13:36 |
| *** vtech has joined #openstack-nova | 13:36 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add a get_by_uuid for aggregates https://review.openstack.org/284895 | 13:36 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add name to ResourceProvider object https://review.openstack.org/281945 | 13:36 |
| *** aysyd has quit IRC | 13:37 | |
| sahid | gibi: hum actually i splitted them back | 13:38 |
| *** Bjoern has joined #openstack-nova | 13:38 | |
| sahid | gibi: we have discussed with dansmith about the issue you have suffered and we should to understand how that happened | 13:39 |
| *** achanda has quit IRC | 13:39 | |
| *** zz_dimtruck is now known as dimtruck | 13:39 | |
| *** mjura has quit IRC | 13:39 | |
| sahid | gibi: can you try to make an other test and see if you can to reproduce it ? | 13:39 |
| *** rodrigods has quit IRC | 13:39 | |
| *** rodrigods has joined #openstack-nova | 13:39 | |
| *** diogogmt has joined #openstack-nova | 13:40 | |
| openstackgerrit | Chuck Carmack proposed openstack/nova: WIP: Change the nova tempest blacklist to use to idempotent ids https://review.openstack.org/298437 | 13:40 |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP resource-pools/placement api https://review.openstack.org/293104 | 13:41 |
| *** aysyd has joined #openstack-nova | 13:41 | |
| gibi | sahid: Ok, I see. I will re-test it in my devstack and report back on the review | 13:42 |
| *** diogogmt has quit IRC | 13:46 | |
| sdague | johnthetubaguy: ok, I guess I'm starting with just an email to the list about current state and moving forward | 13:46 |
| *** markvoelker has joined #openstack-nova | 13:47 | |
| *** mriedem has joined #openstack-nova | 13:47 | |
| johnthetubaguy | sdague: OK, that sounds good, I honestly think its worth having a joint design summit session with glance | 13:47 |
| *** haukebruno has quit IRC | 13:48 | |
| mriedem | we have that planned already, at least it's in the nova schedule | 13:48 |
| johnthetubaguy | it just feels like we keep coming away with different ideas of what has been agreed each time | 13:48 |
| johnthetubaguy | mriedem: coolness | 13:48 |
| mriedem | nikhil: has an etherpad with details | 13:48 |
| johnthetubaguy | ah, right, I remember that coming up now | 13:48 |
| mriedem | otherwise i just joined so have no idea what you're talking about :) | 13:49 |
| *** EinstCrazy has joined #openstack-nova | 13:50 | |
| *** e0ne has joined #openstack-nova | 13:51 | |
| *** markvoelker has quit IRC | 13:51 | |
| *** mtanino has joined #openstack-nova | 13:51 | |
| *** lascii is now known as lutf-8 | 13:52 | |
| sdague | johnthetubaguy: I agree | 13:53 |
| sdague | mriedem: yeh, but the concern is the etherpad is kind of steps 10 - 20 | 13:53 |
| bizarrodan | mriedem: the api never catches flavornotfoundbyname, only flavornotfound which is the parent.. so why do you think it matters what I raise there? | 13:53 |
| *** esracelik has quit IRC | 13:54 | |
| sdague | it kind of assumes the 8 patches up right now will land in their current state, and I don't think that they will (and that was basically the issue last time around) | 13:54 |
| mriedem | bizarrodan: because in the case that flavor_id is None, it's raising FlavorNotFound(flavorid=None) which is going to make the error message kind of useless | 13:54 |
| mriedem | 'flavor not found with id None' | 13:54 |
| mriedem | vs 'flvaor not foud with name foo' | 13:58 |
| bizarrodan | ah, okay | 13:58 |
| *** mdrabe has joined #openstack-nova | 13:58 | |
| *** aysyd has quit IRC | 13:58 | |
| *** pmathews has quit IRC | 13:58 | |
| mriedem | sdague: sure, but the etherpad does say to sort that out, | 13:58 |
| sdague | mriedem: sure | 13:58 |
| mriedem | sdague: there are like 4 items in that etherpad and i don't think we're going to get past #1 in a design session | 13:58 |
| mriedem | which is the one we care about | 13:58 |
| sdague | right | 13:58 |
| sdague | so, I feel like unless we get a bunch of this sorted out prior to austin, we are unlikely to get much productivity out of that session | 13:58 |
| *** parora has quit IRC | 13:58 | |
| *** openstack has quit IRC | 13:58 | |
| *** openstack has joined #openstack-nova | 14:00 | |
| artom | bizarrodan, when you get a chance can you give me your opinion on my last two comments on https://review.openstack.org/#/c/264015/22/nova/objects/virt_device_metadata.py ? | 14:01 |
| *** haomaiwang has quit IRC | 14:01 | |
| artom | (The registered objects stuff) | 14:01 |
| *** artom is now known as notartom | 14:01 | |
| sdague | I was mostly trying to point to infrastructure patches that would help move things forward regardless | 14:01 |
| *** daemontool_ is now known as daemontool | 14:01 | |
| nikhil | Thanks for your email sdague, I guess I need some code digging to do to get the context. | 14:01 |
| mriedem | sdague: yup, just read it | 14:01 |
| *** haomaiwang has joined #openstack-nova | 14:01 | |
| mriedem | thanks for sending that | 14:01 |
| nikhil | mriedem: johnthetubaguy: here's the etherpad link https://etherpad.openstack.org/p/nova-glance-newton | 14:01 |
| nikhil | +1 on the cross-session | 14:01 |
| mriedem | nikhil: http://lists.openstack.org/pipermail/openstack-dev/2016-April/091145.html | 14:02 |
| nikhil | mriedem: ack, I am planning to sync with interest glance folks and do my own digging | 14:03 |
| mriedem | nikhil: can we agree to get the spec cleaned up and re-proposed by middle of next week (wed at the latest)? | 14:04 |
| *** erlon has quit IRC | 14:04 | |
| *** FL1SK has joined #openstack-nova | 14:04 | |
| *** yamahata has joined #openstack-nova | 14:05 | |
| nikhil | I just read up on the follow up log, makes sense to start with the email. | 14:05 |
| *** erlon has joined #openstack-nova | 14:05 | |
| *** pushkaru has joined #openstack-nova | 14:06 | |
| nikhil | mriedem: I was hoping that a wip spec be up already but mike is/was caught in internal work and he wanted to propose. I will propose one early next week otherwise . | 14:06 |
| *** sdake has joined #openstack-nova | 14:06 | |
| nikhil | mriedem: and to clarify a bit there, you want to preserve the history of the mitaka spec, correct? | 14:06 |
| mriedem | nikhil: yes | 14:07 |
| nikhil | cool | 14:07 |
| mriedem | nikhil: https://specs.openstack.org/openstack/nova-specs/readme.html#previously-approved-specifications | 14:07 |
| *** e0ne has quit IRC | 14:07 | |
| mriedem | nikhil: based on sdague's email, i expect that the re-proposed spec for newton is going to have a dependency on some function getting into glanceclient | 14:08 |
| mriedem | that is, moving it from nova/image/glance to glanceclient | 14:08 |
| nikhil | mriedem: yes and that's why I'm a bit unsure on the changes we need and can commit | 14:08 |
| nikhil | mike was too excited that we did not have to make client changes | 14:08 |
| nikhil | mriedem: I think the proposal sounds great. | 14:09 |
| nikhil | I am mostly worried about the scope of client changes | 14:09 |
| nikhil | just want to avoid anything that will not be common to nova, cinder, ironic etc. | 14:10 |
| nikhil | from first glance, email doesn't seem to indicate that | 14:10 |
| mriedem | version discovery should be in the client anyway https://github.com/openstack/nova/blob/3cdaa30566c17a2add5d9163a0693c97dc1d065b/nova/image/glance.py#L108 | 14:10 |
| nikhil | makes sense | 14:11 |
| ndipanov | mriedem, helloooo | 14:11 |
| mriedem | ndipanov: just +Wed it | 14:11 |
| ndipanov | mriedem, I was totally gonna ask a different thing | 14:11 |
| mriedem | i have a question for you, does https://review.openstack.org/#/c/262586/ complete that blueprint? | 14:11 |
| ndipanov | jk - thanks | 14:11 |
| ndipanov | mriedem, yes | 14:12 |
| mriedem | sweet | 14:12 |
| nikhil | mriedem: sdague: thanks for the feedback, we'll get the ball rolling :) | 14:12 |
| mriedem | i'll close the bp when that change merges today | 14:12 |
| mriedem | nikhil: thanks | 14:12 |
| *** coreywright has quit IRC | 14:14 | |
| *** dimtruck is now known as zz_dimtruck | 14:15 | |
| *** burgerk has joined #openstack-nova | 14:18 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: tests: make XMLMatches work with Python3 https://review.openstack.org/300509 | 14:18 |
| sdague | nikhil: right, I'm fine with things like what image metadata nova uses for what being in nova | 14:18 |
| *** aysyd has quit IRC | 14:18 | |
| sdague | but the discovery stuff is super generic | 14:18 |
| sdague | I feel like: from glanceclient.common should have been a red flag | 14:18 |
| nikhil | sdague: gotcha, and agreed | 14:18 |
| mriedem | cores: easy +W https://review.openstack.org/#/c/294512/ | 14:19 |
| lutf-8 | sdague: can you take a look at https://review.openstack.org/#/c/284905/ at some point. It addresses a concern of yours | 14:20 |
| *** aysyd has joined #openstack-nova | 14:20 | |
| vladikr | http://openstack-in-production.blogspot.ca/2015/03/nova-quota-usage-synchronization.html | 14:20 |
| mriedem | vladikr: what about it? | 14:21 |
| *** pradk has quit IRC | 14:21 | |
| mriedem | vladikr: ccarmack is working that spec | 14:21 |
| mriedem | https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/refresh-quotas-usage | 14:22 |
| vladikr | mriedem, sorry, wrong channel really.. was looking on a bug related to it | 14:22 |
| vladikr | mriedem, thanks | 14:22 |
| mriedem | vladikr: ah, ok | 14:22 |
| *** cdent has quit IRC | 14:22 | |
| sdague | nikhil: https://review.openstack.org/#/c/259097/18/nova/image/glance.py is one of the places where things are stickier, as I'd really like to avoid all that logic on the nova side, as it's pretty substantial | 14:24 |
| sdague | we should probably be at a state to discuss bits like that in detail by austin | 14:24 |
| mriedem | yikes | 14:26 |
| mriedem | if that lives in nova, it should at least be using a factory with subclasses for v1 and v2 impls | 14:26 |
| mriedem | with common code in a base class | 14:26 |
| nikhil | sdague: I see | 14:26 |
| *** efoley has joined #openstack-nova | 14:26 | |
| nikhil | I was thinking that's the conversation john started on the compat layer | 14:27 |
| nikhil | but let's avoid the compat name to prevent further confusion | 14:27 |
| nikhil | I will note that as a research iteam for me (across os projects to see such sticker bits related to glance stuff) | 14:27 |
| sdague | nikhil: right, and this is part of where things broke down. Because ideally common opperations like that should be in glanceclient in my mind | 14:27 |
