*** dannywilson has quit IRC | 00:01 | |
*** ijw has quit IRC | 00:02 | |
*** ijw has joined #openstack-nova | 00:03 | |
gmann_ | mtreinish: in later case Devstack have to run Juno jobs runs for v2 only | 00:03 |
---|---|---|
*** Sukhdev has quit IRC | 00:03 | |
*** mutoulbj has quit IRC | 00:03 | |
*** mtanino has quit IRC | 00:13 | |
*** tsekiyama has quit IRC | 00:15 | |
mtreinish | gmann_: well I think changing the default in tempest is a bit risky because of existing deployments | 00:17 |
mtreinish | even if they're kilo or newer there is no guarantee they're running with 2.1 and things used to just work for them | 00:17 |
mtreinish | gmann_: what I'd really like to see is making 2.1 the compute endpoint type by default in the catalog | 00:17 |
mtreinish | and make the 2.0 endpoint the special case one | 00:18 |
gmann_ | mtreinish: yes, Tempest one might be risky, | 00:19 |
gmann_ | mtreinish: I will update the patches (in current proposed patch it plays with api-paste.ini for making v21 default) for making v21 default through catalog side. | 00:21 |
gmann_ | and see how it goes, i pending for long time and i remember you have mentioned many time :) | 00:21 |
gmann_ | and i forgot to follow up on those, Sorry for that. I will do those by today or max by tomorrow | 00:22 |
mtreinish | no worries, it's not like it's super critical. It'll just be good to get it sorted so it's a little less confusing for people | 00:23 |
*** salv-orlando has quit IRC | 00:24 | |
gmann_ | mtreinish: yea | 00:25 |
*** salv-orlando has joined #openstack-nova | 00:26 | |
*** achanda_ has quit IRC | 00:26 | |
*** ijw_ has joined #openstack-nova | 00:31 | |
*** annashen has joined #openstack-nova | 00:32 | |
*** vilobhmm2 has quit IRC | 00:32 | |
*** annashen has quit IRC | 00:33 | |
*** salv-orlando has quit IRC | 00:34 | |
*** ijw has quit IRC | 00:34 | |
*** mdrabe has joined #openstack-nova | 00:35 | |
*** artom has quit IRC | 00:36 | |
*** ujjain has quit IRC | 00:36 | |
*** ujjain has joined #openstack-nova | 00:37 | |
*** ujjain has quit IRC | 00:37 | |
*** ujjain has joined #openstack-nova | 00:37 | |
*** andreykurilin__ has quit IRC | 00:37 | |
n0ano | mriedem_away, ping me when you're back on-line | 00:41 |
*** pixelbeat_ has joined #openstack-nova | 00:42 | |
*** yonglihe has joined #openstack-nova | 00:43 | |
*** newbie_2 has quit IRC | 00:49 | |
*** angdraug has joined #openstack-nova | 00:49 | |
oomichi | alex_xu_: hi, morning | 00:50 |
alex_xu_ | oomichi: hi, morning | 00:50 |
*** ijw_ has quit IRC | 00:50 | |
*** ijw has joined #openstack-nova | 00:51 | |
oomichi | alex_xu_: a little question about api-wg: do you know current status of approved but not-merged patches? | 00:51 |
*** signed8bit_ZZZzz is now known as signed8bit | 00:51 | |
oomichi | alex_xu_: example is https://review.openstack.org/#/c/185180/ | 00:51 |
*** davideagnello has quit IRC | 00:52 | |
oomichi | alex_xu_: the final deadline of these patches are over already, but we cannot merge them yet | 00:52 |
alex_xu_ | oomichi: I confuse on that also, but I think it means already merge | 00:52 |
*** baoli has joined #openstack-nova | 00:52 | |
*** newbie_2 has joined #openstack-nova | 00:53 | |
oomichi | alex_xu_: yeah, I also have the same feeling. do we have an irc channel for api-wg? | 00:53 |
alex_xu_ | oomichi: it's worth ask in api-wg channel | 00:53 |
*** angdraug has quit IRC | 00:53 | |
alex_xu_ | oomichi: #openstack-api | 00:53 |
*** davideagnello has joined #openstack-nova | 00:54 | |
oomichi | alex_xu_: thanks, I will try it :) | 00:54 |
alex_xu_ | oomichi: np | 00:54 |
*** newbie_2 has quit IRC | 01:00 | |
*** Longgeek has joined #openstack-nova | 01:00 | |
*** signed8bit is now known as signed8bit_ZZZzz | 01:02 | |
*** mutoulbj has joined #openstack-nova | 01:03 | |
*** ssurana has quit IRC | 01:03 | |
anish | mriedem_away: dansmith failed pep8 before, can I get the +2 again https://review.openstack.org/#/c/193414/ | 01:04 |
*** ssurana has joined #openstack-nova | 01:04 | |
*** ssurana has quit IRC | 01:08 | |
*** mutoulbj has quit IRC | 01:09 | |
*** casusbelli has joined #openstack-nova | 01:12 | |
*** liusheng has joined #openstack-nova | 01:12 | |
*** casusbel_ has quit IRC | 01:14 | |
*** mdrabe has quit IRC | 01:14 | |
*** openstack has joined #openstack-nova | 01:24 | |
*** yingjun has joined #openstack-nova | 01:25 | |
*** annegentle has quit IRC | 01:26 | |
*** yamahata has quit IRC | 01:26 | |
*** asettle has joined #openstack-nova | 01:27 | |
*** raildo has quit IRC | 01:30 | |
*** ijw has quit IRC | 01:32 | |
*** mestery_ is now known as mestery | 01:34 | |
*** patrickeast has quit IRC | 01:35 | |
*** ijw has joined #openstack-nova | 01:36 | |
*** baoli_ has joined #openstack-nova | 01:36 | |
*** annegentle has joined #openstack-nova | 01:37 | |
*** jdandrea has quit IRC | 01:38 | |
*** tpatil has quit IRC | 01:38 | |
*** baoli has quit IRC | 01:38 | |
*** signed8bit_ZZZzz is now known as signed8bit | 01:41 | |
*** annegentle has quit IRC | 01:42 | |
*** kaisers1 has quit IRC | 01:44 | |
*** ijw has quit IRC | 01:49 | |
*** park_ has joined #openstack-nova | 01:50 | |
*** rdo has quit IRC | 01:51 | |
*** signed8bit is now known as signed8bit_ZZZzz | 01:52 | |
*** IanGovett has joined #openstack-nova | 01:54 | |
*** signed8bit_ZZZzz is now known as signed8bit | 01:56 | |
*** MentalRay__ has joined #openstack-nova | 01:57 | |
*** dims_ has quit IRC | 01:58 | |
*** yingjun has quit IRC | 01:58 | |
*** zul has joined #openstack-nova | 02:00 | |
*** IanGovett has quit IRC | 02:01 | |
*** doude has quit IRC | 02:04 | |
*** dave-mccowan has joined #openstack-nova | 02:05 | |
*** signed8bit has quit IRC | 02:06 | |
*** doude has joined #openstack-nova | 02:07 | |
*** doude has quit IRC | 02:07 | |
*** doude has joined #openstack-nova | 02:07 | |
*** annegentle has joined #openstack-nova | 02:10 | |
*** shuquan has joined #openstack-nova | 02:13 | |
*** annegentle has quit IRC | 02:13 | |
*** dims__ has joined #openstack-nova | 02:18 | |
*** markvoelker has quit IRC | 02:22 | |
*** baoli_ has quit IRC | 02:23 | |
*** ijw has joined #openstack-nova | 02:24 | |
*** wwriverrat has quit IRC | 02:25 | |
*** unicell has quit IRC | 02:26 | |
*** vilobhmm has joined #openstack-nova | 02:26 | |
*** yingjun has joined #openstack-nova | 02:26 | |
*** yingjun has quit IRC | 02:27 | |
*** yingjun has joined #openstack-nova | 02:27 | |
*** vilobhmm1 has joined #openstack-nova | 02:28 | |
*** jdurgin has quit IRC | 02:28 | |
*** yingjun_ has joined #openstack-nova | 02:29 | |
*** yingjun has quit IRC | 02:29 | |
*** Longgeek has joined #openstack-nova | 02:30 | |
*** dims__ has quit IRC | 02:30 | |
*** vilobhmm has quit IRC | 02:30 | |
*** bkopilov has quit IRC | 02:31 | |
*** jyuso1 has quit IRC | 02:32 | |
openstackgerrit | Jeffrey Zhang proposed openstack/nova: Support all_tenants search_opts for neutron https://review.openstack.org/187657 | 02:34 |
*** yamahata has joined #openstack-nova | 02:34 | |
*** baoli has joined #openstack-nova | 02:35 | |
*** Longgeek has quit IRC | 02:35 | |
*** yamahata has quit IRC | 02:36 | |
*** yamahata has joined #openstack-nova | 02:37 | |
*** diogogmt has joined #openstack-nova | 02:43 | |
*** sorrison_laptop has joined #openstack-nova | 02:45 | |
*** baoli has quit IRC | 02:47 | |
*** baoli has joined #openstack-nova | 02:49 | |
*** yamahata has quit IRC | 02:50 | |
*** hakimo_ has joined #openstack-nova | 02:53 | |
*** gryf_ has quit IRC | 02:54 | |
*** hakimo has quit IRC | 02:54 | |
*** gryf has joined #openstack-nova | 02:55 | |
*** ijw has quit IRC | 02:56 | |
*** dave-mccowan has quit IRC | 02:56 | |
*** shuquan has quit IRC | 03:00 | |
*** yingjun_ has quit IRC | 03:03 | |
*** andrewbogott_afk has quit IRC | 03:08 | |
*** otter768 has quit IRC | 03:08 | |
openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: [WIP] Use explicit greenthread.sleep instead of time.sleep https://review.openstack.org/202365 | 03:09 |
*** casusbelli has quit IRC | 03:11 | |
*** casusbelli has joined #openstack-nova | 03:11 | |
*** bradjones has quit IRC | 03:14 | |
*** tpeoples has joined #openstack-nova | 03:14 | |
*** casusbelli has quit IRC | 03:18 | |
*** casusbelli has joined #openstack-nova | 03:19 | |
*** bradjones has joined #openstack-nova | 03:19 | |
*** bradjones has quit IRC | 03:19 | |
*** bradjones has joined #openstack-nova | 03:19 | |
*** park_heijlong has quit IRC | 03:23 | |
*** annashen has joined #openstack-nova | 03:23 | |
*** annashen has quit IRC | 03:28 | |
*** pixelbeat_ has quit IRC | 03:28 | |
*** xyang1 has quit IRC | 03:30 | |
*** jyuso1 has joined #openstack-nova | 03:32 | |
*** baoli has quit IRC | 03:33 | |
*** markvoelker has joined #openstack-nova | 03:33 | |
*** dannywilson has joined #openstack-nova | 03:34 | |
*** markvoelker has quit IRC | 03:34 | |
*** markvoelker has joined #openstack-nova | 03:34 | |
*** deepthi has joined #openstack-nova | 03:35 | |
openstackgerrit | xing-yang proposed openstack/nova-specs: EMC ScaleIO Data Client Shim Libvirt Volume Driver https://review.openstack.org/181941 | 03:35 |
*** wwriverrat has joined #openstack-nova | 03:36 | |
*** markvoelker_ has joined #openstack-nova | 03:37 | |
*** markvoelker has quit IRC | 03:39 | |
*** y_sawai has joined #openstack-nova | 03:40 | |
*** gryf has quit IRC | 03:41 | |
*** Nikolay_St has quit IRC | 03:41 | |
*** unicell has joined #openstack-nova | 03:43 | |
*** briancurtin has quit IRC | 03:45 | |
*** scottda has quit IRC | 03:45 | |
*** wanghao has quit IRC | 03:45 | |
*** serverascode has quit IRC | 03:45 | |
*** zhiyan has quit IRC | 03:45 | |
*** angular_mike_ has quit IRC | 03:45 | |
*** rex_lee has quit IRC | 03:46 | |
*** y_sawai has quit IRC | 03:46 | |
*** h00327910_ has quit IRC | 03:46 | |
*** y_sawai has joined #openstack-nova | 03:46 | |
*** erlon has quit IRC | 03:46 | |
*** ramishra has quit IRC | 03:46 | |
*** boris-42 has quit IRC | 03:46 | |
*** y_sawai has quit IRC | 03:47 | |
*** aweeks has quit IRC | 03:47 | |
openstackgerrit | xing-yang proposed openstack/nova: Add os-brick based LibvirtVolumeDriver for ScaleIO https://review.openstack.org/194454 | 03:47 |
*** coolsvap|away is now known as coolsvap | 03:47 | |
openstackgerrit | Ren Qiaowei proposed openstack/nova: Remove unnecessary executable permission https://review.openstack.org/202376 | 03:48 |
*** unicell has quit IRC | 03:48 | |
*** unicell1 has joined #openstack-nova | 03:48 | |
*** wanghao has joined #openstack-nova | 03:48 | |
*** scottda has joined #openstack-nova | 03:48 | |
*** kaisers1 has joined #openstack-nova | 03:49 | |
*** garyk has quit IRC | 03:49 | |
*** dougwig has quit IRC | 03:50 | |
*** StevenK has quit IRC | 03:51 | |
*** freyes has quit IRC | 03:52 | |
*** dannywilson has quit IRC | 03:52 | |
*** dannywilson has joined #openstack-nova | 03:52 | |
openstackgerrit | xing-yang proposed openstack/nova-specs: EMC ScaleIO Data Client Shim Libvirt Volume Driver https://review.openstack.org/181941 | 03:53 |
*** zhiyan has joined #openstack-nova | 03:54 | |
*** StevenK has joined #openstack-nova | 03:55 | |
*** rajesht has joined #openstack-nova | 03:55 | |
*** briancurtin has joined #openstack-nova | 03:56 | |
*** serverascode has joined #openstack-nova | 03:58 | |
*** bkopilov has joined #openstack-nova | 03:59 | |
*** rex_lee has joined #openstack-nova | 04:02 | |
*** achanda has joined #openstack-nova | 04:02 | |
*** shuquan has joined #openstack-nova | 04:03 | |
openstackgerrit | wanghao proposed openstack/nova: Image/block device mapping combination is not valid https://review.openstack.org/171984 | 04:03 |
*** aweeks has joined #openstack-nova | 04:04 | |
*** oomichi has quit IRC | 04:05 | |
*** lyanchih has quit IRC | 04:06 | |
*** boris-42 has joined #openstack-nova | 04:08 | |
*** erlon has joined #openstack-nova | 04:08 | |
*** annashen has joined #openstack-nova | 04:08 | |
*** ijw has joined #openstack-nova | 04:10 | |
*** h00327910_ has joined #openstack-nova | 04:11 | |
*** hemna has joined #openstack-nova | 04:11 | |
*** kiran-r has joined #openstack-nova | 04:15 | |
*** shuquan has quit IRC | 04:15 | |
*** belmoreira has joined #openstack-nova | 04:15 | |
*** jwcroppe has joined #openstack-nova | 04:16 | |
*** Longgeek has joined #openstack-nova | 04:17 | |
*** tonytan4ever has quit IRC | 04:18 | |
*** angular_mike_ has joined #openstack-nova | 04:22 | |
*** Longgeek has quit IRC | 04:25 | |
*** Longgeek has joined #openstack-nova | 04:25 | |
*** ssurana has joined #openstack-nova | 04:27 | |
*** belmoreira has quit IRC | 04:28 | |
*** annashen has quit IRC | 04:28 | |
*** dougwig has joined #openstack-nova | 04:29 | |
*** ramishra has joined #openstack-nova | 04:29 | |
openstackgerrit | Zhenyu Zheng proposed openstack/nova: Resize: Return to initial state instead of ACTIVE https://review.openstack.org/202379 | 04:29 |
*** tonytan4ever has joined #openstack-nova | 04:29 | |
*** hemna has quit IRC | 04:30 | |
*** shyu has quit IRC | 04:32 | |
*** shadyabhi has quit IRC | 04:33 | |
openstackgerrit | Zhenyu Zheng proposed openstack/nova: Resize: Return to initial state instead of ACTIVE https://review.openstack.org/202379 | 04:34 |
openstackgerrit | Shiina, Hironori proposed openstack/nova: Add an API to inject an NMI to an instance https://review.openstack.org/202380 | 04:37 |
*** zzzeek has joined #openstack-nova | 04:37 | |
*** freyes has joined #openstack-nova | 04:40 | |
*** irenab has quit IRC | 04:40 | |
*** shyama has joined #openstack-nova | 04:41 | |
*** MentalRay__ has quit IRC | 04:41 | |
*** Nikolay_St has joined #openstack-nova | 04:42 | |
*** kiran-r has quit IRC | 04:47 | |
*** y_sawai has joined #openstack-nova | 04:47 | |
openstackgerrit | Ren Qiaowei proposed openstack/nova: Remove unnecessary executable permission https://review.openstack.org/202376 | 04:49 |
*** markvoelker has joined #openstack-nova | 04:49 | |
*** garyk has joined #openstack-nova | 04:51 | |
*** zzzeek has quit IRC | 04:52 | |
*** doude has quit IRC | 04:52 | |
*** y_sawai has quit IRC | 04:52 | |
*** markvoelker_ has quit IRC | 04:53 | |
*** ssurana has quit IRC | 04:53 | |
*** Jeffrey4l_ has quit IRC | 04:57 | |
*** Mike_D_laptop has joined #openstack-nova | 04:59 | |
*** tonytan4ever has quit IRC | 05:00 | |
*** SourabhP has joined #openstack-nova | 05:02 | |
*** Mike_D_laptop has quit IRC | 05:04 | |
*** otter768 has joined #openstack-nova | 05:09 | |
*** kiran-r has joined #openstack-nova | 05:09 | |
*** ssurana has joined #openstack-nova | 05:10 | |
*** Jeffrey4l_ has joined #openstack-nova | 05:10 | |
openstackgerrit | Feodor Tersin proposed openstack/nova: Fix FlavorDiskTooSmall boot fault for volume-backed instances https://review.openstack.org/196569 | 05:11 |
*** veena has joined #openstack-nova | 05:12 | |
*** otter768 has quit IRC | 05:14 | |
*** deepakcs has joined #openstack-nova | 05:15 | |
deepakcs | Hello, can any cores pls bless https://review.openstack.org/#/c/194538/4 (It already has a lot of +1s) | 05:16 |
