| *** otter768 has quit IRC | 00:00 | |
| *** IanGovett has quit IRC | 00:00 | |
| *** annegentle has joined #openstack-nova | 00:01 | |
| *** achanda has quit IRC | 00:01 | |
| openstackgerrit | Joshua Hesketh proposed openstack/nova: Wedge DB migrations if flavor migrations are not complete https://review.openstack.org/174480 | 00:02 |
|---|---|---|
| *** markvoelker has joined #openstack-nova | 00:02 | |
| *** annegentle has quit IRC | 00:06 | |
| *** dsanders has joined #openstack-nova | 00:06 | |
| *** markvoelker has quit IRC | 00:08 | |
| *** dsanders1 has joined #openstack-nova | 00:09 | |
| *** dsanders has quit IRC | 00:09 | |
| *** park_heijlong has joined #openstack-nova | 00:12 | |
| *** tsekiyam_ has joined #openstack-nova | 00:14 | |
| *** Kevin_Zheng has joined #openstack-nova | 00:18 | |
| *** tsekiyama has quit IRC | 00:18 | |
| *** tsekiyam_ has quit IRC | 00:19 | |
| *** mtanino has quit IRC | 00:19 | |
| *** davideagnello has quit IRC | 00:20 | |
| *** VW_ has joined #openstack-nova | 00:21 | |
| *** achanda has joined #openstack-nova | 00:23 | |
| *** david-lyle has quit IRC | 00:24 | |
| *** VW_ has quit IRC | 00:26 | |
| *** davideagnello has joined #openstack-nova | 00:27 | |
| *** igordcard has quit IRC | 00:27 | |
| *** david-lyle has joined #openstack-nova | 00:27 | |
| *** achanda has quit IRC | 00:28 | |
| *** Marga_ has quit IRC | 00:32 | |
| *** vilobhmm11 has joined #openstack-nova | 00:32 | |
| *** vilobhmm11 has quit IRC | 00:33 | |
| *** vilobhmm11 has joined #openstack-nova | 00:33 | |
| *** davideagnello has quit IRC | 00:33 | |
| *** vilobhmm11 has quit IRC | 00:34 | |
| *** vilobhmm12 has joined #openstack-nova | 00:34 | |
| *** emagana has quit IRC | 00:36 | |
| *** vilobhmm1 has quit IRC | 00:36 | |
| *** yingjun has joined #openstack-nova | 00:37 | |
| *** jaypipes has quit IRC | 00:37 | |
| *** park_heijlong has quit IRC | 00:38 | |
| *** armax has quit IRC | 00:41 | |
| *** neelashah has joined #openstack-nova | 00:42 | |
| *** Marga_ has joined #openstack-nova | 00:46 | |
| *** promulo_ has joined #openstack-nova | 00:50 | |
| *** promulo__ has quit IRC | 00:54 | |
| *** Marga_ has quit IRC | 00:57 | |
| *** marun has quit IRC | 00:57 | |
| *** browne has quit IRC | 01:00 | |
| openstackgerrit | Eli Qiao proposed openstack/nova: Neutron: Check port binding status https://review.openstack.org/156112 | 01:03 |
| *** annegentle has joined #openstack-nova | 01:03 | |
| *** annashen has joined #openstack-nova | 01:05 | |
| *** tjones1 has quit IRC | 01:06 | |
| *** kaisers has quit IRC | 01:07 | |
| *** annegentle has quit IRC | 01:08 | |
| *** Marga_ has joined #openstack-nova | 01:09 | |
| *** iamjarvo has joined #openstack-nova | 01:10 | |
| *** annashen has quit IRC | 01:10 | |
| *** iamjarvo has joined #openstack-nova | 01:10 | |
| *** patrickeast has quit IRC | 01:11 | |
| *** gyee has quit IRC | 01:12 | |
| *** dsanders1 has quit IRC | 01:16 | |
| *** baoli has joined #openstack-nova | 01:17 | |
| *** baoli_ has joined #openstack-nova | 01:21 | |
| *** kaisers has joined #openstack-nova | 01:22 | |
| *** baoli has quit IRC | 01:24 | |
| *** baoli_ has quit IRC | 01:26 | |
| *** Nic has quit IRC | 01:26 | |
| *** zhangtralon has joined #openstack-nova | 01:27 | |
| *** annashen has joined #openstack-nova | 01:30 | |
| *** popw has joined #openstack-nova | 01:30 | |
| *** rushil has quit IRC | 01:32 | |
| openstackgerrit | Lianhao Lu proposed openstack/nova: Convert bandwidth_usage related timestamp to UTC native datetime https://review.openstack.org/173213 | 01:32 |
| *** vilobhmm12 has quit IRC | 01:33 | |
| *** baoli has joined #openstack-nova | 01:34 | |
| *** erkules_ has joined #openstack-nova | 01:38 | |
| *** flwang has quit IRC | 01:38 | |
| *** erkules has quit IRC | 01:40 | |
| *** iamjarvo has quit IRC | 01:42 | |
| *** annashen has quit IRC | 01:43 | |
| *** park_heijlong has joined #openstack-nova | 01:47 | |
| openstackgerrit | Eli Qiao proposed openstack/nova: Object: Fix incorrect parameter set in flavor save_extra_specs https://review.openstack.org/176157 | 01:48 |
| *** zzzeek has quit IRC | 01:48 | |
| *** changbl has joined #openstack-nova | 01:48 | |
| *** VW_ has joined #openstack-nova | 01:48 | |
| *** VW_ has quit IRC | 01:50 | |
| *** flwang has joined #openstack-nova | 01:50 | |
| *** VW_ has joined #openstack-nova | 01:50 | |
| *** flwang has left #openstack-nova | 01:51 | |
| *** zzzeek has joined #openstack-nova | 01:54 | |
| *** zzzeek has quit IRC | 01:54 | |
| *** BadCub_Away is now known as BadCub | 01:56 | |
| *** otter768 has joined #openstack-nova | 01:56 | |
| *** isd1 has joined #openstack-nova | 01:56 | |
| *** gtt116__ has quit IRC | 01:59 | |
| *** gtt116 has joined #openstack-nova | 01:59 | |
| *** popw has quit IRC | 02:00 | |
| *** dboik has joined #openstack-nova | 02:00 | |
| *** annegentle has joined #openstack-nova | 02:00 | |
| *** popw has joined #openstack-nova | 02:00 | |
| *** otter768 has quit IRC | 02:01 | |
| *** unicell has quit IRC | 02:01 | |
| *** park_heijlong has quit IRC | 02:01 | |
| *** browne has joined #openstack-nova | 02:01 | |
| *** markvoelker has joined #openstack-nova | 02:01 | |
| *** harlowja is now known as harlowja_away | 02:03 | |
| *** tojuvone has joined #openstack-nova | 02:03 | |
| *** dboik has quit IRC | 02:05 | |
| *** annashen has joined #openstack-nova | 02:09 | |
| *** annegentle has quit IRC | 02:09 | |
| *** xyang1 has quit IRC | 02:09 | |
| *** pixelb has quit IRC | 02:11 | |
| *** annashen has quit IRC | 02:15 | |
| *** park_heijlong has joined #openstack-nova | 02:16 | |
| *** yingjun has quit IRC | 02:18 | |
| *** park_heijlong has quit IRC | 02:21 | |
| *** zhangtralon has quit IRC | 02:23 | |
| *** kaisers has quit IRC | 02:24 | |
| *** annegentle has joined #openstack-nova | 02:27 | |
| *** alexpilotti has quit IRC | 02:27 | |
| *** Marga_ has quit IRC | 02:32 | |
| *** park_heijlong has joined #openstack-nova | 02:33 | |
| *** iamjarvo has joined #openstack-nova | 02:33 | |
| *** iamjarvo has joined #openstack-nova | 02:34 | |
| *** neelashah has quit IRC | 02:36 | |
| *** jamielennox is now known as jamielennox|away | 02:39 | |
| *** kaisers has joined #openstack-nova | 02:39 | |
| *** VW_ has quit IRC | 02:48 | |
| *** davideagnello has joined #openstack-nova | 02:53 | |
| *** VW_ has joined #openstack-nova | 02:58 | |
| *** VW__ has joined #openstack-nova | 02:59 | |
| *** VW__ has quit IRC | 03:00 | |
| *** VW_ has quit IRC | 03:00 | |
| *** VW_ has joined #openstack-nova | 03:00 | |
| *** annegentle has quit IRC | 03:10 | |
| *** annegentle has joined #openstack-nova | 03:13 | |
| *** joefides_ has quit IRC | 03:14 | |
| *** joefides has joined #openstack-nova | 03:15 | |
| *** zhangtralon has joined #openstack-nova | 03:15 | |
| *** annashen has joined #openstack-nova | 03:16 | |
| *** popw has quit IRC | 03:16 | |
| *** popw has joined #openstack-nova | 03:17 | |
| *** annegentle has quit IRC | 03:20 | |
| *** annashen has quit IRC | 03:21 | |
| *** annegentle has joined #openstack-nova | 03:21 | |
| *** davideagnello has quit IRC | 03:25 | |
| *** annegentle has quit IRC | 03:26 | |
| *** iamjarvo has quit IRC | 03:27 | |
| *** baoli has quit IRC | 03:34 | |
| *** dboik has joined #openstack-nova | 03:36 | |
| *** dboik_ has joined #openstack-nova | 03:37 | |
| *** dboik has quit IRC | 03:40 | |
| *** hightall has joined #openstack-nova | 03:46 | |
| *** deepthi has joined #openstack-nova | 03:48 | |
| *** annashen has joined #openstack-nova | 03:48 | |
| *** arnaud___ has joined #openstack-nova | 03:50 | |
| *** dave-mccowan has quit IRC | 03:50 | |
| *** MVenesio has quit IRC | 03:51 | |
| *** dboik_ has quit IRC | 03:54 | |
| *** otter768 has joined #openstack-nova | 03:57 | |
| *** zhangtralon has quit IRC | 03:58 | |
| *** otter768 has quit IRC | 04:02 | |
| *** VW__ has joined #openstack-nova | 04:02 | |
| *** fawadkhaliq has joined #openstack-nova | 04:02 | |
| *** dikonoor has joined #openstack-nova | 04:05 | |
| *** wverdugo500 has joined #openstack-nova | 04:05 | |
| *** VW_ has quit IRC | 04:05 | |
| *** krtaylor has quit IRC | 04:06 | |
| *** thomasem has quit IRC | 04:07 | |
| dikonoor | alex_xu : Just saw John's comment on https://review.openstack.org/#/c/175739/ | 04:07 |
| *** dsanders has joined #openstack-nova | 04:07 | |
| *** thomasem has joined #openstack-nova | 04:09 | |
| dikonoor | alex_xu: Like we discussed earlier, the elevated context should not affect the policy enforcement as the enforcement occurs much before the eleveate context gets applied i e in authorize (as you pointed out the other day)..and the context used by authorize is the correct context | 04:10 |
| dikonoor | alex_xu: I don't see John online. I assume he's in another time zone.Is there any other core who can take a look? | 04:10 |
| *** krtaylor has joined #openstack-nova | 04:10 | |
| *** Kevin_Zheng has quit IRC | 04:12 | |
| *** VW__ has quit IRC | 04:15 | |
| *** patrickeast has joined #openstack-nova | 04:15 | |
| *** iamjarvo has joined #openstack-nova | 04:17 | |
| dikonoor | sdague: could you take a look at https://review.openstack.org/#/c/175739/. This is closely related to https://review.openstack.org/#/c/150352/ (which you have reviewed) | 04:19 |
| *** TobiasE has quit IRC | 04:22 | |
| *** annegentle has joined #openstack-nova | 04:22 | |
| *** annashen has quit IRC | 04:25 | |
| *** annegentle has quit IRC | 04:28 | |
| *** sdake has joined #openstack-nova | 04:29 | |
| *** pradipta has joined #openstack-nova | 04:31 | |
| *** sdake_ has joined #openstack-nova | 04:32 | |
| *** dsanders has quit IRC | 04:33 | |
| *** sdake has quit IRC | 04:33 | |
| *** hightall has quit IRC | 04:34 | |
| *** unicell has joined #openstack-nova | 04:36 | |
| *** panbalag has left #openstack-nova | 04:38 | |
| *** hightall has joined #openstack-nova | 04:41 | |
| *** vilobhmm1 has joined #openstack-nova | 04:42 | |
| *** fawadkhaliq has quit IRC | 04:43 | |
| *** kiran-r has joined #openstack-nova | 04:50 | |
| *** patrickeast has quit IRC | 05:01 | |
| *** iamjarvo has quit IRC | 05:03 | |
| *** iamjarvo has joined #openstack-nova | 05:04 | |
| *** iamjarvo has quit IRC | 05:08 | |
| *** penick has joined #openstack-nova | 05:10 | |
| *** fawadkhaliq has joined #openstack-nova | 05:16 | |
| *** nkrinner has joined #openstack-nova | 05:20 | |
| *** emagana has joined #openstack-nova | 05:20 | |
| *** dsanders has joined #openstack-nova | 05:22 | |
| *** annegentle has joined #openstack-nova | 05:24 | |
| *** annashen has joined #openstack-nova | 05:26 | |
| *** annegentle has quit IRC | 05:29 | |
| *** Marga_ has joined #openstack-nova | 05:29 | |
| *** Marga_ has quit IRC | 05:29 | |
| *** yamahata has joined #openstack-nova | 05:29 | |
| *** Marga_ has joined #openstack-nova | 05:30 | |
| *** dsanders has quit IRC | 05:30 | |
| *** coolsvap|afk is now known as coolsvap | 05:31 | |
| *** annashen has quit IRC | 05:32 | |
| openstackgerrit | Joshua Hesketh proposed openstack/nova: Fix migrate_flavor_data string substitution https://review.openstack.org/176172 | 05:32 |
| *** bkopilov has quit IRC | 05:35 | |
| *** penick has quit IRC | 05:36 | |
| *** bkopilov has joined #openstack-nova | 05:37 | |
| *** vilobhmm11 has joined #openstack-nova | 05:41 | |
| *** sdake_ has quit IRC | 05:42 | |
| *** vilobhmm1 has quit IRC | 05:44 | |
| *** haigang has joined #openstack-nova | 05:44 | |
| *** park_heijlong has quit IRC | 05:45 | |
| *** vilobhmm11 has quit IRC | 05:48 | |
| *** achanda has joined #openstack-nova | 05:49 | |
| *** ajayaa has joined #openstack-nova | 05:52 | |
| *** vilobhmm1 has joined #openstack-nova | 05:54 | |
| *** otter768 has joined #openstack-nova | 05:58 | |
| *** dsanders has joined #openstack-nova | 05:59 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge V2 and V2.1 keypair functional tests https://review.openstack.org/176177 | 05:59 |
| *** dsanders has quit IRC | 06:02 | |
| *** otter768 has quit IRC | 06:03 | |
| *** isd1 has quit IRC | 06:04 | |
| *** arnaud___ has quit IRC | 06:05 | |
| *** cfriesen has quit IRC | 06:07 | |
| *** markvoelker has quit IRC | 06:10 | |
| *** dsanders has joined #openstack-nova | 06:12 | |
| openstackgerrit | Eli Qiao proposed openstack/nova: Object: Merge db accesses for extra field of instance https://review.openstack.org/160076 | 06:15 |
| *** ildikov has quit IRC | 06:15 | |
| *** eliqiao1 has joined #openstack-nova | 06:16 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex https://review.openstack.org/171943 | 06:18 |
| *** achanda has quit IRC | 06:20 | |
| *** achanda has joined #openstack-nova | 06:21 | |
| *** abhishekk has joined #openstack-nova | 06:23 | |
| *** vilobhmm1 has quit IRC | 06:24 | |
| *** annegentle has joined #openstack-nova | 06:25 | |
| *** davideagnello has joined #openstack-nova | 06:26 | |
| *** lpetrut has joined #openstack-nova | 06:27 | |
| *** annegentle has quit IRC | 06:29 | |
| *** davideagnello has quit IRC | 06:31 | |
| openstackgerrit | Eli Qiao proposed openstack/nova: Object: Merge db accesses for extra field of instance https://review.openstack.org/160076 | 06:32 |
| openstackgerrit | Eli Qiao proposed openstack/nova: Object: Merge db accesses for extra field of instance https://review.openstack.org/160076 | 06:33 |
| *** sdake has joined #openstack-nova | 06:35 | |
| *** Murali has joined #openstack-nova | 06:37 | |
| *** Marga_ has quit IRC | 06:38 | |
| *** markvoelker has joined #openstack-nova | 06:40 | |
| *** sahid has joined #openstack-nova | 06:43 | |
| openstackgerrit | Stephen Ahn proposed openstack/nova: libvirt: Rollback status when cold migrate failed https://review.openstack.org/144261 | 06:44 |
| *** thomasem has quit IRC | 06:44 | |
| *** markvoelker has quit IRC | 06:45 | |
| *** kaisers has quit IRC | 06:50 | |
| *** emagana has quit IRC | 06:53 | |
| alex_xu | dikonoor: yea, I replied the comment, but John will wakeup after few hours (I guess at least 2 hours) | 06:56 |
| *** browne has quit IRC | 06:58 | |
| dikonoor | alex_xu : ok..will wait for him..Is there any other core you would know who can take a look (and give comments) . I added Sean Dague as he reviewed the path of yours.. | 06:58 |
| *** rajesht has joined #openstack-nova | 06:59 | |
| dikonoor | alex_xu: Aso, I see check-tempest-dsvm-cells failing on a swift test case (apart from pci-test and hyper-v-ci) | 06:59 |
| *** asalkeld has joined #openstack-nova | 06:59 | |
| asalkeld | hi all, when are we getting a new novaclient release? | 07:00 |
| alex_xu | dikonoor: it's fine I think, those are non-voting | 07:00 |
| asalkeld | heat's gate is broken because of : https://bugs.launchpad.net/python-novaclient/+bug/1437244 | 07:00 |
| openstack | Launchpad bug 1437244 in python-novaclient "novaclient error when creating floating ip bulk" [Medium,Fix committed] - Assigned to Eugeniya Kudryashova (ekudryashova) | 07:00 |
| asalkeld | http://logs.openstack.org/75/176175/2/check/gate-heat-python27/00ac0db/console.html | 07:01 |
| asalkeld | it's installing python_novaclient-2.24.0, we need a 2.25 | 07:02 |
| *** gtt116_ has joined #openstack-nova | 07:02 | |
| *** dsanders has quit IRC | 07:03 | |
| openstackgerrit | Deepak C Shetty proposed openstack/nova: WIP: Nova part of volume snapshot improvements https://review.openstack.org/172813 | 07:03 |
| *** kaisers has joined #openstack-nova | 07:03 | |
| *** gtt116 has quit IRC | 07:05 | |
| *** mrodden has quit IRC | 07:06 | |
| *** apuimedo_ has joined #openstack-nova | 07:07 | |
| *** mrodden has joined #openstack-nova | 07:07 | |
| *** romainh has joined #openstack-nova | 07:10 | |
| *** kaisers has quit IRC | 07:11 | |
| *** jyuso1 has joined #openstack-nova | 07:12 | |
| *** doude has joined #openstack-nova | 07:12 | |
| *** coolsvap is now known as coolsvap|afk | 07:13 | |
| *** thomasem has joined #openstack-nova | 07:14 | |
| *** yingjun has joined #openstack-nova | 07:14 | |
| *** pask81 has joined #openstack-nova | 07:16 | |
| gilliard | Morning! | 07:16 |
| *** mrodden has quit IRC | 07:20 | |
| *** dsanders has joined #openstack-nova | 07:21 | |
| eliqiao | check-grenade-dsvm is broken :( | 07:21 |
| abhishekk | rajesh: you around? | 07:21 |
| *** jichen has joined #openstack-nova | 07:21 | |
| *** mrodden has joined #openstack-nova | 07:21 | |
| gilliard | eliqiao: bummer. What's wrong? | 07:22 |
| eliqiao | gilliard: http://logs.openstack.org/57/176157/1/check/check-grenade-dsvm/e4904f7/console.html https://bugs.launchpad.net/openstack-gate/+bug/1446847 | 07:24 |
| openstack | Launchpad bug 1446847 in OpenStack-Gate "stable/juno gate is busted on stevedore/cliff releases on 4/21" [Undecided,New] | 07:24 |
| gilliard | oooooh a dependency release has broken the gate. | 07:24 |
| *** annegentle has joined #openstack-nova | 07:25 | |
| *** kaisers has joined #openstack-nova | 07:26 | |
| *** yingjun_ has joined #openstack-nova | 07:27 | |
| *** yingjun_ has quit IRC | 07:27 | |
| *** achanda has quit IRC | 07:29 | |
| *** liusheng has joined #openstack-nova | 07:30 | |
| *** annegentle has quit IRC | 07:31 | |
| *** hanzhang has joined #openstack-nova | 07:31 | |
| *** achanda_ has joined #openstack-nova | 07:32 | |
| *** Longgeek has joined #openstack-nova | 07:35 | |
| *** sdake has quit IRC | 07:35 | |
| *** sudipto has joined #openstack-nova | 07:37 | |
| *** dsanders has quit IRC | 07:37 | |
| sudipto | bauzas, Good morning! | 07:37 |
| *** kaisers has quit IRC | 07:39 | |
| *** markvoelker has joined #openstack-nova | 07:41 | |
| *** zhangtralon has joined #openstack-nova | 07:42 | |
| *** jcoufal has joined #openstack-nova | 07:43 | |
| *** tralon_ has joined #openstack-nova | 07:43 | |
| *** andreykurilin__ has joined #openstack-nova | 07:44 | |
| *** achanda_ has quit IRC | 07:45 | |
| *** markvoelker has quit IRC | 07:46 | |
| *** sdake has joined #openstack-nova | 07:48 | |
| hanzhang | hi all, who can help review this patch:https://review.openstack.org/#/c/175742/ | 07:49 |
| *** jistr has joined #openstack-nova | 07:50 | |
| *** josecastroleon has joined #openstack-nova | 07:52 | |
| openstackgerrit | Eli Qiao proposed openstack/nova: Set power state to None when evaucate failed do to recreate not support https://review.openstack.org/173150 | 07:54 |
| *** kaisers has joined #openstack-nova | 07:55 | |
| *** otter768 has joined #openstack-nova | 07:59 | |
| *** lucasagomes has joined #openstack-nova | 08:00 | |
| *** yingjun has quit IRC | 08:00 | |
| *** sdake has quit IRC | 08:00 | |
| ndipanov | hanzhang, that code is so bad that I think we should just leave it alone... | 08:00 |
| *** josecastroleon has quit IRC | 08:01 | |
| *** yingjun has joined #openstack-nova | 08:02 | |
| openstackgerrit | Bartosz Fic proposed openstack/nova: Live migration failure in API doesn't set vm_state to Error https://review.openstack.org/168916 | 08:03 |
| *** otter768 has quit IRC | 08:04 | |
| *** yamahata has quit IRC | 08:04 | |
| *** alexpilotti has joined #openstack-nova | 08:04 | |
| openstackgerrit | Angus Salkeld proposed openstack/python-novaclient: Fix Floating ip __repr__() https://review.openstack.org/176228 | 08:05 |
| *** matrohon has joined #openstack-nova | 08:06 | |
| *** matrohon has quit IRC | 08:06 | |
| *** matrohon has joined #openstack-nova | 08:06 | |
| *** josecastroleon_ has joined #openstack-nova | 08:07 | |
| *** yingjun has quit IRC | 08:08 | |
| *** josecastroleon_ has quit IRC | 08:10 | |
| *** ildikov has joined #openstack-nova | 08:10 | |
| *** Nikolay_St has joined #openstack-nova | 08:10 | |
| hanzhang | I want to know why? | 08:13 |
| *** rushiagr_away is now known as rushiagr | 08:13 | |
| *** ildikov has quit IRC | 08:13 | |
| hanzhang | ndipanov, can you explain why? | 08:14 |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for keypair_* https://review.openstack.org/176231 | 08:17 |
| *** atuvenie_ has joined #openstack-nova | 08:18 | |
| *** ildikov has joined #openstack-nova | 08:19 | |
| ndipanov | hanzhang, sure | 08:22 |
| ndipanov | it's racy | 08:22 |
| ndipanov | you can't check quotas on the neutron side without reserving them | 08:23 |
| ndipanov | otherwise that information is useless | 08:23 |
| *** abhishekk has quit IRC | 08:23 | |
| ndipanov | because it may be out of date at any later point | 08:23 |
| ndipanov | if you check the nova-net implementation of that method - it would do nothing and defer to the quota mechanism that will do reservations | 08:24 |