| *** coreywright has joined #openstack-nova | 14:27 | |
| sdague | even if it's not on Client itself | 14:28 |
| *** rmevans has joined #openstack-nova | 14:28 | |
| sarafraj | good morning | 14:28 |
| sdague | glanceclient.compat could have some methods to provide common approaches to some of the basic things like this | 14:28 |
| mriedem | what is this mythical glanceclient.compat that keeps being talked about? | 14:29 |
| nikhil | mriedem: I think you're right to avoid that discussion. It may be different ideas in different minds, I think. | 14:30 |
| sdague | mriedem: because glanceclient.client is basically raw manipulation of the rest data (as are most of the clients) someone needs to write a glue layer for common ops | 14:30 |
| nikhil | ref. compat layer | 14:30 |
| sdague | either every caller | 14:30 |
| sdague | or somewhere common | 14:30 |
| *** knikolla has joined #openstack-nova | 14:31 | |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Make neutronapi get_floating*() methods return objects https://review.openstack.org/115749 | 14:31 |
| mriedem | ok, but that doesn't exist today right? | 14:31 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Add tests for floating_ip private functions https://review.openstack.org/267768 | 14:31 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Remove DictCompat from FixedIP https://review.openstack.org/265951 | 14:31 |
| sdague | mriedem: correct | 14:31 |
| sdague | and that has been part of the tension on these patches | 14:31 |
| sdague | the giant amount of glance structure internals being pushed into nova | 14:32 |
| mriedem | sdague: ok, my point in https://review.openstack.org/#/c/259097/ was (1) split the create/update changes into 2 patches and (2) create a common base class that does common things, and subclass that for special v1/v2 impl | 14:32 |
| mriedem | so it's not all sitting in one giant module with if/else conditions all over the place | 14:32 |
| sdague | mriedem: sure that's part of it | 14:32 |
| mriedem | if there is a common base class in nova, maybe we start there and then push that down into glanceclient.compat at some point | 14:32 |
| mriedem | idk | 14:32 |
| sdague | sure | 14:33 |
| sdague | regardless, there is some stuff that couple happen up front that would make a bunch of this better | 14:33 |
| sdague | which was kind of my email, at least that clears up some stuff so what remains is a little easier to get in your head all at once | 14:33 |
| *** jeblair_ is now known as jeblair | 14:33 | |
| nikhil | makes sense (baby steps) | 14:34 |
| gibi | sahid: I reproduced the problem, left a comment in the review. I have to leave soon but will be available on Monday | 14:35 |
| *** vishwanathj has joined #openstack-nova | 14:37 | |
| *** pwnall has joined #openstack-nova | 14:39 | |
| sdague | mordred: you are about to own this blueprint if you aren't careful :) | 14:40 |
| *** zz_dimtruck is now known as dimtruck | 14:40 | |
| *** pmathews has joined #openstack-nova | 14:41 | |
| *** markvoelker has joined #openstack-nova | 14:41 | |
| *** ametts has joined #openstack-nova | 14:42 | |
| *** pmathews1 has joined #openstack-nova | 14:42 | |
| nikhil | haha | 14:45 |
| nikhil | +! | 14:45 |
| nikhil | +1 :) | 14:45 |
| *** pmathews has quit IRC | 14:45 | |
| *** jistr has joined #openstack-nova | 14:46 | |
| *** markvoelker has quit IRC | 14:47 | |
| *** pmathews1 has quit IRC | 14:48 | |
| *** nkrinner has quit IRC | 14:48 | |
| openstackgerrit | Tom Patzig proposed openstack/nova-specs: Add "disable root_disk" option to flavors https://review.openstack.org/267673 | 14:50 |
| *** cdent has joined #openstack-nova | 14:50 | |
| *** gjayavelu has joined #openstack-nova | 14:52 | |
| *** cfriesen has joined #openstack-nova | 14:53 | |
| *** mpavone has quit IRC | 14:53 | |
| *** mylu has joined #openstack-nova | 14:53 | |
| *** pmathews has joined #openstack-nova | 14:54 | |
| *** woodster_ has joined #openstack-nova | 14:54 | |
| *** Marga_ has joined #openstack-nova | 14:55 | |
| mriedem | sdague: bizarrodan: https://review.openstack.org/#/c/300284/ | 14:55 |
| mriedem | that's blocking liberty nova changes | 14:55 |
| sdague | +A | 14:56 |
| mriedem | thanks | 14:56 |
| *** rdopiera has quit IRC | 14:56 | |
| *** e0ne has joined #openstack-nova | 14:57 | |
| sahid | mriedem: what do we do for https://review.openstack.org/#/c/284095/ ? I think we do not have real good solution | 14:57 |
| openstackgerrit | Tom Patzig proposed openstack/nova: Feature: Add option to disable root_disk in flavor https://review.openstack.org/280301 | 14:57 |
| *** Marga_ has quit IRC | 14:57 | |
| sahid | i want to report a bug as we know there is a mistake reported by the api in this situation but i do not think we can do something more currently | 14:57 |
| mriedem | sdague: did you see my replies in the ML thread about the port_security_enabled=False stuff with neutron? | 14:58 |
| mriedem | sahid: i don't think we can break the API and then just report a bug saying we'll fix it later | 14:58 |
| *** Marga_ has joined #openstack-nova | 14:58 | |
| sdague | mriedem: yes, though I'm not sure I have the whole thing in my right now | 14:58 |
| mriedem | sahid: although arguably the api is already broken | 14:58 |
| *** delattec has joined #openstack-nova | 14:58 | |
| sahid | mriedem: i don't think we break the api | 14:58 |
| mriedem | sdague: is linking back to servers/<server uuid>/os-security-groups crazy? | 14:59 |
| *** Marga_ has quit IRC | 14:59 | |
| *** Marga_ has joined #openstack-nova | 15:00 | |
| mriedem | that's what we do today for flavors | 15:00 |
| openstackgerrit | Tom Patzig proposed openstack/python-novaclient: Feature: Add option to disable root_disk in flavor https://review.openstack.org/300541 | 15:00 |
| sdague | mriedem: linking back is not crazy | 15:00 |
| *** haomaiwang has quit IRC | 15:01 | |
| sdague | it's probably quite sensible | 15:01 |
| mriedem | sdague: that requires a microversion though right? | 15:01 |
| *** haomaiwa_ has joined #openstack-nova | 15:01 | |
| *** thorst has quit IRC | 15:01 | |
| sdague | yes | 15:01 |
| sdague | security groups are per tenant though, not per server | 15:02 |
| sdague | the container might be slightly different | 15:02 |
| *** thorst has joined #openstack-nova | 15:02 | |
| *** cdelatte has quit IRC | 15:02 | |
| sdague | I think we'd have to think through the new representation | 15:02 |
| sdague | the concept seems fine though | 15:02 |
| mriedem | and would we only do that for neutron | 15:02 |
| mriedem | or also nova-network | 15:03 |
| sdague | almost all the time you are getting info about security groups we need more than the name | 15:03 |
| mriedem | it'd probably be better if it were both in a microversoin | 15:03 |
| sdague | I would do it all the time | 15:03 |
| *** dedery has joined #openstack-nova | 15:03 | |
| mriedem | yeah, i like that approach, it's better than the rest api having to talk to the network api about getting back what might or might not be the list of security groups | 15:04 |
| mriedem | that felt very clunky to me | 15:04 |
| *** dedery has quit IRC | 15:04 | |
| *** dedery has joined #openstack-nova | 15:04 | |
| notartom | bizarrodan, sorry to bug you again, but when you get a chance can you give me your opinion on my last two comments on https://review.openstack.org/#/c/264015/22/nova/objects/virt_device_metadata.py ? | 15:04 |
| bizarrodan | notartom: yep, it's in the queue | 15:05 |
| *** EinstCrazy has quit IRC | 15:05 | |
| mriedem | sdague: then the question is, are we ok with fixing the bug that sahid is working on, knowing that the rest api is going to be returning security-groups=['default'] even though that's possibly wrong | 15:05 |
| notartom | bizarrodan, much thanks :) | 15:05 |
| *** markvoelker has joined #openstack-nova | 15:05 | |
| mriedem | and then fixing that with a microversion later | 15:05 |
| sdague | does the bug fix make us more wrong than we are today? | 15:05 |
| mriedem | today you can't boot a server into a network that has port_security_enabled=False, period | 15:06 |
| *** prateek has quit IRC | 15:06 | |
| mriedem | http://paste.openstack.org/show/491809/ | 15:06 |
| *** links has joined #openstack-nova | 15:07 | |
| *** josecastroleon has quit IRC | 15:07 | |
| mriedem | i'm inclined to say being able to boot a server is more important that the possibly wrong rest api response in this case | 15:08 |
| mriedem | knowing we can fix that later with a microversion | 15:08 |
| mriedem | and the nova secgroup-list API gives you the correct info | 15:08 |
| mriedem | *list-secgroup | 15:08 |
| sdague | mriedem: yeh, that seems like a tradeoff that is worth taking | 15:08 |
| sdague | I'd say that we reno it for a known issue | 15:09 |
| mriedem | yeah | 15:09 |
| sdague | I think with a reno, I can accept bad API data for fixing the bug | 15:09 |
| mriedem | the other probably less hard question is, do we like sahid's fix https://review.openstack.org/#/c/284095/ more than ftersin's fix https://review.openstack.org/#/c/173204/ | 15:10 |
| mriedem | the difference is sahid just ignores the 'default' security group in the neutronv2 api, in ftersin's change it changes the compute api to not even send the default security group for neutronv2 (if it wasn't requested) | 15:10 |
| *** mpavone has joined #openstack-nova | 15:10 | |
| mriedem | i think ftersin's is cleaner | 15:11 |
| mriedem | simply because i'd like to avoid adding any more logic to neutronv2.api.allocate_for_instance if we can help it | 15:12 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Change BuildRequest to contain a serialized instance https://review.openstack.org/298455 | 15:12 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add Allocation and AllocationList objects https://review.openstack.org/282442 | 15:12 |
| mriedem | sahid: would you be ok with that? | 15:13 |
| mriedem | sahid: you'd probably need to cleanup ftersin's patch | 15:13 |
| mriedem | add a release note, etc | 15:13 |
| sahid | mriedem: i do not know the patch from ftersin, i can only support my patch | 15:14 |
| openstackgerrit | Dan Smith proposed openstack/nova: Further hack up the n.t.unit.db.fakes module of horribleness https://review.openstack.org/296106 | 15:14 |
| openstackgerrit | Dan Smith proposed openstack/nova: Handle pre-migration flavor creation failures in the crusty old API https://review.openstack.org/299625 | 15:14 |