*** _gryf has joined #openstack-nova | 05:20 | |
openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: [WIP] Use explicit greenthread.sleep instead of time.sleep https://review.openstack.org/202365 | 05:30 |
*** _kiran_ has joined #openstack-nova | 05:33 | |
*** asettle has quit IRC | 05:34 | |
*** kiranr has joined #openstack-nova | 05:35 | |
*** kiranr has quit IRC | 05:35 | |
*** lyanchih has joined #openstack-nova | 05:36 | |
_gryf | johnthetubaguy, hi | 05:37 |
*** kiran-r has quit IRC | 05:37 | |
*** _kiran_ has quit IRC | 05:39 | |
*** sudipto has joined #openstack-nova | 05:39 | |
*** SourabhP has quit IRC | 05:42 | |
*** shuquan has joined #openstack-nova | 05:42 | |
openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for instance_get_all_by_host_and_not_type https://review.openstack.org/202389 | 05:42 |
*** RuiChen has quit IRC | 05:42 | |
*** vishwana_ has joined #openstack-nova | 05:43 | |
*** RuiChen has joined #openstack-nova | 05:43 | |
*** moshele has joined #openstack-nova | 05:43 | |
*** irenab has joined #openstack-nova | 05:44 | |
deepakcs | johnthetubaguy: ^^ | 05:44 |
*** vishwanathj has quit IRC | 05:47 | |
*** y_sawai has joined #openstack-nova | 05:48 | |
*** y_sawai_ has joined #openstack-nova | 05:48 | |
*** ankit has joined #openstack-nova | 05:49 | |
*** shuquan has quit IRC | 05:49 | |
*** Murali has joined #openstack-nova | 05:50 | |
rajesht | mikal, alaski: you around ?? | 05:50 |
openstackgerrit | Ghanshyam Mann proposed openstack/nova: Show 'locked_by' information in server Show & List Detail https://review.openstack.org/202391 | 05:50 |
*** itzikb_ has joined #openstack-nova | 05:52 | |
*** y_sawai has quit IRC | 05:52 | |
*** shuquan has joined #openstack-nova | 05:53 | |
*** wanghua has joined #openstack-nova | 05:53 | |
*** dboik has joined #openstack-nova | 05:57 | |
*** shuquan has quit IRC | 05:57 | |
*** shuquan has joined #openstack-nova | 05:58 | |
*** dboik has quit IRC | 06:02 | |
*** jmccrory has quit IRC | 06:02 | |
*** nkrinner_afk is now known as nkrinner | 06:07 | |
*** shuquan has quit IRC | 06:08 | |
*** jmccrory has joined #openstack-nova | 06:09 | |
*** shuquan has joined #openstack-nova | 06:10 | |
*** Sukhdev has joined #openstack-nova | 06:11 | |
*** scheuran has joined #openstack-nova | 06:11 | |
*** cfriesen has quit IRC | 06:12 | |
*** Nikolay_St has quit IRC | 06:12 | |
*** annashen has joined #openstack-nova | 06:12 | |
*** y_sawai_ has quit IRC | 06:13 | |
*** moshele has quit IRC | 06:13 | |
*** ildikov has quit IRC | 06:18 | |
*** mpaolino has joined #openstack-nova | 06:19 | |
*** mpaolino has quit IRC | 06:25 | |
*** belmoreira has joined #openstack-nova | 06:25 | |
*** dannywilson has quit IRC | 06:26 | |
*** asettle has joined #openstack-nova | 06:26 | |
openstackgerrit | Abhishek Kekane proposed openstack/nova: libvirt: Kill rsync/scp processes before deleting instance https://review.openstack.org/192986 | 06:28 |
*** vilobhmm1 has quit IRC | 06:31 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex https://review.openstack.org/197213 | 06:31 |
*** jichen has joined #openstack-nova | 06:37 | |
*** annashen has quit IRC | 06:37 | |
*** jichen has quit IRC | 06:38 | |
*** shyu has joined #openstack-nova | 06:39 | |
*** jichen has joined #openstack-nova | 06:39 | |
*** abhishekk has joined #openstack-nova | 06:39 | |
*** shuquan has quit IRC | 06:43 | |
*** sahid has joined #openstack-nova | 06:43 | |
*** asettle has quit IRC | 06:45 | |
*** Sukhdev has quit IRC | 06:45 | |
*** oro_ has joined #openstack-nova | 06:47 | |
*** oro has joined #openstack-nova | 06:47 | |
*** shuquan has joined #openstack-nova | 06:47 | |
*** abhishekk has quit IRC | 06:49 | |
*** jdurgin has joined #openstack-nova | 06:50 | |
*** jdurgin has quit IRC | 06:50 | |
*** Longgeek has quit IRC | 06:51 | |
openstackgerrit | Sean McCully proposed openstack/nova: Creates disk.fallocate_test https://review.openstack.org/188249 | 06:52 |
*** Longgeek has joined #openstack-nova | 06:52 | |
*** Kevin_Zheng has quit IRC | 06:55 | |
*** Jhelf has joined #openstack-nova | 06:58 | |
*** alex_klimov has joined #openstack-nova | 06:59 | |
*** moshele has joined #openstack-nova | 07:03 | |
*** ildikov has joined #openstack-nova | 07:04 | |
*** markus_z has joined #openstack-nova | 07:05 | |
*** bkopilov has quit IRC | 07:08 | |
*** otter768 has joined #openstack-nova | 07:10 | |
*** rdopiera has joined #openstack-nova | 07:10 | |
*** andrearosa has joined #openstack-nova | 07:11 | |
*** otter768 has quit IRC | 07:15 | |
*** Jeffrey4l_ is now known as Jeffrey4l | 07:17 | |
*** romainh has joined #openstack-nova | 07:20 | |
*** lyanchih has quit IRC | 07:21 | |
*** Nikolay_St has joined #openstack-nova | 07:21 | |
*** bkopilov has joined #openstack-nova | 07:21 | |
*** Jeffrey4l has quit IRC | 07:22 | |
*** lyanchih has joined #openstack-nova | 07:22 | |
*** foexle has joined #openstack-nova | 07:27 | |
*** lan has quit IRC | 07:33 | |
*** lpetrut has joined #openstack-nova | 07:34 | |
*** Jeffrey4l has joined #openstack-nova | 07:35 | |
*** sorrison_laptop has quit IRC | 07:37 | |
*** Maike has joined #openstack-nova | 07:38 | |
*** jlanoux has joined #openstack-nova | 07:41 | |
*** pixelbeat_ has joined #openstack-nova | 07:42 | |
*** achanda has quit IRC | 07:46 | |
*** eglynn-officeafk has quit IRC | 07:50 | |
*** johnma has quit IRC | 07:52 | |
*** oro_ has quit IRC | 07:55 | |
*** oro has quit IRC | 07:55 | |
openstackgerrit | heijlong proposed openstack/nova: Add uuid into SecurityGroup object https://review.openstack.org/201970 | 07:56 |
*** lucasagomes has joined #openstack-nova | 08:00 | |
*** pixelbeat_ has quit IRC | 08:01 | |
*** markvoelker has quit IRC | 08:04 | |
*** armax has quit IRC | 08:06 | |
*** ssurana has quit IRC | 08:06 | |
openstackgerrit | IWAMOTO Toshihiro proposed openstack/nova: Omit RescheduledException in instance_fault.message https://review.openstack.org/165751 | 08:07 |
deepakcs | Hello, can any cores pls bless https://review.openstack.org/#/c/194538/4 (It already has a lot of +1s) | 08:08 |
*** park_heijlong has joined #openstack-nova | 08:09 | |
*** markvoelker has joined #openstack-nova | 08:11 | |
openstackgerrit | Ghanshyam Mann proposed openstack/nova: Add top,bottom & change version tests for microversion 2.4 https://review.openstack.org/202431 | 08:14 |
openstackgerrit | Sean McCully proposed openstack/nova: Creates disk.fallocate_test https://review.openstack.org/188249 | 08:15 |
*** y_sawai has joined #openstack-nova | 08:15 | |
*** markvoelker has quit IRC | 08:16 | |
*** y_sawai has quit IRC | 08:16 | |
*** jistr has joined #openstack-nova | 08:16 | |
*** markvoelker has joined #openstack-nova | 08:17 | |
*** ssurana has joined #openstack-nova | 08:18 | |
*** shyama has quit IRC | 08:20 | |
*** markvoelker has quit IRC | 08:25 | |
openstackgerrit | Victor Stinner proposed openstack/nova: Fix Python 3 issues in nova.utils and nova.tests https://review.openstack.org/197982 | 08:25 |
*** derekh has joined #openstack-nova | 08:27 | |
*** shuquan has quit IRC | 08:28 | |
*** shuquan has joined #openstack-nova | 08:28 | |
*** haypo has joined #openstack-nova | 08:28 | |
*** redbeard has joined #openstack-nova | 08:29 | |
*** markvoelker has joined #openstack-nova | 08:29 | |
haypo | johnthetubaguy, sdague: hi. i had to update my python3 fix https://review.openstack.org/#/c/197982/ because of another fix related to py3 (written by sdague to fix a random failure of py34) | 08:29 |
openstackgerrit | lyanchih proposed openstack/nova: neutron - race conditions in security-group additions https://review.openstack.org/202436 | 08:31 |
*** boris-42 has quit IRC | 08:32 | |
*** markvoelker has quit IRC | 08:33 | |
*** oro_ has joined #openstack-nova | 08:35 | |
*** oro has joined #openstack-nova | 08:35 | |
openstackgerrit | Andreas Scheuring proposed openstack/nova: Add macvtap as libvirt virtual interface (vif) type to Nova's libvirt driver https://review.openstack.org/182283 | 08:35 |
*** aarefiev has joined #openstack-nova | 08:35 | |
*** jecarey has quit IRC | 08:36 | |
*** alex_klimov has quit IRC | 08:38 | |
*** yassine has joined #openstack-nova | 08:40 | |
*** oro_ has quit IRC | 08:41 | |
*** exploreshaifali has joined #openstack-nova | 08:41 | |
*** kmARC has joined #openstack-nova | 08:42 | |
*** markvoelker has joined #openstack-nova | 08:44 | |
*** rdo has joined #openstack-nova | 08:45 | |
*** exploreshaifali has quit IRC | 08:46 | |
*** salv-orlando has joined #openstack-nova | 08:46 | |
*** rajesht has quit IRC | 08:47 | |
*** markvoelker has quit IRC | 08:48 | |
*** aix has joined #openstack-nova | 08:48 | |
*** shyu has quit IRC | 08:48 | |
johnthetubaguy | deepakcs: do you have all the code up for review for that spec now? | 08:50 |
johnthetubaguy | deepakcs: I am meaning this one: https://blueprints.launchpad.net/nova/+spec/volume-snapshot-improvements | 08:50 |
*** belmoreira has quit IRC | 08:51 | |
deepakcs | johnthetubaguy: yes @ https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/volume-snapshot-improvements,n,z | 08:51 |
deepakcs | johnthetubaguy: working on review comments for the "needed by" patch | 08:51 |
*** ssurana has quit IRC | 08:52 | |
openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: do relative block rebase only with non-null base https://review.openstack.org/202442 | 08:52 |
johnthetubaguy | deepakcs: can you please mark it as NeedsCodeReview, that should get it more attention I hope | 08:52 |
deepakcs | johnthetubaguy: maybe i missed something, how to mark it as NeedsCodeReview ? | 08:53 |
johnthetubaguy | deepakcs: where it says started, you should be able to move it to NeedsCodeReview | 08:53 |
johnthetubaguy | there should be a little pencil next to it | 08:53 |
deepakcs | johnthetubaguy: ah in the lp link, i was looking at the patch link :) | 08:54 |
deepakcs | johnthetubaguy: done, my bad, didn't realise i had to do that in lp | 08:55 |
*** shuquan has quit IRC | 08:55 | |
johnthetubaguy | deepakcs: no worries, it just helps tell people when you are done, otherwise its hard to tell | 08:55 |
deepakcs | johnthetubaguy: agree, I had my patches in WIP for some time, before i moved them to 'ready for review', but forgot abt the LP part | 08:56 |
*** park_ has quit IRC | 08:56 | |
johnthetubaguy | deepakcs: easy done, no worries, all fixed now | 08:57 |
*** gszasz has joined #openstack-nova | 08:57 | |
deepakcs | johnthetubaguy: :) | 08:58 |
*** markvoelker has joined #openstack-nova | 08:58 | |
*** Jeffrey4l has quit IRC | 08:58 | |
*** foexle has quit IRC | 09:00 | |
*** foexle has joined #openstack-nova | 09:01 | |
nagyz_ | i know not the best channel but does any one of you have a link to recent keystoneclient middleware documentation? I'm trying to understand what's needed and what's not needed in the authtoken configuration as it seems to be a mess and differ project by project | 09:01 |
*** jichen has quit IRC | 09:01 | |
*** doude has joined #openstack-nova | 09:03 | |
*** doude has quit IRC | 09:03 | |
*** doude has joined #openstack-nova | 09:03 | |
*** markvoelker has quit IRC | 09:03 | |
*** shuquan has joined #openstack-nova | 09:05 | |
*** belmoreira has joined #openstack-nova | 09:05 | |
*** yonglihe has quit IRC | 09:06 | |
*** Jhelf has quit IRC | 09:06 | |
*** alex_klimov has joined #openstack-nova | 09:06 | |
openstackgerrit | Ghanshyam Mann proposed openstack/nova: Show 'locked_by' information in server Show & List Detail https://review.openstack.org/202391 | 09:07 |
*** lyanchih has quit IRC | 09:08 | |
*** lyanchih has joined #openstack-nova | 09:09 | |
*** otter768 has joined #openstack-nova | 09:11 | |
*** prashantvarma has joined #openstack-nova | 09:11 | |
openstackgerrit | Ghanshyam Mann proposed openstack/nova: Show 'locked_by' information in server Show & List Detail https://review.openstack.org/202391 | 09:12 |
*** tpeoples has quit IRC | 09:12 | |
*** bkopilov has quit IRC | 09:14 | |
*** ndipanov has joined #openstack-nova | 09:14 | |
*** otter768 has quit IRC | 09:16 | |
openstackgerrit | lyanchih proposed openstack/nova: NUMATopologyFilter raise exception and not continue filter next node https://review.openstack.org/200630 | 09:16 |
*** Jeffrey4l has joined #openstack-nova | 09:17 | |
*** pixelbeat_ has joined #openstack-nova | 09:19 | |
openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: do relative block rebase only with non-null base https://review.openstack.org/202442 | 09:22 |
openstackgerrit | WenPing Zhang proposed openstack/nova: blueprint support-yfs-storage-protocol https://review.openstack.org/202459 | 09:25 |
alex_xu_ | andreykurilin: one more explain for 'get_current', you can try request to '/2.1' or '/2' instead of '/', you can get current endpoint's version info, then you needn't find out which version info belong to your current endpoint with '/' | 09:27 |
*** alex_xu_ is now known as alex_xu | 09:27 | |
*** bkopilov has joined #openstack-nova | 09:28 | |
andreykurilin | alex_xu: hi! ok. I'll try it now | 09:29 |
alex_xu | andreykurilin: cool, thanks! | 09:29 |
andreykurilin | alex_xu: btw, I don't see your mark for https://review.openstack.org/#/c/167408/ :) | 09:29 |
alex_xu | andreykurilin: yea, I want to see more how thing works more in later patch before give my point for that one | 09:30 |
andreykurilin | alex_xu: you can check it with https://review.openstack.org/#/c/136458/ , since other patches add only support of "latest" | 09:31 |
*** rubasov has joined #openstack-nova | 09:31 | |
alex_xu | andreykurilin: yea, that patch need update? | 09:32 |
alex_xu | andreykurilin: I didn't saw the detail yet, just saw the unittest failed | 09:33 |
andreykurilin | alex_xu: only recheck is needed due too issue with mock | 09:33 |
alex_xu | andreykurilin: ok, cool, will try that patch later | 09:33 |
andreykurilin | which was fixed in master | 09:33 |
andreykurilin | alex_xu: nice | 09:33 |
*** prashantvarma has quit IRC | 09:34 | |
*** shyama has joined #openstack-nova | 09:35 | |
openstackgerrit | John Garbutt proposed openstack/nova: Adding support for InfiniBand SR-IOV vif type https://review.openstack.org/187052 | 09:37 |
*** annashen has joined #openstack-nova | 09:38 | |
*** salv-orlando has quit IRC | 09:40 | |
*** annashen has quit IRC | 09:42 | |
*** sudipto has quit IRC | 09:42 | |
*** sudipto has joined #openstack-nova | 09:48 | |
*** kjelly has quit IRC | 09:48 | |
*** kjelly has joined #openstack-nova | 09:48 | |
*** kjelly has quit IRC | 09:48 | |
*** deepthi has quit IRC | 09:50 | |
johnthetubaguy | sahid: I am just looking at https://blueprints.launchpad.net/nova/+spec/convert-consoles-to-objects as there is no code up for review yet, I think that needs to be delayed until M, let me know if I am missing something | 09:50 |
sdague | haypo: +A | 09:50 |
sahid | johnthetubaguy: the spec has been approved 10days ago so yes. | 09:52 |
johnthetubaguy | sahid: yeah, I think we need to keep aiming to get specs approved before the summit, its kinda always been the intention, but we have never really managed it | 09:54 |