| *** sandywalsh has quit IRC | 08:24 | |
| *** asalkeld has left #openstack-nova | 08:24 | |
| *** tralon_ has quit IRC | 08:24 | |
| *** sandywalsh has joined #openstack-nova | 08:24 | |
| *** derekh has joined #openstack-nova | 08:25 | |
| *** eglynn has joined #openstack-nova | 08:25 | |
| *** eglynn is now known as eglynn-pto | 08:26 | |
| *** annegentle has joined #openstack-nova | 08:26 | |
| *** yingjun has joined #openstack-nova | 08:26 | |
| *** wverdugo500 has quit IRC | 08:27 | |
| *** oro has quit IRC | 08:27 | |
| *** krtaylor has quit IRC | 08:29 | |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for keypair_* https://review.openstack.org/176231 | 08:29 |
| *** annashen has joined #openstack-nova | 08:30 | |
| *** thomasem has quit IRC | 08:30 | |
| *** annegentle has quit IRC | 08:31 | |
| *** rook has quit IRC | 08:31 | |
| *** Administrator_ has joined #openstack-nova | 08:32 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:32 | |
| hanzhang | ndipanov, I agree this code is bad, but when a tenant has lots of VMs, 'list_ports' is expensive, if port quota is -1, it does not need check? | 08:32 |
| *** haigang has quit IRC | 08:32 | |
| *** rook has joined #openstack-nova | 08:32 | |
| *** med_ has quit IRC | 08:32 | |
| *** pixelb has joined #openstack-nova | 08:33 | |
| *** thomasem has joined #openstack-nova | 08:34 | |
| *** annashen has quit IRC | 08:34 | |
| *** krtaylor has joined #openstack-nova | 08:34 | |
| *** med_ has joined #openstack-nova | 08:35 | |
| *** med_ has quit IRC | 08:36 | |
| *** med_ has joined #openstack-nova | 08:36 | |
| *** fawadkhaliq has quit IRC | 08:36 | |
| openstackgerrit | ShaoHe Feng proposed openstack/nova-specs: Attach/detach SR-IOV interface https://review.openstack.org/139910 | 08:38 |
| *** wverdugo500 has joined #openstack-nova | 08:39 | |
| openstackgerrit | Bartosz Fic proposed openstack/nova: Live migration failure in API doesn't set vm_state to Error https://review.openstack.org/168916 | 08:40 |
| *** coolsvap|afk is now known as coolsvap | 08:40 | |
| *** haigang has joined #openstack-nova | 08:41 | |
| *** markvoelker has joined #openstack-nova | 08:42 | |
| *** gtt116 has joined #openstack-nova | 08:45 | |
| *** gtt116_ has quit IRC | 08:45 | |
| *** sudipto has quit IRC | 08:45 | |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for keypair https://review.openstack.org/176231 | 08:45 |
| *** haigang has quit IRC | 08:46 | |
| *** haigang has joined #openstack-nova | 08:47 | |
| *** markvoelker has quit IRC | 08:47 | |
| *** inhumani` has joined #openstack-nova | 08:49 | |
| *** andreykurilin__ has quit IRC | 08:49 | |
| *** inhumani` has left #openstack-nova | 08:50 | |
| *** inhumanitas` has quit IRC | 08:50 | |
| *** andreykurilin__ has joined #openstack-nova | 08:50 | |
| *** coolsvap is now known as coolsvap|afk | 08:51 | |
| *** hightall has quit IRC | 08:52 | |
| *** haigang has quit IRC | 08:52 | |
| *** yingjun has quit IRC | 08:54 | |
| *** lpetrut has quit IRC | 08:55 | |
| *** gmann has quit IRC | 08:55 | |
| *** zz_johnthetubagu is now known as johnthetubaguy | 08:55 | |
| *** inhumanitas has joined #openstack-nova | 08:56 | |
| *** GnomeKing has joined #openstack-nova | 08:57 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: Allow discrete blocks of online pCPUs for pinning https://review.openstack.org/173187 | 08:57 |
| *** sudipto has joined #openstack-nova | 08:57 | |
| *** GnomeKing is now known as Guest28882 | 08:58 | |
| *** haigang has joined #openstack-nova | 08:58 | |
| dikonoor | johnthetubaguy: this is around your review comments on https://review.openstack.org/#/c/175739/ . Alex_xu and I have commented on top of them. Could you take a look | 09:00 |
| *** sudipto has quit IRC | 09:02 | |
| *** yingjun has joined #openstack-nova | 09:03 | |
| openstackgerrit | Alex Xu proposed openstack/nova: Add missing policy for limits extension https://review.openstack.org/173178 | 09:03 |
| *** igordcard has joined #openstack-nova | 09:05 | |
| *** arnaud___ has joined #openstack-nova | 09:06 | |
| *** nkrinner has quit IRC | 09:07 | |
| *** fawadkhaliq has joined #openstack-nova | 09:07 | |
| *** nkrinner has joined #openstack-nova | 09:09 | |
| *** arnaud___ has quit IRC | 09:10 | |
| *** jlanoux has joined #openstack-nova | 09:11 | |
| johnthetubaguy | dikonoor: looking | 09:13 |
| dikonoor | johnthetubaguy:ok | 09:13 |
| *** zzz_pczesno is now known as pczesno | 09:15 | |
| johnthetubaguy | dikonoor: alex_xu: sorry, I was missing this bit, I see what you are seeing in the code now :) https://github.com/openstack/nova/blob/master/nova/compute/flavors.py#L185 | 09:16 |
| alex_xu | johnthetubaguy: thanks for review :) | 09:16 |
| alex_xu | dikonoor: and thanks for the fix :) | 09:16 |
| *** rushiagr is now known as rushiagr_away | 09:17 | |
| *** sudipto has joined #openstack-nova | 09:18 | |
| *** Administrator_ has quit IRC | 09:19 | |
| dikonoor | johnthetubaguy , alex_xu : Thanks for review and approval. Is there some other core who can take a look | 09:20 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Sync latest code from oslo-incubator https://review.openstack.org/150058 | 09:21 |
| *** afazekas_ has joined #openstack-nova | 09:21 | |
| dikonoor | johnthetubaguy: Need one more tick mark :) | 09:21 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: unused imported option default_ephemeral_format https://review.openstack.org/176236 | 09:22 |
| bauzas | johnthetubaguy: alex_xu: just wondering if we could just rename get_admin_context() to something like get_dumb_empty_admin_context() to make things clearer :) | 09:23 |
| johnthetubaguy | dikonoor: its got the correct priority, etc, so hopefully folks will join | 09:23 |
| dikonoor | johnthetubaguy: ok, then i will wait | 09:24 |
| johnthetubaguy | bauzas: we should probably just stop using it so much! | 09:24 |
| alex_xu | bauzas: there still have code to call get_admin_context(), in the last step, we can remove all the elavated context in all the code. Then probably we can remove get_admin_context() directly | 09:24 |
| johnthetubaguy | alex_xu: +1 | 09:24 |
| bauzas | johnthetubaguy: was just kidding :) | 09:25 |
| bauzas | alex_xu: yeah that's the best path, a single context.RequestContext(None, None) is better IMHO | 09:26 |
| alex_xu | bauzas: yea | 09:26 |
| bauzas | alex_xu: and you have my blessing for removing get_admin_context() :) | 09:27 |
| *** annegentle has joined #openstack-nova | 09:27 | |
| alex_xu | bauzas: ok cool will add you as reviewer when remove it, add to my todo list (it probably months later) :) | 09:28 |
| *** ankit_ag has joined #openstack-nova | 09:29 | |
| bauzas | alex_xu: what's the status with removing all DB-related context checks ? | 09:29 |
| bauzas | alex_xu: I missed the top-level view on that one | 09:29 |
| bauzas | alex_xu: I saw some series merged by Kilo but is there anything left ? | 09:30 |
| alex_xu | bauzas: half of them finished, will continue them in L | 09:30 |
| bauzas | oh ok | 09:30 |
| bauzas | alexyeah, add me in CC please | 09:30 |
| bauzas | oops | 09:30 |
| bauzas | alex_xu: ^ | 09:30 |
| johnthetubaguy | alex_xu: we got you all approved for liberty now right? | 09:30 |
| alex_xu | bauzas: yea | 09:31 |
| alex_xu | johnthetubaguy: you mean the spec? | 09:31 |
| johnthetubaguy | alex_xu: well spec and blueprint | 09:31 |
| johnthetubaguy | alex_xu: seems like good stuff to get done before the summit, thats all I am thinking | 09:31 |
| *** annegentle has quit IRC | 09:31 | |
| bauzas | johnthetubaguy: oh btw. I guess that deadlines will be the sames as for Kilo for specs merges right ? | 09:32 |
| bauzas | johnthetubaguy: I mean, L-1 for non-prio and L-2 for prio ? | 09:32 |
| alex_xu | johnthetubaguy: yea, the main spec merged, one more I will work on is https://review.openstack.org/127863 | 09:32 |
| bauzas | johnthetubaguy: or is this something you want to discuss at the summit ? | 09:32 |
| johnthetubaguy | bauzas: not sure, probably | 09:32 |
| alex_xu | johnthetubaguy: yea, try to catch up | 09:32 |
| *** lpetrut has joined #openstack-nova | 09:32 | |
| bauzas | johnthetubaguy: ack | 09:32 |
| bauzas | johnthetubaguy: I'm just remembering your tick-tock related section in some emails ;) | 09:33 |
| johnthetubaguy | bauzas: yeah, need to get agreement on the priorities and see how much space is left | 09:33 |
| bauzas | :) | 09:34 |
| *** rushiagr_away is now known as rushiagr | 09:37 | |
| *** kaisers has quit IRC | 09:40 | |
| *** markvoelker has joined #openstack-nova | 09:43 | |
| *** kaisers has joined #openstack-nova | 09:43 | |
| *** yingjun has quit IRC | 09:44 | |
| *** erkules_ is now known as erkules | 09:44 | |
| *** erkules has joined #openstack-nova | 09:44 | |
| *** dguitarbite has joined #openstack-nova | 09:45 | |
| *** markvoelker has quit IRC | 09:47 | |
| *** zhangtralon has quit IRC | 09:51 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: enhance method to return pointer_model from image prop https://review.openstack.org/174854 | 09:52 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduce new method to guest tablet device https://review.openstack.org/174853 | 09:52 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: deprecate use_usb_tablet in favor of pointer_model https://review.openstack.org/176242 | 09:52 |
| garyk | bauzas: question - i restarted a host - the scheduler sees that it is up and running, but fails to select it, there is enough resources (it is the only host in the setup). i am trying to debug this - maybe you hit the same issue. looks like a criticil scheduling issue in my opinoin. it is with the latest code | 09:58 |
| bauzas | garyk: which filter is saying 0 hosts ? | 09:59 |
| *** aix has joined #openstack-nova | 10:00 | |
| *** jogo has quit IRC | 10:01 | |
| *** oro has joined #openstack-nova | 10:01 | |
| johnthetubaguy | garyk: watch out for clock drift if you have multiple machines, that frequently bytes some dev setups of mine. | 10:02 |
| garyk | bauzas: ram filter | 10:02 |
| garyk | when i disable that one the next filter... | 10:03 |
| garyk | johnthetubaguy: its on the same devstack setup | 10:03 |
| *** eandersson has joined #openstack-nova | 10:03 | |
| bauzas | garyk: strange, what are saying debug logs ? | 10:03 |
| openstackgerrit | Angus Salkeld proposed openstack/python-novaclient: Fix Floating ip __repr__() https://review.openstack.org/176228 | 10:03 |
| bauzas | garyk: turns out you probably have an issue with the resource tracker | 10:03 |
| garyk | :) | 10:04 |
| *** andreykurilin__ has quit IRC | 10:04 | |
| garyk | i will continue to debug. | 10:04 |
| bauzas | garyk: because filters look at HostState which is populated by the HostManager reading DB | 10:04 |
| bauzas | unless you're using the CachingScheduler | 10:04 |
| sudipto | sahid, thanks for the recheck :) | 10:04 |
| bauzas | so, check also your compute_nodes table | 10:04 |
| bauzas | garyk: ^ | 10:04 |
| garyk | bauzas: no just the default out of the box one in devstack | 10:05 |
| sudipto | bauzas, you have a minute? :) | 10:05 |
| sudipto | bauzas, sorry for intruding into a conversation. (If i am) | 10:05 |
| *** jogo has joined #openstack-nova | 10:06 | |
| sahid | sudipto: np your change is useful i would like to see it merged | 10:07 |
| sudipto | sahid, thanks. I am not able to understand want comment in there and hence wanted bauzas thoughts on it. | 10:07 |
| sudipto | *one comment | 10:08 |
| ankit_ag | johnthetubaguy: Hi, there is a patch under review https://review.openstack.org/#/c/173296/2 where I am returning the non deleted instances in server list if deleted filter is set to false or any invalid value. | 10:09 |
| ankit_ag | johnthetubaguy: I got a comment that we should use strict=True while converting string to boolean but IMO it will change the api spec. Is it ok to make api spec changes at this last moment, Please let me know your opinion. | 10:09 |
| bauzas | sudipto: I'm quite busy this morning by all the reviews I have but you're my 2nd priority :) | 10:11 |
| bauzas | garyk: that I understood, I just wonder what goes wrong | 10:11 |
| sudipto | bauzas, np - anytime you get some time. | 10:12 |
| garyk | bauzas: i am still looking | 10:12 |
| johnthetubaguy | ankit_ag: we can update the spec if it is the correct thing to do | 10:13 |
| johnthetubaguy | ankit_ag: its different for v2 and v2.1 at this point, possibly | 10:14 |
| bauzas | is Lingxian Kong here on IRC ? want to discuss about https://review.openstack.org/#/c/163842/11/nova/compute/resource_tracker.py,cm | 10:14 |
| bauzas | oh man, I so would like to get an IRC roster... | 10:15 |
| ankit_ag | johnthetubaguy: currenty delete filter has issues with both v2 and v2.1 apis. It always returns deleted instances either deleted filter is set to true/false or any invalid values | 10:16 |
| ankit_ag | johnthetubaguy: so if it is acceptable to make api spec changes at this moment, I will use strict=True in my patch | 10:17 |
| eandersson | Anyone seen issues with nova-nonvncproxy getting stuck endlessly here: https://github.com/openstack/nova/blob/stable/kilo/nova/console/websocketproxy.py#L138 | 10:17 |
| johnthetubaguy | ankit_ag: so just looking at the patch its not that straight forward | 10:18 |
| johnthetubaguy | ankit_ag: I was thinking you mean your spec for a feature, but this is a bug fix, so I guess you mean the API "specification" | 10:18 |
| openstackgerrit | Noel Nelson Dsouza proposed openstack/nova: Nova can stop instance when Fake/Actual compute driver used https://review.openstack.org/174746 | 10:18 |
| ankit_ag | johnthetubaguy: yes, its a bug fix | 10:18 |
| sudipto | garyk, I'd be interested to know about the problem as well :) | 10:19 |
| *** IanGovett has joined #openstack-nova | 10:20 | |
| sudipto | PaulMurray, Hi, I am not sure if you are around, just wanted to ask you - a question regarding the monitors in nova - whether they can be made optional? | 10:21 |
| sudipto | PaulMurray, I guess nova.conf is one option? | 10:22 |
| alex_xu | johnthetubaguy: I think we still can provide strict validation for v2.1 even for query string before your spec get agreement https://review.openstack.org/173243 , right? | 10:22 |
| johnthetubaguy | alex_xu: ankit_ag mentioned about your comment here: https://review.openstack.org/#/c/173296/2 but that seems to violate our policy here: https://wiki.openstack.org/wiki/APIChangeGuidelines#Generally_Not_Acceptable | 10:23 |
| *** asalkeld has joined #openstack-nova | 10:23 | |
| ankit_ag | johnthetubaguy: Thank you :) | 10:23 |
| alex_xu | johnthetubaguy: yea, after rethinking, I think we shouldn't change for v2 | 10:23 |
| johnthetubaguy | alex_xu: no worries, I feel its a microversion bump for v2.1 as well though? | 10:24 |
| johnthetubaguy | alex_xu: because that has now been considered "released" | 10:24 |
| alex_xu | johnthetubaguy: yea, if skip that for now is more easy. | 10:25 |
| johnthetubaguy | alex_xu: I mean i like the idea of returning bad request, just needs a version bump I think | 10:26 |
| alex_xu | johnthetubaguy: and I think won't bump microversion just for this single fix, untill something bigger change for server api | 10:26 |
| johnthetubaguy | alex_xu: we need to tidy up those rules on version bumps at some point | 10:26 |
| johnthetubaguy | alex_xu: its a backwards incompatible change, thats my worry here | 10:26 |
| alex_xu | johnthetubaguy: yes, that is incompatible change | 10:27 |
| *** Guest28882 is now known as BobBall | 10:27 | |
| alex_xu | johnthetubaguy: I think we should doc something about what can be changed by microversion | 10:27 |
| *** annegentle has joined #openstack-nova | 10:28 | |
| alex_xu | ankit_ag: sorry for the wrong direction, :( I think johnthetubaguy was right | 10:28 |
| johnthetubaguy | alex_xu: +1 for a dev ref on this | 10:29 |
| johnthetubaguy | alex_xu: no worries, mistakes are a good way to learn, thanks to ankit_ag for pushing hard on this :) | 10:30 |
| alex_xu | johnthetubaguy: Chris think we can fix some small thing in one version bump, as in this patch https://review.openstack.org/#/c/140313/ | 10:30 |
| alex_xu | johnthetubaguy: thanks | 10:30 |
| alex_xu | johnthetubaguy: ok, I will think about dev ref for this | 10:30 |
| ankit_ag | alex_xu: np, I am also like your idea about returning error for invalid requests but the only issue was api spec changes | 10:30 |
| ankit_ag | alex_xu: thank you for your help | 10:31 |
| alex_xu | ankit_ag: thanks too :) | 10:31 |
| johnthetubaguy | alex_xu: yeah, need thought, I think bumping the version has to be cheap, and can be for new things and fixes, but it needs more thought to pin it down | 10:31 |
| *** asalkeld has quit IRC | 10:32 | |
| alex_xu | johnthetubaguy: ok, let work out dev ref for those, then we can get agreement on this. To avoid some confuse for review in the future | 10:32 |
| alex_xu | s/let/let me/ | 10:32 |
| *** annegentle has quit IRC | 10:33 | |
| *** sudipto has quit IRC | 10:33 | |
| johnthetubaguy | alex_xu: awesome, thanks, probably need to cover allowed changes to v2.0 | 10:34 |
| alex_xu | johnthetubaguy: ok, got it, thanks | 10:34 |
| *** jlanoux has quit IRC | 10:35 | |
| eandersson | So anyone had issues with NoVNC Console and ACE/probes? It seems to cause a deadlock for us. | 10:38 |
| eliqiao | johnthetubaguy: alex_xu: just have a quick look at your conversions, so what 's the last agreement for the microversion for backward-incompatible changes if the previous version api has someting incorrect return value etc? | 10:39 |
| eliqiao | take this spec as an example https://review.openstack.org/#/c/163275/ ,alex_xu: gives me -1 because I bring a backward-incompatible changes. but what I am thinking is , acutally the api | 10:41 |
| eliqiao | the api's return code is not so accurate | 10:41 |
| alex_xu | eliqiao: yea, I recheck accroding https://review.openstack.org/#/c/140313/ we allowed correct the return code. | 10:42 |
| eliqiao | alex_xu: cool, thanks , I see you are drafting some dev ref for that, I am +1 on it. thanks for the effort. | 10:43 |
| johnthetubaguy | alex_xu: I would want to bump if we could though, but lets see | 10:43 |
| *** markvoelker has joined #openstack-nova | 10:43 | |
| alex_xu | eliqiao: johnthetubaguy yea, let's get agreement on dev ref | 10:43 |
| dikonoor | alex_xu , eliqiao : I just raised a new bug https://review.openstack.org/#/c/175334/ . In my conversation yesterday with alex_xu , he had mentioned that this might be fixed as part of https://review.openstack.org/#/c/175334/ or as part of another change that comes under bp https://blueprints.launchpad.net/nova/+spec/nova-api-policy-final-part. I found that https://review.openstack.org/#/c/175334/ is addressing problems in /hosts.py and my bug is | 10:46 |
| dikonoor | in hypervisors.py | 10:46 |
| *** Marga_ has joined #openstack-nova | 10:46 | |
| dikonoor | alex_xu , eliqiao : Bug https://bugs.launchpad.net/nova/+bug/1447084 is blocking one of my use cases and it can be fixed by removing the requires_admin_context check in the nova sqlalchemy api layer. | 10:48 |
| openstack | Launchpad bug 1447084 in OpenStack Compute (nova) "view hypervisor details should be controlled by policy.json" [Undecided,New] | 10:48 |
| *** rushiagr is now known as rushiagr_away | 10:48 | |
| *** markvoelker has quit IRC | 10:48 | |
| dikonoor | eliqiao : Would you want to pick this one up . Alex_xu mntioned yesterday that your other changes are targetted for Liberty. Is there any possibility that if fixed , this one can make it for Kilo ? | 10:49 |
| eliqiao | dikonoor: do you mean backport this to kilo ? | 10:50 |
| eliqiao | dikonoor: I see you are using v2 api, right? in my patch, we don't want to break v2 backwards-compatible. | 10:52 |
| eliqiao | dikonoor: so even with my patch, you will still get AdminRequired exception since we remove it from sqlalchemy api layer, but move it to REST api layer. | 10:53 |
| alex_xu | dikonoor: yea, we won't fix that for v2. Maybe I didn't describe that clearly, maybe this can resolve your confuse https://github.com/openstack/nova-specs/blob/master/specs/liberty/approved/nova-api-policy.rst | 10:54 |
| *** deepthi has quit IRC | 10:54 | |
| *** redbeard has joined #openstack-nova | 10:56 | |
| dikonoor | eliqiao , alex_xu : irrespective of at which layer the check is made , at the end of the day the permissions should be controlled by policy.json and admin permissions should not be mandated. Even in Kilo , till few days back non-admins could successfully make this rest call | 10:56 |