| mriedem | sahid: it's pretty simple https://review.openstack.org/#/c/173204/ | 15:14 |
| openstackgerrit | Dan Smith proposed openstack/nova: Block flavor creation until main database is empty https://review.openstack.org/295310 | 15:14 |
| sahid | wich i know is work | 15:14 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add flavor migration routine https://review.openstack.org/294771 | 15:14 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make Flavor create() and destroy() work against API DB https://review.openstack.org/294758 | 15:15 |
| openstackgerrit | Dan Smith proposed openstack/nova: Remove flavor seeding from the base migration https://review.openstack.org/300127 | 15:15 |
| mriedem | sahid: his patch basically only adds the 'default' security group for the nova-network security group driver, if no security group is requested when booting the instance | 15:15 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add name to ResourceProvider object https://review.openstack.org/281945 | 15:15 |
| mriedem | with his change you could still get yourself into trouble if you try booting an instance into a network that doesn't support port security and you provide a security group explicitly, that will fail | 15:16 |
| mriedem | sahid: with your patch ^ should work | 15:17 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add a get_by_uuid for aggregates https://review.openstack.org/284895 | 15:17 |
| sahid | mriedem: my patch is ready, you have tested it, i have tested it, jaypipes +2ed - i can't support that patch without an extra work to verify all cases | 15:17 |
| sahid | or you should to provide to me strong reason | 15:18 |
| sahid | s/that/his | 15:18 |
| mriedem | my only reason was i wanted to avoid more logic in the neutron allocate_for_instance method, which is a monster, | 15:18 |
| mriedem | but like i said, i think your patch might be better in case the user provides a security group but is using a network that doesn't allow them | 15:19 |
| mriedem | with your change we'd just ignore the security group requested | 15:19 |
| sahid | mriedem: it is but it's also one file updated, it will be easy for us to backport it | 15:19 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add ResourcePool and ResourcePoolList objects https://review.openstack.org/284963 | 15:19 |
| sahid | mriedem: yes true | 15:19 |
| *** browne has joined #openstack-nova | 15:20 | |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP resource-pools/placement api https://review.openstack.org/293104 | 15:21 |
| *** thorst has quit IRC | 15:22 | |
| *** mylu has quit IRC | 15:24 | |
| mriedem | sdague: btw, list-secgroup is on the server, not the project, see http://paste.openstack.org/show/492742/ | 15:26 |
| doffm | The entirety of austin is sold out. I think 13000 people must be turning up at this summit. | 15:28 |
| *** mylu has joined #openstack-nova | 15:28 | |
| *** thorst has joined #openstack-nova | 15:30 | |
| markus_z | Having a camping area with tents would be interesting :) | 15:30 |
| sdague | mriedem: ok, yeh, that's probably fine | 15:31 |
| *** e0ne has quit IRC | 15:31 | |
| *** matrohon has quit IRC | 15:31 | |
| *** pmathews1 has joined #openstack-nova | 15:31 | |
| doffm | markus_z: That sounds fun, the openstack outdoor festival. | 15:32 |
| markus_z | doffm: hehe. Without music bands but stable broadband wifi. I'm in. | 15:34 |
| * markus_z is leaving for today. See you folks next week. | 15:34 | |
| *** trams has joined #openstack-nova | 15:35 | |
| *** pmathews has quit IRC | 15:35 | |
| *** haplo37 has joined #openstack-nova | 15:35 | |
| trams | Hello! trying to get nova-conductor to use a CA to verify glance connections? | 15:35 |
| trams | DEFAULT/ca_file is set and works with curl, openssl etc | 15:36 |
| *** pav0 has quit IRC | 15:36 | |
| *** pmathews1 has quit IRC | 15:36 | |
| *** achanda has joined #openstack-nova | 15:37 | |
| trams | Using liberty | 15:37 |
| *** cdent has quit IRC | 15:37 | |
| *** mylu has quit IRC | 15:38 | |
| ndipanov | . | 15:38 |
| *** pmathews has joined #openstack-nova | 15:39 | |
| openstackgerrit | Paul Murray proposed openstack/nova: Remove extensible resource tracking https://review.openstack.org/300420 | 15:39 |
| trams | can't get it to work, but all other tls connections work, ie rabbit, glaera, keystone etc | 15:39 |
| *** ociuhandu has quit IRC | 15:39 | |
| *** mylu has joined #openstack-nova | 15:40 | |
| *** tjones_zzzz has joined #openstack-nova | 15:40 | |
| *** armax has joined #openstack-nova | 15:41 | |
| *** jdurgin1 has joined #openstack-nova | 15:41 | |
| *** achanda has quit IRC | 15:42 | |
| *** jlanoux has quit IRC | 15:42 | |
| *** fawadkhaliq has joined #openstack-nova | 15:42 | |
| *** jlanoux has joined #openstack-nova | 15:43 | |
| mriedem | trams: you're looking for the [ssl]/ca_file option | 15:44 |
| mriedem | not DEFAULT | 15:44 |
| mriedem | sahid: ok, i'll write up a reply to the ML thread with the plan | 15:44 |
| *** krtaylor has joined #openstack-nova | 15:44 | |
| *** Guest90877 has quit IRC | 15:44 | |
| trams | mriedem: that enables client cert validation according to the docs | 15:46 |
| sahid | mriedem: you want me to add a reno txt, right ? which category should i use ? | 15:46 |
| sahid | it's not really clear for me | 15:47 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add tag column to vifs and bdm https://review.openstack.org/264014 | 15:47 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add metadata objects for device tagging https://review.openstack.org/264015 | 15:47 |
| trams | atleast according to: http://docs.openstack.org/liberty/config-reference/content/list-of-compute-config-options.html | 15:47 |
| openstackgerrit | Dan Smith proposed openstack/nova: Device tagging user visible changes https://review.openstack.org/264018 | 15:47 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add compute manager device tagging support https://review.openstack.org/264017 | 15:47 |
| openstackgerrit | Dan Smith proposed openstack/nova: Libvirt driver implementation of device tagging https://review.openstack.org/264016 | 15:47 |
| trams | (StrOpt) CA certificate file to use to verify connecting clients. | 15:47 |
| bizarrodan | vladikr: notartom: ^ no o.vo changes required | 15:47 |
| trams | mriedem: is the docs wrong? | 15:50 |
| mriedem | sahid: use 'issues' i think http://docs.openstack.org/developer/reno/usage.html | 15:51 |
| mriedem | that's for known issues | 15:51 |
| openstackgerrit | sahid proposed openstack/nova: network: make nova to handle port_security_enabled=False https://review.openstack.org/284095 | 15:51 |
| *** Mackie2k has quit IRC | 15:51 | |
| sahid | mriedem: ah yes | 15:51 |
| mriedem | sahid: http://lists.openstack.org/pipermail/openstack-dev/2016-April/091162.html | 15:51 |
| mriedem | trams: which docs? | 15:52 |
| mriedem | trams: this is the code https://github.com/openstack/nova/blob/stable/liberty/nova/image/glance.py#L133 | 15:52 |
| trams | mriedem: http://docs.openstack.org/liberty/config-reference/content/list-of-compute-config-options.html | 15:53 |
| trams | mriedem: thanks. | 15:53 |
| *** mdrabe has quit IRC | 15:53 | |
| *** sridharg has quit IRC | 15:53 | |
| mriedem | trams: yeah, the docs are super confusing here, because there are 3 options | 15:54 |
| mriedem | DEFAULT.ca_file, DEFAULT.ssl_ca_file, ssl.ca_file | 15:54 |
| openstackgerrit | Paul Murray proposed openstack/nova: Remove extensible resource tracking https://review.openstack.org/300420 | 15:54 |
| openstackgerrit | Gage Hugo proposed openstack/nova: Verify pci_passthrough_whitelist fields https://review.openstack.org/300219 | 15:54 |
| *** links has quit IRC | 15:55 | |
| trams | mriedem: right, adn which one do i want? both DEFAULT.ssl_ca_file and ssl.ca_file implies client auth | 15:55 |
| mriedem | DEFAULT.ssl_ca_file is used for the wsgi app | 15:55 |
| mriedem | ssl.ca_file is used for at least glance and keystone auth | 15:55 |
| mriedem | DEFAULT.ca_file is used in nova.crypto | 15:56 |
| mriedem | trams: if the values are the same for all, then i'd set them to the same thing | 15:56 |
| trams | right, but i do not want client cert auth | 15:56 |
| mriedem | i'll send an email to the dev list about this | 15:56 |
| trams | which it does imply | 15:56 |
| mriedem | sorry, the docs suck, that needs to be cleaned up, you'd have to check the code where the options are used | 15:57 |
| notartom | bizarrodan, thanks, but how did you do it? | 15:57 |
| openstackgerrit | Claudiu Belu proposed openstack/nova: compute: fixes python 3 related unit tests https://review.openstack.org/300262 | 15:57 |
| notartom | Oh, just register it in the tests >< | 15:57 |
| *** penick has joined #openstack-nova | 15:57 | |
| notartom | That's cheating, you said you didn't want to register it ;) | 15:57 |
| notartom | Anyways, thanks for that :) | 15:58 |
| bizarrodan | notartom: well, (a) we do that in a couple other places and (b) I didn't want to register it because of the RPC impacts at runtime | 15:58 |
| bizarrodan | but regardless, it should unblock you | 15:58 |
| *** haplo37 has quit IRC | 15:58 | |
| *** haplo37 has joined #openstack-nova | 15:59 | |
| *** zenoway has quit IRC | 15:59 | |
| notartom | Yep, does that nicely. | 15:59 |
| notartom | bizarrodan, out of curiosity, what are the impacts of registering it at runtime? | 15:59 |
| openstackgerrit | sahid proposed openstack/nova: network: make nova to handle port_security_enabled=False https://review.openstack.org/284095 | 16:00 |
| bizarrodan | notartom: if it's registered, it means we can send it over rpc, but more importantly, we'll accept it over rpc, try to deserialize it, etc | 16:00 |
| bizarrodan | notartom: an there is zero reason for us to do that, just expanded surface for weirdness | 16:00 |
| *** haomaiwa_ has quit IRC | 16:01 | |
| notartom | bizarrodan, ah, indeed, no reason a base class should ever be sent over the wire as itself and not a subclass... | 16:01 |
| notartom | See, Python should totally be like Java with its abstract classes :D | 16:01 |
| *** haomaiwang has joined #openstack-nova | 16:01 | |
| sahid | mriedem: cool, good heads-up | 16:01 |
| *** Swami has joined #openstack-nova | 16:03 | |
| *** amotoki has quit IRC | 16:04 | |
| *** irenab has quit IRC | 16:04 | |
| *** irenab_ has joined #openstack-nova | 16:04 | |
| *** apoorvad has joined #openstack-nova | 16:06 | |
| *** mdrabe has joined #openstack-nova | 16:07 | |