johnthetubaguy | the burn out at the end of the release doesn't help, but hopefully the backlog being open should help this | 09:54 |
*** ociuhandu has joined #openstack-nova | 09:55 | |
openstackgerrit | Jeffrey Zhang proposed openstack/nova: Add total_vms field to ComputeNode https://review.openstack.org/202472 | 09:57 |
*** Jeffrey4l has quit IRC | 09:57 | |
openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: [WIP] Use explicit greenthread.sleep instead of time.sleep https://review.openstack.org/202365 | 09:58 |
haypo | sdague: cool, thanks | 09:58 |
haypo | sdague: did you see my comments on dims patch for reseeding OpenSSL RNG after fork? i proposed to use /dev/urandom instead | 09:58 |
*** aix has quit IRC | 09:58 | |
haypo | sdague: it's now recommended by most security guys | 09:59 |
sdague | for which thing? | 09:59 |
sdague | really, security people are recommending urandom for seeding openssl? | 09:59 |
sdague | that seems terrible | 09:59 |
haypo | sdague: no. they recommand to only use /dev/urandom, not use OpenSSL for random | 10:00 |
andreykurilin | alex_xu: oh... HTTPClient is not friendly-written:( there is no ability to send get request to "url/v21", it can send only to "url/v21/<token>" :( so we need to list all versions when HTTPClient is used and rewrite it in future | 10:00 |
sdague | where is the entropy being consumed? | 10:00 |
*** shuquan has quit IRC | 10:01 | |
sdague | because if the answer is for keypairs, then I think the answer should be a definitive no | 10:01 |
andreykurilin | alex_xu: OR we can make simple "request.get", since we don't need authentication stuff | 10:01 |
haypo | sdague: the patch, https://review.openstack.org/#/c/191581/ | 10:01 |
sdague | because otherwise you get this - https://www.schneier.com/blog/archives/2014/01/security_risks_9.html | 10:02 |
haypo | sdague: it's crypto/utils.py in oslo incubator. i don't know how it is used | 10:02 |
haypo | sdague: i understood that /dev/urandom is good enough to generate keys | 10:02 |
sdague | right, that's what scares the crap out of me | 10:02 |
sdague | haypo: please provide a reference for that fact | 10:02 |
*** claudiub has joined #openstack-nova | 10:03 | |
*** zhenguo has quit IRC | 10:03 | |
haypo | sdague: i gave links: https://review.openstack.org/#/c/191581/2/openstack/common/crypto/utils.py & https://review.openstack.org/#/c/191581/5/openstack/common/crypto/utils.py | 10:04 |
sudipto | johnthetubaguy, is the rule to not have features in which are half way through implementation - even though it might be non-priority? :( It sounds a bit unfair for people who have got their patches half way through the review :( | 10:05 |
*** hshiina has joined #openstack-nova | 10:05 | |
johnthetubaguy | sudipto: so the idea is to optimise for completed blueprints by 30th July | 10:05 |
sdague | haypo: so that assumes that this is on systems that aren't fresh created and have accumulated enough entropy in their buffers | 10:05 |
johnthetubaguy | sudipto: I am starting with blueprints with no code up for review | 10:05 |
sdague | which, seems like a poor assumption in cloud | 10:06 |
sdague | if these are freshly deployed compute nodes | 10:06 |
johnthetubaguy | sdague: yeah I think thats why we have nova-cert | 10:06 |
sdague | so the problem is, we have no idea where this function is going to be used | 10:06 |
sudipto | johnthetubaguy, ok - so if a spec has half the code in now, and let's say half the code done in the next couple of days (before July 30th) - will it still be considered? | 10:07 |
sdague | johnthetubaguy: sure, but I'm concerned about hiding this fact behind oslo layers | 10:07 |
johnthetubaguy | sdague: yeah, true, its easy to use by accident in the wrong place | 10:07 |
johnthetubaguy | sudipto: maybe, we should talk about your specific case, technically, the rule is, get all your code up before today, we had 104 approved blueprints this morning, most of the low priority ones do have code up for review already, so we want to give them preference | 10:09 |
sudipto | johnthetubaguy, ok. | 10:09 |
*** afazekas has joined #openstack-nova | 10:10 | |
sdague | right, having been ansilarily involved in a firedrill where all the certs on a range of embed systems were brute forcable, I'd like us to not put that bug into our system | 10:10 |
johnthetubaguy | sudipto: to set expectations, I can't see us merging more that 50% of the blueprints that are already in NeedsCodeReview before 30th, we just don't have enough folks reviewing code right now :( | 10:11 |
sdague | haypo: this was the incident - http://www.bishopfox.com/blog/2012/03/ssl-key-generation-weaknesses/ | 10:11 |
sudipto | johnthetubaguy, i have 2 code drops scheduled. My first code has been up for review since July 7th, and finally - we have it in some shape to be merged. But when I heard about the deadline, i am not sure if i should go ahead with the second code drop... | 10:11 |
johnthetubaguy | sudipto: what is the second code drop? | 10:12 |
*** ociuhandu has quit IRC | 10:12 | |
sdague | and the root cause, was these systems did not have sufficient entropy. | 10:12 |
johnthetubaguy | sudipto: I guess I am trying to work out why its not up for review already | 10:13 |
*** liusheng has quit IRC | 10:13 | |
*** liusheng has joined #openstack-nova | 10:13 | |
sudipto | johnthetubaguy, ok, as you might know the recent compute monitor framework went for a change - it's using the nova objects now - and for my blueprint, i needed to introduce a nova object that would support reporting of bw on NUMA nodes. So this was code drop 1. So we reached a consensus on the object 2 days back. Then the second code drop was meant to be for the reporting bit (the metrics - that would be actually fetched via PCP)... | 10:14 |
haypo | sdague: can you please add a comment giving your two links? i don't have the bandwidth yet to digest these two interesting articles | 10:14 |
sudipto | johnthetubaguy, I am wondering, if it's a good idea to have the code drop 2 now .... | 10:14 |
sdague | haypo: sure | 10:15 |
johnthetubaguy | sudipto: if you haven't started that code yet, possibly not worth it | 10:15 |
sdague | haypo: honestly, this is a case where I think getting the OpenStack Security team involved in this review would be good. They exist for a reason. If I'm being overly paranoid and they say so, I'm fine with it. | 10:15 |
sudipto | johnthetubaguy, honestly, i have the code ready - however, I am yet to finish on the UTs... | 10:15 |
sdague | But I also know it's a common mistake, and rng for crypto is hard | 10:16 |
johnthetubaguy | sudipto: OK, understood, I mean you can try, just upload what you have, and mark it as WIP, thats a middle ground, but honestly, its getting very close to the wire at this point :( | 10:16 |
sudipto | johnthetubaguy, ok that's very reasonable. Just marred by the situation i guess :( | 10:17 |
*** garyk has quit IRC | 10:18 | |
_gryf | johnthetubaguy, hi | 10:18 |
_gryf | johnthetubaguy, i've changed the status of the BP to NeedsCodeReview this morning (since the implementation is done and half of the patches are already accepted) | 10:19 |
johnthetubaguy | sudipto: yeah, this freeze is all about getting the things we agreed as a priority done (like objects), and getting as much value for our users by getting completed blueprints merged, but yeah, this is not a happy time in the released | 10:19 |
_gryf | johnthetubaguy, do it is enough to make it still i l release instead of m? | 10:20 |
johnthetubaguy | _gryf: sounds good, yeah NeedsCodeReview = all the code us submitted for review | 10:20 |
_gryf | johnthetubaguy, cool, since it would be scary to wait entire cycle to make it merge :) | 10:21 |
johnthetubaguy | _gryf: its just a case of getting the code merged before the 30th July | 10:21 |
johnthetubaguy | the big message here is, the more reviews everyone does of each others code, the more likely we can get more blueprints merged | 10:22 |
_gryf | johnthetubaguy, ack. Who can I ask for review changes which touches the topics of service groups and REST API? | 10:22 |
_gryf | johnthetubaguy, (i mean the core team) | 10:23 |
_gryf | johnthetubaguy, since one of my pacth already have a bunch of +1 | 10:23 |
*** dims_ has joined #openstack-nova | 10:24 | |
johnthetubaguy | _gryf: its a waiting game at this point, but for the REST API, there is the API subteam with some great reviewers who can help | 10:25 |
_gryf | johnthetubaguy, ok, thanx for the info | 10:26 |
*** _gryf has quit IRC | 10:26 | |
*** pixelbeat_ has quit IRC | 10:27 | |
*** aix has joined #openstack-nova | 10:28 | |
*** _gryf has joined #openstack-nova | 10:30 | |
openstackgerrit | Sudipta Biswas proposed openstack/nova: Allow NUMA based reporting for Monitors https://review.openstack.org/198972 | 10:32 |
*** afazekas has quit IRC | 10:35 | |
*** Jeffrey4l has joined #openstack-nova | 10:36 | |
*** dims_ has quit IRC | 10:39 | |
*** salv-orlando has joined #openstack-nova | 10:40 | |
*** lyanchih has quit IRC | 10:43 | |
*** exploreshaifali has joined #openstack-nova | 10:43 | |
openstackgerrit | WenPing Zhang proposed openstack/nova: blueprint support-yfs-storage-protocol https://review.openstack.org/202459 | 10:45 |
*** shyama has quit IRC | 10:46 | |
*** shyama_ has joined #openstack-nova | 10:46 | |
*** shyama_ is now known as shyama | 10:46 | |
*** sudipto has quit IRC | 10:46 | |
*** afazekas has joined #openstack-nova | 10:48 | |
*** garyk has joined #openstack-nova | 10:54 | |
*** IanGovett has joined #openstack-nova | 10:55 | |
*** coolsvap has quit IRC | 10:58 | |
*** coolsvap has joined #openstack-nova | 10:58 | |
johnthetubaguy | park_heijlong: just looking at the state of this blueprint, I don't think all the code is up for review yet? https://blueprints.launchpad.net/nova/+spec/add-id-to-security-groups-for-server-show | 10:59 |
ndipanov | so found this little gremlin https://bugs.launchpad.net/nova/+bug/1475254 | 10:59 |
openstack | Launchpad bug 1475254 in OpenStack Compute (nova) "NovaObjectSerializer cannot handle backporting a nested object" [High,Confirmed] | 10:59 |
ndipanov | morale of the story - humans should not raise exceptions in recursive functions | 11:01 |
*** sudipto has joined #openstack-nova | 11:02 | |
*** mwagner_lap has quit IRC | 11:05 | |
openstackgerrit | Paul Murray proposed openstack/nova: Remove conductor api from resource tracker https://review.openstack.org/202501 | 11:08 |
*** otter768 has joined #openstack-nova | 11:12 | |
openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: do relative block rebase only with non-null base https://review.openstack.org/202442 | 11:12 |
* alex_xu sending text to park_heijlong, notify him... | 11:12 | |
*** belmoreira has quit IRC | 11:15 | |
openstackgerrit | Paul Murray proposed openstack/nova: Remove conductor api from resource tracker https://review.openstack.org/202501 | 11:15 |
*** xgsa has quit IRC | 11:16 | |
*** xgsa has joined #openstack-nova | 11:16 | |
openstackgerrit | sahid proposed openstack/nova: hardware: take into account pages size when computing memory https://review.openstack.org/202504 | 11:16 |
sahid | ndipanov: can you take a look at this, it's still work in progress https://review.openstack.org/202504 | 11:16 |
*** otter768 has quit IRC | 11:17 | |
sahid | it's about how we compute memory usage when host have kind of pages | 11:17 |
*** IanGovett has quit IRC | 11:17 | |
openstackgerrit | Ghanshyam Mann proposed openstack/nova: Show 'locked_by' information in server Show & List Detail https://review.openstack.org/202391 | 11:21 |
*** xgsa has quit IRC | 11:23 | |
*** xgsa has joined #openstack-nova | 11:24 | |
*** pixelbeat_ has joined #openstack-nova | 11:24 | |
*** lyanchih has joined #openstack-nova | 11:24 | |
*** lucasagomes is now known as lucas-hungry | 11:25 | |
*** veena has quit IRC | 11:26 | |
openstackgerrit | Sergey Vilgelm proposed openstack/nova: Switch to the oslo_utils.fileutils https://review.openstack.org/199575 | 11:35 |
*** GLaupre has quit IRC | 11:37 | |
bauzas | johnthetubaguy: edleafe: I see the email for the scheduler, I'll try to reply for it today | 11:40 |
*** belmoreira has joined #openstack-nova | 11:40 | |
bauzas | johnthetubaguy: edleafe: tl;dr: anyone can do anything out of tree, and then coming back for discussing | 11:40 |
bauzas | johnthetubaguy: edleafe: but we need to be very clear on modifying the current in-tree schedulers | 11:41 |
johnthetubaguy | bauzas: yep, but I appreciate people like edleafe being up ideas for discussion, thats really healthy | 11:41 |
bauzas | johnthetubaguy: sorry, I pinged edleafe now even if I know he's AFK due to his TZ | 11:41 |
*** jichen has joined #openstack-nova | 11:43 | |
*** aysyd has joined #openstack-nova | 11:45 | |
*** thorst has joined #openstack-nova | 11:47 | |
garyk | is there any chance that one or more cores can please look at https://review.openstack.org/#/c/188789/ - there are tons of patches waiting for this one to get in. this is starting the metadata object support. please i implore you | 11:54 |
*** Longgeek has quit IRC | 11:54 | |
*** IanGovett has joined #openstack-nova | 11:58 | |
*** dprince has joined #openstack-nova | 12:02 | |
*** ptm_away is now known as PaulMurray | 12:03 | |
PaulMurray | ndipanov, have you seen https://bugs.launchpad.net/nova/+bug/1275675 | 12:04 |
openstack | Launchpad bug 1275675 in OpenStack Compute (nova) "Version change in ObjectField does not work with back-levelling" [Wishlist,Confirmed] - Assigned to Paul Murray (pmurray) | 12:04 |
*** bkopilov has quit IRC | 12:04 | |
PaulMurray | ndipanov, and https://review.openstack.org/#/c/78605/ | 12:04 |
* PaulMurray didn't assign that to himself?? | 12:05 | |
*** Longgeek has joined #openstack-nova | 12:05 | |
* PaulMurray must have clicked on something | 12:05 | |
openstackgerrit | Jordan Pittier proposed openstack/nova: libvirt:when deleting snapshot use qemu-img to blockRebase if domain is stopped https://review.openstack.org/192736 | 12:07 |
openstackgerrit | Jordan Pittier proposed openstack/nova: libvirt:when deleting snapshot use qemu-img to blockRebase if domain is stopped https://review.openstack.org/192736 | 12:08 |
*** lyanchih has quit IRC | 12:08 | |
*** vladikr has joined #openstack-nova | 12:11 | |
* kashyap notes randomly: if you're using `gertty` and like to view the Gate job logs on CLI, invoke it this way: BROWSER=lynx gertty (thanks fungi). | 12:12 | |
*** deepakcs has quit IRC | 12:13 | |
*** dims_ has joined #openstack-nova | 12:14 | |
*** tellesnobrega_ has joined #openstack-nova | 12:16 | |
*** tellesnobrega_ has quit IRC | 12:16 | |
dims_ | sdague: good morning, i ran devstack with nova + oslo.service here (https://review.openstack.org/#/c/202075/) - grenade jobs seem to be doing ok. | 12:17 |
*** nkrinner is now known as nkrinner_afk | 12:18 | |
*** jcoufal has joined #openstack-nova | 12:19 | |
ndipanov | PaulMurray, no | 12:20 |