| eliqiao | dikonoor: you can use non-admins to call that rest in kilo ? | 10:57 |
| dikonoor | eliqiao , alex_xu : Yes , till few days back , we could | 10:57 |
| dikonoor | eliqiao , alex_xu : and the behavior has changed recently . | 10:58 |
| eliqiao | dikonoor: interesting... | 10:58 |
| *** haigang has quit IRC | 10:58 | |
| dikonoor | eliqiao , alex_xu : It might have worked previously because of an elevated context somewhere (though I couldn't figure out where it was) . So the kilo behavior remians broken at the moment | 10:58 |
| *** IanGovett has quit IRC | 10:58 | |
| openstackgerrit | Sean Dague proposed openstack/python-novaclient: fix FloatingIP repr https://review.openstack.org/176252 | 10:59 |
| alex_xu | dikonoor: let me check... | 10:59 |
| *** erlon has joined #openstack-nova | 10:59 | |
| eliqiao | check this I3ff3976d24aef108a990fb8ff64afd29c96a1e2f | 11:00 |
| eliqiao | bug# 1427351 | 11:00 |
| dikonoor | eliqiao , alex_xu : https://review.openstack.org/#/c/160506/21/nova/api/openstack/compute/contrib/hypervisors.py ? | 11:01 |
| dikonoor | eliqiao , alex_xu : or some other file? | 11:01 |
| *** ildikov has quit IRC | 11:01 | |
| *** PaulMurray has quit IRC | 11:03 | |
| eliqiao | dikonoor: yes, I think so . | 11:04 |
| eliqiao | dikonoor: sorry, not correct.. | 11:04 |
| *** xgsa has joined #openstack-nova | 11:06 | |
| dikonoor | eliqiao : I went through the changes in that file sometime back trying to figure out which commit has changed the behavior but I don't think the changes in that file could have | 11:06 |
| *** aix has quit IRC | 11:07 | |
| eliqiao | dikonoor: yes, me neither.. | 11:08 |
| *** haigang has joined #openstack-nova | 11:08 | |
| openstackgerrit | Radoslav Gerganov proposed openstack/nova: Consolidate the APIs for getting consoles https://review.openstack.org/148509 | 11:10 |
| dikonoor | eliqiao : So the kilo and liberty both behaviors should support non-admins and that's the reason I have opened the defect. I did go through the above changeset before I opened this defect. | 11:11 |
| alex_xu | eliqiao: dikonoor that sounds like introduced by I3ff3976d24aef108a990fb8ff64afd29c96a1e2f | 11:12 |
| dikonoor | alex_xu : which file? | 11:12 |
| alex_xu | dikonoor: this file https://review.openstack.org/#/c/160506/21/nova/api/openstack/compute/contrib/hypervisors.py | 11:14 |
| *** IanGovett has joined #openstack-nova | 11:14 | |
| *** ildikov has joined #openstack-nova | 11:16 | |
| dikonoor | alex_xu: I see the error in this flow .So I am not sure how the above changes could have caused it.2015-04-19 21:34:22.194 23179 TRACE nova.api.openstack.compute.contrib.hypervisors Traceback (most recent call last): | 11:17 |
| dikonoor | 2015-04-19 21:34:22.194 23179 TRACE nova.api.openstack.compute.contrib.hypervisors File "/usr/lib/python2.7/site-packages/nova/api/openstack/compute/contrib/hypervisors.py", line 147, in show | 11:17 |
| dikonoor | 2015-04-19 21:34:22.194 23179 TRACE nova.api.openstack.compute.contrib.hypervisors service = self.host_api.service_get_by_compute_host(context, hyp.host) | 11:17 |
| dikonoor | 2015-04-19 21:34:22.194 23179 TRACE nova.api.openstack.compute.contrib.hypervisors File "/usr/lib/python2.7/site-packages/nova/compute/api.py", line 3451, in service_get_by_compute_host | 11:17 |
| dikonoor | 2015-04-19 21:34:22.194 23179 TRACE nova.api.openstack.compute.contrib.hypervisors return objects.Service.get_by_compute_host(context, host_name) | 11:17 |
| *** jlanoux has joined #openstack-nova | 11:18 | |
| dikonoor | alex_xu : wait..let me check again | 11:18 |
| *** afazekas_ has quit IRC | 11:18 | |
| dikonoor | alex_xu : objects.Service.get_by_host_and_binary got replace by host_api.service_get_by_compute_host(context, hyp.host ..that's the only change I see and objects.Service.get_by_host_and_binary also has a requires_admin check at the db layer | 11:20 |
| dikonoor | alex_xu : this could have changed the behavior, just that I am not able to clearly understand how the change in method has caused it | 11:21 |
| *** redbeard has quit IRC | 11:23 | |
| bauzas | dikonoor: I was the one who changed the method, what is your problem ? | 11:24 |
| bauzas | dikonoor: is it https://bugs.launchpad.net/nova/+bug/1447084 ? | 11:25 |
| openstack | Launchpad bug 1447084 in OpenStack Compute (nova) "view hypervisor details should be controlled by policy.json" [Undecided,New] | 11:25 |
| sdague | johnthetubaguy: https://review.openstack.org/#/c/176252/ is something we should probably fast approve and release | 11:26 |
| sdague | heat is getting blocked on it | 11:26 |
| dikonoor | dikonoor : hi bauzas..yeah.. the view hypervisor details used to work for non-admins till sometime back | 11:26 |
| sdague | it's a bad repr in python novaclient | 11:26 |
| dikonoor | bauzas: hi bauzas..yeah.. the view hypervisor details used to work for non-admins till sometime back | 11:26 |
| alex_xu | dikonoor: it's break more early https://review.openstack.org/#/c/150568/13/nova/api/openstack/compute/contrib/hypervisors.py | 11:26 |
| alex_xu | dikonoor: bauzas , that can be fixed by this patch https://review.openstack.org/#/c/160089/ | 11:27 |
| bauzas | alex_xu: agreed, we just need to have the context check by the API | 11:28 |
| *** annegentle has joined #openstack-nova | 11:28 | |
| *** tojuvone has quit IRC | 11:28 | |
| alex_xu | dikonoor: bauzas and I think we needn't fix that for kilo, the other api in the hypervisor won't works for non-admin user, so the policy.json is totally un-configuable for the user. So I think nobody really use hypervisor api for non-admin user | 11:29 |
| *** pradipta has quit IRC | 11:29 | |
| *** ivasev has joined #openstack-nova | 11:30 | |
| johnthetubaguy | sdague: is that heat on kilo being blcoked? not sure we are allowed to release any more libs until after RC is done | 11:30 |
| sdague | it's heat on master | 11:30 |
| sdague | when the big library release flushed yesterday, it flushed out a bunch of bugs as well | 11:31 |
| sdague | so this would be a 2.24.1 | 11:31 |
| alex_xu | s/bauzas and I/bauzas: And I/.... | 11:31 |
| dikonoor | alex_xu : I agree it's not very configurable in Kilo but we use it in our environment with a non-admin (to view hypervisor details)..It would have been OK if it wasn't working earlier..but it was and now it's not. | 11:31 |
| johnthetubaguy | sdague: right, we would have to back port to the stable branch and release that I guess | 11:32 |
| bauzas | dikonoor: alex_xu: well I'm probably not happy with some API endpoints like the hypervisor ones to be callable by non-admin people | 11:32 |
| *** annashen has joined #openstack-nova | 11:32 | |
| sdague | johnthetubaguy: no, I don't think so | 11:32 |
| sdague | 2.24.0 is the liberty release | 11:32 |
| bauzas | dikonoor: well, that was just a security flaw before IMHO | 11:33 |
| sdague | this issue was never in a release until 2.24.0 | 11:33 |
| johnthetubaguy | sdague: gotcha | 11:33 |
| johnthetubaguy | sdague: I guess this bug is critical then? | 11:33 |
| bauzas | dikonoor: I don't see any reason why an end-user would get the hypervisor fields | 11:33 |
| sdague | johnthetubaguy: yeh | 11:33 |
| *** afazekas_ has joined #openstack-nova | 11:33 | |
| sdague | bauzas: there are more use cases for OpenStack than public cloud | 11:34 |
| *** annegentle has quit IRC | 11:34 | |
| bauzas | sdague: so the policy.json file is here to help | 11:34 |
| bauzas | sdague: don't get me wrong, I'm not saying we should not allow policy.json to allow this endpoint to be callable unless it's admin, just that it was not checked before | 11:35 |
| dikonoor | bauzas: sdague: yeah, so the decision should be based on the policy json file | 11:35 |
| sdague | bauzas: ok | 11:35 |
| bauzas | dikonoor: which should be checked at the API level | 11:35 |
| bauzas | so, by adding a new call within the API method, I just introduced a new check | 11:36 |
| *** annashen has quit IRC | 11:37 | |
| bauzas | dikonoor: but this check was require_admin_context which was calling the context check | 11:37 |
| bauzas | dikonoor: so alex_xu's patch is helping that by removing all those checks at the DB level | 11:37 |
| bauzas | dikonoor: but that should still be valid if nova.context.require_admin_context() returns True | 11:38 |
| sdague | bauzas: so recheck on https://review.openstack.org/#/c/160089/ is going to fail until a few other patches land | 11:38 |
| bauzas | sdague: oh good to know, thnaks | 11:38 |
| bauzas | sdague: do you know which ones ? | 11:39 |
| bauzas | actually, I can just e-r my bad | 11:39 |
| * bauzas checking | 11:39 | |
| dikonoor | bauzas : https://review.openstack.org/#/c/160089/ | 11:39 |
| bauzas | dikonoor: agreed but what bugs are raising those jobs -1 ? | 11:40 |
| bauzas | that, I'll check on my own | 11:40 |
| * alex_xu lose the context.... | 11:40 | |
| bauzas | nvm | 11:41 |
| bauzas | alex_xu: it was unrelated to our previous convo | 11:41 |
| dikonoor | bauzas : sdague: this patch removes the check for admin and then the permissions will be controlled by policy.json at least for the view hypervisor ..but Alex patch is for Liberty | 11:41 |
| *** TobiasE has joined #openstack-nova | 11:41 | |
| bauzas | dikonoor: nope, it won't change the check, just the level where it's done | 11:41 |
| alex_xu | dikonoor: if we want to fix that bug, this should be another fix. We shouldn't backport https://review.openstack.org/#/c/160089/ this patch | 11:42 |
| dikonoor | bauzas : I don't see any changes in hypervisor.py as part of that changeset | 11:42 |
| bauzas | dikonoor: http://git.openstack.org/cgit/openstack/nova/tree/nova/db/sqlalchemy/api.py#n205 | 11:42 |
| alex_xu | I mean if we want to fix that bug in kilo | 11:42 |
| bauzas | dikonoor: ^ that it the check done when you call db.service_get_by_host() | 11:43 |
| sdague | alex_xu: I agree, the old behavior should be fixed in kilo, especially because it's a stack trace | 11:43 |
| *** wuhg has joined #openstack-nova | 11:43 | |
| sdague | but the new behavior probably shouldn't backport | 11:43 |
| *** mpaolino has joined #openstack-nova | 11:44 | |
| bauzas | sdague: unless I'm wrong, there was no policy check on the hypervisors endpoint | 11:44 |
| alex_xu | sdague: ok, for kilo fix is more easy, we can just pass the admin context down to the code. | 11:44 |
| dikonoor | alex_xu : sdague: I agree the new behavior doesn't need to be backported but the old behavior needs to be fixed in Kilo | 11:44 |
| *** markvoelker has joined #openstack-nova | 11:44 | |
| sdague | bauzas: there was never a permissions check? | 11:44 |
| bauzas | sdague: oh sec, said something bad | 11:45 |
| *** markvoelker has quit IRC | 11:45 | |
| bauzas | nah, it was checked against policy.json still | 11:45 |
| alex_xu | sdague: bauzas there is default rule 'rule:admin_api' for hypervisor | 11:45 |
| *** markvoelker has joined #openstack-nova | 11:45 | |
| dikonoor | sdague : there's a permissions check at authrorize(context) | 11:45 |
| bauzas | alex_xu: yeah that's what I was going to tell | 11:45 |
| bauzas | alex_xu: so the problem is that the authorize() check is not returning the same as context.require_admin_context() for an user which has an admin role... | 11:47 |
| alex_xu | emm... let me describe the problem, try to get everyone on the same page | 11:48 |
| bauzas | alex_xu: because if both say the same, then there is no problem... | 11:48 |
| andreykurilin | sdague: hi! Since latest release of novaclient broke heat and in near future there will be new one release, it would be nice to sync latest oslo code to remove all warning messages(related oslo namespace). If you have a bit of free time, please look at https://review.openstack.org/#/c/150058/ | 11:49 |
| bauzas | alex_xu: go ahead | 11:49 |
| dikonoor | bauzas: the authorize() check works for a non-admin context where the rule says role:non-admin ..So the check against policy.json goes through for a non-admin and breaks at @require_admin_context() | 11:49 |
| bauzas | dikonoor: that I understood | 11:50 |
| sdague | andreykurilin: it is better to fix the one bug in question and get a good release out | 11:50 |
| andreykurilin | sdague: :( | 11:50 |
| alex_xu | In the old code, dikonoor set the policy rule for hypervisor to "compute_extension:hypervisors": "", then dikonoor can access hypervisor api with non-admin user | 11:50 |
| bauzas | dikonoor: but then the context.require_admin_context() should be consistent IMHO | 11:50 |
| alex_xu | After later, the extra db call 'objects.Service.get_by_host_and_topic' which have admin hard-code checks add into the hypervisor's show api | 11:51 |
| bauzas | alex_xu: by digging into context.require_admin_context(), I can see it just checks if the user has an admin role | 11:51 |
| dikonoor | alex_xu : "compute_extension:hypervisors": "" would mean all non-admins can access it..My rule specifically defines that roles admin and XXX can access it | 11:51 |
| alex_xu | then dikonoor can't use non-admin to access hypervisor'show again | 11:51 |
| alex_xu | dikonoor: yea, similar to your case, right? | 11:52 |
| *** ganso_ has joined #openstack-nova | 11:52 | |
| sdague | johnthetubaguy: ok https://review.openstack.org/#/c/176252/ has passed tests (logs are just uploading for the last one) | 11:52 |
| alex_xu | bauzas: yes, that is hard-code admin check in db call | 11:52 |
| sdague | you want to +A it | 11:52 |
| sdague | then we can get a release out as soon as it merges | 11:52 |
| *** baoli has joined #openstack-nova | 11:53 | |
| bauzas | alex_xu: ok, I see the problem, that's because the policy check just looks at what keystone is telling us (if the user is admin or no) | 11:53 |
| alex_xu | bauzas: yes | 11:53 |
| bauzas | alex_xu: I see | 11:54 |
| alex_xu | bauzas: so we can fix that in kilo by pass an evalated context to the db call | 11:54 |
| sdague | johnthetubaguy: ok, I just +Aed so we can get this out quickly | 11:54 |
| bauzas | alex_xu: so given a cloud operator is fool enough to free up all boundaries by just getting rid of all policy.json rules, it should just go | 11:54 |
| sdague | johnthetubaguy: however, when you get back - https://review.openstack.org/#/c/175010/ would be nice to approve (the downgrade remove) | 11:55 |
| bauzas | alex_xu: well, is it possible to merge something in a stable branch that is not cherry-picked from master ? | 11:55 |
| dikonoor | alex_xu : I agree with the elevated context ..at least that's how most of the other api code has been working | 11:55 |
| *** ildikov has quit IRC | 11:56 | |
| bauzas | alex_xu: I just wonder if backporting the master patch that you did is not better ? | 11:56 |
| alex_xu | bauzas: yea, we shouldn't cherry-pick from master, the master fix is more risk for the kilo now. | 11:56 |
| bauzas | mmm, we should ask for some stable core gurus | 11:56 |
| alex_xu | bauzas: https://review.openstack.org/#/c/160089/ is danger patch, there are more than one place call service_get_by_compute_host | 11:57 |
| bauzas | agreed | 11:57 |
| alex_xu | bauzas: too much risk back port it at this time | 11:57 |
| johnthetubaguy | sdague: thanks, yeah, good to complete our first blueprint of liberty | 11:58 |
| *** aix has joined #openstack-nova | 11:58 | |
| bauzas | alex_xu: well, just wondering then why it's more safier to do it for Liberty, like 10 commits after :) | 11:59 |
| *** Kevin_Zheng has quit IRC | 11:59 | |
| bauzas | alex_xu: if you don't feel confident with backporting your change to kilo, then why are you more confident with merging it to trunk ? | 11:59 |
| alex_xu | bauzas: yea, give more chance to find we have something wrong :) | 12:00 |
| bauzas | alex_xu: again, I just think it would be hard time to explain why you need a specific change for Kilo and not just backport what's fixing the bug on the master branch | 12:00 |
| alex_xu | bauzas: I think we talk about fix that for kilo rc? | 12:00 |
| *** otter768 has joined #openstack-nova | 12:00 | |
| bauzas | alex_xu: yup | 12:00 |
| *** haigang has quit IRC | 12:02 | |
| alex_xu | bauzas: emm....one more thing is https://review.openstack.org/#/c/150568 is not only make hypervisor show configurable, it's also make other api method configurable | 12:02 |
| alex_xu | bauzas: the fix for the kilo is only need hypervisor show configuable | 12:02 |
| *** haigang has joined #openstack-nova | 12:03 | |
| bauzas | agreed | 12:03 |
| bauzas | but still, I don't get why we need to fix only hypervisor_shox | 12:03 |
| bauzas | show | 12:03 |
| bauzas | alex_xu: relaxing the DB check seems the best solution to me | 12:03 |
| bauzas | alex_xu: because that check is not verifying the policy rules | 12:04 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Adds support for x509 certificates as keypairs https://review.openstack.org/136458 | 12:04 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 1 https://review.openstack.org/152569 | 12:04 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Deprecate v1.1 and remove v3 https://review.openstack.org/169378 | 12:04 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 2 https://review.openstack.org/167408 | 12:04 |
| *** mmedvede has quit IRC | 12:04 | |
| bauzas | alex_xu: and btw. that makes me wonder if that's the good way (in terms of policy) | 12:04 |
| bauzas | alex_xu: ie. I don't get the point of having require_admin_context() if all of that can be done per endpoint | 12:05 |
| *** otter768 has quit IRC | 12:05 | |
| bauzas | alex_xu: you got it ? | 12:05 |
| alex_xu | bauzas: you mean why we have require_admin_context() before? | 12:05 |
| bauzas | alex_xu: why we keep that, instead | 12:06 |
| bauzas | alex_xu: because the example is good enough : say that the endpoint rule is just "anyone can access it", why should we keep inline a check for an explicit admin role ? | 12:07 |
| *** lucasagomes is now known as lucas-hungry | 12:07 | |
| *** ajayaa has quit IRC | 12:07 | |
| alex_xu | bauzas: that is the clean up we want to do, but we can't finish that in kilo? | 12:07 |
| bauzas | alex_xu: forget about Kilo | 12:08 |
| bauzas | alex_xu: I'm talking about Liberty :) | 12:08 |
| bauzas | alex_xu: in your patch, you just left some explicit context.require_admin_context() calls | 12:08 |
| alex_xu | sorry...I lost in english again... | 12:08 |
| bauzas | sec then | 12:08 |
| bauzas | alex_xu: https://review.openstack.org/#/c/160089/10/nova/api/openstack/compute/contrib/hosts.py,cm | 12:08 |
| *** ildikov has joined #openstack-nova | 12:09 | |
| bauzas | alex_xu: L125 (right pane) | 12:09 |
| *** raildo has joined #openstack-nova | 12:09 | |
| alex_xu | bauzas: that for just keep v2 behavior as before | 12:09 |
| bauzas | alex_xu: which was broken then | 12:09 |
| alex_xu | bauzas: sorry, what was broken? | 12:09 |
| bauzas | alex_xu: because then, you agree on having an endpoint callable by an end-user which should still raise a 401 ? | 12:10 |
| bauzas | alex_xu: if the user is not admin ? | 12:10 |
| alex_xu | bauzas: yes | 12:10 |
| bauzas | sounds like a bug to me | 12:11 |
| *** yingjun has joined #openstack-nova | 12:11 | |
| alex_xu | bauzas: emm...the old code raise adminrequired excepiton from the db code | 12:11 |
| bauzas | not exactly a behaviour we want to keep | 12:11 |
| bauzas | agreed | 12:11 |
| bauzas | but it was just a bug... | 12:11 |
| alex_xu | bauzas: for now, we want to v2 keep behaviour as before | 12:11 |
| alex_xu | bauzas: but for v2.1, we want to fix that | 12:12 |
| bauzas | so we want to keep v2 raising bugs rather than fixing them ? | 12:12 |
| *** haigang has quit IRC | 12:12 | |
| *** VW_ has joined #openstack-nova | 12:12 | |
| *** dave-mccowan has joined #openstack-nova | 12:12 | |
| *** dprince has joined #openstack-nova | 12:13 | |
| bauzas | alex_xu: when you say v2.1, I rather assume v2.X, because v2.1 == v2.0 | 12:13 |
| bauzas | X>1 | 12:13 |
| *** fawadkhaliq has quit IRC | 12:13 | |
| bauzas | alex_xu: lemme check the API guidelines because I just wonder why we keep a bug just for keeping a wrong behaviour | 12:14 |
| alex_xu | bauzas: yes, that's problem for a long time. if we want to fix that for v2, we need tell the user ensure they didn't update their policy rule to empty to avoid the permission related after code updated | 12:14 |
| *** mmedvede has joined #openstack-nova | 12:14 | |