| mriedem | trams: http://lists.openstack.org/pipermail/openstack-dev/2016-April/091167.html | 16:07 |
| *** ssurana has joined #openstack-nova | 16:08 | |
| *** ihrachys has quit IRC | 16:09 | |
| *** dedery has quit IRC | 16:11 | |
| auggy | bizarrodan: i read your name as bizarre rodan https://en.wikipedia.org/wiki/Rodan | 16:12 |
| trams | mriedem: thanks | 16:12 |
| trams | mriedem: i think you pointed me to dead code btw. cant make it trigger when riddlingit with LOG statements | 16:12 |
| *** mriedem is now known as mriedem_lunch | 16:12 | |
| *** zzzeek has quit IRC | 16:13 | |
| bizarrodan | auggy: think less godzilla and more seinfeld-superman-superdan | 16:13 |
| sdague | mriedem_lunch / danpb / johnthetubaguy - https://review.openstack.org/#/c/300427/2/api-ref/ext/rest_parameters.py minor error handler addition to make the parser more robust | 16:13 |
| *** zzzeek has joined #openstack-nova | 16:14 | |
| *** efoley has quit IRC | 16:14 | |
| *** sfinucan has quit IRC | 16:15 | |
| *** mylu has quit IRC | 16:15 | |
| *** jlanoux has quit IRC | 16:17 | |
| doffm | This code is dead Jim. | 16:19 |
| *** jlanoux has joined #openstack-nova | 16:20 | |
| rlrossit | bizarrodan: why does the floating ip object have both fixed_ip_id and fixed_ip as fields? Shouldn't fixed_ip_id boil down to: fixed_ip.id if fixed_ip else None? | 16:21 |
| *** sahid has quit IRC | 16:21 | |
| bizarrodan | rlrossit: because we need to persist the id so we can look up the object if we need to without querying the database twice | 16:22 |
| rlrossit | bizarrodan: ah which is why that isn't lazy loadable, and why that keeps blowing up on your change from way back when :( | 16:23 |
| *** mdrabe has quit IRC | 16:23 | |
| *** mdrabe has joined #openstack-nova | 16:24 | |
| *** e0ne has joined #openstack-nova | 16:24 | |
| *** salv-orlando has joined #openstack-nova | 16:25 | |
| *** danpb has quit IRC | 16:25 | |
| *** jlanoux has quit IRC | 16:25 | |
| *** mdrabe has quit IRC | 16:26 | |
| *** mdrabe has joined #openstack-nova | 16:26 | |
| *** mdrabe has quit IRC | 16:28 | |
| *** mdrabe has joined #openstack-nova | 16:28 | |
| *** e0ne has quit IRC | 16:29 | |
| *** e0ne has joined #openstack-nova | 16:29 | |
| *** Marga_ has quit IRC | 16:30 | |
| *** openstackgerrit has quit IRC | 16:31 | |
| *** rcernin has quit IRC | 16:32 | |
| cfriesen | I'm running "tox" in nova-specs and it's complaining that some sections are missing from my new spec, except that the sections are not missing--they show up in the resulting html output. Anyone have any ideas? | 16:37 |
| *** achanda has joined #openstack-nova | 16:38 | |
| *** e0ne has quit IRC | 16:39 | |
| *** rk4n has quit IRC | 16:40 | |
| *** rook-lappy has quit IRC | 16:40 | |
| *** pushkaru has quit IRC | 16:41 | |
| *** pushkaru has joined #openstack-nova | 16:41 | |
| *** rcarrillocruz has joined #openstack-nova | 16:43 | |
| *** rcarrillocruz has left #openstack-nova | 16:43 | |
| *** rcarrillocruz has quit IRC | 16:43 | |
| auggy | maybe tox is playing an april fool's joke on you? | 16:43 |
| *** rook-lappy has joined #openstack-nova | 16:44 | |
| *** pcaruana has quit IRC | 16:44 | |
| *** achanda has quit IRC | 16:45 | |
| *** openstackgerrit has joined #openstack-nova | 16:45 | |
| auggy | are all the indenting/line breaks/underlines etc in place? | 16:46 |
| *** mmedvede_ has quit IRC | 16:46 | |
| openstackgerrit | Chuck Carmack proposed openstack/nova: DB API changes for the nova-manage quota_usage_refresh command https://review.openstack.org/299441 | 16:46 |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Libvirt driver implementation of device tagging https://review.openstack.org/264016 | 16:47 |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Add compute manager device tagging support https://review.openstack.org/264017 | 16:48 |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Device tagging user visible changes https://review.openstack.org/264018 | 16:49 |
| *** fawadkhaliq has quit IRC | 16:50 | |
| *** edmondsw has quit IRC | 16:50 | |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Libvirt driver implementation of device tagging https://review.openstack.org/264016 | 16:50 |
| *** fawadkhaliq has joined #openstack-nova | 16:50 | |
| * notartom is failing at rebases | 16:52 | |
| *** efoley has joined #openstack-nova | 16:53 | |
| *** ssurana has quit IRC | 16:54 | |
| mordred | sdague: I hear you. :) honestly, I'd be more than happy to help in a significant way | 16:54 |
| *** clenimar has quit IRC | 16:54 | |
| openstackgerrit | Chuck Carmack proposed openstack/nova: DB API changes for the nova-manage quota_usage_refresh command https://review.openstack.org/299441 | 16:54 |
| mordred | sdague: I _may_ have accidentally grown a fairly large amountof knowledge on the subject | 16:54 |
| *** mpavone has quit IRC | 16:55 | |
| *** e0ne has joined #openstack-nova | 16:55 | |
| *** nic has joined #openstack-nova | 16:57 | |
| cfriesen | auggy: as far as I can tell, all the formatting stuff matches with the template and other specs. And like I said, I can see them in the resulting HTML output. | 16:59 |
| *** p01nt3r75 has quit IRC | 16:59 | |
| *** delattec has quit IRC | 17:00 | |
| *** haomaiwang has quit IRC | 17:01 | |
| openstackgerrit | Sujitha proposed openstack/nova: config options: Centralize vmware section https://review.openstack.org/299002 | 17:01 |
| *** haomaiwang has joined #openstack-nova | 17:01 | |
| *** rmevans has quit IRC | 17:02 | |
| *** Marga_ has joined #openstack-nova | 17:02 | |
| *** eharney has quit IRC | 17:02 | |
| *** aysyd has quit IRC | 17:05 | |
| *** ssurana has joined #openstack-nova | 17:07 | |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Add compute manager device tagging support https://review.openstack.org/264017 | 17:08 |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Device tagging user visible changes https://review.openstack.org/264018 | 17:09 |
| openstackgerrit | Chuck Carmack proposed openstack/nova: WIP: Change the nova tempest blacklist to use to idempotent ids https://review.openstack.org/298437 | 17:09 |
| *** yamahata has quit IRC | 17:11 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Change BuildRequest to contain a serialized instance https://review.openstack.org/298455 | 17:11 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Put more into compute.api._populate_instance_for_create https://review.openstack.org/300580 | 17:12 |
| auggy | cfriesen: if you have a change up in gerrit i could give it a try and see if i see anything | 17:12 |
| openstackgerrit | Zhenguo Niu proposed openstack/nova-specs: Add Ironic Shellinabox HTTP proxy https://review.openstack.org/300582 | 17:13 |
| *** mgoddard has quit IRC | 17:14 | |
| *** ndipanov has quit IRC | 17:14 | |
| *** e0ne has quit IRC | 17:15 | |
| *** lpetrut has quit IRC | 17:16 | |
| *** Marga_ has quit IRC | 17:17 | |
| *** Marga_ has joined #openstack-nova | 17:17 | |
| *** davideagnello has joined #openstack-nova | 17:19 | |
| *** cdent has joined #openstack-nova | 17:20 | |
| *** Bjoern has quit IRC | 17:20 | |
| *** eharney has joined #openstack-nova | 17:22 | |
| *** unicell has joined #openstack-nova | 17:23 | |
| *** mriedem_lunch is now known as mriedem | 17:24 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Move config options from nova/objects/network.py https://review.openstack.org/269240 | 17:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the linuxnet options (4) https://review.openstack.org/275423 | 17:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the linuxnet options (2) https://review.openstack.org/271489 | 17:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the linuxnet options (3) https://review.openstack.org/272231 | 17:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Move config options from nova/network/linux_net.py https://review.openstack.org/270441 | 17:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the linuxnet options (1) https://review.openstack.org/270959 | 17:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve help text for the network object options https://review.openstack.org/269313 | 17:24 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: [extensions] cache discovered extensions https://review.openstack.org/287131 | 17:29 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: [tests] initialize client objects inside setUp https://review.openstack.org/300032 | 17:29 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Cache discover_version method https://review.openstack.org/299862 | 17:29 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Add support of caching https://review.openstack.org/299860 | 17:29 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Move all extensions from contrib dir https://review.openstack.org/280726 | 17:29 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Switch to 2.1 default api_version in v2.Client https://review.openstack.org/300586 | 17:29 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Use common find_server from v2.shell https://review.openstack.org/300587 | 17:29 |
| andreykurilin | mriedem: hi! | 17:29 |
| cfriesen | auggy: thanks for the offer, gotta run it by some folks internally first. | 17:29 |
| andreykurilin | mriedem: Do you free time to discuss novaclient?) | 17:30 |
| auggy | cfriesen: alright, np, sorry i can't offer more assistance :( | 17:30 |
| *** pmathews has quit IRC | 17:32 | |
| *** mriedem has quit IRC | 17:33 | |
| *** mriedem has joined #openstack-nova | 17:34 | |
| *** delattec has joined #openstack-nova | 17:35 | |
| *** salv-orlando has quit IRC | 17:36 | |
| mriedem | andreykurilin: hi | 17:36 |
| andreykurilin | mriedem: I have a big change and it would nice to bump major version if it will be accept | 17:37 |
| mriedem | andreykurilin: this https://review.openstack.org/#/c/300586/ ? | 17:37 |
| *** ljxiash has quit IRC | 17:38 | |
| *** delatte has joined #openstack-nova | 17:38 | |
| andreykurilin | mriedem: no:) | 17:38 |
| andreykurilin | mriedem: it is small and should not affect a lot of people and deployments | 17:38 |
| andreykurilin | mriedem: https://review.openstack.org/280726 | 17:38 |
| *** ljxiash has joined #openstack-nova | 17:38 | |
| andreykurilin | this inbe | 17:38 |
| *** yamahata has joined #openstack-nova | 17:38 | |
| andreykurilin | *this one | 17:38 |
| *** delattec has quit IRC | 17:40 | |
| *** haomaiwang has quit IRC | 17:40 | |
| *** vishwanathj has quit IRC | 17:40 | |
| *** burgerk has quit IRC | 17:40 | |