ndipanov | but your fix is not correct I think | 12:20 |
ndipanov | now that I look at it | 12:20 |
ndipanov | adding tests now | 12:20 |
ndipanov | will post a fix soon | 12:21 |
ndipanov | but I think it's the same thing | 12:21 |
*** redbeard has quit IRC | 12:22 | |
ndipanov | PaulMurray, it's the same bug | 12:22 |
ndipanov | PaulMurray, how come you just abandoned the fix | 12:24 |
ndipanov | it's a pretty bad bug | 12:24 |
sdague | dims_: so, actually it's gate-devstack-dsvm-updown that we care about | 12:26 |
sdague | the grenade jobs won't see it until the next release, because kilo is still the old code | 12:26 |
sdague | but, I see we have a bunch of runs there | 12:27 |
sdague | so I'm good with it | 12:27 |
*** doude has quit IRC | 12:28 | |
*** READ10 has joined #openstack-nova | 12:29 | |
*** markvoelker has joined #openstack-nova | 12:29 | |
dims_ | sdague: ack. thanks! | 12:29 |
*** doude has joined #openstack-nova | 12:29 | |
*** dims_ has quit IRC | 12:30 | |
*** shuquan has joined #openstack-nova | 12:30 | |
*** salv-orlando has quit IRC | 12:31 | |
*** dims__ has joined #openstack-nova | 12:32 | |
johnthetubaguy | sdague: do you know where we document what "dsvm-updown" does? I have lost track a bit | 12:33 |
sdague | johnthetubaguy: it's not all that documented, it's a devstack functional job | 12:33 |
johnthetubaguy | sdague: it was the "updown" bit, is that turning the services off an on again? | 12:34 |
*** markvoelker has quit IRC | 12:34 | |
sdague | it's turning them off again | 12:34 |
sdague | so bring everything up, and shut it all down | 12:35 |
*** shuquan has quit IRC | 12:35 | |
sdague | to make sure that shutting down works | 12:35 |
johnthetubaguy | cool, that makes sense | 12:35 |
johnthetubaguy | its not updownup | 12:35 |
PaulMurray | ndipanov, I abandoned the fix because I got into one of those "is this what we really want to do" discussions and it was down as a wish........... | 12:35 |
sdague | so that there isn't a big run around at release when we try to flip grenade to a new stable, and then we find out kill on services is broken | 12:36 |
sdague | which... was the kilo situation | 12:36 |
ndipanov | PaulMurray, bs - it's a horrible bug | 12:36 |
PaulMurray | ndipanov, also the line at the time was this shouldn't happen so its more like a runtime sanity check | 12:36 |
ndipanov | also bs | 12:36 |
johnthetubaguy | sdague: +1 its awesome to add that test in | 12:36 |
PaulMurray | ndipanov, but I do agree with you | 12:36 |
ndipanov | it is happening to me now | 12:36 |
ndipanov | we can't land a fix because of this | 12:36 |
ndipanov | I will mark my bug as duplicate and post a fix | 12:36 |
ndipanov | a fix is very simple actually | 12:36 |
ndipanov | the fix that is | 12:37 |
PaulMurray | ndipanov, cool - I'll take a look. Actually I am about to go on holiday for a couple of weeks - so I might not be around for long | 12:37 |
PaulMurray | end of week that is | 12:38 |
*** redbeard has joined #openstack-nova | 12:38 | |
*** bfic has joined #openstack-nova | 12:38 | |
*** sudipto has quit IRC | 12:39 | |
*** dims__ has quit IRC | 12:39 | |
*** IanGovett has quit IRC | 12:39 | |
*** IanGovett1 has joined #openstack-nova | 12:39 | |
*** park_ has joined #openstack-nova | 12:40 | |
ndipanov | yeah I'll post this in like 20 min | 12:40 |
*** park__ has joined #openstack-nova | 12:40 | |
ndipanov | just to get the tests working PaulMurray | 12:40 |
*** lpetrut has quit IRC | 12:41 | |
*** park_ has quit IRC | 12:41 | |
*** park__ has quit IRC | 12:41 | |
*** park_ has joined #openstack-nova | 12:41 | |
*** park_ has quit IRC | 12:42 | |
*** park_heijlong_ has joined #openstack-nova | 12:42 | |
*** markvoelker has joined #openstack-nova | 12:43 | |
*** IanGovett has joined #openstack-nova | 12:44 | |
*** shuquan has joined #openstack-nova | 12:44 | |
*** krtaylor has quit IRC | 12:44 | |
*** IanGovett1 has quit IRC | 12:45 | |
park_heijlong_ | johnthetubaguy: hi, for the bp https://blueprints.launchpad.net/nova/+spec/add-id-to-security-groups-for-server-show, there is only 1 patch left, I could finish it this night | 12:46 |
park_heijlong_ | johnthetubaguy: could I catch the L release freeze..? | 12:47 |
johnthetubaguy | park_heijlong_: OK, if you can get that up, drop me an email, and I will see if we can sneak that in. The big problem, is thats its really close to the wire to get enough reviews to get it merged by 30th July | 12:47 |
*** kylek3h has joined #openstack-nova | 12:48 | |
*** markvoelker has quit IRC | 12:48 | |
*** cpschult has joined #openstack-nova | 12:48 | |
park_heijlong_ | johnthetubaguy: yep, make sense, I will try best to... thank you in advance | 12:49 |
*** lucas-hungry is now known as lucasagomes | 12:49 | |
*** IanGovett1 has joined #openstack-nova | 12:49 | |
johnthetubaguy | sdague: I think we spoke before about documenting these ideas, would love a quick review of some of this, if you get chance: https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule#How_can_I_get_my_code_merged_faster.3F | 12:49 |
*** jcoufal has quit IRC | 12:50 | |
*** IanGovett has quit IRC | 12:50 | |
*** krtaylor has joined #openstack-nova | 12:51 | |
*** mjturek1 has joined #openstack-nova | 12:52 | |
*** salv-orlando has joined #openstack-nova | 12:53 | |
*** vladikr has quit IRC | 12:54 | |
*** lpetrut has joined #openstack-nova | 12:54 | |
markus_z | What's your strategy to deprecate (private) methods which get deleted in a following patch set? | 12:54 |
markus_z | I couldn't found a decorator for that. | 12:54 |
*** IanGovett has joined #openstack-nova | 12:54 | |
markus_z | s/found/find | 12:55 |
*** jcoufal has joined #openstack-nova | 12:55 | |
*** diogogmt has quit IRC | 12:55 | |
*** bfic has quit IRC | 12:55 | |
*** diogogmt has joined #openstack-nova | 12:56 | |
*** IanGovett1 has quit IRC | 12:56 | |
edleafe | tonyb: still around? You just missed me earlier | 12:57 |
*** markvoelker has joined #openstack-nova | 12:57 | |
*** thorst has quit IRC | 12:58 | |
johnthetubaguy | edleafe: saw your mail about the scheduler, sounds interesting, just wondered if you saw my attempted write up of the ideas I raised at the summit: https://review.openstack.org/#/c/191914/ | 12:59 |
*** jcoufal has quit IRC | 12:59 | |
edleafe | johnthetubaguy: sure, I've read it, and it covers what I remember being discussed at the summit | 13:00 |
*** zhenguo has joined #openstack-nova | 13:01 | |
*** markvoelker has quit IRC | 13:02 | |
*** mwagner_lap has joined #openstack-nova | 13:02 | |
johnthetubaguy | edleafe: cool, I am just being cheaky and seeing if I could get someone to try the kafka idea (its basically a rehash of the no-db-scheduler blueprint) | 13:02 |
edleafe | johnthetubaguy: I didn't add my ideas because we hadn't talked about them | 13:02 |
johnthetubaguy | edleafe: another backlog spec is welcome, but trying them out is probably more useful | 13:03 |
edleafe | johnthetubaguy: one thing I know is that the community can argue forever on an idea | 13:03 |
johnthetubaguy | edleafe: certainly, I see the bottleneck as the data sync, so its the right bit to focus on, for max benefit | 13:03 |
edleafe | johnthetubaguy: I wanted to try to present some numbers first | 13:03 |
edleafe | johnthetubaguy: it's also key to separation, oddly enough | 13:04 |
edleafe | johnthetubaguy: if this is successful, there is no reason that it wouldn't be available to cinder, neutron, etc | 13:04 |
*** bfic has joined #openstack-nova | 13:04 | |
*** thorst has joined #openstack-nova | 13:04 | |
johnthetubaguy | edleafe: so we will want a single block of state for that I think, but thats a different discussion | 13:05 |
edleafe | johnthetubaguy: yes | 13:05 |
bauzas | edleafe: I'm trying to provide you a reply by today | 13:05 |
*** markvoelker has joined #openstack-nova | 13:05 | |
edleafe | johnthetubaguy: what I hope to get out of the midcycle are some goals to shoot for | 13:05 |
bauzas | edleafe: to be clear, I like to see some out-of-tree drivers for that | 13:05 |
johnthetubaguy | edleafe: I am trying to sketch out the direction of where we are thinking in devref, so we can get more folks involved: https://review.openstack.org/#/c/192260/ | 13:06 |
bauzas | edleafe: but I wonder how to do that correctly in-tree | 13:06 |
*** mdrabe has joined #openstack-nova | 13:06 | |
edleafe | bauzas: this is simply a POC experiment | 13:06 |
bauzas | edleafe: if so, why do you need some help/view ? | 13:06 |
*** markvoelker_ has joined #openstack-nova | 13:06 | |
edleafe | bauzas: it is a waste of time to debate how to integrate it until we know that it's worth the effort | 13:06 |
bauzas | edleafe: well, anyone can POC a new driver | 13:07 |
edleafe | bauzas: I need to know what sort of results would be definitive | 13:07 |
johnthetubaguy | edleafe: yeah, thats a good goal, hopefully this is a starting point: https://review.openstack.org/#/c/192260/ | 13:07 |
edleafe | bauzas: it's not a driver | 13:07 |
bauzas | edleafe: so, johnthetubaguy's backlog spec is there for that | 13:07 |
edleafe | bauzas: it's a different approach to sharing data | 13:07 |
bauzas | edleafe: I know Cassandra , but I would like to understand why | 13:08 |
johnthetubaguy | edleafe: so you can plug that as a new driver, I think this pattern would work: https://github.com/openstack/nova/blob/master/nova/scheduler/caching_scheduler.py | 13:08 |
bauzas | edleafe: because here, we don't need a technology, rather a design disussion | 13:08 |
openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Add support of latest version in CLI https://review.openstack.org/184281 | 13:08 |
openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Add "get_current" method to VersionManager https://review.openstack.org/202296 | 13:08 |
openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 1 https://review.openstack.org/152569 | 13:08 |
openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Set "latest" as default compute api version https://review.openstack.org/184658 | 13:08 |
openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 2 https://review.openstack.org/167408 | 13:08 |
edleafe | bauzas: because I have seen it perform in situations that had the same issues as we are having | 13:08 |
*** itzikb_ has quit IRC | 13:09 | |
edleafe | bauzas: the technology enables the design | 13:09 |
bauzas | edleafe: please don't | 13:09 |
bauzas | edleafe: see that http://sbauza.github.io/2015/04/20/#/4 ? | 13:09 |
edleafe | bauzas: we have been trying to re-invent much of that technology using other things | 13:09 |
johnthetubaguy | bauzas: its just one implementation of a different approach, giving it a whirl in a prototype is fine I think | 13:09 |
bauzas | edleafe: anyone can work on a new driver and see how it works | 13:09 |
*** rlrossit has joined #openstack-nova | 13:09 | |
*** markvoelker has quit IRC | 13:10 | |
*** yingjun has joined #openstack-nova | 13:10 | |
bauzas | edleafe: you only have to implement a very limited interface for your own scheduler | 13:10 |
bauzas | edleafe: and that can be called thanks to an opt | 13:10 |
edleafe | bauzas: you don't seem to understand that this isn't simply "another driver" | 13:11 |
bauzas | edleafe: so I'm really +1 on giving you room for that | 13:11 |
openstackgerrit | garyk proposed openstack/nova: Neutron: configure requested amount of fixed_ips https://review.openstack.org/163864 | 13:11 |
garyk | mriedem_away: you have a sec ti talk about https://github.com/openstack/nova/commit/1ba72081715869fe6d0ffba3436fb09ef9499335? | 13:11 |
edleafe | bauzas: another driver would have to consume the same data in the same way, and emit it in the same way | 13:11 |
garyk | the change prevents us from deleting an instance … | 13:11 |
bauzas | edleafe: you're talking about a backend, right? | 13:12 |
*** vladikr has joined #openstack-nova | 13:12 | |
bauzas | edleafe: incorrect | 13:12 |
bauzas | edleafe: another driver can provide its own way to consume resources | 13:12 |
edleafe | bauzas: all I'm asking is "What results would impress you so much that it would be worth considering this type of change?" | 13:12 |
bauzas | edleafe: be sure that lots of alternatives exist out-of-tree | 13:12 |
edleafe | bauzas: this would require changing how resources are reported, too | 13:12 |
openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Adds support for x509 certificates as keypairs https://review.openstack.org/136458 | 13:13 |
*** otter768 has joined #openstack-nova | 13:13 | |
edleafe | bauzas: don't fixate on a particular solution | 13:13 |
bauzas | edleafe: my take on that is to say - anyone can do anything he wants | 13:13 |
*** coolsvap is now known as coolsvap|away | 13:13 | |
edleafe | bauzas: tell me what I could show you that would be significant | 13:13 |
bauzas | edleafe: but if that requires in-tree implementation, take it in small steps | 13:13 |
edleafe | bauzas: other experiments could use the same metrics | 13:13 |
edleafe | bauzas: this might meet those expectations, but others could, too | 13:13 |
bauzas | edleafe: right, what we need to consider is the migration path from the current situation to the one you wabnt | 13:14 |
edleafe | bauzas: you're getting ahead of things again | 13:14 |
bauzas | edleafe: that's exactly why we're working on fixing stuff and reducing debt | 13:14 |
edleafe | discussing what small steps we should take should come *after* we know it's going to be worth it, not before | 13:14 |
*** bnemec has joined #openstack-nova | 13:14 | |
bauzas | edleafe: so, play with the POC, use it even downstream | 13:14 |
johnthetubaguy | edleafe: for me, look at doing 1000 builds in parallel with a setup where you have space for many 1100 builds, and do that as quickly as possible, with minimal retries, at least thats the kind of thing I am worried about | 13:15 |
edleafe | bauzas: you seem to want to iron out all the details before we even know that this will be an improvement | 13:15 |
edleafe | johnthetubaguy: thank you! | 13:15 |
edleafe | johnthetubaguy: finally someone understands what I asked! :) | 13:15 |
bauzas | okay, I appreciate your tone :) | 13:16 |
* bauzas looking back to code | 13:16 | |
johnthetubaguy | edleafe: ah, so I miss-read your email then, I didn't get thats what you wanted, but yeah, thats the use case that sucks right now :) | 13:16 |
*** dims_ has joined #openstack-nova | 13:16 | |
mnaser | i've had this patch for around a week now with 2 +1s.. any cores free to have a look over? it's a pretty straight forward patch.. https://review.openstack.org/#/c/199516/ | 13:16 |
johnthetubaguy | edleafe: parallelism in the current schedule means more bad descisions | 13:16 |
edleafe | johnthetubaguy: I know that I can experiment out-of-tree without anyone's approval | 13:17 |
edleafe | johnthetubaguy: I only want to know what sort of results would make the community want to consider possibly taking it in-tree | 13:17 |