| bauzas | alex_xu: erm, I don't get it | 12:15 |
| bauzas | alex_xu: policy.json file is config-driven right ? | 12:15 |
| alex_xu | bauzas: then in the discussion of policy blueprint, we think we can fix that for v2.1. we decide to not bother the v2 api user | 12:15 |
| *** jaypipes has joined #openstack-nova | 12:15 | |
| bauzas | alex_xu: so if the operator is just opening the door for a specific endpoint, I can assume he knows what he does and expect it to be working ? | 12:15 |
| *** yingjun has quit IRC | 12:15 | |
| bauzas | alex_xu: is this something merged, or still in discussion ? | 12:16 |
| alex_xu | bauzas: yes, but we assume user can do anything.... | 12:16 |
| *** yingjun has joined #openstack-nova | 12:16 | |
| bauzas | alex_xu: agreed, so let's the policy file be the single source of truth and get rid of that require_admin_check() which are nasty :D | 12:16 |
| alex_xu | bauzas: yes, the decision was merged in the spec | 12:17 |
| *** ociuhandu has joined #openstack-nova | 12:17 | |
| bauzas | alex_xu: https://wiki.openstack.org/wiki/APIChangeGuidelines#Generally_Considered_OK | 12:17 |
| bauzas | " Fixing a bug so that a request which resulted in an error response before is now successful" | 12:17 |
| *** VW_ has quit IRC | 12:17 | |
| bauzas | again, I don't get why we're not fixing it for v2 | 12:18 |
| alex_xu | bauzas: no....it isn't about api change rule. it more conern on the operator upgrade the code | 12:18 |
| *** rfolco has joined #openstack-nova | 12:19 | |
| bauzas | alex_xu: ... and which was previously buggy now works ? hurrah ! | 12:19 |
| alex_xu | bauzas: we didn't want to bother the v2 user, because we think v2 will be instead of v2.1 very soon also | 12:20 |
| *** baoli has quit IRC | 12:21 | |
| alex_xu | bauzas: works in v2.1, not v2 | 12:21 |
| *** aloga_ has joined #openstack-nova | 12:24 | |
| bauzas | alex_xu: so we are not expecting to fix that bug on the stable branches then ? | 12:26 |
| alex_xu | bauzas: we fixed it for stable branch | 12:27 |
| bauzas | alex_xu: and do you know if that problem has been bug filled ? | 12:27 |
| bauzas | ie. if there is a bug ticket | 12:28 |
| alex_xu | bauzas: here https://bugs.launchpad.net/nova/+bug/1447084 | 12:28 |
| openstack | Launchpad bug 1447084 in OpenStack Compute (nova) "view hypervisor details should be controlled by policy.json" [High,Confirmed] | 12:28 |
| bauzas | alex_xu: do you have the link for the stable patch ? | 12:28 |
| alex_xu | bauzas: there isn't patch yet | 12:28 |
| bauzas | alex_xu: that bugfile is way too restrictive IMHO | 12:28 |
| *** neelashah has joined #openstack-nova | 12:29 | |
| jaypipes | morning all | 12:29 |
| *** annegentle has joined #openstack-nova | 12:29 | |
| alex_xu | bauzas: you mean? | 12:30 |
| bauzas | alex_xu: I mean that the general problem is not described | 12:30 |
| gilliard | Hi jaypipes | 12:30 |
| *** mwagner_lap has quit IRC | 12:31 | |
| bauzas | alex_xu: ie. that non-admin users can't access API endpoints even if policy.json rules are allowing them | 12:31 |
| alex_xu | bauzas: yea, that's better | 12:31 |
| jaypipes | gilliard: marooning :) | 12:31 |
| bauzas | jaypipes: ^ if you can just scroll back, I would be interested in your opinion on that | 12:31 |
| bauzas | jaypipes: I know you're having an API WG hat, hence me hassling you, yay | 12:32 |
| alex_xu | dikonoor: would you working on it https://bugs.launchpad.net/nova/+bug/1447084? | 12:32 |
| openstack | Launchpad bug 1447084 in OpenStack Compute (nova) "view hypervisor details should be controlled by policy.json" [High,Confirmed] | 12:32 |
| *** iamjarvo has joined #openstack-nova | 12:32 | |
| *** iamjarvo has quit IRC | 12:33 | |
| alex_xu | a lot of scroll back | 12:33 |
| *** iamjarvo has joined #openstack-nova | 12:33 | |
| *** joefides has quit IRC | 12:33 | |
| jaypipes | bauzas: I joined I think a bit in the middle of the conversation between you and alex_xu, but I think I got the gist of it. | 12:34 |
| jaypipes | bauzas: did you have a specific question on it? | 12:35 |
| bauzas | jaypipes: it's about all fixing that to both v2 and stable branches | 12:35 |
| bauzas | *all about | 12:35 |
| bauzas | because I don't understand why something broken couldn't be fixed at least in v2 | 12:36 |
| jaypipes | bauzas: yes, I don't see any reason not to fix that in the stable branches. I don't necessarily consider it a High priority bug (I think Medium or Low), and I would also like to see in the bug description some information about what is returned to the end user. The only thing it shows is the traceback in the nova-api log, which obviously doesn't get back to the end user. | 12:37 |
| bauzas | jaypipes: I just think the bug report is too narrow, we probably need to make it broader | 12:38 |
| alex_xu | let me update more | 12:39 |
| jaypipes | bauzas: in which way? I mean, it needs information about the HTTP response code that is given back, but otherwise, it's pretty good. | 12:39 |
| *** rushiagr_away is now known as rushiagr | 12:39 | |
| bauzas | jaypipes: the bug specifically reports a problem about an endpoint method, I would just consider the upper level | 12:39 |
| *** claudiub has joined #openstack-nova | 12:40 | |
| jaypipes | bauzas: sorry, not quite following you... | 12:40 |
| bauzas | jaypipes: which is that nova.context.require_admin_context() just shoots what's not admin, even if the endpoint rule in the policy file is saying "go for it" | 12:40 |
| jaypipes | bauzas: isn't the policy.json file handled at an upper level, but there needs to be an entry in it for the os-hypervisors API extension part? | 12:40 |
| bauzas | jaypipes: there is already one | 12:41 |
| jaypipes | ohhhh... | 12:41 |
| bauzas | jaypipes: there is a call to authorize') | 12:41 |
| bauzas | jaypipes: but that kinky nasty require_admin_context() just doesn't care about if it's valid or no | 12:41 |
| jaypipes | bauzas: I see. Then, yes, certainly, please add that information to the bug report :) Sorry, I wasn't clear on that | 12:42 |
| bauzas | lemme do some housekeeping | 12:42 |
| jaypipes | bauzas: so, change the summary of the bug to: "require_admin_context() does not account for policy.json rulesets" | 12:42 |
| bauzas | exactly ! | 12:42 |
| jaypipes | alex_xu: ^^ | 12:42 |
| jaypipes | work for you? | 12:42 |
| bauzas | my approach would be to just shoot it in the head | 12:42 |
| bauzas | ie. just stop removing that check | 12:43 |
| bauzas | oooo | 12:43 |
| bauzas | just stop doing that check | 12:43 |
| bauzas | or just stop by removing that check even | 12:43 |
| alex_xu | emm....don't get the point of the summary... | 12:43 |
| bauzas | alex_xu: that's what we discussed => policy.json is the single source of truth | 12:44 |
| bauzas | alex_xu: there is no need to check for contexts | 12:44 |
| alex_xu | bauzas: ok...I still think that is complex fix for rc | 12:46 |
| bauzas | alex_xu: okay, I understand your concerns, what are you afraid of ? | 12:47 |
| alex_xu | bauzas: ok, what we should do after remove check for those api https://review.openstack.org/#/c/160089/10/nova/api/openstack/compute/contrib/hosts.py | 12:48 |
| alex_xu | bauzas: add nova_context.require_admin_context(context) for hosts api also? | 12:48 |
| bauzas | alex_xu: we still have the authorize() check no ? | 12:48 |
| alex_xu | bauzas: yes, but the behavior is changed, hosts api didn't allow non-admin user access it even the policy rule is empty | 12:49 |
| alex_xu | before | 12:51 |
| bauzas | alex_xu: that I understand but I considered it as a bug | 12:52 |
| bauzas | jaypipes: could you please confirm ? :) | 12:52 |
| alex_xu | bauzas: ok if we didn't add "nova_context.require_admin_context(context)" we should notify the user update their policy.json if they set the rule empty before, because after update the code, if the rule is empty, then the anyone can access it | 12:53 |
| bauzas | alex_xu: but this is the exact expected behaviour ! :)Ã | 12:53 |
| alex_xu | bauzas: and this different with our agreement on the spec https://github.com/openstack/nova-specs/blob/master/specs/liberty/approved/nova-api-policy.rst | 12:53 |
| * jaypipes reads spec | 12:54 | |
| bauzas | alex_xu: lemme check the review | 12:54 |
| alex_xu | bauzas: we said we don't want to bother v2 user for that. those problem will be fixed at v2.1. | 12:54 |
| alex_xu | and v2.1 will be replace v2 very soon | 12:54 |
| bauzas | alex_xu: not for stable branches AFAIK :) | 12:54 |
| bauzas | and v2.1 == v2 right ? | 12:54 |
| bauzas | so fixing it in v2.1 requires fixing it in v2.0, unless it's on a separate version bump right ? | 12:55 |
| alex_xu | bauzas: v2 is the api in compute/contrib, v2.1 is the api in the compute/plugins/v3.... | 12:55 |
| bauzas | alex_xu: I know all of that | 12:56 |
| bauzas | alex_xu: the point being ? | 12:56 |
| jaypipes | alex_xu: OK, so if I am reading this correctly, I believe that *if* you fix that bug by implementing nova-api-policy blueprint, then *no*, that would not be appropriate to backport to Kilo, due to it being a major change. | 12:56 |
| alex_xu | bauzas: ok, I got you point.... | 12:56 |
| alex_xu | bauzas: hah, jaypipes on my side now ;) | 12:56 |
| *** baoli has joined #openstack-nova | 12:57 | |
| bauzas | the major change being to remove the decorators ? | 12:57 |
| bauzas | (sorry if you feel myself a bit harsh, that's not my intention) | 12:58 |
| jaypipes | alex_xu: well, if you *only* fixed that particular bug -- i.e. the specific os-hypervisors extension -- then I suppose one might argue for backporting. But if you intend to make that bug into a generic "require_admin_context() should take policy.json into account", that would likely be too big of a change to backport. | 12:58 |
| bauzas | but I want to understand why it's such a major change to just free up rules | 12:58 |
| jaypipes | alex_xu: bit of a dilemma. :) | 12:58 |
| *** eharney has joined #openstack-nova | 12:58 | |
| bauzas | jaypipes: I think we don't need to fix require_admin_context(), we just need to kill it | 12:59 |
| alex_xu | jaypipes: yea | 12:59 |
| jaypipes | bauzas: so, I agree. however that is not alex_xu's main priority here :) alex_xu's main priority is to identify something that can be backported into Kilo, since he is more concerned with the stable branch than master for this bug fix ;) | 13:00 |
| jaypipes | alex_xu: correct? | 13:00 |
| alex_xu | jaypipes: yes | 13:00 |
| jaypipes | either that, or it's just way too late for alex_xu and way to early for me ;) | 13:01 |
| bauzas | jaypipes: lol | 13:01 |
| bauzas | that's just perfect timing for me, I just took a nap on my patio | 13:01 |
| bauzas | s/patio/backyard | 13:01 |
| *** busterswt has joined #openstack-nova | 13:01 | |
| alex_xu | there are some fog before my eye | 13:02 |
| bauzas | alex_xu: ok here is what I propose | 13:02 |
| jaypipes | alex_xu: so, i would say a compromise on this would be to keep that bug description as it is, submit a patch to master that *just* fixes the os-hypervisors extension and that bug, backport it to kilo rc2, and then work on nova-api-policy to fix the world. | 13:02 |
| bauzas | jaypipes: and open a bug for tracking that | 13:02 |
| bauzas | jaypipes: you just beated me at the typing | 13:02 |
| jaypipes | bauzas: why? the blueprint already exists... | 13:02 |
| jaypipes | nova-api-policy | 13:03 |
| alex_xu | jaypipes: bauzas , so I should stop this merge https://review.openstack.org/160089? | 13:03 |
| bauzas | jaypipes: because even if I always say that bugs are improvement possibilities, that's a bug | 13:03 |
| *** ociuhandu has quit IRC | 13:03 | |
| *** joefides has joined #openstack-nova | 13:04 | |
| bauzas | jaypipes: the bug being that operators can't modify the rules for some endpoints | 13:04 |
| *** ildikov has quit IRC | 13:04 | |
| bauzas | jaypipes: and they have zero explanation why it just doesn't work | 13:04 |
| jaypipes | guys, unfortunately, I need to run out now... | 13:04 |
| jaypipes | perhaps this is worth a quick email to openstack-dev? | 13:05 |
| bauzas | jaypipes: sounds valuable | 13:05 |
| alex_xu | jaypipes: ok, thanks for your time | 13:05 |
| bauzas | go off guys, I'm opening a bug FWIW and I'm leaving the world knowing that | 13:05 |
| *** lucas-hungry is now known as lucasagomes | 13:05 | |
| bauzas | alex_xu: jaypipes: thanks for your time, appreciated | 13:06 |
| alex_xu | bauzas: you too:) | 13:06 |
| *** igordcard has quit IRC | 13:07 | |
| alex_xu | bauzas: would you want to write email? or you hope me write one? | 13:07 |
| bauzas | alex_xu: I'm on it | 13:08 |
| bauzas | alex_xu: go off | 13:08 |
| alex_xu | bauzas: thanks a lot :) | 13:08 |
| bauzas | alex_xu: Vancouver btw. ? | 13:08 |
| alex_xu | bauzas: yes this time :) | 13:09 |
| bauzas | alex_xu: do you have possibility to be there ? | 13:09 |
| bauzas | awesome | 13:09 |
| *** zzzeek has joined #openstack-nova | 13:09 | |
| *** ferest has joined #openstack-nova | 13:11 | |
| *** rushiagr is now known as rushiagr_away | 13:12 | |
| *** afazekas_ has quit IRC | 13:13 | |
| *** ociuhandu has joined #openstack-nova | 13:18 | |
| *** Murali has quit IRC | 13:18 | |
| *** yamahata has joined #openstack-nova | 13:19 | |
| *** thomasem has quit IRC | 13:19 | |
| *** ildikov has joined #openstack-nova | 13:19 | |
| *** annegentle has quit IRC | 13:20 | |
| *** thomasem has joined #openstack-nova | 13:20 | |
| *** ajayaa has joined #openstack-nova | 13:22 | |
| *** mriedem_away is now known as mriedem | 13:22 | |
| sdague | johnthetubaguy: ok, I just cut 2.24.1, hopefully correctly | 13:22 |
| johnthetubaguy | sdague: OK, thanks, not sure how to do all that myself, I figured we would just call it 2.25 in case there were any features in there, but thats fine I guess | 13:23 |
| mriedem | 2.25 implies new features or requirements updated | 13:23 |
| mriedem | which could blow caps on stable | 13:23 |
| sdague | I just googled and followed mriedem's directions | 13:24 |
| mriedem | i have directions? | 13:25 |
| johnthetubaguy | mriedem: OK, cool | 13:25 |
| johnthetubaguy | I need to go read all that at some point | 13:25 |
| sdague | https://wiki.openstack.org/wiki/Nova/Client_Release_Process | 13:25 |
| sdague | your name is on that page | 13:25 |
| *** VW_ has joined #openstack-nova | 13:26 | |
| mriedem | oh, that blurb :) the rest is someone else, probably russellb or mikal | 13:26 |
| mriedem | i will gladly take credit though | 13:26 |
| mriedem | if we set the series correctly in novaclient we could just use those fancy release tools scripts | 13:26 |
| sdague | mriedem: I used the release tools | 13:27 |
| sdague | are there other ones? | 13:27 |
| mriedem | sdague: you used this? https://wiki.openstack.org/wiki/Oslo/ReleaseProcess#Installing_Release_Tools | 13:27 |
| sdague | I used dolph's thing | 13:28 |
| sdague | so I guess not | 13:28 |
| mriedem | ah | 13:28 |
| larsks | I have what is probably a dumb question: in NovaProxyRequestHandlerBase (in nova/console/websocketproxy.py), where does the "msg" attribute come from (e.g., https://github.com/openstack/nova/blob/master/nova/console/websocketproxy.py#L120)? | 13:28 |
| mriedem | yeah, ^ is the oslo release tools | 13:28 |
| *** annegentle has joined #openstack-nova | 13:29 | |
| *** jecarey has joined #openstack-nova | 13:29 | |
| mriedem | sdague: so does https://review.openstack.org/#/c/176113/ unwedge the gate? | 13:29 |
| sdague | yeh, we're good for now | 13:30 |
| mriedem | sweet | 13:30 |
| mriedem | larsks: you might ask sahid or directxman12 | 13:30 |
| mriedem | larsks: probably from websockify.ProxyRequestHandler since it's used as a mixin | 13:31 |
| mriedem | which isn't at all confusing | 13:31 |
| *** annegentle has quit IRC | 13:32 | |
| sahid | larsks: you mean the method? | 13:32 |
| sahid | msg? | 13:32 |
| sahid | to print a log | 13:32 |
| *** annegentle has joined #openstack-nova | 13:32 | |
| larsks | sahid: someone was reporting this traceback: http://paste.openstack.org/show/205086/ | 13:33 |
| larsks | I have a vague recollection of a recent bug or mailing list traffic on the same thing, but I can't find it. | 13:33 |
| sahid | i see | 13:33 |
| sahid | you probably have to update your version of websokify | 13:33 |
| larsks | I was trying to track down what was happening, but I've been stymied in trying to figure out exactly where self.msg is coming from... | 13:33 |
| *** andreykurilin__ has joined #openstack-nova | 13:33 | |
| *** annashen has joined #openstack-nova | 13:33 | |
| sahid | larsks: which version are you using? | 13:34 |
| bauzas | jaypipes: alex_xu: http://lists.openstack.org/pipermail/openstack-dev/2015-April/062282.html | 13:34 |
| sahid | larsks: https://bugzilla.redhat.com/show_bug.cgi?id=1208175 | 13:34 |
| openstack | bugzilla.redhat.com bug 1208175 in openstack-nova "Can't connect to console via Horizon - ends up error code 1006" [High,Post] - Assigned to sferdjao | 13:34 |
| larsks | This was for someone else. I'm not sure, but it sounds like python 2.6, so probably RHEL 6 or similar and whatever that comes with... | 13:34 |
| mriedem | i've had issues with the websockify code on rhel 6.5 | 13:34 |
| mriedem | b/c of how urlparse works there | 13:35 |
| mriedem | maybe not related here though | 13:35 |
| sahid | mriedem: i don't think so | 13:35 |
| mriedem | but on rhel 6.5, urlparse won't parse a schema of 'ws' so you can get a NoneType | 13:35 |
| *** vladikr has joined #openstack-nova | 13:35 | |
| larsks | sahid: thanks for the bz link. | 13:35 |
| sahid | actually it's related to a specific version supported by redhat | 13:35 |
| mriedem | redhat patches it's python package doesn't it? | 13:35 |
| *** rushil has joined #openstack-nova | 13:36 | |
| mriedem | what version of websockify is required? in stable/juno we have this range: https://github.com/openstack/requirements/blob/stable/juno/global-requirements.txt#L144 | 13:37 |
| mriedem | https://github.com/kanaka/websockify/blob/v0.6.0/websockify/websocket.py#L270 | 13:39 |
| *** annashen has quit IRC | 13:39 | |
| BobBall | mriedem: Do you Have The Power(tm) to set a CI system as voting? | 13:41 |
| mriedem | BobBall: i do | 13:41 |
| mriedem | xenserver ci | 13:41 |
| mriedem | ? | 13:41 |
| * bauzas remembers the lovely love of websockify pulling numpy | 13:41 | |
| BobBall | Wonderful. Yes, XenServer CI. | 13:41 |
| mriedem | BobBall: done https://review.openstack.org/#/admin/groups/511,members | 13:42 |
| sahid | mriedem: the issue is on icehouse | 13:42 |
| BobBall | But, we'd also like XenProject CI to be enabled as well :) http://jenkins.openstack.xenproject.org/job/dsvm-tempest-xen/buildTimeTrend is showing a nice green state (with a few failures, of course, as are normally expected) | 13:42 |
| mriedem | sahid: oh.. https://github.com/openstack/nova/blob/stable/icehouse/requirements.txt#L30 | 13:42 |
| BobBall | was talking with johnthetubaguy earlier and he's happy for now - but clearly if there are other suggestions for what we should do to make sure everyone is happy we'll do them too | 13:42 |
| mriedem | BobBall: i only see Rackspace XenServer CI as a gerrit member | 13:43 |
| mriedem | we only disabled the one, i'm assuming that's the same account for both? | 13:43 |
| BobBall | No; XenProject CI is a new CI account | 13:43 |
| BobBall | I wonder if I need to do something else with Infra first then | 13:43 |
| sahid | mriedem: yes and we probably publish 5.1 https://github.com/kanaka/websockify/blob/v0.5.1/websockify/websocket.py | 13:44 |
| sahid | i mean ship | 13:44 |
| *** ociuhandu has quit IRC | 13:44 | |
| *** xyang1 has joined #openstack-nova | 13:45 | |
| mriedem | sahid: hmm, seems like it'd be fine | 13:46 |
| mriedem | https://github.com/kanaka/websockify/blob/v0.5.1/websockify/websocket.py#L390 | 13:46 |
| *** iamjarvo has quit IRC | 13:46 | |
| mriedem | https://github.com/openstack/nova/blob/2014.1.4/nova/console/websocketproxy.py#L107 | 13:46 |
| *** ociuhandu has joined #openstack-nova | 13:46 | |
| mriedem | unless red hat patched one of those | 13:46 |
| mriedem | like backported some websocketproxy code from juno to icehouse for it's rhosp stuff, idk | 13:46 |
| sahid | mriedem: no if you look the proptoype of the method is not good | 13:46 |
| *** thangp has joined #openstack-nova | 13:46 | |
| sahid | we are expecting an extra argument | 13:47 |