| openstackgerrit | Hemanth Makkapati proposed openstack/nova: config options: centralize xenserver vm_utils opts https://review.openstack.org/298875 | 17:40 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Move all extensions from contrib dir https://review.openstack.org/280726 | 17:41 |
| *** haomaiwang has joined #openstack-nova | 17:41 | |
| *** READ10 has quit IRC | 17:42 | |
| *** ljxiash has quit IRC | 17:42 | |
| *** salv-orlando has joined #openstack-nova | 17:45 | |
| johnthetubaguy | cfriesen: are they the correct heading level, its probably a little picky about some of that | 17:45 |
| johnthetubaguy | cfriesen: oops, bad reading of scrollback | 17:45 |
| *** ccarmack has quit IRC | 17:46 | |
| *** MVenesio has joined #openstack-nova | 17:48 | |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Libvirt driver implementation of device tagging https://review.openstack.org/264016 | 17:51 |
| auggy | johnthetubaguy: yeah i think without seeing the text in question it's probably hard to say | 17:51 |
| *** fawadkhaliq has quit IRC | 17:52 | |
| *** fawadkhaliq has joined #openstack-nova | 17:52 | |
| *** markvoelker has quit IRC | 17:52 | |
| mriedem | andreykurilin: simply changing the default from v2.0 to v2.1 is probably a major version bump | 17:53 |
| mriedem | since it's an api change | 17:53 |
| *** jdurgin1 has quit IRC | 17:53 | |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Device tagging user visible changes https://review.openstack.org/264018 | 17:54 |
| mriedem | doffm: rlrossit: ctrath: i've got black tar in here all ready to go | 17:54 |
| *** gjayavelu has quit IRC | 17:55 | |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Add compute manager device tagging support https://review.openstack.org/264017 | 17:55 |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Device tagging user visible changes https://review.openstack.org/264018 | 17:56 |
| *** markvoelker has joined #openstack-nova | 17:56 | |
| *** vishwanathj has joined #openstack-nova | 17:57 | |
| *** jistr has quit IRC | 17:58 | |
| *** PaulMurray has quit IRC | 17:59 | |
| andreykurilin__ | mriedem: that version is used only if somebody uses novaclient in wrong way. Nobody should use novaclient.v2.client.Client as entry point and we have a warning for such cases, so, imo, we can change your that version without bumping major version | 17:59 |
| rlrossit | mriedem: do we all get to share a needle? | 17:59 |
| openstackgerrit | Dan Smith proposed openstack/nova: Remove flavor seeding from the base migration https://review.openstack.org/300127 | 17:59 |
| andreykurilin__ | *we can change that version | 18:00 |
| mriedem | andreykurilin: dropping deprecated stuff doesn't mean it's not an api change | 18:00 |
| mriedem | the volume CLIs in novaclient were deprecated for 2 releases and we should be removing those now, but that will trigger a major version bump | 18:00 |
| mriedem | speaking of which...i'm going to remove those now | 18:00 |
| *** haomaiwang has quit IRC | 18:01 | |
| *** haomaiwang has joined #openstack-nova | 18:01 | |
| *** ccarmack has joined #openstack-nova | 18:01 | |
| *** ctrath has quit IRC | 18:02 | |
| *** harlowja has quit IRC | 18:02 | |
| andreykurilin__ | mriedem but it was a wrong non-documented way:) ok, I'm not against bumping version for it, but I wanted to discuss the change about contrib dir | 18:03 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Create Instance from BuildRequest if not in a cell https://review.openstack.org/263927 | 18:03 |
| *** rlrossit has quit IRC | 18:03 | |
| *** jed56 has quit IRC | 18:03 | |
| andreykurilin__ | mriedem yeah, I agreed that we are ready two remove deprecated volume commands | 18:03 |
| mriedem | andreykurilin: so the contrib dir thing is more to follow what we have in the server for not doing contrib with legacy v2? | 18:04 |
| mriedem | i.e. no extensions? | 18:04 |
| *** rmevans has joined #openstack-nova | 18:05 | |
| *** ctrath has joined #openstack-nova | 18:05 | |
| openstackgerrit | Sujitha proposed openstack/nova: config options: Centralize upgrade_levels section https://review.openstack.org/297381 | 18:05 |
| *** pmathews has joined #openstack-nova | 18:05 | |
| bizarrodan | is laski hiding under some equally-cool nick today? | 18:05 |
| mriedem | bizarrodan: meat lutf-8 | 18:07 |
| mriedem | *meet | 18:07 |
| bizarrodan | wow, that's epic good | 18:08 |
| lutf-8 | :) | 18:08 |
| bizarrodan | lutf-8: so that patch, with that fixture, passes unit tests | 18:08 |
| bizarrodan | after failing 6500 of them with just removing the flavor | 18:08 |
| bizarrodan | which is awesome | 18:08 |
| bizarrodan | so I think just getting the devstack change up will get us what we need | 18:08 |
| bizarrodan | I thought it was going to be a lot harder | 18:08 |
| lutf-8 | that's great | 18:09 |
| andreykurilin__ | mriedem I suppose we should leave support of extensions for novaclient, since we don't deprecate it yet, but it looks like extensions from contrib dir are not extensions in case of v2.1 and, btw, all of them are loaded automatically via shell in case of even 2.0 | 18:09 |
| lutf-8 | bizarrodan: yeah, just getting devstack passing seems like what we need. then just announcements and doc updates | 18:10 |
| bizarrodan | I done already announced it | 18:11 |
| lutf-8 | oh yeah, I guess I saw that | 18:11 |
| lutf-8 | great | 18:11 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: network: make nova to handle port_security_enabled=False https://review.openstack.org/284095 | 18:14 |
| openstackgerrit | Dan Smith proposed openstack/nova: Further hack up the n.t.unit.db.fakes module of horribleness https://review.openstack.org/296106 | 18:15 |
| openstackgerrit | Dan Smith proposed openstack/nova: Handle pre-migration flavor creation failures in the crusty old API https://review.openstack.org/299625 | 18:15 |
| openstackgerrit | Dan Smith proposed openstack/nova: Block flavor creation until main database is empty https://review.openstack.org/295310 | 18:15 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add flavor migration routine https://review.openstack.org/294771 | 18:15 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make Flavor create() and destroy() work against API DB https://review.openstack.org/294758 | 18:15 |
| openstackgerrit | Dan Smith proposed openstack/nova: Remove flavor seeding from the base migration https://review.openstack.org/300127 | 18:16 |
| bizarrodan | just fixing the test I wrote that was broken :D | 18:16 |
| *** zenoway has joined #openstack-nova | 18:16 | |
| *** daemontool has quit IRC | 18:18 | |
| *** rcrit has quit IRC | 18:18 | |
| *** mvk_ has quit IRC | 18:19 | |
| *** zenoway has quit IRC | 18:21 | |
| *** baoli_ has quit IRC | 18:22 | |
| *** vishwanathj has quit IRC | 18:22 | |
| *** eil397 has joined #openstack-nova | 18:23 | |
| *** rlrossit has joined #openstack-nova | 18:25 | |
| *** p01nt3r75 has joined #openstack-nova | 18:26 | |
| openstackgerrit | Karen Bradshaw proposed openstack/nova: [WIP] Create parameters.yaml per operation https://review.openstack.org/300264 | 18:26 |
| mriedem | lutf-8: i'm +2 on https://review.openstack.org/#/c/294758/ now | 18:27 |
| lutf-8 | mriedem: excellent. I wasn't going to touch it again until you were happy | 18:28 |
| mriedem | bizarrodan: so what happened here ? https://review.openstack.org/#/c/299625/ it looks like the functional test stuff is gone now | 18:28 |
| bizarrodan | mriedem: ah, yeah I covered up sdague's push | 18:29 |
| bizarrodan | mriedem: I'll fix in a few | 18:30 |
| mriedem | ok, everything below looks good | 18:31 |
| *** markvoelker has quit IRC | 18:31 | |
| *** mdrabe has quit IRC | 18:31 | |
| *** mdrabe has joined #openstack-nova | 18:32 | |
| *** edmondsw has joined #openstack-nova | 18:32 | |
| bizarrodan | lutf-8: I can drop my -2 and let you send it when you're ready if that's cool | 18:34 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Use common find_server from v2.shell https://review.openstack.org/300587 | 18:35 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Move all extensions from contrib dir https://review.openstack.org/280726 | 18:35 |
| *** baoli has joined #openstack-nova | 18:35 | |
| lutf-8 | bizarrodan: it's everything though https://review.openstack.org/#/c/295310/30 that needs to get in right? | 18:36 |
| bizarrodan | lutf-8: correct | 18:36 |
| bizarrodan | dropped my -2 | 18:37 |
| lutf-8 | okay. after the test jobs finish up let's look at getting that in | 18:37 |
| bizarrodan | woot | 18:37 |
| armax | mriedem: ping | 18:41 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Create Instance from BuildRequest if not in a cell https://review.openstack.org/263927 | 18:44 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Change BuildRequest to contain a serialized instance https://review.openstack.org/298455 | 18:44 |
| mriedem | armax: pong | 18:45 |
| *** eil397 has left #openstack-nova | 18:46 | |
| armax | mriedem: I put a Neutron section on this etherpad https://etherpad.openstack.org/p/newton-nova-priorities-tracking | 18:46 |
| armax | mriedem: I hope it’s ok | 18:46 |
| *** manjeets has left #openstack-nova | 18:46 | |
| mriedem | armax: that's fine | 18:46 |
| armax | mriedem: btw did we settle for a joint session at the summit? I recall we said Wed morning, but I am not sure we finalized | 18:46 |
| mriedem | armax: i have wed from 11:00-11:40 penciled in | 18:47 |
| mriedem | armax: in here https://etherpad.openstack.org/p/newton-nova-summit-ideas at the bottom | 18:47 |
| armax | mriedem: ack | 18:47 |
| mriedem | armax: you can get us after 2 hours of talking about jay's scheduler blueprints | 18:48 |
| armax | mriedem: makes sense | 18:48 |
| mriedem | armax: we could also move to thursday if needed, but i figure if routed networks is going to come up, that should probably be after jay's scheduler sessions | 18:48 |
| mriedem | we're doing a lot of the cross-project sessiosn on thursday | 18:49 |
| mriedem | and nova-specific stuff on wed | 18:49 |
| armax | mriedem: agreed, I think wed morning is best | 18:49 |
| mriedem | ok | 18:49 |
| mriedem | plan is to finalize the nova schedule on tuesday 4/12 | 18:49 |
| armax | mriedem: noted in my calendar, I hope to get to a final schedule for neutron during next week | 18:50 |
| *** smatzek has quit IRC | 18:51 | |
| *** Sukhdev has joined #openstack-nova | 18:53 | |
| *** david-lyle has quit IRC | 18:53 | |
| *** aysyd has joined #openstack-nova | 18:53 | |