*** cpschult has quit IRC | 13:17 | |
*** otter768 has quit IRC | 13:17 | |
johnthetubaguy | edleafe: yeah, I understand that now | 13:17 |
johnthetubaguy | edleafe: the two reviews should help describe that: https://review.openstack.org/200001 https://review.openstack.org/#/c/191914 if they don't then need more work | 13:18 |
johnthetubaguy | edleafe: to be clear I am fairly sure they need a lot more work, but I would love some ideas on how to get what you need from them | 13:19 |
edleafe | johnthetubaguy: thanks. I'll be sure to absorb them before the midcycle | 13:19 |
*** jichen_ has joined #openstack-nova | 13:19 | |
edleafe | johnthetubaguy: I like playing with ideas without focusing on why they *won't* work | 13:20 |
johnthetubaguy | edleafe: the first one is the more important one, getting a devref to describe the scheduler work | 13:20 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: Fix the incorrect PciDeviceList version number https://review.openstack.org/201280 | 13:20 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: objects: fix deserialization with nested backleveling https://review.openstack.org/202554 | 13:20 |
ndipanov | dansmith, mriedem_away ^^ | 13:20 |
*** jichen has quit IRC | 13:20 | |
*** jichen_ is now known as jichen | 13:20 | |
johnthetubaguy | edleafe: its a really useful approach for performance work | 13:20 |
ndipanov | we should really land the first patch asap - I will propose a kilo backport too | 13:21 |
ndipanov | PaulMurray, ^^ | 13:21 |
nihilifer | bauzas: do you have some time for review today? | 13:21 |
*** ivasev has joined #openstack-nova | 13:21 | |
johnthetubaguy | edleafe: the stuff I tried and prototyped last year (that generated the CachingScheduler) is what has lead me in this current direction | 13:21 |
edleafe | johnthetubaguy: ok, I'll dig into it | 13:21 |
johnthetubaguy | edleafe: fully recommend the crazy protoype, it was extremely educational! | 13:22 |
nihilifer | bauzas: just want to know whether the approach in https://review.openstack.org/#/c/195783/ is now good for you | 13:22 |
edleafe | johnthetubaguy: :) | 13:22 |
andreykurilin | sdague, johnthetubaguy, alex_xu: hi! I fixed all patches in novaclient based on our yesterday's discussion. | 13:22 |
andreykurilin | *the day before yesterday | 13:22 |
bauzas | edleafe: were you at the session about scaling out the scheduler ? https://etherpad.openstack.org/p/YVR-nova-scalling-out-scheduler-for-cells | 13:23 |
edleafe | bauzas: of course | 13:23 |
*** zzzeek has joined #openstack-nova | 13:24 | |
*** ctrath has joined #openstack-nova | 13:24 | |
bauzas | edleafe: okay, so I feel that something is unclear to me | 13:24 |
bauzas | edleafe: by reading again your idea, you're thinking about changing how we report those metrics so we can scale | 13:24 |
haypo | johnthetubaguy: sorry, but i'm not sure that i understood the status "NeedsCoreReview" for https://blueprints.launchpad.net/nova/+spec/nova-python3 -- can I continue to submit python3 patches next weeks? or should I wait until november? | 13:24 |
johnthetubaguy | haypo: you should wait, really, at least after the end of today, I think you should wait | 13:25 |
bauzas | edleafe: but that sounds very close to what we discussed during that session | 13:25 |
bauzas | edleafe: and where the outcome was "we need to define first our thoughts" | 13:25 |
edleafe | bauzas: well, yes, it certainly sparked my thinking | 13:26 |
edleafe | bauzas: when I first created cells ("zones" back then) I had modeled them using Cassandra | 13:26 |
johnthetubaguy | haypo: the basic idea is that we need to get what we have up for review merged by 30th June, after that point, we try to just focus on the priority stuff | 13:26 |
edleafe | bauzas: that got shot down by the "MySQL Everywhere" sentiment | 13:27 |
openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: do relative block rebase only with non-null base https://review.openstack.org/202442 | 13:27 |
bauzas | edleafe: I really want to stop talking about implementations :) | 13:27 |
edleafe | bauzas: all the problems of the original zones design came from trying to force one design into another implementation | 13:27 |
bauzas | edleafe: what we need is a way to provide resources to a distributed system | 13:27 |
*** gyee has joined #openstack-nova | 13:27 | |
*** edmondsw has joined #openstack-nova | 13:28 | |
edleafe | bauzas: hmmmm... what could possibly do that? :) | 13:28 |
edleafe | bauzas: let's please stop arguing about one approach over another, until we have some results | 13:28 |
edleafe | bauzas: let's make choices based on data, not feelings or biases | 13:29 |
edleafe | bauzas: if I run this experiment and the numbers don't impress, I'll be the first person to say "don't do this" | 13:29 |
edleafe | bauzas: but I think it's a bad idea for long-term development to shoot down experiments before they are run | 13:30 |
*** ankit has quit IRC | 13:30 | |
*** yingjun has quit IRC | 13:31 | |
*** yingjun has joined #openstack-nova | 13:31 | |
bauzas | edleafe: again, I really want that experiment | 13:32 |
bauzas | edleafe: I even proposed one https://etherpad.openstack.org/p/liberty-nova-scalable-scheduler | 13:32 |
*** jlearman has joined #openstack-nova | 13:33 | |
bauzas | edleafe: so that's pretty cool to see you're volunteer for working on that | 13:33 |
edleafe | bauzas: did you ever get a chance to try it? | 13:33 |
bauzas | edleafe: and your proposal is very close to what johnthetubaguy proposed with Kafka | 13:33 |
*** jcoufal has joined #openstack-nova | 13:33 | |
edleafe | bauzas: so if we can get an experimental setup in place, we could eventually try out several approaches and see which is strongest | 13:34 |
edleafe | bauzas: but that's still getting ahead of ourselves | 13:34 |
bauzas | edleafe: my proposal was focusing on having a distributed lock for consuming resources | 13:34 |
bauzas | edleafe: yours and johnthetubaguy are more about how we get those resources | 13:35 |
*** jaypipes has joined #openstack-nova | 13:35 | |
bauzas | edleafe: that's where I think it doesn't really need something to be done to be tested | 13:36 |
*** yingjun has quit IRC | 13:36 | |
edleafe | bauzas: yes: how they are reported, compared, and claimed | 13:36 |
johnthetubaguy | bauzas: I did put both in my spec, thats one of the issues I need to fix in that spec actually | 13:36 |
*** thangp has joined #openstack-nova | 13:36 | |
bauzas | edleafe: what I'm missing with your proposal is how it interacts with the existing design | 13:37 |
*** bkopilov has joined #openstack-nova | 13:37 | |
bauzas | edleafe: I understood you want more than a drvier | 13:37 |
*** lyanchih has joined #openstack-nova | 13:37 | |
edleafe | bauzas: not sure I understand | 13:37 |
bauzas | edleafe: you can't really do a big-bang, right? | 13:37 |
edleafe | bauzas: meaning what? | 13:38 |
bauzas | edleafe: you have the existing architecture, the scheduler client reporting stats by writing a DB row, the scheduler reading that once and consuming that, and you have the RT claims | 13:38 |
bauzas | edleafe: you're planning to touch the 3 above things | 13:39 |
bauzas | edleafe: IIUC | 13:39 |
edleafe | bauzas: yes, it's all open to change in the experiment | 13:39 |
bauzas | edleafe: but what I'm missing is how you guarantee that we keep a continuous deployment | 13:40 |
bauzas | well, not 'CD', but rather a continuous delivery | 13:40 |
johnthetubaguy | bauzas: no of that really matters in this context right? | 13:40 |
edleafe | bauzas: ah, so you're getting back to the in-tree discussion | 13:40 |
*** tjones1 has joined #openstack-nova | 13:40 | |
edleafe | bauzas: we can talk about that, but only after it's clear that we want to go down that path | 13:41 |
bauzas | edleafe: well, isn't that the goal ? | 13:41 |
ndipanov | dansmith, what up with -2s? | 13:41 |
ndipanov | I can fix the patch | 13:41 |
*** tjones1 has left #openstack-nova | 13:41 | |
ndipanov | if we don't fix that we can't fix the version skip | 13:41 |
edleafe | bauzas: from my proposal: | 13:41 |
edleafe | "I would ask that we acknowledge for now that: | 13:41 |
dansmith | ndipanov: the patch proposes to fundamentally change the behavior, yes? | 13:41 |
edleafe | a) it will be disruptive and painful to switch something like this at | 13:41 |
edleafe | this point in Nova's development | 13:41 |
edleafe | b) it would have to provide *significant* improvement to make such a | 13:41 |
ndipanov | dansmith, pls move to -1 | 13:42 |
edleafe | change worthwhile | 13:42 |
ndipanov | dansmith, let me think if this will fix our case | 13:42 |
edleafe | bauzas: I know that any such change is going to be difficult | 13:42 |
edleafe | bauzas: I don't know that it will be worthwhile | 13:42 |
ndipanov | I think it would | 13:42 |
*** rajesht has joined #openstack-nova | 13:42 | |
edleafe | bauzas: the experiment is simply to answer the second point | 13:42 |
*** ferest has joined #openstack-nova | 13:44 | |
bauzas | edleafe: okay, then you wonder what is considered "significant", then ? | 13:44 |
bauzas | in other words, you need to define success KPIs ? | 13:44 |
ndipanov | dansmith, well it will not work in this case | 13:44 |
edleafe | bauzas: exactly | 13:44 |
bauzas | sorry, success criterias | 13:44 |
ndipanov | it would cause an infinet loop I think | 13:44 |
ndipanov | backport instance <- get the same thing <- call for backport <- get the same thing ... | 13:45 |
bauzas | edleafe: well, we don't have so many figures upstream, and I feel the best people to reply to that are operatoes | 13:45 |
bauzas | edleafe: you lost me by providing a technical description to a simple question : what do you consider an an improvement for the scheduler ? | 13:45 |
bauzas | edleafe: and like I said, the ones who know the best about that are ops | 13:46 |
ndipanov | dansmith, we could fix this with a two step approach | 13:46 |
dansmith | ndipanov: not if you make the change in the one that fixes the version, or if you drop the relationship first | 13:46 |
*** jdandrea has joined #openstack-nova | 13:46 | |
ndipanov | not sure I follow | 13:46 |
* bauzas is just melting on his chair | 13:47 | |
*** ferest has quit IRC | 13:47 | |
*** unicell has joined #openstack-nova | 13:47 | |
*** jlearman has quit IRC | 13:47 | |
bauzas | edleafe: I remember some good chat I had with a RAX operations guy at the summit | 13:48 |
*** rajesht has quit IRC | 13:48 | |
bauzas | edleafe: this guy was having good insights for helping how to improve the scheduler | 13:48 |
*** unicell1 has quit IRC | 13:48 | |
bauzas | someone from alaski's team IIRC | 13:48 |
*** exploreshaifali has quit IRC | 13:48 | |
edleafe | bauzas: I'd certainly like to help the RAX use case | 13:48 |
ndipanov | dansmith, we could merge a patch wehre we never sends 1.2 to kilo nodes | 13:48 |
ndipanov | then fix it in kilo | 13:48 |
ndipanov | then fix it in master again to send it to kilo nodes | 13:49 |
bauzas | edleafe: what I'm trying to explain is that if you need help for that, fire an email to openstack-operators@ and people would be glad to help you | 13:49 |
edleafe | bauzas: I might also need some of their cloud to test across global deployments | 13:49 |
ndipanov | but that serializer code is still busted | 13:49 |
*** irenab has quit IRC | 13:49 | |
bauzas | edleafe: just mention that you're planning to work on an experiment to identify what could be the benefits but you need to weigh the benefits | 13:49 |
bauzas | edleafe: not sure the ML thread on -dev@ is that clear too - I'll try to reply it asap | 13:50 |
*** bfic has quit IRC | 13:50 | |
edleafe | bauzas: sure. These are the sorts of things I want to come away from the midcycle with, so that I can start to create a plan | 13:50 |
bauzas | edleafe: not sure the midcycle can help more that that | 13:51 |
bauzas | than | 13:51 |
alaski | bauzas: not sure who you would have talked to, it could have been any of the ops folks as they're all pretty opinionated about the scheduler :) | 13:51 |
*** jlearman has joined #openstack-nova | 13:51 | |
edleafe | alaski: ya think? :) | 13:51 |
*** jcoufal has quit IRC | 13:52 | |
alaski | it's hard not to focus on the pain :) | 13:52 |
*** wanghua has quit IRC | 13:52 | |
alaski | edleafe: johnthetubaguy already outlined one use case, but more generally our interest is in reducing races when we're near capacity | 13:53 |
edleafe | alaski: good to know | 13:53 |
edleafe | alaski: I'll probably need your help designing a test that would show that | 13:53 |
alaski | in the current scheduler we have too much contention over a subset of the remaining capacity. our weighers don't do a great job of spreading requests | 13:53 |
dansmith | ndipanov: if you drop the relationship from master first, so that it will not send 1.2, then write a patch to fix the serializer to ask for the proper top version, add pcidevicelist 1.2, then re-add the relationship you should be good right? | 13:54 |
edleafe | alaski: "your" meaning RAX folks | 13:54 |
alaski | edleafe: yes | 13:54 |
johnthetubaguy | edleafe: alaski: yeah +1, the above scenario I was talking about focuses on that | 13:54 |
ndipanov | drop the relationship completely ? | 13:54 |
ndipanov | or just make sure to never send 1.2 to kilo nodes | 13:54 |
dansmith | ndipanov: it's not getting us anything right now, right? | 13:54 |
alaski | edleafe: bursts of builds are another class of problem, but it sort of boils down to the same thing. too much contention on a subset of available capacity | 13:55 |
*** dboik has joined #openstack-nova | 13:55 | |
ndipanov | what do you mean by drop the relationship? | 13:55 |
*** dboik has quit IRC | 13:55 | |
ndipanov | exactly | 13:55 |
*** dboik has joined #openstack-nova | 13:55 | |
ndipanov | dansmith, ^ | 13:55 |
*** mriedem_away is now known as mriedem | 13:56 | |
dansmith | ndipanov: remove it | 13:56 |
*** artom has joined #openstack-nova | 13:56 | |
*** yingjun has joined #openstack-nova | 13:56 | |
ndipanov | like delete the field from the instance object? | 13:57 |
*** signed8bit has joined #openstack-nova | 13:57 | |
dansmith | from pcidevicelist | 13:58 |
bauzas | alaski: that's one of the reasons I was working on a distributed lock, because that's mostly where the conditions at the limits are hit | 13:58 |
*** jichen has quit IRC | 13:58 | |
* ndipanov confused | 13:58 | |
* bauzas is probably writing Franglish | 13:58 | |
*** cfriesen has joined #openstack-nova | 13:58 | |
*** jichen has joined #openstack-nova | 13:58 | |
*** shuquan has quit IRC | 13:59 | |
*** redbeard has quit IRC | 13:59 | |
* edleafe is going to get more coffee | 14:00 | |
openstackgerrit | Shiina, Hironori proposed openstack/nova: Add a dirver API to inject an NMI https://review.openstack.org/202380 | 14:00 |
*** dboik_ has joined #openstack-nova | 14:00 | |
*** yamahata has joined #openstack-nova | 14:01 | |
*** thangp has quit IRC | 14:02 | |
ndipanov | dansmith, when you have some time pls explain what you meant in a bit more words... :) | 14:02 |