| mriedem | sahid: this is not icehouse level code in nova: File "/usr/lib/python2.6/site-packages/nova/console/websocketproxy.py", line 116, in new_client self.msg(_('connect info: %s'), str(connect_info)) | 13:47 |
| sahid | oh sorry yes | 13:47 |
| mriedem | sahid: so my guess is red hat backported something it shouldn't have | 13:47 |
| sahid | :) | 13:47 |
| mriedem | and broke stable | 13:47 |
| *** yingjun has quit IRC | 13:48 | |
| mriedem | well, stable for rhosp users | 13:48 |
| *** mwagner_lap has joined #openstack-nova | 13:48 | |
| *** Nikolay_St has quit IRC | 13:48 | |
| *** burt has joined #openstack-nova | 13:49 | |
| *** romainh has left #openstack-nova | 13:49 | |
| *** amotoki has joined #openstack-nova | 13:49 | |
| BobBall | mriedem: XenProject-CI is openstack-ci@xenproject.org - not sure how to get the userid. I wonder if the gerrit members being shown are only the old-style accounts rather than the new openid ones? | 13:50 |
| BobBall | There have been issues finding it before | 13:50 |
| mriedem | BobBall: got it https://review.openstack.org/#/admin/groups/511,members | 13:50 |
| mriedem | does that look correct now? | 13:50 |
| *** iamjarvo has joined #openstack-nova | 13:52 | |
| *** iamjarvo has quit IRC | 13:53 | |
| *** iamjarvo has joined #openstack-nova | 13:53 | |
| BobBall | Heh; sorry; not quite. I didn't realise that johnthetubaguy had set up his own XenServer CI account :) The XenServer CI (that runs on Rackspace) that we need voting for is "Citrix XenServer CI" | 13:54 |
| *** iamjarvo has quit IRC | 13:54 | |
| *** iamjarvo has joined #openstack-nova | 13:54 | |
| mriedem | BobBall: specifici email? and should i remove johnthetubaguy's? | 13:54 |
| BobBall | I've not seen johnthetubaguy's voting but the Citrix XenServer CI doesn't have an email associated with it (like the VMWare NSX CI account) | 13:55 |
| johnthetubaguy | yeah, that ones a bit dead, the rackspace one | 13:55 |
| *** tonytan4ever has joined #openstack-nova | 13:55 | |
| mriedem | ok, removed that other one, added Citrix XenServer CI | 13:56 |
| johnthetubaguy | I never got one build working, and pulled onto something else | 13:56 |
| mriedem | https://review.openstack.org/#/admin/groups/511,members good now? | 13:56 |
| johnthetubaguy | mriedem: cools | 13:56 |
| BobBall | That's perfect! thanks mriedem. | 13:56 |
| mriedem | btw, is there a reason there are 2 xen ci's? | 13:56 |
| mriedem | why does it run on stable? | 13:56 |
| mriedem | it frustrates my email filtering | 13:56 |
| *** annegentle has quit IRC | 13:57 | |
| *** johnthetubaguy is now known as zz_johnthetubagu | 13:57 | |
| BobBall | Which one runs on Stable? Reason is that the citrix CI is using our own orchestration (not zuul) and depends on a custom nodepool. We're planning on getting that fixed, so they merge into one, but it's likely to take a while. The XenProject CI runs libvirt+xen vs the Citrix XenServer CI which runs, you guessed it... XenServer :) | 13:57 |
| *** redbeard has joined #openstack-nova | 13:58 | |
| *** iamjarvo has quit IRC | 13:58 | |
| *** jcoufal has quit IRC | 13:59 | |
| mriedem | BobBall: https://review.openstack.org/#/q/reviewer:openstack-ci%2540xenproject.org+status:open,n,z | 13:59 |
| mriedem | xenproject ci is running on stable | 13:59 |
| mriedem | and doesn't need to be | 13:59 |
| *** annegentle has joined #openstack-nova | 13:59 | |
| *** apmelton_away is now known as apmelton | 13:59 | |
| BobBall | Drat. I was sure I'd fixed that... Clearly my zuul config is broken. *Investigates*. Thanks. | 13:59 |
| mriedem | np | 13:59 |
| *** nkrinner has quit IRC | 14:01 | |
| *** cfriesen has joined #openstack-nova | 14:01 | |
| *** tsekiyama has joined #openstack-nova | 14:01 | |
| *** dboik has joined #openstack-nova | 14:01 | |
| *** otter768 has joined #openstack-nova | 14:01 | |
| *** Marga_ has quit IRC | 14:02 | |
| kashyap | Urgh, Launchpad's text formatting is crazy, I think I needed to give four spaces instead of two. Wonder if anyone can deduce a reproducer from the description - https://bugs.launchpad.net/nova/+bug/1442401 | 14:02 |
| openstack | Launchpad bug 1442401 in OpenStack Compute (nova) "When resizing a VM, ssh will fail unless libvirt/driver.py is patched" [Undecided,New] | 14:02 |
| *** popw has quit IRC | 14:02 | |
| kashyap | I just tested w/ today's Git (commits mentioned in the bug) and can't reproduce (tested w/ a simple resize up). | 14:03 |
| mriedem | kashyap: that's reported for icehouse? | 14:03 |
| openstackgerrit | Dan Smith proposed openstack/nova-specs: Add admin-query-any-keypair.rst https://review.openstack.org/175579 | 14:03 |
| kashyap | mriedem: Damn, I'm blind. Since I test from Git, I seemed to blindly assume the reporter did so too. | 14:04 |
| mriedem | kashyap: i'm only assuming since "I'm not quite sure the live_migration_uri is still used in Icehouse but it used to be used." | 14:04 |
| mriedem | and we test live migration in the gate now | 14:04 |
| mriedem | and we test resize with shared storage in the ceph job | 14:05 |
| mriedem | i'll mark as incomplete asking for specifics | 14:05 |
| openstackgerrit | Dan Smith proposed openstack/nova-specs: Add admin-query-any-keypair.rst https://review.openstack.org/175579 | 14:05 |
| kashyap | mriedem: Oh right - it's the aiopcpu (I can't help but read it as Physical CPU, but it's actually reads as "AIO plus a Compute node") | 14:05 |
| *** otter768 has quit IRC | 14:06 | |
| kashyap | mriedem: Yep, I did mark it, and asked him the details. | 14:06 |
| mriedem | me too :) | 14:06 |
| mriedem | double whammy | 14:06 |
| kashyap | For the life of me, I don't get it why people (even advanced users) never take time to even outline a reproducer | 14:07 |
| kashyap | It's more or less: "Here, trace, take it" :-) | 14:07 |
| mriedem | that's what 'incomplete' is for | 14:07 |
| dansmith | alex_xu: still around? | 14:07 |
| bauzas | dansmith: I think I hassled him so much that he's now just out of battery | 14:08 |
| dansmith | bauzas: nice going :/ | 14:08 |
| *** artom has joined #openstack-nova | 14:09 | |
| alex_xu | dansmith: bauzas still wake up :) | 14:12 |
| alex_xu | bauzas: thanks for mail | 14:12 |
| dansmith | alex_xu: actually, I think I got it figured out.. I had a question about the policy bit | 14:12 |
| dansmith | alex_xu: however, I'm not sure where in policy.json to add this | 14:12 |
| *** haomaiwang has quit IRC | 14:12 | |
| dansmith | alex_xu: is the v3 extension os-keypairs or just keypairs? | 14:12 |
| *** ankit_ag has quit IRC | 14:13 | |
| dansmith | oh, you said it's just keypairs | 14:13 |
| *** haomaiwang has joined #openstack-nova | 14:13 | |
| alex_xu | dansmith: this rule "os_compute_api:os-keypairs:index" | 14:13 |
| dansmith | alex_xu: oh, that's different than what you said in the review | 14:14 |
| alex_xu | dansmith: oops, let me check what i said... | 14:14 |
| dansmith | alex_xu: and I don't see any other "role:admin" in there, did you mean "rule:admin_api" ? | 14:14 |
| dansmith | nor any $(user_id)s type things | 14:15 |
| * dansmith doesn't know how this policy stuff works | 14:15 | |
| bauzas | dansmith: welcome in the oslo policy world | 14:15 |
| alex_xu | dansmith: it should be 'rule:admin_api' | 14:16 |
| dansmith | okay | 14:16 |
| *** lpetrut has quit IRC | 14:16 | |
| alex_xu | dansmith: 'user_id:%(user_id)s' means check the target's user_id item | 14:16 |
| bauzas | alex_xu: btw. I'm still following the admin_or_user rule but I seriously doubt that the user part can work :) | 14:16 |
| *** nelsnelson has joined #openstack-nova | 14:16 | |
| dansmith | alex_xu: okay, but I don't see that anywhere else in the policy file.. is this really the first such example? | 14:16 |
| bauzas | alex_xu: which means nothing to me | 14:16 |
| alex_xu | dansmith: bauzas 'user_id:%(user_id)s' is similar to 'project_id:%(project_id)s' | 14:17 |
| bauzas | alex_xu: I know and that's just a noop thing | 14:17 |
| dansmith | yeah | 14:17 |
| alex_xu | dansmith: yea, that is first use it for user-owner | 14:17 |
| alex_xu | first time to use for user-owner | 14:17 |
| dansmith | alex_xu: okay, gonna have to trust you I guess :D | 14:17 |
| *** tsekiyama has quit IRC | 14:17 | |
| alex_xu | dansmith: thanks, I tested the code, it works | 14:18 |
| *** salv-orlando has quit IRC | 14:18 | |
| alex_xu | :) | 14:18 |
| dansmith | heh, okay | 14:18 |
| bauzas | alex_xu: https://github.com/openstack/oslo.policy/blob/master/oslo_policy/_checks.py#L280 | 14:18 |
| *** tonytan4ever has quit IRC | 14:18 | |
| alex_xu | bauzas: it should be this | 14:19 |
| *** atuvenie_ has quit IRC | 14:19 | |
| *** ferest has quit IRC | 14:20 | |
| bauzas | alex_xu: yeah, but the matching just checks that you have like an user_id or a tenant_id in your context, not what it should look | 14:20 |
| *** tsekiyama has joined #openstack-nova | 14:20 | |
| *** tsekiyama has quit IRC | 14:20 | |
| bauzas | (the target being the context) | 14:20 |
| *** tsekiyama has joined #openstack-nova | 14:21 | |
| *** mtanino has joined #openstack-nova | 14:21 | |
| bauzas | for example, it doesn't prevent you to get some data from tenant 1 if your creds are saying that you're tenant 2 | 14:21 |
| bauzas | alex_xu: it just checks that you have a tenant_id IIUC | 14:22 |
| bauzas | alex_xu: all of that segregation is done at the DB level (ie. filter by tenant_id) not in the policy enforcement | 14:22 |
| bauzas | alex_xu: so that's like a noop thing | 14:22 |
| alex_xu | bauzas: yes | 14:23 |
| alex_xu | bauzas: the project owner is done by db | 14:23 |
| BobBall | mriedem: I think that it's not voting on stable now; if you see any more comments on stable please let me know (I'll also check again in a day or two. | 14:23 |
| bauzas | so, what's the point of keeping it ? | 14:23 |
| mriedem | BobBall: thanks | 14:23 |
| dansmith | bauzas: this seems to work, FWIW | 14:23 |
| bauzas | it just makes confusion | 14:23 |
| dansmith | well, actually, | 14:23 |
| dansmith | I don't have a test for a non-admin user listing another | 14:23 |
| dansmith | but it doesn't allow a normal user to run the thing | 14:24 |
| bauzas | dansmith: that's the admin thing who prevents that | 14:24 |
| bauzas | dansmith: not the 'user' part of it | 14:24 |
| *** smccully has quit IRC | 14:24 | |
| dansmith | bauzas: well, the "or user_id:" part isn't letting a normal user list them just because they have a user_id | 14:24 |
| *** smccully has joined #openstack-nova | 14:24 | |
| *** aloga_ has quit IRC | 14:25 | |
| dansmith | isn't that what you were worried about? | 14:25 |
| bauzas | "is_admin:True or project_id:%(project_id)s" | 14:25 |
| dansmith | I think they're both part of the fake project, right? | 14:25 |
| bauzas | so I'm just saying that project_id:%(project_id)s is useless provided you're keystone auth'd | 14:25 |
| alex_xu | bauzas: it is useful dansmith's spec | 14:25 |
| alex_xu | bauzas: but for some api, it is useless. | 14:26 |
| bauzas | alex_xu: okay, then I'm speaking about something I didn't see | 14:26 |
| bauzas | alex_xu: which spec ? | 14:26 |
| alex_xu | bauzas: https://review.openstack.org/#/c/175579/3/specs/liberty/approved/admin-query-any-keypair.rst | 14:27 |
| *** sudipto has joined #openstack-nova | 14:28 | |
| *** READ10 has joined #openstack-nova | 14:28 | |
| bauzas | alex_xu: okay I see now, because you're explicitely specifying a target, contrary to the other endpoints | 14:28 |
| *** Marga_ has joined #openstack-nova | 14:29 | |
| sudipto | mriedem, Hi, regarding your UT comment on https://review.openstack.org/#/c/173187 . I had made a comment on the previous patchset. The host.get_online_cpus is operating off the data initialized in the fakelibvirt driver. It appeared to me that the data is coming from the fake Connection class which is initlalizing the HostInfo class - that by default returns only 2 active CPUS - online_cpus = set([0, 1]) - Do you want me to override this behavi | 14:29 |
| sudipto | or and ask this to return explicit online set ids? Any examples on how I could do that? | 14:29 |
| *** iamjarvo has joined #openstack-nova | 14:29 | |
| *** iamjarvo has quit IRC | 14:29 | |
| *** wuhg has quit IRC | 14:29 | |
| bauzas | alex_xu: okay, I then have to apologize, I was referring to 99% of Nova code which are not specifying a target :) | 14:29 |
| *** iamjarvo has joined #openstack-nova | 14:30 | |
| *** iamjarvo has quit IRC | 14:30 | |
| dansmith | alex_xu: on the api_samples, do you really think it's necessary to add an api_samples test for this? | 14:30 |
| dansmith | it doesn't change the output of the method, and I'm not really seeing much in the way of examples for it | 14:30 |
| sahid | mriedem: https://review.openstack.org/#/c/170110/ | 14:30 |
| mriedem | sudipto: sec, didn't see that comment | 14:30 |
| *** iamjarvo has joined #openstack-nova | 14:30 | |
| *** iamjarvo has quit IRC | 14:30 | |
| alex_xu | bauzas: yes, without target it also works | 14:31 |
| bauzas | mriedem: sounds like we need to isolate the test then :/ | 14:31 |
| *** iamjarvo has joined #openstack-nova | 14:31 | |
| *** iamjarvo has quit IRC | 14:31 | |
| bauzas | alex_xu: it works sure, but like 1==1 | 14:31 |
| openstackgerrit | Dan Smith proposed openstack/nova: Adding user_id handling to keypair index api call https://review.openstack.org/70485 | 14:31 |
| dansmith | alex_xu: ^ | 14:32 |
| mriedem | sudipto: can you mock out what the fakelibvirt connection is returning? | 14:32 |
| alex_xu | dansmith: A little not sure, but I remember there are some sample test for the some parameter, but let me check later | 14:32 |
| bauzas | mriedem: I saw your evenings' comments, I need more time to read those through | 14:32 |
| dansmith | alex_xu: okay, well, there is the code, feel free to suggest in there | 14:32 |
| *** iamjarvo has joined #openstack-nova | 14:32 | |
| dansmith | alex_xu: no rush, of course | 14:32 |
| *** iamjarvo has quit IRC | 14:32 | |
| *** atuvenie has joined #openstack-nova | 14:32 | |
| alex_xu | dansmith: ok, thanks | 14:32 |
| *** iamjarvo has joined #openstack-nova | 14:33 | |
| *** iamjarvo has quit IRC | 14:33 | |
| sudipto | mriedem, do you mean this call: (cpus, cpu_map, online) = self.get_connection().getCPUMap() | 14:33 |
| sudipto | ? | 14:33 |
| mriedem | sudipto: i'm looking it up quick, sec | 14:33 |
| sudipto | mriedem, ok | 14:33 |
| *** lpetrut has joined #openstack-nova | 14:33 | |
| *** iamjarvo has joined #openstack-nova | 14:33 | |
| *** iamjarvo has quit IRC | 14:33 | |
| alex_xu | bauzas: yea, it is 1==1 | 14:34 |
| *** Nikolay_St has joined #openstack-nova | 14:34 | |
| *** iamjarvo has joined #openstack-nova | 14:34 | |
| *** iamjarvo has quit IRC | 14:34 | |
| *** iamjarvo has joined #openstack-nova | 14:35 | |
| *** Nikolay_St has quit IRC | 14:35 | |
| garyk | anyone know if the grenade check has been fixed - or is this still broken? | 14:35 |
| *** iamjarvo has quit IRC | 14:35 | |
| mriedem | sudipto: so the test gets it's value from here: http://git.openstack.org/cgit/openstack/nova/tree/nova/tests/unit/virt/libvirt/fakelibvirt.py#n818 | 14:35 |
| mriedem | sudipto: where is self.host_info.cpus set? | 14:35 |
| bauzas | garyk: I was just looking at it | 14:35 |
| mriedem | garyk: fixed | 14:35 |
| *** iamjarvo has joined #openstack-nova | 14:36 | |
| bauzas | was it glance ? | 14:36 |
| *** iamjarvo has quit IRC | 14:36 | |
| *** iamjarvo has joined #openstack-nova | 14:36 | |
| *** unicell1 has joined #openstack-nova | 14:36 | |
| *** iamjarvo has quit IRC | 14:36 | |
| mriedem | sudipto: oh i see http://git.openstack.org/cgit/openstack/nova/tree/nova/tests/unit/virt/libvirt/fakelibvirt.py#n153 | 14:37 |
| mriedem | defaults to 2 in HostInfo in fakelibvirt | 14:37 |
| sudipto | mriedem, yeah, i was gathering it back from the code. | 14:37 |
| *** unicell has quit IRC | 14:37 | |
| *** iamjarvo has joined #openstack-nova | 14:37 | |
| mriedem | so we should be able to override that | 14:37 |
| garyk | mriedem: thanks | 14:37 |
| *** iamjarvo has quit IRC | 14:37 | |
| garyk | mriedem: do we need to rebase or just 'recheck' | 14:38 |
| mriedem | recheck | 14:38 |
| garyk | gracias | 14:38 |
| *** iamjarvo has joined #openstack-nova | 14:38 | |
| *** iamjarvo has quit IRC | 14:38 | |
| mriedem | sudipto: sec, i think i care more about nova.virt.hardware.get_vcpu_pin_set | 14:39 |
| mriedem | but i have 3 people in m year atm | 14:39 |
| *** iamjarvo has joined #openstack-nova | 14:39 | |
| sudipto | mriedem, Ok I will wait. | 14:39 |
| *** iamjarvo has joined #openstack-nova | 14:39 | |
| *** iamjarvo has quit IRC | 14:40 | |
| *** iamjarvo has joined #openstack-nova | 14:40 | |
| *** iamjarvo has quit IRC | 14:41 | |
| *** artom has quit IRC | 14:41 | |
| *** artom_ has joined #openstack-nova | 14:41 | |
| *** iamjarvo has joined #openstack-nova | 14:41 | |
| *** iamjarvo has quit IRC | 14:41 | |
| mriedem | sudipto: commented, i was wrong, i meant don't mock out hardware.get_vcpu_pin_set | 14:42 |
| *** iamjarvo has joined #openstack-nova | 14:42 | |
| mriedem | mocking out Host.get_online_cpus() is fine | 14:42 |
| sudipto | mriedem, Alrite, i will remove those mocks. | 14:42 |
| *** iamjarvo has joined #openstack-nova | 14:43 | |
| *** sdake has joined #openstack-nova | 14:43 | |
| *** iamjarvo has quit IRC | 14:43 | |
| mriedem | sudipto: also i think we should sort those lists in the error message, would be more clear imo | 14:43 |
| *** yamahata has quit IRC | 14:43 | |
| sudipto | mriedem, yeah i agree, i will change that as well. | 14:44 |
| *** iamjarvo has joined #openstack-nova | 14:44 | |
| *** yamahata has joined #openstack-nova | 14:44 | |
| mriedem | then i'm +2 so hard | 14:44 |
| *** sdake_ has joined #openstack-nova | 14:44 | |
| *** sdake has quit IRC | 14:48 | |
| *** iamjarvo has quit IRC | 14:49 | |
| *** tonytan4ever has joined #openstack-nova | 14:49 | |
| *** READ10 has quit IRC | 14:49 | |
| openstackgerrit | Brent Eagles proposed openstack/nova-specs: Wrap Neutron client calls https://review.openstack.org/141108 | 14:49 |
| *** browne has joined #openstack-nova | 14:50 | |
| openstackgerrit | Brent Eagles proposed openstack/nova-specs: Wrap Neutron client calls https://review.openstack.org/141108 | 14:50 |
| bauzas | mriedem: on the unittest, while I understand your point, don't you think it's maybe cleaner to keep the method isolation ? | 14:51 |
| mriedem | bauzas: in most cases i'd say yes, but not here | 14:51 |
| mriedem | get_vcpu_pin_set is a utility that parses the config option which we're setting in the driver test, | 14:52 |
| bauzas | mriedem: sure, we could just remove the mock and use the dependent methods, but I think it's just not really a good pattern and the reason is not really important | 14:52 |
| mriedem | and the driver code is tightly coupled to how that option is parsed | 14:52 |
| mriedem | so we could mock it out and be hunky dorey but if the parsing returns something unexpected our unit test is just giving false positives | 14:52 |
| bauzas | mriedem: yeah, I know about that, my concern was just that we have so many places that are doing like the same that we could just avoid reproduce that here | 14:53 |
| *** READ10 has joined #openstack-nova | 14:53 | |
| mriedem | bauzas: this isn't the same thing as mocking out a call to conductor in a compute api unit test | 14:53 |
| *** aloga_ has joined #openstack-nova | 14:53 | |
| bauzas | mriedem: well I had so much pleasure with the Resource Tracker UTs that I'm now vaccinated against that pattern | 14:53 |
| bauzas | mriedem: but I don't want to bikeshed | 14:54 |
| mriedem | if the utility had to hit the database or something, that's a different story maye | 14:54 |
| mriedem | *maybe | 14:54 |
| mriedem | but the utility method is only one level deep, it doesn't touch any other services | 14:54 |
| mriedem | so it works or it doesn't | 14:54 |
| *** davideagnello has joined #openstack-nova | 14:54 | |
| bauzas | agreed, that's just an helper that we mock out | 14:54 |
| mriedem | and if it doesn't, we don't want to be mocking that out in the driver test | 14:54 |
| bauzas | well, okaaaay | 14:54 |
| mriedem | r'ar! :) | 14:55 |
| bauzas | sudipto: happy with releasing a new pathc ? | 14:55 |
| *** tonytan4ever has quit IRC | 14:55 | |
| mriedem | he's already working on it i think | 14:55 |
| bauzas | :) | 14:55 |
| *** dsanders has joined #openstack-nova | 14:55 | |
| *** ajayaa has quit IRC | 14:55 | |
| bauzas | alaski: what do you think of having a cells job voting ? | 14:56 |