| *** markvoelker has joined #openstack-nova | 18:53 | |
| *** markvoelker has quit IRC | 18:53 | |
| *** markvoelker has joined #openstack-nova | 18:54 | |
| *** markvoelker has quit IRC | 18:54 | |
| *** markvoelker has joined #openstack-nova | 18:54 | |
| mriedem | armax: cool, i moved it back a bit since next week the TC votes on the tuesday cross-project sessions, which might impact some of the nova sessions | 18:54 |
| armax | mriedem: good point | 18:55 |
| openstackgerrit | Kam Nasim proposed openstack/nova: Fixes exception during a resize of a flavor with swap to one with no swap. The root cause of this failure was that the swap record for an instance in the block_device_mapping table in db is being used by get_disk_mapping() to create the swap disk mapping https://review.openstack.org/300616 | 18:56 |
| *** david-lyle has joined #openstack-nova | 18:59 | |
| *** Goneri has quit IRC | 18:59 | |
| *** Sukhdev has quit IRC | 19:00 | |
| *** haomaiwang has quit IRC | 19:01 | |
| *** haomaiwang has joined #openstack-nova | 19:01 | |
| *** fawadkhaliq has quit IRC | 19:05 | |
| *** fawadkhaliq has joined #openstack-nova | 19:06 | |
| *** fawadkhaliq has quit IRC | 19:07 | |
| *** salv-orlando has quit IRC | 19:07 | |
| *** dimtruck is now known as zz_dimtruck | 19:12 | |
| *** fawadkhaliq has joined #openstack-nova | 19:13 | |
| *** baoli has quit IRC | 19:19 | |
| *** rk4n has joined #openstack-nova | 19:19 | |
| *** tbachman has quit IRC | 19:20 | |
| *** tbachman has joined #openstack-nova | 19:24 | |
| bizarrodan | lutf-8: so, I think we're good to go, although jenkins has done its thing where it thinks there are merge conflicts when there can't have been | 19:27 |
| sdague | dims: do you remember why this line was in here - https://github.com/openstack/nova/commit/79b280479c864d7c0f16e6b9fffd30662f2394c6#diff-c4a9886c5c7a2095bdcfeb79fd53f316R131 | 19:28 |
| bizarrodan | lutf-8: they all passed tests in their form at some point | 19:28 |
| bizarrodan | I'll trigger up some rechecks | 19:28 |
| *** baoli has joined #openstack-nova | 19:29 | |
| lawn_bauwer | bizarrodan: given I'm okay with https://review.openstack.org/#/c/294758/22 can I +2/+W and merge all the series ? | 19:33 |
| mriedem | bizarrodan: rlrossit has stepped in a pile of turducken objects https://review.openstack.org/#/c/115749/9/nova/network/neutronv2/api.py | 19:33 |
| mriedem | could be epic | 19:33 |
| bizarrodan | lawn_bauwer: I'm going to let lutf-8 +W it | 19:33 |
| bizarrodan | lawn_bauwer: but happy to have your +2 on there too | 19:34 |
| lawn_bauwer | heh okay | 19:34 |
| lawn_bauwer | done | 19:34 |
| bizarrodan | thanks! | 19:34 |
| lawn_bauwer | do you guys use fishes for April 1st ? | 19:35 |
| *** cleong has quit IRC | 19:35 | |
| lawn_bauwer | because saying "thanks for the fish" in a review on April 1st would be bad if you guys don't use fishes for pranks | 19:35 |
| bizarrodan | I've never heard of it, but it's a french thing right? | 19:35 |
| sdague | lawn_bauwer: I'm pretty sure that's purely a french thing | 19:35 |
| lawn_bauwer | sec | 19:36 |
| lawn_bauwer | https://en.wikipedia.org/wiki/April_Fools'_Day#April_fish | 19:36 |
| *** smatzek has joined #openstack-nova | 19:36 | |
| *** harlowja has joined #openstack-nova | 19:37 | |
| lawn_bauwer | okay, my comment is then a purely French :( | 19:37 |
| mriedem | there is a story about led zeppelin and a red snapper specifically | 19:37 |
| mriedem | but i don't think that was a joke | 19:37 |
| lawn_bauwer | https://hamillfrenchblog.files.wordpress.com/2014/03/poisson-davril.jpg | 19:37 |
| lawn_bauwer | ^ that kind of prank | 19:37 |
| sdague | lawn_bauwer: I would just have assumed you meant this - https://en.wikipedia.org/wiki/So_Long,_and_Thanks_for_All_the_Fish | 19:37 |
| lawn_bauwer | when you're a student | 19:37 |
| lawn_bauwer | sdague: yeah that too | 19:38 |
| auggy | lawn_bauwer: isn't it poisson d'avril there? | 19:38 |
| lawn_bauwer | sdague: I tried to use this phrase from Arthur Dent by an April 1st because of the fish relationship | 19:38 |
| lawn_bauwer | see, guys, I'm not commenting on a cheese ! | 19:38 |
| auggy | :D | 19:38 |
| lawn_bauwer | auggy: zactly | 19:38 |
| auggy | oh yeah i need an april fool's friday nick | 19:39 |
| auggy | i'm so inspired by bizarrodan | 19:39 |
| *** auggy is now known as auggy_rodin | 19:39 | |
| * auggy_rodin strikes a thinker pose | 19:39 | |
| lawn_bauwer | woah | 19:39 |
| auggy_rodin | i read dan's nick as bizarre rodan, like godzilla haha and then i was like yeah wasn't that the sculptor | 19:40 |
| auggy_rodin | turns out his first name was augustine | 19:40 |
| *** eglynn has joined #openstack-nova | 19:40 | |
| * auggy_rodin likes chocolate fish better | 19:41 | |
| auggy_rodin | oh also this is my wedding anniversary | 19:41 |
| auggy_rodin | when my husband and i set the wedding date, people thought we were joking :D | 19:42 |
| *** fawadkhaliq has quit IRC | 19:43 | |
| lawn_bauwer | I just hope the Tesla 3 is not an April 1st joke... | 19:43 |
| dims | sdague : so, if i remember right, is_enabled was the only way at that time to register config options in that module, then we added register_opts in sslutils.py, forgot to cleanup Nova | 19:43 |
| *** baoli has quit IRC | 19:43 | |
| auggy_rodin | lawn_bauwer: inorite | 19:43 |
| mriedem | where is jay pipes/ | 19:43 |
| mriedem | ? | 19:43 |
| mriedem | does he have a special nick? | 19:43 |
| bizarrodan | I think he has a plane ticket | 19:43 |
| *** mvk_ has joined #openstack-nova | 19:43 | |
| dims | mriedem : he's flying back home from mother ship | 19:43 |
| mriedem | gdi, of course | 19:43 |
| *** fawadkhaliq has joined #openstack-nova | 19:44 | |
| sdague | dims: ok, so is_enabled has other side effects as well | 19:44 |
| dims | the runtime error(s)? | 19:44 |
| *** fawadkhaliq has quit IRC | 19:45 | |
| *** fawadkhaliq has joined #openstack-nova | 19:46 | |
| *** smatzek has quit IRC | 19:47 | |
| cfriesen | what's the expected resize behaviour when using local storage for /etc/nova/instances, but using RBD for the actual instance storage? Would we expect to create a duplicate RBD image prior to resizing the disk the same way we do for image files stored locally? | 19:47 |
| sdague | dims: yeh | 19:47 |
| sdague | which i guess no one has yet complained about | 19:48 |
| *** gjayavelu has joined #openstack-nova | 19:48 | |
| *** salv-orlando has joined #openstack-nova | 19:48 | |
| sdague | the naming of that method is... very unexpected | 19:48 |
| sdague | for what it does | 19:48 |
| lawn_bauwer | cfriesen: you mean having local directories for instances, but those being run on a shared storage ? | 19:50 |
| lawn_bauwer | cfriesen: that's just messing up Nova | 19:50 |
| lawn_bauwer | I wouldn't recommend that | 19:50 |
| cfriesen | lawn_bauwer: I mean local directories for /etc/nova/instances, but with libvirt images_type=rbd | 19:51 |
| lawn_bauwer | yeah that's what I meant | 19:52 |
| lawn_bauwer | for me it's /var/lib/nova but anyway... | 19:52 |
| *** Sukhdev has joined #openstack-nova | 19:52 | |
| cfriesen | Sorry, yeah. | 19:52 |
| *** rtheis has quit IRC | 19:52 | |
| lawn_bauwer | that's not something operators should do | 19:52 |
| *** rtheis has joined #openstack-nova | 19:54 | |
| cfriesen | so what's the recommended way to use images_type=rbd ? | 19:54 |
| mriedem | bizarrodan: can you subclass an ovo object and override one of it's fields to change the type? | 19:54 |
| mriedem | like the 'id' field | 19:54 |
| lawn_bauwer | cfriesen: to have those directories on a shared storage, not local | 19:55 |
| *** baoli has joined #openstack-nova | 19:55 | |
| bizarrodan | mriedem: it would be pretty evil | 19:57 |
| bizarrodan | mriedem: it would need to have a different obj_name for RPC compatibility | 19:57 |
| bizarrodan | just like any subclass with a different behavior | 19:57 |
| rlrossit | bizarrodan: way to rain on our parade | 19:57 |
| mriedem | bizarrodan: yeah, rlrossit and i were talking about subclassing FloatingIP and overriding the id field in https://review.openstack.org/#/c/115801/ | 19:57 |
| mriedem | rather than add an external_id field | 19:57 |
| bizarrodan | rlrossit: I could cover your parade with something else if you want | 19:58 |
| bizarrodan | yeah | 19:58 |
| *** zenoway has joined #openstack-nova | 19:58 | |
| cfriesen | lawn_bauwer: You mean the mountpoint for "instances_path" needs to be shared across all compute nodes when using images_type=rbd? Is this documented anywhere? | 19:58 |
| lawn_bauwer | cfriesen: yeah, that's what I'm trying to find out | 19:58 |
| cfriesen | lawn_bauwer: because the various walkthroughs on using ceph don't seem to specify that. | 19:59 |
| lawn_bauwer | cfriesen: agreed but that's something we don't support internally | 19:59 |
| *** rk4n has quit IRC | 20:00 | |
| *** gjayavelu has quit IRC | 20:00 | |
| cfriesen | lawn_bauwer: for what it's worth, we've been playing with it for a while and it mostly works except that we've come across a problem with resize | 20:00 |
| *** haomaiwang has quit IRC | 20:01 | |
| cfriesen | That said...even with shared storage for the resize case how would the rbd backend get duplicated (to allow for reverting the resize)? | 20:01 |
| *** haomaiwang has joined #openstack-nova | 20:01 | |
| openstackgerrit | Matt Riedemann proposed openstack/python-novaclient: Remove deprecated volume(snapshot) commands/bindings https://review.openstack.org/300636 | 20:02 |
| mriedem | who wants to remove a bunch of code? ^ | 20:02 |
| rlrossit | bizarrodan: wait by different obj_name you just mean a different object right? | 20:02 |
| * rlrossit forgets most of o.vo | 20:02 | |
| *** zz_dimtruck is now known as dimtruck | 20:03 | |
| bizarrodan | rlrossit: the obj_name is effectively a schema name.. an interface.. it says "I am a $name object" which means for a given version and name, you know the behavior of that thing | 20:03 |
| rlrossit | yeah so if I subclass it to make a NeutronFloatingIP I can do whatever I want with it right? | 20:03 |
| *** apoorvad has quit IRC | 20:04 | |
| bizarrodan | so FloatingIP has a given behavior at version 1.5, which means if your thing does a different thing than that (i.e. has a string id) then it needs to be called something else | 20:04 |
| *** fawadkhaliq has quit IRC | 20:04 | |
| mriedem | we'd call it NeutronFloatingIP | 20:04 |
| mriedem | i think | 20:04 |
| bizarrodan | yes, but you can't make its obj_name return FloatingIP | 20:04 |