*** Kennan has quit IRC | 14:02 | |
*** alex_klimov has quit IRC | 14:03 | |
*** Kennan has joined #openstack-nova | 14:03 | |
*** dboik has quit IRC | 14:04 | |
alaski | bauzas: yes, a distributed lock would help. But I think that gets to edleafes point that some of this work is rebuilding a system like cassandra which brings that stuff in "for free" | 14:06 |
bauzas | alaski: hence johnthetubaguy's backlog spec :p | 14:06 |
bauzas | alaski: https://review.openstack.org/#/c/191914/ | 14:06 |
*** thangp has joined #openstack-nova | 14:07 | |
johnthetubaguy | Cassandra could be a cool alternative, for the reasons alaski mentions, its worth a quick punt | 14:07 |
*** cpschult has joined #openstack-nova | 14:08 | |
bauzas | heh, are we sure Greek-ish stuff are the best ? | 14:08 |
* bauzas jk | 14:08 | |
dansmith | ndipanov: first, we need to fix the serializer so it asks for the proper top-level version and backport that, yes? until that's done, the rest is moot | 14:09 |
ndipanov | nope | 14:09 |
ndipanov | if we do that | 14:09 |
ndipanov | then we would get an infinite loop | 14:09 |
ndipanov | since kilo serializer causes this | 14:09 |
ndipanov | so if we fix it and backport it | 14:10 |
ndipanov | then fix the version on master | 14:10 |
alaski | johnthetubaguy: bauzas to be clear I have no affinity towards Cassandra. Kafka or any other system would be just as interesting to try | 14:10 |
ndipanov | that fix will cause kilo serializer to keep sending back the instance when it fails on PciDeviceList | 14:10 |
ndipanov | and keep getting the same thing back | 14:10 |
dansmith | ndipanov: you mean because kilo doesn't know about the new pcidevice or pcidevicelist, right? | 14:11 |
alaski | bauzas: what I do really like is the consideration of how computes are reporting resources and how it might change, as right now it's heavily biased towards a filtering scheduler implementation, as it should given that it's what we have | 14:11 |
*** wxy has joined #openstack-nova | 14:11 | |
johnthetubaguy | alaski: FWIW, same here, I kinda like the idea of falling back to mysql, too, assuming thats simple, which it might well not be... | 14:11 |
*** jecarey has joined #openstack-nova | 14:11 | |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Specify current directory during processutils.execute using new cwd param https://review.openstack.org/191604 | 14:12 |
ndipanov | yes dansmith | 14:12 |
alaski | johnthetubaguy: what I really want to see would be a mysql implementation and a cassandra and a kafka and a ... implementataion that could all work very differently. which means the important bit is figuring out how to report resources | 14:13 |
alaski | johnthetubaguy: and not have a mysql fallback for a kafka implementation, just accept that they're different things | 14:13 |
*** mutoulbj has joined #openstack-nova | 14:13 | |
*** wxy_ has joined #openstack-nova | 14:14 | |
*** redbeard has joined #openstack-nova | 14:14 | |
johnthetubaguy | alaski: yeah true, "fallback" might best being "use the old system" well current system, I guess | 14:14 |
*** alex_klimov has joined #openstack-nova | 14:14 | |
*** wxy_ has left #openstack-nova | 14:15 | |
*** sudipto has joined #openstack-nova | 14:15 | |
dansmith | ndipanov: so if you fix the version in master, but don't add the relationship from instance to the new pcidevicelist, and then backport that change to kilo, then you can fix the relationship in master so we start sending the thing and kilo will accept it, yes? | 14:15 |
* bauzas regrets his non-presence at the midcycle, given that convo | 14:15 | |
*** claudiub has quit IRC | 14:15 | |
ndipanov | dansmith, yes that's what was my plan B | 14:15 |
*** alex_klimov has quit IRC | 14:15 | |
bauzas | alaski: johnthetubaguy: well, what I want to avoid is pulling some hard dependency which would make us a bit unsatisfied | 14:15 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Enable python34 tests for nova/tests/unit/scheduler/test*.py https://review.openstack.org/187968 | 14:15 |
bauzas | alaski: johnthetubaguy: take the SG drivers as an example | 14:16 |
bauzas | alaski: johnthetubaguy: so any improvement using a different and cool backend is +1 to me, but that shouldn't be the reason for the change | 14:16 |
*** mdrabe has quit IRC | 14:16 | |
*** ijuwang has joined #openstack-nova | 14:16 | |
bauzas | alaski: johnthetubaguy: we should conceptually fix our problems and have different SLAs based on backend features | 14:17 |
bauzas | exactly like a virt driver | 14:17 |
*** wxy has quit IRC | 14:17 | |
bauzas | all of that being MHO of course | 14:17 |
johnthetubaguy | bauzas: I think we agree, but sometimes you just need to try out ideas | 14:18 |
alaski | yeah, that's basically my thought as well | 14:18 |
edleafe | johnthetubaguy: +1 | 14:18 |
openstackgerrit | jichenjc proposed openstack/nova: Cleanup HTTPRequest for security_groups test https://review.openstack.org/182630 | 14:19 |
openstackgerrit | jichenjc proposed openstack/nova: Change group controller of V2 test cases https://review.openstack.org/182613 | 14:19 |
openstackgerrit | jichenjc proposed openstack/nova: Add schema check for security_groups https://review.openstack.org/182614 | 14:19 |
bauzas | johnthetubaguy: well, I do agree but while I consider it as valid, I also consider that sometimes we need to fix our interfaces first :) | 14:19 |
*** fifieldt_ has quit IRC | 14:19 | |
bauzas | johnthetubaguy: tbh, I really like POCs and I wish I could do one - but I also need to reduce our debt :) | 14:19 |
bauzas | just becase the current model is pretty broken | 14:20 |
*** xyang2 has joined #openstack-nova | 14:20 | |
alaski | bauzas: before bringing it in tree the interfaces would need to be fixed, but that's a ways off. and a PoC could help us understand what the interfaces should look like | 14:20 |
johnthetubaguy | alaski: +1 | 14:20 |
ndipanov | dansmith, but we should fix the deserialization regardless... | 14:21 |
*** unicell has quit IRC | 14:21 | |
bauzas | alaski: I do admit that :) | 14:21 |
dansmith | ndipanov: agreed, might as well get that in the pipeline | 14:22 |
bauzas | I feel sorry to be seen as conservative, while I really consider PoCs as good | 14:22 |
ndipanov | dansmith, but as we've seen just asking for a blanket top level version may not work and can bite us again | 14:22 |
ndipanov | so maybe we need something a bit more robust | 14:22 |
ndipanov | without causing potentially N RPC calls | 14:22 |
ndipanov | maybe walk the tree, see what's unsupported - make a single request | 14:23 |
dansmith | ndipanov: sure, so do you want to work on that in o.vo and port it over to nova before we fix the bug, or do you want to make it ask the proper toplevel right now so we can proceed? | 14:23 |
*** markvoelker_ has quit IRC | 14:24 | |
ndipanov | first I will fix the versions | 14:25 |
ndipanov | so even a simple version fix would need to go o.vo first, right? | 14:26 |
*** mriedem has quit IRC | 14:27 | |
jlearman | Hi, newb here, just getting myself set up to help out (I'll probably start by upgrading Nova UT testcases from mox to mock). I'm planning to use virtualbox with ubuntu on my Windows laptop, but haven't started down that path yet. Any reason that won't work, or is there a better suggestion for Windows? | 14:27 |
*** alex_klimov has joined #openstack-nova | 14:27 | |
*** mriedem has joined #openstack-nova | 14:27 | |
*** jichen has quit IRC | 14:29 | |
*** annegentle has joined #openstack-nova | 14:30 | |
*** jichen has joined #openstack-nova | 14:30 | |
dansmith | ndipanov: you mean asking for the proper toplevel, right? | 14:30 |
dansmith | ndipanov: if you want to do that in nova, I'll mirror it on the o.vo side if it looks clean and we can do it in parallel | 14:30 |
ndipanov | dansmith, since it's not on the critical path for this version thing | 14:31 |
ndipanov | I can do it in o.vo | 14:32 |
*** rfolco has quit IRC | 14:32 | |
*** sudipto has quit IRC | 14:32 | |
dansmith | it's not? maybe because you're getting an instance fault anyway? | 14:32 |
*** yingjun has quit IRC | 14:32 | |
ndipanov | so like I said | 14:32 |
ndipanov | well I think it's better to fix the versions first even if with 2 patches | 14:33 |
ndipanov | and then properly fix the serializer | 14:33 |
ndipanov | (walk the tree, backport all that are too new in one go) | 14:34 |
dansmith | oh, you mean that | 14:34 |
ndipanov | as opposed to just backport the latest top level | 14:34 |
ndipanov | because as is demonstrated by this issue | 14:35 |
dansmith | I want to fix the current serializer asking for the proper toplevel ASAP, because that's what we're currently depending on | 14:35 |
ndipanov | it still leaves a few edge cases open | 14:35 |
ndipanov | ok but that's still broken for the situation we are currently in | 14:35 |
openstackgerrit | Matthew Booth proposed openstack/nova: Implement compare-and-swap for instance update https://review.openstack.org/202593 | 14:35 |
dansmith | so I'll fix the toplevel issue | 14:35 |
ndipanov | and we still need a 2 step fix for the versions | 14:35 |
dansmith | yeah, I understand, but the toplevel version thing affects everyone | 14:36 |
alex_xu | andreykurilin: hi, when you begin vacation?, may I update your patch when your vacation? | 14:36 |
ndipanov | ok | 14:36 |
ndipanov | but it's still not the proper fix - that's all I'm saying | 14:36 |
dansmith | it's not going to fix you for this, I get that | 14:36 |
openstackgerrit | Shiina, Hironori proposed openstack/nova: Add a compute RPC API to inject an NMI https://review.openstack.org/202595 | 14:37 |
ndipanov | but yes less wrong than blatantly wrong is better for sure | 14:37 |
*** mdrabe has joined #openstack-nova | 14:37 | |
openstackgerrit | Mike Dorman proposed openstack/nova: Add console allowed origins setting https://review.openstack.org/201677 | 14:37 |
*** tonytan4ever has joined #openstack-nova | 14:38 | |
mdbooth | jaypipes mriedem johnthetubaguy: This version adds a retry when it detects the race instead of raising a different exception: https://review.openstack.org/#/c/202593/ | 14:38 |
jaypipes | mdbooth: yep, it's in the queu. | 14:38 |
jaypipes | queue... | 14:38 |
mdbooth | I thought you americans called them lines | 14:38 |
dansmith | mdbooth: not for this :) | 14:39 |
* mdbooth is confused :) | 14:39 | |
dansmith | heh | 14:39 |
johnthetubaguy | mdbooth: ah, cool, thanks for pushing on that | 14:40 |
*** lpetrut has quit IRC | 14:41 | |
ndipanov | so dansmith I also saw that we keep asking for that backport in one case for block device | 14:42 |
ndipanov | ah but that can happen when Block device mapping is sent over RPC | 14:43 |
ndipanov | cool cool | 14:43 |
ndipanov | nevermind | 14:43 |
*** jichen_ has joined #openstack-nova | 14:45 | |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtISCSIVolumeDriver into it's own module https://review.openstack.org/202301 | 14:46 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtISERVolumeDriver into it's own module https://review.openstack.org/197311 | 14:46 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtNETVolumeDriver into it's own module https://review.openstack.org/197341 | 14:46 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtNFSVolumeDriver into it's own module https://review.openstack.org/196859 | 14:46 |
*** jichen has quit IRC | 14:48 | |
*** jichen_ is now known as jichen | 14:48 | |
openstackgerrit | Shiina, Hironori proposed openstack/nova: Add a cells RPC API to inject an NMI https://review.openstack.org/202605 | 14:49 |
*** READ10 has quit IRC | 14:49 | |
*** rfolco has joined #openstack-nova | 14:51 | |
*** markvoelker has joined #openstack-nova | 14:53 | |
openstackgerrit | jichenjc proposed openstack/nova: Check flavor type before add tenant access https://review.openstack.org/124338 | 14:53 |
*** eglynn has joined #openstack-nova | 14:55 | |
*** jichen has quit IRC | 14:58 | |
*** Longgeek has quit IRC | 14:58 | |
*** Jeffrey4l has quit IRC | 14:58 | |
openstackgerrit | Nikola Dipanov proposed openstack/nova: Fix the incorrect PciDeviceList version number https://review.openstack.org/201280 | 14:58 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: objects: send PciDeviceList 1.2 to all code that can handle it https://review.openstack.org/202616 | 14:58 |
openstackgerrit | Shiina, Hironori proposed openstack/nova: Add a Nova API to inject an NMI https://review.openstack.org/202617 | 14:59 |
*** redbeard has quit IRC | 15:00 | |
*** Longgeek has joined #openstack-nova | 15:00 | |
*** ociuhandu has joined #openstack-nova | 15:00 | |
edleafe | jlearman: Yeah, that sounds like a good setup | 15:01 |
*** READ10 has joined #openstack-nova | 15:01 | |
andreykurilin | alex_xu: from Monday:) but I'll available in the evening first 3 days. feel free to change them, if I'll be unavailable:) | 15:02 |
openstackgerrit | Sudipta Biswas proposed openstack/nova: Allow NUMA based reporting for Monitors https://review.openstack.org/198972 | 15:02 |
openstackgerrit | Sudipta Biswas proposed openstack/nova: Base enablement for memory bandwidth monitor https://review.openstack.org/202623 | 15:02 |
*** karimb has joined #openstack-nova | 15:02 | |
andreykurilin | alex_xu: *from Monday and till 31 July | 15:03 |
jaypipes | dansmith: super easy one: https://review.openstack.org/#/c/202501/ | 15:03 |
dansmith | jaypipes: will look in a bit | 15:03 |
*** Murali has quit IRC | 15:03 | |
jroll | jaypipes: hey fyi I revved the ironic multiple compute spec | 15:04 |
jroll | going to try to get initial code before next week | 15:04 |
*** mriedem is now known as mriedem_away | 15:04 | |
jaypipes | jroll: coolio. thank you sir! | 15:04 |
*** sudipto has joined #openstack-nova | 15:04 | |
jroll | np | 15:04 |
*** alex_klimov has quit IRC | 15:06 | |
*** coolsvap|away is now known as coolsvap | 15:06 | |
*** Longgeek has quit IRC | 15:07 | |
*** tpeoples has joined #openstack-nova | 15:08 | |
*** mnestratov has quit IRC | 15:09 | |
*** lyanchih has quit IRC | 15:09 | |
*** sudipto_ has joined #openstack-nova | 15:09 | |
*** amotoki has joined #openstack-nova | 15:10 | |
*** vladikr has quit IRC | 15:10 | |
*** Jeffrey4l has joined #openstack-nova | 15:11 | |
*** unicell has joined #openstack-nova | 15:12 | |
*** sudipto has quit IRC | 15:12 | |
*** belmoreira has quit IRC | 15:13 | |
*** otter768 has joined #openstack-nova | 15:14 | |
*** redbeard has joined #openstack-nova | 15:15 | |
*** mutoulbj has quit IRC | 15:16 | |
*** hshiina has left #openstack-nova | 15:16 | |
*** dannywilson has joined #openstack-nova | 15:16 | |
*** mnestratov has joined #openstack-nova | 15:17 | |
openstackgerrit | Andrea Rosa proposed openstack/nova: Manage an empty machine-id file https://review.openstack.org/202640 | 15:18 |