| sudipto | mriedem, bauzas : yeah i am done making the change :) | 14:56 |
| mriedem | bauzas: he wanted to let it burn in for a few days | 14:56 |
| mriedem | then we make it voting | 14:57 |
| bauzas | sudipto: cool, waiting your new patch before clicking on a button :) | 14:57 |
| alaski | bauzas: I think it's a great idea, if we can go a few days with it being reliable | 14:57 |
| mriedem | people are really chomping at the bit on this cpu pin set change... | 14:57 |
| bauzas | mriedem: cool, I remember you working on logstash right? | 14:57 |
| *** jichen has quit IRC | 14:57 | |
| mriedem | bauzas: i got the cells logs indexed a couple of weeks ago | 14:57 |
| bauzas | awesome :) | 14:57 |
| *** romainh has joined #openstack-nova | 14:58 | |
| bauzas | already a couple of weeks? wah | 14:58 |
| *** ajayaa has joined #openstack-nova | 14:58 | |
| *** aloga_ has quit IRC | 14:58 | |
| *** signed8bit has joined #openstack-nova | 14:59 | |
| *** russellb has quit IRC | 15:00 | |
| *** dsanders has quit IRC | 15:02 | |
| *** dsanders has joined #openstack-nova | 15:02 | |
| mriedem | dansmith: you want this in https://review.openstack.org/#/c/176172/ | 15:02 |
| *** zz_johnthetubagu is now known as johnthetubaguy | 15:03 | |
| *** johnthetubaguy is now known as zz_johnthetubagu | 15:03 | |
| *** zz_johnthetubagu is now known as johnthetubaguy | 15:04 | |
| dansmith | mriedem: stupid print | 15:04 |
| *** russellb has joined #openstack-nova | 15:05 | |
| mriedem | dansmith: jhesketh also reported this https://bugs.launchpad.net/nova/+bug/1447132 | 15:06 |
| openstack | Launchpad bug 1447132 in OpenStack Compute (nova) "nova-manage db migrate_flavor_data doesn't do instances not in instance_extra" [Undecided,New] | 15:06 |
| *** annegentle has quit IRC | 15:06 | |
| *** annegentle has joined #openstack-nova | 15:06 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: Allow discrete blocks of online pCPUs for pinning https://review.openstack.org/173187 | 15:07 |
| dansmith | mriedem: ah | 15:07 |
| *** pixelb has quit IRC | 15:07 | |
| sudipto | bauzas, mriedem : hopefully i have all the comments covered now. | 15:07 |
| dansmith | mriedem: I guess we should have done this part with t-h midway through kilo | 15:07 |
| mriedem | *shrug* :/ | 15:08 |
| *** iamjarvo has joined #openstack-nova | 15:08 | |
| mriedem | dansmith: you can't make an omelet without breaking a few eggs right? :) | 15:08 |
| *** dsanders has quit IRC | 15:09 | |
| mriedem | sahid: +2 on https://review.openstack.org/#/c/170110/ | 15:10 |
| mriedem | _init_instance is a monster | 15:10 |
| *** apuimedo_ has quit IRC | 15:11 | |
| *** yassine_ has joined #openstack-nova | 15:12 | |
| sahid | thank you | 15:13 |
| dansmith | mriedem: heh, well, yeah, but we're going to have to backport these fixes | 15:13 |
| *** annegentle has quit IRC | 15:13 | |
| mriedem | dansmith: sure | 15:13 |
| mriedem | dansmith: that's what i'm +2 on stable now for :) | 15:13 |
| dansmith | hah | 15:14 |
| sahid | mriedem: you have the bit +2 on stable/juno? | 15:14 |
| dansmith | that was a very self-serving todo I took :) | 15:14 |
| mriedem | sahid: yeah, as of a few days ago | 15:14 |
| mriedem | sudipto: one tiny question in the testes https://review.openstack.org/#/c/173187/ | 15:15 |
| mriedem | i think there is a duplicate test now | 15:15 |
| sahid | if i can ask you to look a this https://review.openstack.org/#/c/155730/ it already have +2 | 15:15 |
| sudipto | mriedem darn - yeah. | 15:16 |
| mriedem | sudipto: ok, just remove that and then i'm +2 | 15:16 |
| *** dboik_ has joined #openstack-nova | 15:16 | |
| mriedem | sahid: yeah in a bit | 15:16 |
| *** fawadkhaliq has joined #openstack-nova | 15:16 | |
| *** dboik_ has quit IRC | 15:17 | |
| *** romainh has quit IRC | 15:17 | |
| mriedem | still need to read corporate overlord email this morning | 15:17 |
| *** dboik_ has joined #openstack-nova | 15:17 | |
| *** Marga_ has quit IRC | 15:17 | |
| sahid | mriedem: no worries at all | 15:18 |
| *** andreykurilin__ has quit IRC | 15:18 | |
| *** iamjarvo has quit IRC | 15:18 | |
| *** READ10 has quit IRC | 15:19 | |
| dansmith | mriedem: you just happened across this bug, right, you don't know if he's doing anything to address it? | 15:19 |
| mriedem | dansmith: no, jhesketh had it in the ML thread on the topic | 15:20 |
| mriedem | your thread on fixing t-h for the new migration | 15:20 |
| *** dboik has quit IRC | 15:20 | |
| dansmith | ah, hadn't looked yet | 15:20 |
| mriedem | i read 2 ML threads this morning, a record | 15:20 |
| *** rajesht has quit IRC | 15:20 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: Allow discrete blocks of online pCPUs for pinning https://review.openstack.org/173187 | 15:20 |
| *** yassine_ has quit IRC | 15:22 | |
| *** yassine_ has joined #openstack-nova | 15:22 | |
| sudipto | mriedem, thanks! | 15:22 |
| sahid | sudipto: good luck ! | 15:22 |
| sudipto | sahid, thanks! | 15:22 |
| sudipto | dansmith, If i may ask for your thoughts on the review now - whenever you have time. https://review.openstack.org/173187 | 15:23 |
| *** yassine_ has quit IRC | 15:24 | |
| *** READ10 has joined #openstack-nova | 15:24 | |
| *** lefais has joined #openstack-nova | 15:24 | |
| *** lefais has quit IRC | 15:25 | |
| *** yassine_ has joined #openstack-nova | 15:25 | |
| *** armax has joined #openstack-nova | 15:26 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: mark NUMA huge page mappings as shared instead https://review.openstack.org/176360 | 15:30 |
| *** claudiub has quit IRC | 15:32 | |
| dansmith | zzzeek: can you help me with some sqla-fu? | 15:32 |
| zzzeek | dansmith: sure | 15:32 |
| *** oro has quit IRC | 15:32 | |
| sudipto | sahid, mriedem : there's a very small review out here. https://review.openstack.org/#/c/175357/ if you can take a look sometime.... | 15:32 |
| dansmith | zzzeek: https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/api.py#L6043-6045 | 15:33 |
| *** annegentle has joined #openstack-nova | 15:33 | |
| dansmith | zzzeek: that works as long as there is a row for the instance in the extra table | 15:33 |
| *** marun has joined #openstack-nova | 15:33 | |
| dansmith | zzzeek: but there might not be one, so is there something I can do to that query to not exclude those instances? | 15:34 |
| *** kiran-r has quit IRC | 15:34 | |
| zzzeek | dansmith: outerjoin | 15:34 |
| *** rwsu has joined #openstack-nova | 15:34 | |
| dansmith | zzzeek: s/join(/outerjoin(/ ? | 15:35 |
| zzzeek | dansmith: yup | 15:35 |
| zzzeek | outerjoin will include rows where there is no instnace.extra | 15:35 |
| dansmith | zzzeek: hmm, seems too easy, got anything else? | 15:35 |
| zzzeek | dansmith: and since you are already filtering InstanceExtra on a column that is NULL, it will still be NULL | 15:35 |
| dansmith | zzzeek: (no not really) | 15:35 |
| dansmith | zzzeek: awesome | 15:36 |
| zzzeek | dansmith: if you were filtering InstnaceExtra on something non-NULL, then you’d need to add an OR with what to do for an outerjoin, but i dont hitnk thats needed here | 15:36 |
| dansmith | zzzeek: okay excellent | 15:36 |
| dansmith | I'll give 'er a shot | 15:36 |
| zzzeek | dansmith: let me know | 15:36 |
| dansmith | I shall, thanks :) | 15:36 |
| *** annegentle has quit IRC | 15:39 | |
| *** sudipto has quit IRC | 15:40 | |
| *** annegentle has joined #openstack-nova | 15:40 | |
| *** sudipto has joined #openstack-nova | 15:40 | |
| *** Sukhdev has joined #openstack-nova | 15:40 | |
| *** yamahata has quit IRC | 15:43 | |
| *** iamjarvo has joined #openstack-nova | 15:43 | |
| *** iamjarvo has quit IRC | 15:44 | |
| *** iamjarvo has joined #openstack-nova | 15:44 | |
| *** iamjarvo has quit IRC | 15:45 | |
| *** sdake_ has quit IRC | 15:48 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: use importutils instead of python built-in https://review.openstack.org/176372 | 15:50 |
| *** hemnafk is now known as hemna | 15:52 | |
| *** atuvenie has quit IRC | 15:52 | |
| sahid | mriedem: dansmith anychance you give my your thinking about this https://review.openstack.org/#/c/168876/ ? | 15:53 |
| dansmith | sahid: maybe in a bit | 15:53 |
| dansmith | zzzeek: seems to work, thanks for making it easy :) | 15:54 |
| zzzeek | dansmith: cool | 15:54 |
| sahid | dansmith: cool | 15:54 |
| sahid | :) | 15:54 |
| *** romainh has joined #openstack-nova | 15:55 | |
| *** atuvenie has joined #openstack-nova | 15:55 | |
| *** lpetrut has quit IRC | 15:57 | |
| mriedem | sahid: wrapped up in some other stuff atm, have to come back later | 15:57 |
| *** vilobhmm1 has joined #openstack-nova | 15:58 | |
| *** dboik_ has quit IRC | 15:59 | |
| *** dboik has joined #openstack-nova | 16:00 | |
| *** Marga_ has joined #openstack-nova | 16:00 | |
| *** flip214 has joined #openstack-nova | 16:01 | |
| *** otter768 has joined #openstack-nova | 16:02 | |
| *** Hosam has joined #openstack-nova | 16:05 | |
| *** redbeard has quit IRC | 16:07 | |
| *** dprince has quit IRC | 16:07 | |
| *** otter768 has quit IRC | 16:07 | |
| *** matrohon has quit IRC | 16:07 | |
| cfriesen | hi folks...I'm seeing an odd error where nova-compute is spinning trying to log but the log socket isn't connected anymore. However, nova-compute doesn't try to reconnect the socket so it'll keep failing. strace on the nova-compute pid looks like this: | 16:07 |
| *** browne has quit IRC | 16:08 | |
| cfriesen | 1429716352.692420 epoll_ctl(16, EPOLL_CTL_ADD, 3, {EPOLLOUT|EPOLLERR|EPOLLHUP, {u32=3, u64=12834288515915186179}}) = 0 | 16:08 |
| cfriesen | 1429716352.694612 epoll_wait(16, {{EPOLLOUT|EPOLLHUP, {u32=3, u64=12834288515915186179}}}, 1023, 60000) = 1 | 16:08 |
| cfriesen | 1429716352.696764 epoll_ctl(16, EPOLL_CTL_DEL, 3, {EPOLLRDBAND|EPOLLWRBAND|EPOLLHUP|EPOLLRDHUP|0x4348000, {u32=0, u64=12834288515915186176}}) = 0 | 16:08 |
| cfriesen | 1429716352.698915 sendto(3, "<183>nova-compute 2015-04-21 11:17:45.570 11565 DEBUG oslo_messaging._drivers.amqp [req-0257930a-3b50-42b9-ae02-35c4b257517b - - - - -] UNIQUE_ID is 8e844dabd9e942cabe5241f8f5992d3e. _add_unique_id /u"..., 269, 0, NULL, 0) = -1 ENOTCONN (Transport endpoint is not connected) | 16:08 |
| cfriesen | 1429716352.701564 epoll_ctl(16, EPOLL_CTL_ADD, 3, {EPOLLOUT|EPOLLERR|EPOLLHUP, {u32=3, u64=12834288515915186179}}) = 0 | 16:08 |
| cfriesen | 1429716352.704726 epoll_wait(16, {{EPOLLOUT|EPOLLHUP, {u32=3, u64=12834288515915186179}}}, 1023, 60000) = 1 | 16:08 |
| cfriesen | 1429716352.706861 epoll_ctl(16, EPOLL_CTL_DEL, 3, {EPOLLRDBAND|EPOLLWRBAND|EPOLLHUP|EPOLLRDHUP|0x4348000, {u32=0, u64=12834288515915186176}}) = 0 | 16:08 |
| cfriesen | 1429716352.708977 sendto(3, "<183>nova-compute 2015-04-21 11:17:45.570 11565 DEBUG oslo_messaging._drivers.amqp [req-0257930a-3b50-42b9-ae02-35c4b257517b - - - - -] UNIQUE_ID is 8e844dabd9e942cabe5241f8f5992d3e. _add_unique_id /u"..., 269, 0, NULL, 0) = -1 ENOTCONN (Transport endpoint is not connected) | 16:08 |
| cfriesen | (sorry for the blast) | 16:08 |
| mriedem | ew, paste | 16:08 |
| *** marun has quit IRC | 16:09 | |
| cfriesen | mriedem: etherpad or something next time? | 16:10 |
| mriedem | paste.openstack.org | 16:10 |
| *** arnaud___ has joined #openstack-nova | 16:10 | |
| mriedem | or https://gist.github.com/ | 16:10 |
| cfriesen | mriedem: will do, thanks | 16:10 |
| cfriesen | anyone got any ideas? almost seems like bad error handling of the ENOTCONN on the logging socket | 16:11 |
| *** romainh has quit IRC | 16:11 | |
| *** yassine_ has quit IRC | 16:12 | |
| cfriesen | I think this is probably the syslog socket, /var/log/nova/nova-compute.log is size zero | 16:12 |
| *** jistr has quit IRC | 16:12 | |
| cfriesen | if I restart nova-compute it comes back up and logs properly | 16:12 |
| *** lpetrut has joined #openstack-nova | 16:13 | |
| *** dboik has quit IRC | 16:15 | |
| *** unicell1 has quit IRC | 16:18 | |
| mriedem | cfriesen: do you see stuff in /var/log/messages? | 16:19 |
| mriedem | i noticed recently in our internal ci that nova logs were going to both /var/log/nova and /var/log/messages | 16:19 |
| bauzas | mriedem: because you're using syslog ? | 16:20 |
| mriedem | it didn't use to show up in there | 16:20 |
| bauzas | mriedem: I remember a flag about that | 16:20 |
| bauzas | hold on, checking | 16:20 |
| mriedem | could be a rhel 6.5 vs 7.1 thing | 16:20 |
| *** jlanoux has quit IRC | 16:20 | |
| mriedem | if some defaults changed, idk | 16:20 |
| mriedem | doesn't appear to happen in the community runs http://logs.openstack.org/70/175570/1/check/check-tempest-dsvm-full/04ca208/logs/syslog.txt.gz | 16:21 |
| bauzas | mriedem: well, ask a Red Hat boy, idk :) | 16:21 |
| openstackgerrit | Dan Smith proposed openstack/nova: Fix migrate_flavor_data() to catch instances with no instance_extra rows https://review.openstack.org/176387 | 16:21 |
| dansmith | mriedem: ^ | 16:21 |
| dansmith | mriedem: not too bad.. still had to create the row, so it's not just a four-char fix, but.. close :) | 16:21 |
| *** coolsvap|afk has quit IRC | 16:21 | |
| dansmith | oops, left some dbeug in there | 16:21 |
| mriedem | at least they weren't prints | 16:22 |
| dansmith | stacktest doesn't show me prints, so they have to be LOGs now :) | 16:22 |
| bauzas | mriedem: yeah, there is a flag http://docs.openstack.org/admin-guide-cloud/content/section_manage-logs.html | 16:22 |
| *** dboik has joined #openstack-nova | 16:24 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Fix migrate_flavor_data() to catch instances with no instance_extra rows https://review.openstack.org/176387 | 16:24 |
| *** coolsvap|afk has joined #openstack-nova | 16:24 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Fix migrate_flavor_data() to catch instances with no instance_extra rows https://review.openstack.org/176387 | 16:25 |
| *** dprince has joined #openstack-nova | 16:25 | |
| mriedem | bauzas: aha | 16:27 |
| mriedem | oslo.log 0.4.0 changed use_syslog to default to True (from False) | 16:27 |
| bauzas | mriedem: digging into the code to see the definition | 16:27 |
| mriedem | it's defined in oslo.log | 16:27 |
| bauzas | baaam | 16:27 |
| bauzas | mriedem: that was exactly where I was looking for but you beated me | 16:27 |
| mriedem | so now if you have debug=True in nova.conf and use_syslog=True (by default) it's going both places | 16:27 |
| *** annashen has joined #openstack-nova | 16:28 | |
| bauzas | honestly, I was confounding oslo.cfg and oslo.log (/me tired) | 16:28 |
| *** haomaiwang has quit IRC | 16:28 | |
| mriedem | bauzas: you're not allowed to be tired | 16:28 |
| *** haomaiwang has joined #openstack-nova | 16:28 | |
| mriedem | i bet devstack is setting use_syslog=False | 16:28 |
| bauzas | mriedem: speak to the one who will have to babysit tomorrow | 16:28 |
| cfriesen | mriedem: sorry, was writing this up locally. no, in our environment we don't even have /var/log/messages. But syslog is working, other stuff is logging. | 16:29 |
| *** rajesht has joined #openstack-nova | 16:29 | |
| mriedem | SYSLOG=$(trueorfalse False SYSLOG) | 16:29 |
| mriedem | devstack defaults to not use syslog | 16:29 |
| *** Marga_ has quit IRC | 16:29 | |
| mriedem | oslo.log defaults to True | 16:29 |
| mriedem | where are the sane defaults police when you need them? | 16:29 |
| bauzas | cfriesen: would be worth checking your defaults if syslog is enabled or no | 16:30 |
| *** annashen has quit IRC | 16:30 | |
| mriedem | jogo: ^ | 16:30 |
| *** Marga_ has joined #openstack-nova | 16:30 | |
| *** atuvenie has quit IRC | 16:30 | |
| cfriesen | mriedem: use_syslog=True in nova.conf | 16:31 |
| bauzas | cfriesen: then IMHO you should tune your factory to write logs correctly | 16:31 |
| cfriesen | mriedem: bauzas: but that doesn't explain why nova-compute wouldn't close and re-connect the socket after it gets an ENOTCONN on the log socket | 16:32 |
| cfriesen | is that an oslo.log thing? | 16:32 |
| bauzas | mriedem: funny https://github.com/openstack/oslo.log/blob/1.1.0/oslo_log/_options.py#L64 | 16:33 |
| *** Marga_ has quit IRC | 16:33 | |
| *** Marga_ has joined #openstack-nova | 16:34 | |
| dhellmann | cfriesen, bauzas : as far as I can tell from the git logs, the default for use-syslog has never been True in oslo.log | 16:35 |
| mriedem | cfriesen: https://github.com/openstack/oslo-incubator/commit/1188d88a73e75da5c143d0dfad34e10dad0daee0 ? | 16:35 |
| bauzas | cfriesen: yup https://github.com/openstack/nova/blob/master/nova/cmd/all.py#L50 | 16:35 |
| bauzas | dhellmann: yeah that's what I raised ^ | 16:35 |
| dansmith | alaski: I just saw a really really weird thing | 16:36 |
| dhellmann | bauzas: ok, just making sure :-) | 16:36 |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: libvirt: Allow discrete online pCPUs for pinning https://review.openstack.org/173187 | 16:36 |
| bauzas | dhellmann: I haven't checked the incubator tho | 16:36 |
| alaski | dansmith: uh oh | 16:36 |
| dansmith | alaski: yeah, it was a bit concerning: | 16:36 |
| dansmith | alaski: I submitted a patch | 16:36 |
| mriedem | bauzas: wrong about the default, i saw the change log but it was this https://review.openstack.org/#/c/133638/1 | 16:37 |
| dansmith | alaski: it went into the queue | 16:37 |
| *** andreykurilin__ has joined #openstack-nova | 16:37 | |
| dansmith | alaski: and then when it came out, the cells job reported as passed | 16:37 |
| dansmith | alaski: any ideas? | 16:37 |
| alaski | dansmith: hah | 16:37 |
| bauzas | mriedem: yeah that's exactly what I'm checking, old Nova can still uses the incubator | 16:37 |
| dansmith | alaski: :) | 16:37 |
| bauzas | and dims not there, damn PTL ! | 16:37 |
| bauzas | :p | 16:37 |
| alaski | dansmith: now that it's happened, how will you live with it? | 16:37 |
| dhellmann | bauzas: the code in oslo.log has the history from the incubator, but it has moved around so it might be a bit harder to scan | 16:37 |
| bauzas | dhellmann: we probably need to check the Nova code directly | 16:38 |
| cfriesen | bauzas: dhellmann: what's the proper way to cause nova to log to syslog then? | 16:38 |
| dhellmann | bauzas: yep | 16:38 |
| bauzas | dhellmann: I mean the old one | 16:38 |
| *** davideagnello has quit IRC | 16:38 | |
| dhellmann | cfriesen: do you mean other than setting that config flag to true in the config file? | 16:38 |
| bauzas | ok let's sit up in the delorean... | 16:38 |
| *** dboik has quit IRC | 16:39 | |
| sudipto | mriedem, bauzas apologies for asking again but had to incorporate some comments from garyk | 16:39 |
| mriedem | sudipto: yeah, +2 again | 16:39 |
| sudipto | mriedem, thanks | 16:39 |
| cfriesen | mriedem: that seems to only be for the case where we start up the service and syslog isn't running. what if the syslog daemon restarts while nova-compute is running? | 16:39 |
| *** arnaud___ has quit IRC | 16:39 | |
| bauzas | sudipto: I only serve to click buttons | 16:39 |
| sudipto | bauzas, :-) thx. | 16:39 |
| *** mpaolino has quit IRC | 16:41 | |
| mriedem | dansmith: re; the thread with jhesketh, how many actual CD (< 2 weeks) are out in the wild now? | 16:41 |
| mriedem | i thought rax was more like 4 releases per year | 16:41 |
| mriedem | and HP public cloud is no more | 16:41 |
| mriedem | and cern is a big bang upgrade every release right b/c of cells | 16:42 |
| bauzas | eh, juno was already https://github.com/openstack/nova/blob/stable/juno/nova/openstack/common/log.py#L102 | 16:42 |
| openstackgerrit | Jim Rollenhagen proposed openstack/nova: Ironic: pass injected files through to configdrive https://review.openstack.org/176395 | 16:43 |
| bauzas | mmm, sounds like Nova never defaulted to use syslog... | 16:43 |
| *** READ10 has quit IRC | 16:44 | |
| dansmith | mriedem: I thought hp was quite close | 16:47 |
| dansmith | mriedem: but like I said, they could have done nothing so far and they'd be fine | 16:47 |
| dansmith | mriedem: i.e. just opportunistic migrations | 16:47 |
| *** unicell has joined #openstack-nova | 16:48 | |
| *** subscope_ has joined #openstack-nova | 16:48 | |
| *** browne has joined #openstack-nova | 16:50 | |
| jroll | who's handling RC backport stuff? I just proposed a backport for https://bugs.launchpad.net/nova/+bug/1447249 | 16:51 |
| openstack | Launchpad bug 1447249 in OpenStack Compute (nova) "Ironic: injected files not passed through to configdrive" [Undecided,In progress] - Assigned to Jim Rollenhagen (jim-rollenhagen) | 16:51 |
| *** ijw has joined #openstack-nova | 16:51 | |