| mriedem | and override id and fixed_ip_id | 20:04 |
| bizarrodan | okay, then yes | 20:04 |
| rlrossit | would we have it start at the same version as FloatingIP? | 20:04 |
| bizarrodan | no | 20:04 |
| mriedem | or make fixed_ip_id go -> FixedIP -> VirtualInterface -> uuid | 20:04 |
| *** fawadkhaliq has joined #openstack-nova | 20:05 | |
| bizarrodan | but also note that you can't send that to a compute | 20:05 |
| cfriesen | lawn_bauwer: libvirt.driver.LibvirtDriver.migrate_disk_and_power_off calls libvirt_utils.copy_image(), but that doesn't seem to special-case the RBD backend case like I think it should. | 20:05 |
| rlrossit | why not? | 20:05 |
| bizarrodan | so you need to have some way to handle sending the older thing for newton | 20:05 |
| bizarrodan | rlrossit: because a mitaka compute will choke on that object because it didn't know about it | 20:05 |
| mriedem | i don't tihnk we need to send this to compute | 20:05 |
| *** rk4n has joined #openstack-nova | 20:05 | |
| bizarrodan | okay cool | 20:06 |
| mriedem | it's just compute api <-> network api | 20:06 |
| bizarrodan | then maybe don't even register it, just register_if(False) it | 20:06 |
| rlrossit | so wait | 20:06 |
| rlrossit | how do we ever add objects then? if computes can't figure them out when we make them? | 20:06 |
| *** mastermind has joined #openstack-nova | 20:06 | |
| lawn_bauwer | cfriesen: sorry, was digging | 20:07 |
| cfriesen | no worries | 20:07 |
| mriedem | rlrossit: you turn it into a primitive and send it to the older thing | 20:07 |
| mriedem | in the rpc api | 20:07 |
| bizarrodan | yep | 20:07 |
| bizarrodan | that | 20:07 |
| bizarrodan | rlrossit: you missed most of that fun :) | 20:07 |
| bizarrodan | rlrossit: you showed up after the grass was already green and mowed :) | 20:07 |
| *** Sukhdev has quit IRC | 20:08 | |
| lawn_bauwer | cfriesen: so, AFAIR, the problem is that we don't remove directories of moved instances | 20:08 |
| mriedem | bizarrodan: there are still some mortimer turds in that green grass | 20:08 |
| lawn_bauwer | because those are seen as shared | 20:08 |
| bizarrodan | mriedem: yeah, but I know where they are, so I just avoid them | 20:08 |
| cfriesen | lawn_bauwer: before that, I don't think we duplicate the ceph image in order to resize the disks | 20:08 |
| *** gjayavelu has joined #openstack-nova | 20:08 | |
| rlrossit | and let me step on them | 20:08 |
| bizarrodan | I'm fine if you step on them | 20:09 |
| *** zenoway has quit IRC | 20:09 | |
| mriedem | rlrossit: so if NeutronFloatingIP is only going between compute API (rest api) and neutron API, no rpc, then register_if(False) is probably the thing to do | 20:09 |
| mriedem | id would be uuid | 20:09 |
| *** gjayavel_ has joined #openstack-nova | 20:09 | |
| mriedem | and we could put a port_id on it that's a uuid | 20:10 |
| mriedem | in place of fixed_ip_id | 20:10 |
| *** zenoway has joined #openstack-nova | 20:10 | |
| cfriesen | lawn_bauwer: libvirt_utils.copy_image() just copies using cp/rsync/scp, but that doesn't do what we want for a remote ceph device | 20:10 |
| mriedem | or we also override fixed_ip_id and just make it a uuid field | 20:10 |
| rlrossit | do I need conversions also? to map fixed_ip_id to port_id? | 20:10 |
| rlrossit | so the api doesn't need repeat code everywhere | 20:10 |
| *** rk4n has quit IRC | 20:10 | |
| rlrossit | or do double gets rather | 20:10 |
| mriedem | yeah, i'd probably just use fixed_ip_id as the port uuid field, but make it a uuid field | 20:10 |
| *** rmevans has quit IRC | 20:11 | |
| lawn_bauwer | cfriesen: sec, sec, verifying something | 20:11 |
| mriedem | or add a getattr to handle the conversoin in the object | 20:11 |
| * rlrossit wonders if he wants to start this on Friday at 3... | 20:11 | |
| *** angdraug has joined #openstack-nova | 20:11 | |
| mriedem | probably not | 20:11 |
| rlrossit | the PTL has spoken | 20:11 |
| cfriesen | all hail | 20:11 |
| mriedem | unless you live alone and don't have plans for the weekend, and are under 30, then yes, get started | 20:11 |
| *** pushkaru has quit IRC | 20:11 | |
| rlrossit | depends on what constitutes as "plans" | 20:12 |
| * bizarrodan disappears in a cloud of dust | 20:12 | |
| *** mastermind has quit IRC | 20:13 | |
| *** gjayavelu has quit IRC | 20:13 | |
| lutf-8 | bizarrodan: mriedem just got back from something. I see mriedem pushed the big red button | 20:13 |
| bizarrodan | lutf-8: is that the buzzing I hear? | 20:14 |
| lutf-8 | yep. prepare for impact | 20:14 |
| * bizarrodan puts he head between his knees | 20:14 | |
| mriedem | was i not supposed to? | 20:15 |
| mriedem | i saw you guys all playing hot potato | 20:15 |
| bizarrodan | you were | 20:15 |
| lutf-8 | yeah, it's fine. I was about to | 20:16 |
| *** sdake has quit IRC | 20:17 | |
| *** vilobhmm11 has joined #openstack-nova | 20:18 | |
| *** rlrossit has quit IRC | 20:18 | |
| cfriesen | lawn_bauwer: if the expectation is that shared block storage implies shared instance path, then why does LibvirtDriver.instance_on_disk() explicitly check both cases separately? | 20:19 |
| *** gjayavel_ has quit IRC | 20:20 | |
| cfriesen | lawn_bauwer: the comment in that routine specifically talks about shared block storage but not shared filesystem | 20:20 |
| *** vilobhmm111 has joined #openstack-nova | 20:20 | |
| lawn_bauwer | cfriesen: okay, all of that is a pain to review | 20:20 |
| *** pushkaru has joined #openstack-nova | 20:21 | |
| lawn_bauwer | and I agree, I can't find something that specific that I mentioned | 20:21 |
| lawn_bauwer | my experience was rather coming from evacuations | 20:21 |
| lawn_bauwer | when you evacuate an instance, there is a specific check when initing the node that verifies whether the instance was moved and consequently delete local files if so | 20:22 |
| *** vilobhmm11 has quit IRC | 20:22 | |
| lawn_bauwer | that's messed up when you have local disks + instances on shared storage | 20:22 |
| cfriesen | _destroy_evacuated_instances()? | 20:23 |
| openstackgerrit | Diana Clarke proposed openstack/nova: Fix: unable to delete instance when cinder is down https://review.openstack.org/298997 | 20:23 |
| lawn_bauwer | cfriesen: yup | 20:23 |
| *** salv-orl_ has joined #openstack-nova | 20:23 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Move config options from nova/api directory https://review.openstack.org/258180 | 20:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the API options (4) https://review.openstack.org/297237 | 20:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the API options (2) https://review.openstack.org/260587 | 20:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the API options (1) https://review.openstack.org/260586 | 20:24 |
| cfriesen | lawn_bauwer: I think that's okay, actually. It calls "destroy_disks = not (self._is_instance_storage_shared()), and that routine calls self.driver.check_instance_shared_storage_local() | 20:24 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the API options (3) https://review.openstack.org/260588 | 20:24 |
| cfriesen | which in turn calls self.image_backend.backend().is_shared_block_storage() | 20:24 |
| cfriesen | which is different for the rbd backend | 20:25 |
| cfriesen | so in _destroy_evacuated_instances() we end up with destroy_disks as False | 20:25 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the legacy API options https://review.openstack.org/273266 | 20:26 |
| *** salv-orlando has quit IRC | 20:26 | |
| cfriesen | But regardless of that, I think we've got a problem with resizing an rbd-backed instance. I think we need a routine that handles the ceph instructions for duplicating the original disk so that we can immediately revert to it if we revert the resize. | 20:27 |
| lawn_bauwer | cfriesen: possibly | 20:27 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the OSAPI options https://review.openstack.org/274879 | 20:28 |
| cfriesen | or am I missing a call to Image.clone() somewhere | 20:29 |
| cfriesen | nope, don't think so. | 20:30 |
| *** pwnall has quit IRC | 20:32 | |
| lawn_bauwer | cfriesen: you're right, but destroy_disks set to False is problematic when you still have your local directory, fair to say ? | 20:33 |
| *** sdake has joined #openstack-nova | 20:33 | |
| lawn_bauwer | cfriesen: the RBD image driver tells nova to not delete the disk, but it should still delete the local /var/lib/nova/instances dir | 20:33 |
| lawn_bauwer | including the libvirt xml and so on | 20:34 |
| *** mastermind has joined #openstack-nova | 20:34 | |
| cfriesen | lawn_bauwer: I think it'll still do that. see the special-case in LibvirtDriver.cleanup(). Even with destory_disks false it'll still call self.delete_instance_files() | 20:35 |
| *** eglynn has quit IRC | 20:36 | |
| cfriesen | because is_shared_block_storage is true. Or at least that's the intent. | 20:36 |
| openstackgerrit | Diana Clarke proposed openstack/nova: Fix: unable to delete instance when cinder is down https://review.openstack.org/298997 | 20:37 |
| cfriesen | We think there might be a bug there too, actually, for the case where migrate_data is None but the local compute node is using RBD storage. | 20:37 |
| *** zenoway has quit IRC | 20:37 | |
| *** markvoelker has quit IRC | 20:37 | |
| *** inc0 has joined #openstack-nova | 20:37 | |
| inc0 | hey guys, in Liberty upgrade_levels auto wasn't a thing yet right? | 20:38 |
| *** apoorvad has joined #openstack-nova | 20:38 | |
| bizarrodan | correct | 20:38 |
| inc0 | should I pin it to 7 for libertty? | 20:38 |
| *** ljxiash has joined #openstack-nova | 20:40 | |
| *** claudiub has quit IRC | 20:40 | |
| *** browne has quit IRC | 20:41 | |
| *** raildo is now known as raildo-afk | 20:42 | |
| *** ljxiash has quit IRC | 20:44 | |
| *** mylu has joined #openstack-nova | 20:45 | |
| *** browne has joined #openstack-nova | 20:45 | |
| *** gjayavelu has joined #openstack-nova | 20:46 | |
| mriedem | inc0: i believe you set it to 'kilo', see https://github.com/openstack/nova/blob/stable/liberty/nova/compute/rpcapi.py#L307 | 20:50 |
| mriedem | inc0: because your conductor/api should be liberty, and your computes would still be kilo | 20:50 |
| mriedem | so liberty conductor needs to convert liberty things to send to the kilo computes | 20:50 |
| mriedem | until you get all of your kilo computes upgraded to liberty | 20:50 |
| inc0 | mriedem, yeah, my mistake, it should be 12 all the way right? | 20:50 |