*** dguryanov|2 has joined #openstack-nova | 15:18 | |
*** hemna has joined #openstack-nova | 15:18 | |
*** otter768 has quit IRC | 15:18 | |
*** armax has joined #openstack-nova | 15:18 | |
*** afazekas has quit IRC | 15:20 | |
*** hemna has quit IRC | 15:20 | |
dansmith | PaulMurray: around? | 15:20 |
dansmith | jaypipes: on that conductor api thing, I don't get why we need to add the object mocks if we're just removing unused conductor mocks | 15:24 |
dansmith | jaypipes: it's not like the object calls get turned into conductor calls and we can catch them with either type of mock... | 15:25 |
jaypipes | dansmith: hmm. | 15:25 |
dansmith | I guess I'll comment on the review and see if he pops up | 15:26 |
jaypipes | dansmith: so, for the test_no_op_init_compute_node() one, PaulMurray's code is correcting a bad unit test. The create_mock and service_mock objects are being tested on lines 819-820 to ensure that they are not called. Previously, the test was passing because the mocks were the completely wrong thing (the conductor calls), but now, the tests are passing when referring to the correct thing (ComputeNode.create() and Servi | 15:29 |
jaypipes | ce.get_by_compute_host) | 15:29 |
*** rdopiera has quit IRC | 15:29 | |
*** tsekiyama has joined #openstack-nova | 15:30 | |
openstackgerrit | Matthew Booth proposed openstack/nova: Implement compare-and-swap for instance update https://review.openstack.org/202593 | 15:30 |
jaypipes | dansmith: for the test_compute_node_loaded() test, same thing. line 855 is testing that the ComputeNode.create() method is NOT called, which is the correct assertion. the test was previously passing because the wrong thing was being mocked (and of course, the assertion that the mock was not called was passing as a false positive) | 15:30 |
openstackgerrit | Stephen Finucane proposed openstack/nova: Add 'hw_cpu_threads_policy' image metadata prop https://review.openstack.org/202647 | 15:31 |
openstackgerrit | Stephen Finucane proposed openstack/nova: Add 'cpu_policy' and 'cpu_threads_policy' fields https://review.openstack.org/202648 | 15:31 |
openstackgerrit | Stephen Finucane proposed openstack/nova: Add 'hw:cpu_threads_policy=avoid' filtering https://review.openstack.org/202649 | 15:31 |
openstackgerrit | Stephen Finucane proposed openstack/nova: trivial: Add some logs to 'numa_topology_filter' https://review.openstack.org/202650 | 15:31 |
openstackgerrit | Stephen Finucane proposed openstack/nova: Add 'hw:cpu_threads_policy=require' scheduling https://review.openstack.org/202651 | 15:31 |
openstackgerrit | Stephen Finucane proposed openstack/nova: Add 'hw:cpu_threads_policy=separate' scheduling https://review.openstack.org/202652 | 15:31 |
openstackgerrit | Stephen Finucane proposed openstack/nova: Make use of 'InstanceNUMACell.cpu_policy' field https://review.openstack.org/202653 | 15:31 |
openstackgerrit | Stephen Finucane proposed openstack/nova: trivial: Make use of 'CPUAllocation' enum values https://review.openstack.org/202654 | 15:31 |
openstackgerrit | Stephen Finucane proposed openstack/nova: Add 'hw:cpu_threads_policy=isolate' scheduling https://review.openstack.org/202655 | 15:31 |
*** thangp has quit IRC | 15:31 | |
dansmith | jaypipes: my point is, the objects don't make the conductor calls | 15:32 |
dansmith | jaypipes: so if the code is using the objects, it should have been failing before without those mocks | 15:32 |
dansmith | jaypipes: but the patch doesn't change the code, just the test | 15:32 |
jaypipes | dansmith: the patch corrects the tests that were previously referring to conductor calls and providing false positive assertions about not being called. | 15:33 |
openstackgerrit | sahid proposed openstack/nova: hardware: take into account pages size when computing memory https://review.openstack.org/202504 | 15:34 |
*** vladikr has joined #openstack-nova | 15:34 | |
*** gyee has quit IRC | 15:34 | |
jaypipes | dansmith: and the patch simply removes the conducto_api attribute from the resource tracker. | 15:34 |
dansmith | jaypipes: it doesn't. all it changes are what is mocked out, but the new mocks and the old mocks aren't in the same code path | 15:34 |
dansmith | jaypipes: let me pull this down and play with it and maybe I can make my point | 15:34 |
jaypipes | ok | 15:34 |
*** thangp has joined #openstack-nova | 15:34 | |
*** ociuhandu has quit IRC | 15:35 | |
andreykurilin | sdague: Fix for stable/kilo branch of novaclient is ready: https://review.openstack.org/202656 | 15:36 |
sdague | andreykurilin: thanks | 15:36 |
*** scheuran has quit IRC | 15:36 | |
*** Mike_D_laptop has joined #openstack-nova | 15:37 | |
dansmith | jaypipes: I assume it's because we assert that a bunch of things are not called. So we can mock them out for no reason and they're still ... not called :) | 15:37 |
*** andreykurilin__ has joined #openstack-nova | 15:38 | |
*** gyee has joined #openstack-nova | 15:39 | |
*** andreykurilin has quit IRC | 15:39 | |
*** MentalRay__ has joined #openstack-nova | 15:39 | |
*** ig0r_ has quit IRC | 15:40 | |
sdague | dansmith: what happens to the symbolic compute levels (i.e. kilo) if we backport an rpc change to stable kilo? | 15:40 |
sdague | do those get bumped? | 15:40 |
*** rgerganov_ has joined #openstack-nova | 15:41 | |
garyk | mriedem_away: you around? | 15:41 |
*** sudipto_ has quit IRC | 15:41 | |
garyk | we have a serious problem with you have a sec ti talk about https://github.com/openstack/nova/commit/1ba72081715869fe6d0ffba3436fb09ef9499335? | 15:41 |
garyk | not a serious problem with you :) | 15:41 |
garyk | just a serious problem with the change that you made. https://github.com/openstack/nova/commit/1ba72081715869fe6d0ffba3436fb09ef9499335? | 15:42 |
*** dprince has quit IRC | 15:42 | |
garyk | problem is that when an instance gets into error state it can never be deleted… | 15:42 |
dansmith | sdague: well, in general we don't backport rpc stuff unless it's absolutely critical | 15:42 |
dansmith | sdague: if we have to, we do it as a .z version.. updating the symbol probably depends on what the change is.. | 15:43 |
*** ivasev has quit IRC | 15:43 | |
sdague | ok, I was just thinking of our upgrade testing, which uses the symbol to pin | 15:43 |
dansmith | sdague: can I ask what the change is before I commit? :) | 15:44 |
sdague | it came up in doing grenade multinode - https://review.openstack.org/#/c/199091/ | 15:44 |
sdague | I realized that pinning everyone means they might not be using the latest rpc version on stable | 15:45 |
sdague | which is the same issue on the pcpu jobs | 15:45 |
dansmith | sdague: so is this hypothetical? | 15:45 |
sdague | also, man, did I wish this was in the db instead of in the config | 15:45 |
sdague | yes, it's a hypothetical question | 15:45 |
*** vishwanathj has joined #openstack-nova | 15:45 | |
dansmith | okay, so first off: https://review.openstack.org/#/c/201733/ | 15:46 |
dansmith | second, if we backport a thing, we do it as a .z and it has to be like fully compatible with the .0 so the pin doesn't matter as much.. we'd be backporting a fix, not a feature, so that should be doable | 15:46 |
dansmith | moving the pin doesn't really matter because we discard the .z for compatibility checking | 15:47 |
sdague | dansmith: ok, so https://review.openstack.org/#/c/201733/ kind of excites me. You imagining an admin call to set that in the cluster? | 15:47 |
*** vishwana_ has quit IRC | 15:47 | |
*** rgerganov_ has quit IRC | 15:48 | |
dansmith | sdague: no, it'll be determined by code levels, but the services will be able to decide what to do based on the quorum | 15:48 |
dansmith | sdague: so, the new rpc version starts being used when you upgrade the last service | 15:48 |
*** lpetrut has joined #openstack-nova | 15:49 | |
dansmith | sdague: in order to make that work for non-rpc things, we need to register api services, which we don't currently do | 15:49 |
*** markus_z has quit IRC | 15:49 | |
dansmith | but for rpc pinning purposes it should be enough as-is I think | 15:49 |
dansmith | we need to land that in lemming so we can use it in mitasktska-whatever-it-is | 15:49 |
sdague | dansmith: it will remove the need for the config file bit though, right? | 15:50 |
dansmith | yeah | 15:50 |
sdague | I'm ok with automatic quorum | 15:50 |
sdague | though what happens with stale compute_node entries? | 15:50 |
dansmith | it also means that if we start up and we're older than the oldest version in the db, we can punt on startup | 15:50 |
sdague | right, yes, all sounds very reasonable. ++ | 15:50 |
dansmith | sdague: yeah, those'll need to be removed | 15:50 |
sdague | well, I wonder if we provide an admin API call to just force it, which would flag all the non upgraded compute nodes as down | 15:51 |
dansmith | not computenode, but service | 15:51 |
sdague | sure, service | 15:51 |
dansmith | well, that doesn't really work | 15:52 |
*** annegentle has quit IRC | 15:52 | |
dansmith | because if the thing is still up and old, and you just decide it's too old, you need to fence it or shut it down or whatever | 15:52 |
dansmith | otherwise it'll still consume messages from shared queues and such | 15:52 |
sdague | it would under the circumstances that it is actually down and not coming back | 15:52 |
*** annegentle has joined #openstack-nova | 15:52 | |
dansmith | yes | 15:52 |
sdague | but has stale records | 15:52 |
sdague | which I would imagine would happen a reasonable amount | 15:53 |
dansmith | but we can also just make it not start up when it comes back if it's older than the minimum version | 15:53 |
sdague | right, but the service entries don't auto delete right? | 15:53 |
dansmith | or rather, older than the oldest still-running node | 15:53 |
sdague | if the node died | 15:53 |
sdague | and it's not coming back | 15:53 |
dansmith | no, but we have a notion of liveness | 15:53 |
sdague | ok | 15:53 |
sdague | anyway, right, there might be an edge here, but that's all just details and bugs | 15:53 |
dansmith | but yes, we still need to not require operators to manually clean up service records for things that are never coming back for sure | 15:54 |
sdague | ++ on approach | 15:54 |
dansmith | it's all edge cases I think, but yeah :) | 15:54 |
sdague | yeh, but they mostly can be worked through in sane ways | 15:54 |
dansmith | I think so | 15:54 |
openstackgerrit | heijlong proposed openstack/nova: Add microversion to allow uuid in server show response data https://review.openstack.org/202665 | 15:54 |
*** paul-carlton has joined #openstack-nova | 15:56 | |
park_heijlong_ | johnthetubaguy: I finished the last patch for my bp, appreciated for your comments, make you update, thank you | 15:57 |
*** ddaskal has joined #openstack-nova | 15:58 | |
*** ctrath has quit IRC | 15:59 | |
openstackgerrit | Matthew Booth proposed openstack/nova: Thoughts on atomic host/node transfer during evacuate https://review.openstack.org/202667 | 15:59 |
*** jogo has quit IRC | 15:59 | |
mdbooth | dansmith: Update host and node atomically during evacuate: https://review.openstack.org/#/c/202667/1 | 16:00 |
mdbooth | It's a WIP, very untidy, unit test failures, etc | 16:00 |
mdbooth | Interesting bits are in compute/api.py and compute/manager.py | 16:01 |
*** angdraug has joined #openstack-nova | 16:01 | |
mdbooth | Only a couple of lines | 16:01 |
dansmith | okay | 16:02 |
*** ctrath has joined #openstack-nova | 16:02 | |
*** jlanoux has quit IRC | 16:02 | |
mdbooth | dansmith: That's literally all I'm going to do with it for the moment :) | 16:02 |
mdbooth | I have no current plans to develop that, it's just an idea, with code | 16:02 |
*** dboik_ has quit IRC | 16:03 | |
dansmith | seems reasonable | 16:04 |
*** coolsvap is now known as coolsvap|away | 16:05 | |
dansmith | might be easier to make objects track the previous version and just let you say "only do this if $attr is what we previously knew it to be" | 16:05 |
dansmith | i.e. save(consistent=['vm_state', 'task-state', 'host', 'node']) | 16:05 |
*** lpetrut has quit IRC | 16:05 | |
dansmith | well, maybe s/easier/cleaner/ | 16:06 |
mdbooth | Yeah | 16:06 |
mdbooth | Although expected can take a list | 16:06 |
PaulMurray | hi dansmith - you had a question about the mocks | 16:06 |
dansmith | PaulMurray: yeah, I commented | 16:06 |
mdbooth | So task_state: ['REBUILDING', 'SPAWNING'] | 16:06 |
*** park_heijlong_ has quit IRC | 16:06 | |
dansmith | mdbooth: true, but it's a little different than host and node and other general purpose things | 16:07 |
mdbooth | Yup | 16:07 |
tdurakov | bauzas, Hi Sylvain, are you around? | 16:07 |
*** mutoulbj has joined #openstack-nova | 16:07 | |
*** sahid has quit IRC | 16:07 | |
*** ijw has quit IRC | 16:07 | |
bauzas | tdurakov: not really :/ | 16:08 |
bauzas | pretty busy | 16:08 |
*** newbie_2 has joined #openstack-nova | 16:09 | |
tdurakov | bauzas, short question, how scheduler refactoring is going? what about this one https://bugs.launchpad.net/nova/+bug/1214943 ? | 16:09 |
openstack | Launchpad bug 1214943 in OpenStack Compute (nova) "Live migration should use the same memory over subscription logic as instance boot" [High,Confirmed] - Assigned to Sylvain Bauza (sylvain-bauza) | 16:09 |
mdbooth | dansmith: anyway, I know you were robustifying it, thought that could be a useful tool | 16:09 |
mdbooth | Let me know if you think it's worth spending time on | 16:09 |
*** nic has joined #openstack-nova | 16:10 | |
*** tjones1 has joined #openstack-nova | 16:12 | |
*** sudipto has joined #openstack-nova | 16:12 | |
*** oro has quit IRC | 16:13 | |
*** whenry has joined #openstack-nova | 16:13 | |
*** mriedem_away is now known as mriedem | 16:13 | |
*** kmARC has quit IRC | 16:13 | |
mriedem | garyk: that would have been a problem before that revert too | 16:13 |
garyk | mriedem: ah, ok. i think that the issue is that the task state is not being reverted. i am currenly debugging | 16:14 |
mriedem | garyk: if you have a stack trace you could see if whatever method is being hit in the compute manager, if it gets to the compute manager, has the @reverts_task_state decorator | 16:14 |
mriedem | however, if it's only the compute_api.delete that fails and puts it in error state, then yeah that could be a problem | 16:14 |
garyk | mriedem: i am checking that the reverts task works now | 16:15 |
mriedem | garyk: fwiw, ankit re-proposed that change https://review.openstack.org/#/c/194063/ | 16:16 |
*** jogo has joined #openstack-nova | 16:16 | |
mriedem | but this ^ supposedly has the race issue in mind | 16:16 |
bauzas | warning, paste bomb | 16:16 |