| mriedem | jroll: that's not an rc blocker right? | 16:52 |
| *** tonytan4ever has joined #openstack-nova | 16:52 | |
| *** pkoniszewski has joined #openstack-nova | 16:52 | |
| jroll | mriedem: I guess I'm curious who would decide that, it's certainly a bug in kilo | 16:52 |
| mriedem | jroll: looks like more of a kilo-backport-potential than an rc blocker | 16:53 |
| mriedem | jroll: kilo-rc-potential means we should potentially hold up the kilo release for this | 16:53 |
| jroll | mriedem: got it, I agree | 16:53 |
| * jroll fixes that tag | 16:53 | |
| mriedem | already done | 16:53 |
| *** penick has joined #openstack-nova | 16:53 | |
| jroll | heh, thanks | 16:53 |
| *** vilobhmm11 has joined #openstack-nova | 16:54 | |
| *** mpaolino has joined #openstack-nova | 16:54 | |
| *** melwitt has joined #openstack-nova | 16:55 | |
| *** sahid has quit IRC | 16:56 | |
| *** vilobhmm1 has quit IRC | 16:56 | |
| *** harlowja_away is now known as harlowja | 16:59 | |
| *** derekh has quit IRC | 17:00 | |
| melwitt | bauzas: cells meeting if you're around | 17:02 |
| melwitt | nvm | 17:02 |
| bauzas | melwitt: on the phone but here anyway | 17:02 |
| *** andreykurilin__ has quit IRC | 17:03 | |
| *** ildikov has quit IRC | 17:04 | |
| *** pask81 has quit IRC | 17:04 | |
| mriedem | dansmith: question inline https://review.openstack.org/#/c/176387/ | 17:06 |
| *** subscope_ has quit IRC | 17:07 | |
| mriedem | i expected extra.flavor to be None | 17:07 |
| *** subscope_ has joined #openstack-nova | 17:07 | |
| *** salv-orlando has joined #openstack-nova | 17:07 | |
| dansmith | mriedem: replied | 17:09 |
| *** pkoniszewski has quit IRC | 17:11 | |
| mriedem | so the flavor migration in instance.save() sets extra.flavor to something? | 17:11 |
| mriedem | and it'd be whatever is in sys_meta | 17:11 |
| mriedem | ok | 17:11 |
| dansmith | mriedem: yeah | 17:12 |
| mriedem | alright, +2 | 17:12 |
| *** annegentle has quit IRC | 17:13 | |
| dansmith | thanks | 17:13 |
| *** vilobhmm11 has quit IRC | 17:13 | |
| *** unicell has quit IRC | 17:16 | |
| *** unicell1 has joined #openstack-nova | 17:16 | |
| *** russellb has quit IRC | 17:17 | |
| *** tjones1 has joined #openstack-nova | 17:19 | |
| *** davideagnello has joined #openstack-nova | 17:21 | |
| *** isd has joined #openstack-nova | 17:21 | |
| *** russellb has joined #openstack-nova | 17:21 | |
| *** lennyb has joined #openstack-nova | 17:23 | |
| *** VW_ has quit IRC | 17:23 | |
| *** joefides_ has joined #openstack-nova | 17:23 | |
| *** sdake has joined #openstack-nova | 17:24 | |
| *** joefides has quit IRC | 17:26 | |
| *** joefides_ has quit IRC | 17:27 | |
| *** lucasagomes is now known as lucas-dinner | 17:28 | |
| *** dboik has joined #openstack-nova | 17:29 | |
| *** sdake has quit IRC | 17:29 | |
| *** kr4m3r has joined #openstack-nova | 17:29 | |
| *** sdake has joined #openstack-nova | 17:30 | |
| *** dboik has quit IRC | 17:30 | |
| *** annashen has joined #openstack-nova | 17:30 | |
| *** dboik has joined #openstack-nova | 17:31 | |
| *** dboik has quit IRC | 17:31 | |
| kr4m3r | hello all, i have a lab deployed with devstack using nova-network. I have fixed a private network only (no floating) with my local network and all is working good. I set private as 10.0.0.129/27. So, is it possible modify private network to a /26 without change ip allocated on instances now? | 17:31 |
| *** dboik has joined #openstack-nova | 17:31 | |
| mriedem | kr4m3r: please see the channel topic | 17:32 |
| -openstackstatus- NOTICE: gerrit is restarting to clear hung stream-events tasks. any review events between 16:48 and 17:32 utc will need to be rechecked or have their approval votes reapplied to trigger testing in zuul | 17:32 | |
| kr4m3r | mriedem: sorry! | 17:32 |
| *** lpetrut has quit IRC | 17:33 | |
| *** russellb has quit IRC | 17:35 | |
| *** annashen has quit IRC | 17:35 | |
| *** kr4m3r has left #openstack-nova | 17:36 | |
| *** Mike_D_laptop has joined #openstack-nova | 17:36 | |
| dansmith | jaypipes: this is still a thing? https://review.openstack.org/#/c/152689/ | 17:39 |
| *** eandersson has left #openstack-nova | 17:39 | |
| *** russellb has joined #openstack-nova | 17:40 | |
| *** ijw has quit IRC | 17:40 | |
| *** Sukhdev has quit IRC | 17:40 | |
| *** vilobhmm1 has joined #openstack-nova | 17:40 | |
| bauzas | alaski: so I missed the most part of the meeting :( | 17:41 |
| bauzas | alaski: sorry about that, but yeah I already thought about not providing an Instance object | 17:42 |
| bauzas | alaski: instead we should look at what the filters check for instance_props and just leave them in there | 17:42 |
| *** david-lyle has quit IRC | 17:42 | |
| bauzas | alaski: I was just lazy to update this for the spec | 17:42 |
| alaski | bauzas: no problem. I can propose an update | 17:42 |
| bauzas | dansmith: melwitt: ^ | 17:42 |
| alaski | bauzas: also we should split instance_props into component pieces, not a dict | 17:43 |
| jaypipes | dansmith: yeah, I think so. | 17:43 |
| bauzas | alaski: sure, what I suggest you is to review all the filters to see what's needed | 17:43 |
| bauzas | alaski: it was not a dict anyway | 17:43 |
| bauzas | alaski: I mean in the objecty | 17:43 |
| alaski | bauzas: yeah. sorry, meant the current dict | 17:43 |
| bauzas | alaski: instead of that, we need to have X fields, one per filter need | 17:44 |
| alaski | yeah | 17:44 |
| *** lennyb has left #openstack-nova | 17:44 | |
| bauzas | alaski: so that if a filter needs a new one, they just bump a new version | 17:44 |
| alaski | right, version the object not an attribute on it | 17:44 |
| bauzas | alaski: as I said, I was lazy so I didn't check it for the spec | 17:44 |
| bauzas | alaski: like I did for filt_props | 17:45 |
| bauzas | alaski: here http://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/request-spec-object.html#implementation | 17:45 |
| *** mpaolino has quit IRC | 17:45 | |
| bauzas | alaski: I splitted filt_props keys into each fielf | 17:46 |
| dansmith | jaypipes: :( | 17:46 |
| bauzas | alaski: so what you need is to remove the Instance nested field and provide instead a new field per related Instance field | 17:46 |
| alaski | bauzas: rigth | 17:47 |
| alaski | right | 17:47 |
| bauzas | alaski: ping me if you need help | 17:47 |
| bauzas | alaski: in the mean time, I'll work on the implementation | 17:47 |
| bauzas | alaski: so I will probably have to do that too :D | 17:48 |
| bauzas | alaski: I'm on PTO tomorrow btw. | 17:48 |
| jaypipes | dansmith: will look into it shortly. | 17:49 |
| alaski | bauzas: okay, enjoy your PTO | 17:50 |
| *** yamahata has joined #openstack-nova | 17:50 | |
| dansmith | jaypipes: sahid just rebased it I guess | 17:50 |
| dansmith | jaypipes: I'll take another look through, I was just surprised that wasn't in already :( | 17:51 |
| *** atuvenie has joined #openstack-nova | 17:51 | |
| *** achanda has joined #openstack-nova | 17:51 | |
| *** annegentle has joined #openstack-nova | 17:58 | |
| *** annegentle has quit IRC | 18:01 | |
| *** annegentle has joined #openstack-nova | 18:02 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: mark NUMA huge page mappings as shared access https://review.openstack.org/176360 | 18:02 |
| *** yamahata has quit IRC | 18:03 | |
| *** otter768 has joined #openstack-nova | 18:03 | |
| *** johnthetubaguy is now known as zz_johnthetubagu | 18:07 | |
| *** otter768 has quit IRC | 18:08 | |
| *** eharney has quit IRC | 18:11 | |
| *** Mike_D_laptop has quit IRC | 18:15 | |
| *** ijw has joined #openstack-nova | 18:18 | |
| *** aix has quit IRC | 18:21 | |
| *** garyk has quit IRC | 18:21 | |
| *** tonytan_brb has joined #openstack-nova | 18:22 | |
| *** Mike_D_laptop has joined #openstack-nova | 18:22 | |
| *** tonytan4ever has quit IRC | 18:23 | |
| *** lpetrut has joined #openstack-nova | 18:23 | |
| *** alexpilotti has quit IRC | 18:25 | |
| *** Mike_D_laptop has quit IRC | 18:26 | |
| *** marun has joined #openstack-nova | 18:28 | |
| *** jogo has quit IRC | 18:29 | |
| *** sudipto has quit IRC | 18:32 | |
| *** ildikov has joined #openstack-nova | 18:33 | |
| *** yamahata has joined #openstack-nova | 18:33 | |
| *** dikonoor has quit IRC | 18:33 | |
| *** VW_ has joined #openstack-nova | 18:34 | |
| *** rajesht has quit IRC | 18:35 | |
| *** dboik_ has joined #openstack-nova | 18:38 | |
| *** jogo has joined #openstack-nova | 18:40 | |
| *** yamahata has quit IRC | 18:40 | |
| *** dboik has quit IRC | 18:41 | |
| *** joefides has joined #openstack-nova | 18:42 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add config option to disable handling virt lifecycle events https://review.openstack.org/159275 | 18:45 |
| *** eharney has joined #openstack-nova | 18:46 | |
| *** tonytan4ever has joined #openstack-nova | 18:48 | |
| *** ZZelle has quit IRC | 18:48 | |
| *** ZZelle has joined #openstack-nova | 18:49 | |
| *** thedodd has joined #openstack-nova | 18:49 | |
| *** neelashah has quit IRC | 18:50 | |
| *** tonytan_brb has quit IRC | 18:50 | |
| *** rushil has quit IRC | 18:51 | |
| *** ivasev has quit IRC | 18:54 | |
| *** jichen has joined #openstack-nova | 18:55 | |
| *** artom has joined #openstack-nova | 18:57 | |
| *** artom_ has quit IRC | 18:57 | |
| *** lsmola__ has joined #openstack-nova | 19:02 | |
| *** redbeard has joined #openstack-nova | 19:04 | |
| *** ijw_ has joined #openstack-nova | 19:04 | |
| *** neelashah has joined #openstack-nova | 19:06 | |
| *** lsmola_ has quit IRC | 19:06 | |
| *** neelashah1 has joined #openstack-nova | 19:07 | |
| *** ijw has quit IRC | 19:07 | |
| dansmith | this needs eyes: https://review.openstack.org/#/c/176387/ | 19:08 |
| dansmith | we need to have that on the shortlist for backporting so we can make the flavor migration required | 19:08 |
| *** lsmola__ has quit IRC | 19:10 | |
| *** neelashah has quit IRC | 19:10 | |
| *** ArxCruz has joined #openstack-nova | 19:13 | |
| *** annegentle has quit IRC | 19:14 | |
| *** claudiub has joined #openstack-nova | 19:15 | |
| *** edmondsw has joined #openstack-nova | 19:16 | |
| *** xgsa has quit IRC | 19:18 | |
| *** annegentle has joined #openstack-nova | 19:18 | |
| *** ijw_ has quit IRC | 19:19 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add config option to disable handling virt lifecycle events https://review.openstack.org/159275 | 19:21 |
| *** achanda has quit IRC | 19:22 | |
| *** lsmola__ has joined #openstack-nova | 19:22 | |
| *** VW_ has quit IRC | 19:23 | |
| *** jaypipes has quit IRC | 19:24 | |
| *** rushil has joined #openstack-nova | 19:26 | |
| dansmith | clarkb: jogo: what happened to the aiopcpu job? | 19:27 |
| clarkb | dansmith: it is called "multinode" now | 19:27 |
| dansmith | oh dang | 19:27 |
| clarkb | should still be there | 19:27 |
| dansmith | just saw that | 19:27 |
| jogo | dansmith: isn't that a much better name? | 19:28 |
| dansmith | jogo: I just was pattern matching for aiopcpu and didn't see it | 19:28 |
| dansmith | obviously it's a better name :) | 19:28 |
| *** sandywalsh has quit IRC | 19:30 | |
| *** VW_ has joined #openstack-nova | 19:30 | |
| *** sandywalsh has joined #openstack-nova | 19:31 | |
| *** achanda has joined #openstack-nova | 19:31 | |
| *** nkrinner has joined #openstack-nova | 19:37 | |
| *** shakamunyi has joined #openstack-nova | 19:40 | |
| mriedem | sdague: regression in the api policy stuff, can you do a sanity check on this since it's going to be backported - https://review.openstack.org/#/c/175739/ | 19:42 |
| *** mmedvede has quit IRC | 19:43 | |
| *** joefides has quit IRC | 19:44 | |
| redbeard | for debugging: do I have to investigate nova/compute/manager.py or nova/manager.py for checking instance cache related things? | 19:47 |
| mriedem | compute manager | 19:47 |
| sdague | mriedem: that's a regression? | 19:47 |
| mriedem | sdague: well, there is a flavormanage entry in policy.json so it's configurable, | 19:48 |
| mriedem | sdague: which is also making me wonder about os-hosts: https://review.openstack.org/#/c/160088/ | 19:49 |
| sdague | mriedem: it always was? because I thought we had db level admin checks here | 19:49 |
| mriedem | b/c of the poliy file http://git.openstack.org/cgit/openstack/nova/tree/etc/nova/policy.json#n91 | 19:49 |
| mriedem | we did | 19:49 |
| *** ZZelle_ has joined #openstack-nova | 19:49 | |
| mriedem | oh right, so it didn't matter | 19:49 |
| sdague | so it was in the policy file, but ignored | 19:49 |
| mriedem | yeah | 19:49 |
| mriedem | you could say non-admin in policy but the db api would explode w/o admin | 19:49 |
| sdague | correct | 19:49 |
| sdague | we *knew* the policy work wasn't going to be 100% complete for kilo, and people were ok with getting in what we could | 19:50 |
| sdague | so, unless somone can point that this is actually a regression in behavior from juno, I don't think we should backport to kilo | 19:50 |
| sdague | or did we just do a context.elevated() before making the db call? | 19:51 |
| mriedem | yeah, i think tha'ts the difference | 19:51 |
| redbeard | thanks mriedem | 19:51 |
| *** annegentle has quit IRC | 19:52 | |
| *** ociuhandu has quit IRC | 19:52 | |
| mriedem | the nova.compute.flavors code would elevate the context | 19:52 |
| sdague | mriedem: ok, can we dig up that code reference then? | 19:52 |
| mriedem | johnthetubaguy did in the review | 19:52 |
| openstackgerrit | jichenjc proposed openstack/nova: Add missing @require_context https://review.openstack.org/176454 | 19:52 |
| *** atuvenie has quit IRC | 19:53 | |
| bauzas | mriedem: sdague: I wrote a thread today about that | 19:53 |
| mriedem | the commit message here could have used more detail | 19:53 |
| sdague | ok, sold | 19:53 |
| bauzas | mriedem: sdague: FYI http://lists.openstack.org/pipermail/openstack-dev/2015-April/062282.html | 19:54 |
| sdague | mriedem: yes, commit messages on stuff like this needs *a lot* more context | 19:54 |
| sdague | that would have hugely saved time | 19:54 |
| mriedem | bauzas: ah i see hypervisors was also affected from your thread | 19:54 |
| mriedem | i was looking at os-hosts | 19:54 |
| *** openstackgerrit has quit IRC | 19:54 | |
| cfriesen | mriedem: regarding https://review.openstack.org/#/c/159275/ if that change is made, what happens if the instance shuts itself down? Would nova start it back up again? | 19:54 |
| bauzas | mriedem: IMHO, each time we were using require_admin_context(), it was bugging | 19:55 |
| *** mmedvede has joined #openstack-nova | 19:55 | |
| *** openstackgerrit has joined #openstack-nova | 19:55 | |
| bauzas | mriedem: here, it was due to service_get_by_cn() | 19:55 |
| bauzas | mriedem: but frankly, I'm just in favor of removing the require_admin_context() call everywhere because that's just not RBAC | 19:56 |
| mriedem | cfriesen: shuts itself down? | 19:57 |
| bauzas | mriedem: sdague: https://bugs.launchpad.net/nova/+bug/1447164 | 19:57 |
| openstack | Launchpad bug 1447164 in OpenStack Compute (nova) "require_admin_context() does not account for policy.json rulesets" [Medium,New] | 19:57 |
| mriedem | cfriesen: nova never starts an instance back up | 19:57 |
| *** Marga_ has quit IRC | 19:57 | |
| mriedem | cfriesen: in sync_instance_power_state, we only stop instances | 19:57 |
| sdague | so.... can we actually build some non mock based tests for these thing? | 19:58 |
| mriedem | if nova thinks it's active but the hypervisor says it's not, we stop it | 19:58 |
| mriedem | cfriesen: if nova thinks it's stopped but the hypervisor says it's active, we stop it | 19:58 |
| mriedem | which is really confusing | 19:58 |
| cfriesen | mriedem: right, so currently if the guest stops itself that triggers a lifecycle event that results in nova updating the DB to "shutoff" | 19:58 |
| mriedem | yeah | 19:58 |
| mriedem | which the periodic task would also hit every 10 minutes by default | 19:59 |
| mriedem | cfriesen: also note http://git.openstack.org/cgit/openstack/nova/tree/nova/compute/power_state.py#n19 | 19:59 |
| mriedem | "The hypervisor is always considered the authority on the status of a particular VM, and the power_state in the DB should be viewed as a snapshot of the VMs's state in the (recent) past." | 19:59 |
| *** subscope_ has quit IRC | 19:59 | |
| mriedem | that's not true in the compute manager | 19:59 |
| cfriesen | mriedem: so with your change, would we potentially wait 10min to update the db? | 19:59 |
| mriedem | cfriesen: yes, assuming it syncs up | 19:59 |
| cfriesen | mreidem: are you suggesting that we don't _want_ the hypervisor to be the authority? It seems reasonable to me that if a guest halts themselves that the nova state get updated to that effect. Also, if the hypervisor process crashes I think it would be good to update the DB immediately. | 20:01 |
| dansmith | cfriesen: I think he's suggesting that there are lots of problems with these events | 20:02 |
| dansmith | cfriesen: and that if an operator wants to disable them, they should be able to | 20:02 |
| *** ijw has joined #openstack-nova | 20:03 | |
| *** otter768 has joined #openstack-nova | 20:04 | |
| cfriesen | dansmith: yeah, I get that. just seems like a bit of a cop-out. "this is hard, so we won't bother doing it." :) | 20:05 |
| dansmith | cfriesen: not really | 20:05 |
| dansmith | cfriesen: it's a workaround | 20:05 |
| *** ildikov has quit IRC | 20:05 | |
| dansmith | cfriesen: it's a "we suck at getting this right, and it's not critical, so you can turn it off until we figure it out if you'd prefer" | 20:05 |
| mriedem | cfriesen: yes that ^ | 20:05 |
| mriedem | especially around reboot | 20:06 |
| mriedem | there are several bugs and changes tied to this code as mentioned in the commit message | 20:06 |
| *** patrickeast has joined #openstack-nova | 20:06 | |
| cfriesen | dansmith: mriedem: okay, fair enough. | 20:06 |
| mriedem | there have been changes to make the reboot flow less shitty with delays on the events, but those aren't cure alls | 20:06 |
| dansmith | so many that some of us wake up in the middle of the night screaming | 20:06 |
| mriedem | but in the case that your compute service reboots with the host and libvirt is spewing events at the same time, you can get into some odd states that you have to recover from manually when the host comes back up | 20:07 |
| cfriesen | so the takeaway is that if you care about the state of a specific instance then you better have some external monitoring of that instance and not rely on nova to have accurate state? | 20:07 |
| dansmith | no | 20:08 |
| dansmith | the takeaway is that if this is biting you in the ass and you'd rather have ten-minute resolution on your instance states, turn this on | 20:08 |
| mriedem | yes that ^ | 20:08 |
| *** otter768 has quit IRC | 20:09 | |
| dansmith | if you'd rather have high resolution state updates and it's not biting you in the ass, or you're into kinky shit like that, turn it off | 20:09 |
| dansmith | it's pretty simple. | 20:10 |
| mriedem | hyper-v ci was saying they hit this and i think cburgess was also saying he'd seen it possibly | 20:10 |
| mriedem | i know we had a case internally with it | 20:10 |
| mriedem | stress testing reboot with kvm | 20:11 |
| *** matrohon has joined #openstack-nova | 20:11 | |
| *** achanda has quit IRC | 20:12 | |
| *** annegentle has joined #openstack-nova | 20:14 | |
| *** fawadkhaliq has quit IRC | 20:15 | |
| dansmith | sdague: my grenade patch is in the gate, does that mean the branching thing is good now? | 20:17 |
| sdague | dansmith: no | 20:18 |
| dansmith | okay | 20:18 |
| dansmith | thought that had to wait until it was, but I guess it's just not hit until it is? | 20:18 |
| sdague | oh, I should have Depends-On it | 20:18 |
| dansmith | sdague: really trying not to read that as "I should have diapers on" | 20:19 |
| sdague | dansmith: yeh, it was just not running | 20:19 |
| dansmith | okay | 20:20 |
| sdague | so... I just updated it with a Depends-On for https://review.openstack.org/#/c/175391/ | 20:20 |
| *** ildikov has joined #openstack-nova | 20:20 | |
| sdague | which means it .... should .... actually be testing correctly | 20:20 |
| dansmith | okay sweet | 20:20 |
| dansmith | oh? | 20:20 |
| dansmith | so I can recheck my nova patch that depends-on that? | 20:20 |
| sdague | yes | 20:21 |
| sdague | I think the stack should work now | 20:21 |
| dansmith | woot | 20:21 |
| sdague | it's a convoluted stack as it requires keystone fixes as well | 20:21 |