| mriedem | 12? | 20:50 |
| mriedem | 'kilo' | 20:50 |
| mriedem | it's an alias | 20:50 |
| mriedem | https://github.com/openstack/nova/blob/stable/liberty/nova/compute/rpcapi.py#L307 | 20:50 |
| inc0 | ok | 20:51 |
| inc0 | and...for liberty correct value will be | 20:51 |
| inc0 | ? | 20:51 |
| mriedem | 'kilo' | 20:51 |
| inc0 | ( I want to deploy all liberty) | 20:51 |
| inc0 | ok | 20:51 |
| mriedem | then you don't pin | 20:51 |
| mriedem | reading http://www.danplanet.com/blog/2015/10/05/upgrades-in-nova-rpc-apis/ might help | 20:51 |
| mriedem | the pin is for rolling upgrades | 20:51 |
| inc0 | well I need to pin because then we'll upgrade to mitaka | 20:52 |
| mriedem | inc0: if your conductor is mitaka, then you can set it to 'auto' or 'liberty' | 20:53 |
| mriedem | https://github.com/openstack/nova/blob/stable/mitaka/nova/compute/rpcapi.py#L333 | 20:53 |
| inc0 | hold on. I deploy liberty right now, which means as long as I don't want to upgrade it's all good | 20:55 |
| inc0 | but if I upgrade, start with conductors | 20:55 |
| inc0 | then I can have auto level and it will be enough right? | 20:55 |
| *** delatte has quit IRC | 20:56 | |
| *** ctrath has quit IRC | 20:56 | |
| *** mriedem1 has joined #openstack-nova | 20:56 | |
| mriedem1 | yeah, see https://github.com/openstack/nova/blob/stable/liberty/nova/objects/service.py#L33 | 20:56 |
| bizarrodan | inc0: are you talking about upgrading from liberty to mitaka? | 20:56 |
| *** gjayavelu has quit IRC | 20:56 | |
| inc0 | bizarrodan, yup | 20:57 |
| *** zenoway has joined #openstack-nova | 20:57 | |
| bizarrodan | inc0: you can do auto only on mitaka code, so when upgrading your conductors you can then choose auto | 20:57 |
| bizarrodan | but not before | 20:57 |
| *** MVenesio has quit IRC | 20:57 | |
| inc0 | but if I deploy liberty without upgrade_levels at all | 20:57 |
| inc0 | and then upgrade conductors with upgrade_levels=auto | 20:58 |
| *** eharney has quit IRC | 20:58 | |
| bizarrodan | that should be fine | 20:58 |
| inc0 | conductors will detect that computes are still liberty and will do stuff correctly right? | 20:58 |
| inc0 | ok, thanks Dan | 20:58 |
| bizarrodan | that's the idea | 20:58 |
| bizarrodan | that's how it works in upstream grenade for mitaka | 20:58 |
| *** eharney has joined #openstack-nova | 20:59 | |
| *** mriedem has quit IRC | 20:59 | |
| *** mastermind has quit IRC | 20:59 | |
| inc0 | ok, good to know | 20:59 |
| *** haomaiwang has quit IRC | 21:01 | |
| *** haomaiwang has joined #openstack-nova | 21:01 | |
| *** haplo37 has quit IRC | 21:01 | |
| openstackgerrit | Brian Rosmaita proposed openstack/nova-specs: Consistent Success Reporting https://review.openstack.org/248248 | 21:02 |
| *** ssurana has quit IRC | 21:04 | |
| openstackgerrit | Anusha Unnam proposed openstack/nova: Config options: Centralize xenapi driver options https://review.openstack.org/299495 | 21:09 |
| *** devananda has quit IRC | 21:12 | |
| *** devananda has joined #openstack-nova | 21:13 | |
| *** ccarmack has quit IRC | 21:13 | |
| *** cdent has quit IRC | 21:14 | |
| *** thorst has quit IRC | 21:17 | |
| bizarrodan | lutf-8: beat me to it | 21:18 |
| bizarrodan | lutf-8: I was watching it fail and was going to jump on the recheck | 21:18 |
| lutf-8 | I was surprised you hadn't gotten it already | 21:19 |
| *** vtech has quit IRC | 21:19 | |
| bizarrodan | dealing with something else :/ | 21:19 |
| *** rfolco has quit IRC | 21:21 | |
| bizarrodan | lutf-8: we are like 90 minutes away from having all our flavors in the api database | 21:22 |
| bizarrodan | cellsv2 or bust | 21:22 |
| lutf-8 | \o/ | 21:22 |
| bizarrodan | after this and a mandatory cooling down period, I should find something else to bite off | 21:22 |
| bizarrodan | something easier | 21:22 |
| bizarrodan | with nothing to do with flavors. | 21:22 |
| *** inc0 has quit IRC | 21:23 | |
| *** zengine has quit IRC | 21:24 | |
| lutf-8 | there's plenty of other things that could use migrating :) | 21:24 |
| *** ayoung is now known as ayoung-afk | 21:25 | |
| lutf-8 | there's a few things in progress that could potentially use help. much of the big untouched work is around multiple cells support I think | 21:26 |
| *** rtheis has quit IRC | 21:26 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova-specs: Get Me a Network https://review.openstack.org/283206 | 21:27 |
| bizarrodan | yeah, well, it's still friday so I will rest on my current success and worry about more next week | 21:28 |
| mriedem1 | bizarrodan: lutf-8: you could use the success bot to post that flavors are in the API DB once this lands | 21:28 |
| bizarrodan | mriedem1: I could, but I'm not sure I'd want to | 21:29 |
| mriedem1 | you can review my get-me-a-network spec :) | 21:29 |
| *** mriedem1 is now known as mriedem | 21:29 | |
| mriedem | now with fancy jsonschema! | 21:29 |
| *** Sukhdev has joined #openstack-nova | 21:29 | |
| mriedem | melwitt: btw, good questions on the get me a network spec | 21:29 |
| *** akshai_ has quit IRC | 21:30 | |
| *** ayoung-afk is now known as ayoung | 21:30 | |
| *** ayoung is now known as ayoung_afk | 21:30 | |
| *** Sukhdev has quit IRC | 21:31 | |
| *** haplo37 has joined #openstack-nova | 21:31 | |
| *** Goneri has joined #openstack-nova | 21:31 | |
| mriedem | easy bug fix https://bugs.launchpad.net/nova/+bug/1564551 | 21:33 |
| openstack | Launchpad bug 1564551 in OpenStack Compute (nova) "initialize_connection failure can result in 'exceptions.TypeError' object has no attribute 'code'" [Medium,Triaged] | 21:33 |
| *** baoli has quit IRC | 21:36 | |
| *** thorst has joined #openstack-nova | 21:38 | |
| *** mriedem has quit IRC | 21:38 | |
| *** thorst has quit IRC | 21:42 | |
| *** daemontool has joined #openstack-nova | 21:44 | |
| *** ayoung_afk is now known as ayoung | 21:49 | |
| *** pushkaru has quit IRC | 21:49 | |
| *** xyang1 has quit IRC | 21:49 | |
| *** Sukhdev has joined #openstack-nova | 21:52 | |
| *** ivase has quit IRC | 21:55 | |
| *** Sukhdev has quit IRC | 21:55 | |
| *** ivase has joined #openstack-nova | 21:55 | |
| *** sdague has quit IRC | 21:56 | |
| *** Sukhdev has joined #openstack-nova | 22:00 | |
| *** haomaiwang has quit IRC | 22:01 | |
| *** ducttape_ has quit IRC | 22:01 | |
| *** markvoelker_ has joined #openstack-nova | 22:01 | |
| *** haomaiwang has joined #openstack-nova | 22:01 | |
| openstackgerrit | Anusha Unnam proposed openstack/nova: Config options: Centralize xenapi driver options https://review.openstack.org/299495 | 22:02 |
| *** dimtruck is now known as zz_dimtruck | 22:03 | |
| *** daemontool has quit IRC | 22:03 | |
| *** notartom has quit IRC | 22:03 | |
| *** fawadkhaliq has quit IRC | 22:05 | |
| *** fawadkhaliq has joined #openstack-nova | 22:05 | |
| *** pmathews has quit IRC | 22:06 | |
| *** Sukhdev has quit IRC | 22:07 | |
| *** ZZelle_ has joined #openstack-nova | 22:10 | |
| *** figleaf is now known as edleafe | 22:11 | |
| *** lutf-8 is now known as alaski | 22:12 | |
| *** sdake_ has joined #openstack-nova | 22:13 | |
| *** sdake has quit IRC | 22:13 | |
| *** edmondsw has quit IRC | 22:15 | |
| *** vilobhmm11 has joined #openstack-nova | 22:16 | |
| *** vilobhmm111 has quit IRC | 22:17 | |
| openstackgerrit | Sujitha proposed openstack/nova: Config options: Centralize neutron options https://review.openstack.org/300681 | 22:18 |
| *** sdake has joined #openstack-nova | 22:19 | |
| *** sdake_ has quit IRC | 22:21 | |
| *** mdrabe has quit IRC | 22:21 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Create Instance from BuildRequest if not in a cell https://review.openstack.org/263927 | 22:23 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Change BuildRequest to contain a serialized instance https://review.openstack.org/298455 | 22:23 |
| *** Sukhdev has joined #openstack-nova | 22:25 | |
| *** sheel has quit IRC | 22:27 | |
| *** browne has quit IRC | 22:28 | |
| *** ametts has quit IRC | 22:28 | |
| *** achanda has joined #openstack-nova | 22:29 | |
| *** aysyd has quit IRC | 22:31 | |
| *** mylu has quit IRC | 22:32 | |
| *** mylu has joined #openstack-nova | 22:33 | |
| *** sarafraj has left #openstack-nova | 22:35 | |
| *** markvoelker_ has quit IRC | 22:39 | |
| *** Marga_ has quit IRC | 22:39 | |
| bizarrodan | hoo boy, it's in teh gatez | 22:41 |
| openstackgerrit | Jay Pipes proposed openstack/os-vif: linux_bridge: convert over to use privsep module https://review.openstack.org/287725 | 22:41 |
| openstackgerrit | Jay Pipes proposed openstack/os-vif: ovs: move code from plugin into linux_net helper https://review.openstack.org/299405 | 22:43 |
| *** krtaylor has quit IRC | 22:44 | |
| *** fawadkhaliq has quit IRC | 22:44 | |
| *** fawadkhaliq has joined #openstack-nova | 22:45 | |
| *** daemontool has joined #openstack-nova | 22:51 | |
| *** knikolla has quit IRC | 22:53 | |
| *** krtaylor has joined #openstack-nova | 22:56 | |
| *** haomaiwang has quit IRC | 23:01 | |
| *** Marga_ has joined #openstack-nova | 23:01 | |
| *** mgoddard has joined #openstack-nova | 23:01 | |
| *** haomaiwang has joined #openstack-nova | 23:01 | |
| *** mylu has quit IRC | 23:03 | |
| *** fawadkhaliq has quit IRC | 23:04 | |
| *** auggy_rodin is now known as auggy | 23:04 | |
| *** fawadkhaliq has joined #openstack-nova | 23:04 | |
| *** Marga_ has quit IRC | 23:06 | |
| *** pmathews has joined #openstack-nova | 23:07 | |
| *** mgoddard has quit IRC | 23:11 | |
| *** mtanino has quit IRC | 23:13 | |
| *** Swami has quit IRC | 23:13 | |
| *** Marga_ has joined #openstack-nova | 23:14 | |
| *** pmathews has quit IRC | 23:14 | |
| *** yassou has quit IRC | 23:14 | |
| *** armax has quit IRC | 23:17 | |
| *** daemontool has quit IRC | 23:17 | |
| *** daemontool has joined #openstack-nova | 23:18 | |
| *** armax has joined #openstack-nova | 23:19 | |
| *** knikolla has joined #openstack-nova | 23:22 | |
| *** haplo37 has quit IRC | 23:22 | |
| *** Sukhdev has quit IRC | 23:23 | |
| *** erlon has quit IRC | 23:27 | |
| *** angdraug has quit IRC | 23:34 | |
| *** harlowja has quit IRC | 23:36 | |
| *** yamahata has quit IRC | 23:36 | |
| *** zenoway has quit IRC | 23:37 | |
| *** marcusvrn_ has quit IRC | 23:37 | |
| *** daemontool has quit IRC | 23:38 | |
| *** nic has quit IRC | 23:46 | |
| *** zenoway has joined #openstack-nova | 23:49 | |
| *** zenoway has quit IRC | 23:53 | |
| *** ivase_ has joined #openstack-nova | 23:55 | |
| *** efoley has quit IRC | 23:58 | |
| *** ivase has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