| dansmith | I mostly just want to see it run and be green, or fix things if it's broken | 20:21 |
| dansmith | holding up all the other DB stuff is a power trip, so I'm in no rush :P | 20:21 |
| sdague | http://dl.dropbox.com/u/6514884/screenshot_225.png | 20:22 |
| dansmith | cool | 20:22 |
| dansmith | depends-on rocks | 20:22 |
| *** joefides has joined #openstack-nova | 20:25 | |
| *** annegentle has quit IRC | 20:28 | |
| *** annegentle has joined #openstack-nova | 20:29 | |
| *** VW_ has quit IRC | 20:30 | |
| *** rajesht has joined #openstack-nova | 20:31 | |
| *** sandywalsh has quit IRC | 20:33 | |
| *** annashen has joined #openstack-nova | 20:34 | |
| *** sandywalsh has joined #openstack-nova | 20:34 | |
| *** rajesht has quit IRC | 20:36 | |
| *** ijw has quit IRC | 20:36 | |
| *** annashen has quit IRC | 20:39 | |
| *** joefides_ has joined #openstack-nova | 20:39 | |
| *** joefides has quit IRC | 20:42 | |
| mriedem | damn no jaypipes | 20:44 |
| mriedem | no alex_xu | 20:44 |
| mriedem | bauzas: you around? | 20:44 |
| bauzas | mriedem: yup | 20:44 |
| mriedem | is my understanding of tihs correct? https://review.openstack.org/#/c/160089/10//COMMIT_MSG | 20:44 |
| *** ajayaa has quit IRC | 20:44 | |
| bauzas | lemme take more context | 20:44 |
| mriedem | we have an api policy regression in the os-hypervisors api but it's fixed on master with the api policy bp, which we can't backport | 20:44 |
| bauzas | mriedem: oh | 20:45 |
| mriedem | alex_xu: suggested in the bug report to have a separate fix which elevates the context and we backport that to stable/kilo (or just make it on stable/kilo) | 20:45 |
| bauzas | mriedem: yeah that's problematic | 20:45 |
| bauzas | mriedem: I was actually wondering this morning (my time) how it could be acceptable to provide a change not cherry-picked from master | 20:46 |
| mriedem | dansmith: that's allowed in exceptional cases right? ^ | 20:46 |
| bauzas | mriedem: when I just realized all of that is just... bad | 20:46 |
| *** ArxCruz has quit IRC | 20:46 | |
| bauzas | mriedem: because I think all of this is just making operators confused | 20:47 |
| dansmith | mriedem: yeah, I think sometimes we have to do that | 20:47 |
| bauzas | so the question is : could we just remove the context decorator in the stable branches ? | 20:48 |
| mriedem | given https://review.openstack.org/#/c/160089/ isn't merged yet, we could sneak a fix in though | 20:48 |
| mriedem | and backport it | 20:48 |
| *** matrohon has quit IRC | 20:48 | |
| mriedem | bauzas: we're not going to backport https://review.openstack.org/#/c/160089/ | 20:48 |
| mriedem | but i think we could slip a change in underneath that which fixes the regression by passing the elevated context to service_get_by_compute_host, | 20:49 |
| mriedem | then rebase https://review.openstack.org/#/c/160089/ on top of that | 20:49 |
| *** ArxCruz has joined #openstack-nova | 20:49 | |
| bauzas | mmm the big thing is https://bugs.launchpad.net/nova/+bug/1447164 IMHO | 20:49 |
| openstack | Launchpad bug 1447164 in OpenStack Compute (nova) "require_admin_context() does not account for policy.json rulesets" [Medium,New] | 20:49 |
| edmondsw | the more limited bug is https://bugs.launchpad.net/nova/+bug/1447084 | 20:49 |
| openstack | Launchpad bug 1447084 in OpenStack Compute (nova) "view hypervisor details should be controlled by policy.json" [High,Confirmed] - Assigned to Divya K Konoor (dikonoor) | 20:49 |
| edmondsw | the latter being a regression issue | 20:50 |
| bauzas | mriedem: edmondsw: yeah I tried to just take this in http://lists.openstack.org/pipermail/openstack-dev/2015-April/062282.html | 20:50 |
| bauzas | mriedem: sooooo | 20:50 |
| bauzas | mriedem: giving all of that and how crappy it is, we can just provide like an hotfix bug for the stable branch - with a big fat comment explaining why - but I would be in favor of discussing that whole crap and get rid of the context check for *all* Nova | 20:51 |
| mriedem | bauzas: but it doesn't seem we have to just do a single stable-only change, | 20:52 |
| bauzas | mriedem: both can be done, sure, given the former is acceptable by the rules | 20:52 |
| mriedem | we can fix on master and backport, then relay current change on top of that | 20:52 |
| bauzas | mriedem: well, the bug edmondsw mentions is actually a regression | 20:52 |
| mriedem | which would keep the git history consistent | 20:52 |
| mriedem | yeah, we fix the regression and backport | 20:52 |
| mriedem | then rebase https://review.openstack.org/#/c/160089/ on top of it | 20:53 |
| bauzas | mriedem: right, sounds a good option to me - at least for the regression only | 20:53 |
| *** achanda has joined #openstack-nova | 20:53 | |
| bauzas | mriedem: and consider removing those context checks in master only - for both v2.0 and v2.1 IMHO | 20:53 |
| mriedem | hell, https://review.openstack.org/#/c/160089/ might not even matter | 20:54 |
| mriedem | we could just patch the v2 hypervisors api to pass admin context to service_get_by_compute_host and backport that, regardless of https://review.openstack.org/#/c/160089/ - but it would weird... | 20:54 |
| mriedem | *be weird | 20:54 |
| bauzas | yeah... | 20:55 |
| bauzas | by adding a new method to the API, I just walked the left foot in the ... | 20:55 |
| mriedem | 'we have to elevate the context b/c of kilo code, don't ask why we're doing this on master' :) | 20:55 |
| mriedem | well, we should have better end to end testing with non-admin rules in the policy or something | 20:56 |
| mriedem | try hitting these things without and admin context and see what blows up | 20:56 |
| *** dboik_ has quit IRC | 20:56 | |
| *** neelashah1 has quit IRC | 20:56 | |
| bauzas | mriedem: https://bugs.launchpad.net/nova/+bug/1447164 is just up to you for your pleasure :) | 20:57 |
| openstack | Launchpad bug 1447164 in OpenStack Compute (nova) "require_admin_context() does not account for policy.json rulesets" [Medium,New] | 20:57 |
| bauzas | mriedem: my take is just to get rid of it, but that doesn't fix the regression problem in stable | 20:58 |
| mriedem | bauzas: yeah, i think the require_admin_context() was just made to keep the same behavior/restrictions in v2 api for better or worse | 20:58 |
| bauzas | mriedem: yeah but that... sucks ? | 20:58 |
| mriedem | your argument is we don't need that anymore | 20:58 |
| mriedem | i don't disagree... | 20:58 |
| bauzas | yup | 20:58 |
| *** Marga_ has joined #openstack-nova | 20:59 | |
| mriedem | i guess as long as the default rules on those v2 apis which are doing that check default to admin, it shouldn't be an issue | 20:59 |
| *** raildo has quit IRC | 20:59 | |
| bauzas | mriedem: at least the issue has the merit to make me understand a very old issue I had downstream | 20:59 |
| mriedem | bauzas: yes we came across this back in grizzly | 21:00 |
| mriedem | we == ibm | 21:00 |
| bauzas | mriedem: the problem is that you can't assume that operators won't leave the rules to be admin-only | 21:00 |
| *** nkrinner has quit IRC | 21:00 | |
| bauzas | mriedem: we have the same problem with evacuate for example | 21:00 |
| bauzas | mriedem: lemme show you the old bug I have | 21:00 |
| mriedem | https://review.openstack.org/#/c/27023/ | 21:01 |
| mriedem | showing first ^ | 21:01 |
| bauzas | https://bugzilla.redhat.com/show_bug.cgi?id=1164271 | 21:01 |
| openstack | bugzilla.redhat.com bug 1164271 in openstack-nova "nova evacuate not allowed for non-admin even after policy change" [Medium,Assigned] - Assigned to sbauza | 21:01 |
| mriedem | https://bugs.launchpad.net/nova/+bug/1168488 | 21:02 |
| openstack | Launchpad bug 1168488 in OpenStack Compute (nova) "host-list policy irrelevant" [Medium,Won't fix] | 21:02 |
| bauzas | eh 27K was just 1yr ago... | 21:02 |
| bauzas | oh, sylvain, you have to go back to school - 2 years... | 21:02 |
| mriedem | yeah grizzly https://review.openstack.org/#/c/26972/ | 21:02 |
| *** baoli has quit IRC | 21:03 | |
| *** Marga_ has quit IRC | 21:03 | |
| *** VW_ has joined #openstack-nova | 21:03 | |
| bauzas | mriedem: so we are going towards the right direction with alex_xu's patches | 21:03 |
| bauzas | mriedem: because all the DB checks are just insane how | 21:04 |
| bauzas | now evne | 21:04 |
| bauzas | even even | 21:04 |
| bauzas | mriedem: but we just need to make sure that we only trust the policy file with the authorizers, period. :) | 21:04 |
| *** Marga_ has joined #openstack-nova | 21:07 | |
| *** jichen has quit IRC | 21:08 | |
| *** vladikr has quit IRC | 21:10 | |
| *** dboik has joined #openstack-nova | 21:12 | |
| *** dprince has quit IRC | 21:13 | |
| *** rdo has quit IRC | 21:19 | |
| *** oro has joined #openstack-nova | 21:20 | |
| *** rdo has joined #openstack-nova | 21:21 | |
| melwitt | looking at this bug https://bugs.launchpad.net/python-openstackclient/+bug/1434172, does anyone know if security group create is supposed to have "description" as optional? the documentation says it's optional, but there is a unit test that shows it's supposed to be required https://github.com/openstack/nova/blob/master/nova/tests/unit/api/openstack/compute/contrib/test_security_groups.py#L171-L179 | 21:23 |
| openstack | Launchpad bug 1434172 in python-openstackclient "security group create errors without description" [Low,In progress] - Assigned to Dean Troyer (dtroyer) | 21:23 |
| *** mjturek1 has quit IRC | 21:24 | |
| melwitt | so either it's a nova bug or a documentation bug | 21:24 |
| redbeard | could someone point out how exactly instances are _choosen_ for the periodic task which is responsible for "Updated the network info_cache for instance _heal_instance_info_cache "? is it like a round-robin? | 21:26 |
| mriedem | melwitt: looks optional in the v2 api: http://git.openstack.org/cgit/openstack/nova/tree/nova/api/openstack/compute/contrib/security_groups.py#n197 | 21:26 |
| melwitt | mriedem: thanks. that's what I was mentioning, that the docs exactly contradict the code, and I'm not sure which it's supposed to be :( | 21:27 |
| mriedem | melwitt: also optional in the data model http://git.openstack.org/cgit/openstack/nova/tree/nova/db/sqlalchemy/models.py#n683 | 21:27 |
| melwitt | mriedem: ah! that helps. | 21:27 |
| mriedem | there is some validation thing in the security group api | 21:27 |
| mriedem | http://git.openstack.org/cgit/openstack/nova/tree/nova/compute/api.py#n3838 | 21:28 |
| *** thangp has quit IRC | 21:29 | |
| mriedem | melwitt: so i guess it's required for the nova impl | 21:29 |
| mriedem | because http://git.openstack.org/cgit/openstack/nova/tree/nova/api/openstack/compute/contrib/security_groups.py#n201 | 21:29 |
| *** openstackgerrit has quit IRC | 21:29 | |
| *** openstackgerrit has joined #openstack-nova | 21:30 | |
| mriedem | redbeard: it gets all instances, pushes them on a stack, and for each iteration in the task it pops one off and refreshes it's nw info cache | 21:30 |
| mriedem | once the stack is empty, it gets all instances again | 21:30 |
| *** flwang has joined #openstack-nova | 21:31 | |
| redbeard | ah ok. so setting down the interval aren't a good idea to re-refresh the cache for a newly launched instance | 21:31 |
| *** dboik has quit IRC | 21:31 | |
| flwang | mriedem: ping, re nova backup | 21:31 |
| flwang | mriedem: i need your help on the rpc version, https://review.openstack.org/#/c/164494/ | 21:32 |
| redbeard | mriedem, any chance to trigger a refresh for 1 specific instance manually? | 21:33 |
| mriedem | redbeard: i don't think so | 21:34 |
| *** yamahata has joined #openstack-nova | 21:34 | |
| *** neelashah has joined #openstack-nova | 21:36 | |
| *** dboik has joined #openstack-nova | 21:37 | |
| melwitt | mriedem: thanks for the help. I'm guessing it's probably old error in the nova impl then. I'll update the bug | 21:37 |
| *** igordcard has joined #openstack-nova | 21:37 | |
| *** dboik has quit IRC | 21:38 | |
| *** dboik has joined #openstack-nova | 21:38 | |
| *** busterswt has quit IRC | 21:40 | |
| *** tonytan4ever has quit IRC | 21:41 | |
| *** lucas-dinner has quit IRC | 21:41 | |
| *** sdake has quit IRC | 21:44 | |
| *** annegentle has quit IRC | 21:45 | |
| *** snirow has joined #openstack-nova | 21:45 | |
| *** yamahata has quit IRC | 21:46 | |
| *** dprince has joined #openstack-nova | 21:46 | |
| *** snirow has joined #openstack-nova | 21:46 | |
| *** snirow has quit IRC | 21:47 | |
| *** edmondsw has quit IRC | 21:47 | |
| *** snirow has joined #openstack-nova | 21:49 | |
| cfriesen | dhellmann: I've been going over the logging scenario I saw earlier (strace output at http://paste.openstack.org/show/205143/) and from what I can tell it's coming from a line in amqpdriver.py that looks like this: LOG.debug('MSG_ID is %s', msg_id) | 21:50 |
| *** snirow has quit IRC | 21:50 | |
| cfriesen | dhellmann: the wierd thing is that it's calling sendto() as though the socket is connected (with null address and 0 addrlen), but since our syslog only listens to /dev/log as DGRAM it could never have been connected. | 21:51 |
| *** dboik has quit IRC | 21:55 | |
| *** dboik has joined #openstack-nova | 21:55 | |
| *** ndipanov has quit IRC | 21:56 | |
| *** dprince has quit IRC | 21:58 | |
| *** burt has quit IRC | 21:59 | |
| mriedem | lbragstad: entering mastodon mode | 22:00 |
| mriedem | flwang: hit me up tomorrow or leave a comment/question in the review | 22:00 |
| openstackgerrit | melanie witt proposed openstack/nova: WIP: send object to cells instance_update_at_top https://review.openstack.org/176518 | 22:01 |
| lbragstad | mriedem: don't work angry | 22:01 |
| mriedem | i ride dirty and i work angry | 22:01 |
| lbragstad | mriedem: you could put that on the back of a cut | 22:01 |
| *** isd has quit IRC | 22:04 | |
| *** andreykurilin__ has joined #openstack-nova | 22:05 | |
| *** otter768 has joined #openstack-nova | 22:05 | |
| *** lpetrut has quit IRC | 22:09 | |
| *** otter768 has quit IRC | 22:09 | |
| *** signed8b_ has joined #openstack-nova | 22:09 | |
| *** openstackgerrit has quit IRC | 22:11 | |
| *** openstackgerrit has joined #openstack-nova | 22:11 | |
| *** signed8bit has quit IRC | 22:12 | |
| cfriesen | dhellmann: I suspect we hit this python bug: http://bugs.python.org/issue15179 | 22:14 |
| *** dboik has quit IRC | 22:14 | |
| cfriesen | dhellmann: so no issues with oslo code, just buggy python | 22:14 |
| flwang | mriedem: ok, thanks a lot | 22:19 |
| *** jaypipes has joined #openstack-nova | 22:20 | |
| *** erlon has quit IRC | 22:21 | |
| *** annegentle has joined #openstack-nova | 22:22 | |
| *** neelashah has quit IRC | 22:23 | |
| *** ArxCruz has quit IRC | 22:23 | |
| *** GeekyCraig has joined #openstack-nova | 22:24 | |
| *** VW_ has quit IRC | 22:24 | |
| GeekyCraig | I am doing my first multi-node install, and I am having big problems with Nova, in fact, the Nova directory only contained rootwrap.d, and n other files. I made the nova.conf, but now I am getting issues in regards to the nova-api.conf file, which is not in the directory. Any help? | 22:26 |
| mriedem | GeekyCraig: please see channel topic | 22:26 |
| GeekyCraig | oops | 22:26 |
| GeekyCraig | my bad | 22:26 |
| mriedem | np | 22:27 |
| *** VW_ has joined #openstack-nova | 22:28 | |
| *** sdake has joined #openstack-nova | 22:31 | |
| *** BadCub is now known as BadCub_Hiding | 22:31 | |
| mriedem | hrm, yum exits with 0 even if one of the packages in a list isn't installed (not found)? | 22:32 |
| mriedem | that seems bad | 22:32 |
| *** artom_ has joined #openstack-nova | 22:34 | |
| *** artom has quit IRC | 22:34 | |
| clarkb | mriedem: yup | 22:36 |
| mriedem | that sucks | 22:36 |
| clarkb | mriedem: iirc its something like if you install >1 package on the command line (so not transitive things) and at least one succeeds then its return 0 | 22:36 |
| clarkb | mriedem: so you can work around it by installing one package at a time \o/ | 22:36 |
| mriedem | yeah fun | 22:36 |
| mriedem | loops | 22:36 |
| mriedem | push the blame to the caller | 22:37 |
| *** annegentle has quit IRC | 22:45 | |
| *** shakamunyi has quit IRC | 22:46 | |
| *** apmelton is now known as apmelton_ | 22:50 | |
| *** apmelton_ is now known as apmelton_away | 22:50 | |
| *** nelsnelson has quit IRC | 22:51 | |
| *** ZZelle_ has quit IRC | 22:54 | |
| *** penick has quit IRC | 22:55 | |
| *** melwitt has quit IRC | 22:57 | |
| *** andreykurilin__ has quit IRC | 22:59 | |
| *** Hosam has quit IRC | 23:00 | |
| *** tjones2 has joined #openstack-nova | 23:04 | |
| *** tjones1 has quit IRC | 23:05 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: neutron: log hypervisor_macs before raising PortNotUsable https://review.openstack.org/175458 | 23:07 |
| *** signed8b_ is now known as signed8bit_ZZZzz | 23:08 | |
| *** eharney has quit IRC | 23:09 | |
| *** VW_ has quit IRC | 23:12 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: Cleanup of compute api reboot method https://review.openstack.org/127427 | 23:13 |
| *** thedodd has quit IRC | 23:13 | |
| *** fawadkhaliq has joined #openstack-nova | 23:15 | |
| *** ganso_ has quit IRC | 23:17 | |
| *** fawadkhaliq has quit IRC | 23:20 | |
| *** david-lyle has joined #openstack-nova | 23:24 | |
| tjones2 | dansmith: ^^ fixed your comments | 23:24 |
| tjones2 | or addressed them | 23:24 |
| *** markvoelker has quit IRC | 23:26 | |
| *** mriedem is now known as mriedem_away | 23:26 | |
| mikal | . | 23:26 |
| *** amotoki has quit IRC | 23:27 | |
| lifeless | .. | 23:27 |
| *** rushil has quit IRC | 23:30 | |
| *** GeekyCraig has quit IRC | 23:32 | |
| *** hemna is now known as hemnafk | 23:35 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: Cleanup of compute api reboot method https://review.openstack.org/127427 | 23:35 |
| *** Marga_ has quit IRC | 23:36 | |
| *** artom_ has quit IRC | 23:36 | |
| *** annashen has joined #openstack-nova | 23:36 | |
| *** Marga_ has joined #openstack-nova | 23:37 | |
| *** sdake has quit IRC | 23:39 | |
| jroll | any reason we don't send instance.system_metadata for update notifications? | 23:39 |
| *** annashen has quit IRC | 23:41 | |
| *** claudiub has quit IRC | 23:42 | |
| *** Longgeek has quit IRC | 23:42 | |
| mikal | tjones2: hey, I only just noticed I got your name wrong on https://review.openstack.org/#/c/127427/ -- sorry about that | 23:44 |
| tjones2 | haha - no worries :-) | 23:46 |
| mikal | I will do an apology code review right now to make up for it | 23:46 |
| tjones2 | nice! get my name wrong all you want then :-) | 23:46 |
| *** jecarey has quit IRC | 23:47 | |
| dansmith | jhesketh: around? | 23:49 |
| mikal | Heh | 23:50 |
| jhesketh | dansmith: yep | 23:54 |
| dansmith | jhesketh: so, excellent job testing my code! | 23:54 |
| dansmith | jhesketh: did you see the fix for the instances without extra rows? | 23:54 |
| jhesketh | Not yet. I'll take a look in a bit :-) | 23:55 |
| jhesketh | I read your email and have some answers, but I'll look at the new patch first | 23:55 |
| dansmith | jhesketh: cool, it was #1 in the gate until we just realized things are broooken | 23:55 |
| dansmith | in the gate, not the patch | 23:55 |
| dansmith | jhesketh: cool, I'm about to stop for evening and food | 23:55 |
| jhesketh | If you want to hold off until I can run it over our datasets I may be able to provide some feedback | 23:56 |
| *** ijw has joined #openstack-nova | 23:56 | |
| dansmith | jhesketh: but regardless, this is a thing we really want/need to do in kilo, and so backporting things to make it work is reasonable to achieve that goal | 23:56 |
| dansmith | jhesketh: well, unless they figure out the gate thing soonly, it's on hold | 23:56 |
| *** isd has joined #openstack-nova | 23:56 | |
| jhesketh | Sure, I guess that's a discussion for Nova :-) | 23:57 |
| dansmith | jhesketh: I should have pushed to get T-H running this migration mid-kilo when we merged the stuff, but, what can I say? I suck. | 23:57 |
| *** tsekiyam_ has joined #openstack-nova | 23:57 | |
| jhesketh | If I knew about it I could have taken a look, but sadly I struggle to keep up with what happens in nova land sorry | 23:58 |
| dansmith | it's not your fault, for sure | 23:58 |
| dansmith | now is about the earliest we could have had this running in grenade to simulate a real scenario though, so it just kinda is what it is, timing-wise | 23:58 |
| *** tsekiyama has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