bauzas | or rather, gerrit bomb | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Scheduler RPC API select_dests() https://review.openstack.org/199205 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Prepare filters for using RequestSpec object https://review.openstack.org/197912 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec in the ChanceScheduler https://review.openstack.org/199051 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Add RequestSpec methods for primitiving into dicts https://review.openstack.org/187197 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec object in HostManager https://review.openstack.org/191251 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Use the same pci_requests field for all filters and HostManager https://review.openstack.org/191250 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Remove unnecessary method in FilterScheduler https://review.openstack.org/188485 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Consider that all scheduler calls are IO Ops https://review.openstack.org/188486 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Create RequestSpec object https://review.openstack.org/145528 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec object in the FilterScheduler https://review.openstack.org/188487 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Add two fields to ImageMetaProps object https://review.openstack.org/197911 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Affinity filter for RequestSpec https://review.openstack.org/202673 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Aggregate filters for RequestSpec https://review.openstack.org/202674 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify default filters for RequestSpec https://review.openstack.org/202675 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify metric-related filters for RequestSpec https://review.openstack.org/202676 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify NUMA, PCI and num_instances filters for RequestSpec https://review.openstack.org/202677 | 16:17 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify left filters for RequestSpec https://review.openstack.org/202678 | 16:17 |
*** MentalRay__ has quit IRC | 16:17 | |
garyk | mriedem: ok, i'll check that | 16:17 |
* bauzas competing for the Longest Ever Patch Series | 16:17 | |
garyk | mriedem: nah, it still has that dodgy code: if original_task_state in (task_states.DELETING, | 16:18 |
garyk | task_states.SOFT_DELETING): | 16:18 |
*** thorst has quit IRC | 16:18 | |
*** rgerganov_ has joined #openstack-nova | 16:18 | |
*** sambol has joined #openstack-nova | 16:19 | |
*** ZZelle has quit IRC | 16:20 | |
*** chuckcarmack has quit IRC | 16:20 | |
*** ZZelle has joined #openstack-nova | 16:20 | |
*** dboik has joined #openstack-nova | 16:20 | |
*** chuckcarmack has joined #openstack-nova | 16:20 | |
*** zsais has joined #openstack-nova | 16:20 | |
*** sudipto has quit IRC | 16:21 | |
jaypipes | dansmith: right, that's what I was trying (badly) to explain above... | 16:21 |
ericksonsantos | whooaa | 16:22 |
ericksonsantos | (about those changes ^) | 16:22 |
dansmith | jaypipes: you still think the patch is correct? | 16:22 |
jaypipes | dansmith: yes. | 16:23 |
chuckcarmack | dims_: Would it be ok I if worked on https://bugs.launchpad.net/nova/+bug/1434866 as a newbie? Its low hanging fruit. | 16:24 |
openstack | Launchpad bug 1434866 in OpenStack Compute (nova) "nova-cells and nova-idmapshift missing in documentation" [Low,Triaged] | 16:24 |
dims_ | chuckcarmack: +1 go for it :) | 16:24 |
chuckcarmack | dims_: thanks, I'll give it a try | 16:24 |
dansmith | jaypipes: PaulMurray: the patch is making up for the fact that the patch to convert to objects didn't adjust the tests, right? | 16:24 |
garyk | mriedem: the revert task state is called - problem is that this does not actually set the task state to None… man this is so broken | 16:25 |
jaypipes | dansmith: correct. | 16:25 |
dims_ | chuckcarmack: it's not assigned to anyone so it's fair game, i had a review earlier, feel free to use anything from it which you feel useful | 16:25 |
dansmith | jaypipes: so that's a change that has nothing to do with "remove conductor api from resource_tracker" | 16:25 |
dansmith | jaypipes: it's confusing because it seems like we're replacing conductor with objects, but we're not | 16:25 |
chuckcarmack | dims_: yes, you have a good start on it | 16:25 |
dansmith | jaypipes: so IMHO, there should be a separate patch that says "in commit $foo we forgot to change the tests, this fixes that" and then let the conductor removal stand on its own | 16:26 |
jaypipes | dansmith: well, the actual code change does indeed remove the conductor_api attribute frmo the resource tracker. | 16:26 |
*** edmondsw has quit IRC | 16:26 | |
jaypipes | dansmith: sure, that would be fine with me as well. | 16:26 |
dansmith | jaypipes: right, but it has nothing to do with the object mock changes | 16:26 |
jaypipes | sure, understood. | 16:26 |
PaulMurray | dansmith, yes - only caught it when I removed the conductor api from resource tracker because they were using it to set up the mocks | 16:26 |
dansmith | yeah | 16:26 |
*** redbeard has quit IRC | 16:27 | |
PaulMurray | dansmith, the tests work without the mocks - the mocks are there to check it is doing (or not doing) the calls it should (or shouldn't) | 16:27 |
dansmith | if you want to keep them as one patch, please call it out very specifically in the commit message, but I'd still prefer two | 16:27 |
dansmith | PaulMurray: yeah, I know | 16:27 |
PaulMurray | sorry - I should read back | 16:27 |
dansmith | the problem is that it's confusing because it seems like you're removing the conductor usage and replacing it with objects usage, and you're not | 16:27 |
dansmith | and since the test is testing for things that never happen, I think it'd be better to be super explicit about it | 16:28 |
*** lpetrut has joined #openstack-nova | 16:28 | |
*** achanda has joined #openstack-nova | 16:28 | |
dansmith | because later when that test is failing, and this is the last patch that touched it.. we won't remember that the previously patch actually made the change and this one just silently cleaned up the tests | 16:28 |
PaulMurray | well, actually the tests test for something that should never happen. If the code is wrong they might catch it | 16:28 |
*** achanda has quit IRC | 16:29 | |
dansmith | right, isn't that what I said? :) | 16:29 |
*** dims_ has quit IRC | 16:30 | |
PaulMurray | This patch was intended to remove conductor_api from resource tracker. In doing that I found those tests broke because they were wrong - so I corrected them | 16:31 |
*** armax has quit IRC | 16:31 | |
PaulMurray | Do you want me to change the tests too? | 16:32 |
*** dims_ has joined #openstack-nova | 16:32 | |
*** dims_ has quit IRC | 16:32 | |
dansmith | PaulMurray: I think you should fix the tests in patch 1, and remove conductor in patch 2 | 16:32 |
PaulMurray | :) | 16:32 |
PaulMurray | my patches always become 10 patches, even the easy ones | 16:33 |
*** rlrossit has left #openstack-nova | 16:33 | |
dansmith | PaulMurray: and jaypipes and I will +2 them super quick | 16:33 |
dansmith | PaulMurray: well, here's the other option | 16:33 |
dansmith | I can just shut up and you can get someone else to +2 :) | 16:33 |
*** tpeoples has quit IRC | 16:33 | |
*** mwagner_lap has quit IRC | 16:34 | |
PaulMurray | oh all right then, let me see what I can do...... | 16:34 |
*** tonytan4ever has quit IRC | 16:35 | |
dansmith | PaulMurray: just don't worry about it | 16:35 |
*** gyee has quit IRC | 16:35 | |
dansmith | I'll do it myself | 16:35 |
dansmith | it'll take less time than arguing | 16:35 |
dansmith | I already have the thing pulled down anywa | 16:35 |
*** achanda has joined #openstack-nova | 16:35 | |
*** vilobhmm has joined #openstack-nova | 16:36 | |
*** mutoulbj has quit IRC | 16:37 | |
*** redbeard has joined #openstack-nova | 16:37 | |
*** lpetrut has quit IRC | 16:38 | |
edleafe | dansmith: hey, if I argue with you, will you do my work, too? :) | 16:40 |
*** jogo has quit IRC | 16:40 | |
*** hemnafk is now known as hemna | 16:41 | |
*** jogo has joined #openstack-nova | 16:41 | |
PaulMurray | edleafe, you have to get to a special place with dansmith - It took me 2 years of arguing :) | 16:43 |
ericksonsantos | lol | 16:43 |
edleafe | PaulMurray: thanks for the tip! :) | 16:43 |
openstackgerrit | Dan Smith proposed openstack/nova: Remove conductor api from resource tracker https://review.openstack.org/202501 | 16:44 |
openstackgerrit | Dan Smith proposed openstack/nova: Fix test_tracker object mocks https://review.openstack.org/202690 | 16:44 |
*** rwsu has quit IRC | 16:44 | |
PaulMurray | dansmith, thanks a lot - I have to rush to get my car from servicing before they lock me out | 16:44 |
*** nic has quit IRC | 16:44 | |
dansmith | PaulMurray: np | 16:44 |
*** annashen has joined #openstack-nova | 16:45 | |
dansmith | edleafe: at least you know that my motive is just to get it right and not to be nitpicky for the sake of it because it costs me nothing | 16:45 |
edleafe | dansmith: of course - just having a little fun with ya | 16:46 |
*** lucasagomes is now known as lucas-dinner | 16:47 | |
tdurakov | mriedem, hi | 16:49 |
bauzas | jaypipes: FYI https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/request-spec-object,n,z becomes very fat, it would be awesome if you could review some of them :) | 16:49 |
tdurakov | mriedem, could you take a look at https://review.openstack.org/#/c/184495/ ansd subsequent one? | 16:50 |
*** y_sawai has joined #openstack-nova | 16:52 | |
*** thorst has joined #openstack-nova | 16:52 | |
jaypipes | bauzas: will do this evening. | 16:52 |
*** mutoulbj has joined #openstack-nova | 16:52 | |
openstackgerrit | Nikola Dipanov proposed openstack/nova: Allow for migration object to be passed to _move_claim https://review.openstack.org/199679 | 16:53 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: rt: generalize claim code to be useful for other move actions https://review.openstack.org/197138 | 16:53 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: rt: move filtering of mygration by type lower in the call stack https://review.openstack.org/199678 | 16:53 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: compute: Make rebuild_instance RPC take an optional migration https://review.openstack.org/200485 | 16:53 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: Define migration states and transitions and use when changing status https://review.openstack.org/197669 | 16:53 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: Introduce a simple state machine module https://review.openstack.org/197668 | 16:53 |
*** tpeoples has joined #openstack-nova | 16:54 | |
ndipanov | mriedem, https://review.openstack.org/#/c/201280/ | 16:54 |
ndipanov | still waiting for the check though :( | 16:54 |
garyk | mriedem: problem is that there is an exception in the compute node - for example unable to delete disks. the task state is updated correctly. subsequent deletes are not sent to the compute node. still trying to understand why. problem is restarting services hs it working | 16:55 |
ndipanov | but would be good to land this and the kilo backport soon if possible | 16:55 |
sdague | johnthetubaguy: if you are still awake - https://review.openstack.org/#/c/152703 was the provider networks in the config drive thing that infra wanted last cycle, but missed the deadline. Curious on your thinking | 16:55 |
*** y_sawai has quit IRC | 16:55 | |
johnthetubaguy | sdague: ah, yeah | 16:56 |
*** romainh has left #openstack-nova | 16:56 | |
johnthetubaguy | sdague: we could raise the priority in the hope it gets more attention? | 16:57 |
sdague | johnthetubaguy: sure, though, honestly, it's a single review, and you were one of the ones that were interested in that particular case | 16:58 |
*** rgerganov_ has quit IRC | 16:58 | |
sdague | I was poking you as an interested reviewer, not as ptl :) | 16:58 |
*** oro has joined #openstack-nova | 17:00 | |
*** aix has quit IRC | 17:00 | |
*** derekh has quit IRC | 17:00 | |
*** kmARC has joined #openstack-nova | 17:00 | |
*** jistr has quit IRC | 17:00 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/200708 | 17:04 |
*** ildikov has quit IRC | 17:05 | |
*** ssurana has joined #openstack-nova | 17:09 | |
*** jcoufal has joined #openstack-nova | 17:09 | |
openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: do relative block rebase only with non-null base https://review.openstack.org/202442 | 17:10 |
openstackgerrit | Ming Yang proposed openstack/nova: Case-insensitive filters https://review.openstack.org/201725 | 17:13 |
openstackgerrit | Ming Yang proposed openstack/nova: Refactor get_filtered_hosts in filter host manager https://review.openstack.org/192878 | 17:13 |
*** krtaylor has quit IRC | 17:13 | |
*** armax has joined #openstack-nova | 17:14 | |
*** otter768 has joined #openstack-nova | 17:14 | |
openstackgerrit | Dan Smith proposed openstack/nova: Fix test_tracker object mocks https://review.openstack.org/202690 | 17:16 |
openstackgerrit | Dan Smith proposed openstack/nova: Remove conductor api from resource tracker https://review.openstack.org/202501 | 17:17 |
*** Maike has quit IRC | 17:17 | |
*** haypo has left #openstack-nova | 17:18 | |
*** IanGovett1 has joined #openstack-nova | 17:18 | |
*** slong_ has quit IRC | 17:19 | |
*** otter768 has quit IRC | 17:19 | |
*** signed8bit has quit IRC | 17:19 | |
dansmith | jaypipes: just fixed commit message wrapping and added co-authored for PaulMurray: https://review.openstack.org/#/c/202690/2' | 17:19 |
*** SourabhP has joined #openstack-nova | 17:19 | |
*** slong_ has joined #openstack-nova | 17:19 | |
*** IanGovett has quit IRC | 17:19 | |
*** signed8bit has joined #openstack-nova | 17:20 | |
jaypipes | dansmith: pushed. | 17:20 |
*** xyang2 has quit IRC | 17:20 | |
*** xyang2 has joined #openstack-nova | 17:21 | |
*** achanda has quit IRC | 17:25 | |
*** mtanino has joined #openstack-nova | 17:25 | |
bauzas | jaypipes: ack, thanks | 17:25 |
*** rfolco has quit IRC | 17:26 | |
*** y_sawai has joined #openstack-nova | 17:27 | |
*** vilobhmm has quit IRC | 17:28 | |
*** vilobhmm has joined #openstack-nova | 17:28 | |
*** karimb has quit IRC | 17:30 | |
*** yamahata has quit IRC | 17:31 | |
openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: resize ploop disks with prl_disk_tool https://review.openstack.org/182390 | 17:32 |
*** dims_ has joined #openstack-nova | 17:32 | |
*** annegentle has quit IRC | 17:33 | |
*** armax has quit IRC | 17:34 | |
*** exploreshaifali has joined #openstack-nova | 17:34 | |
*** achanda has joined #openstack-nova | 17:35 | |
openstackgerrit | Vilobh Meshram proposed openstack/nova: WIP : Servicegroup foundational refactoring for Control Plane https://review.openstack.org/202714 | 17:36 |
johnthetubaguy | sdague: ah, sorry, yeah, was attempting to multitask and failed | 17:36 |
*** casusbelli has quit IRC | 17:36 | |
openstackgerrit | Moshe Levi proposed openstack/nova: Adding support for InfiniBand SR-IOV vif type https://review.openstack.org/187052 | 17:38 |
*** dims_ has quit IRC | 17:39 | |
*** achanda has quit IRC | 17:40 | |
*** annashen has quit IRC | 17:40 | |
*** davideagnello has quit IRC | 17:40 | |
vilobhmm | johnthetubaguy : ping | 17:41 |
*** annashen has joined #openstack-nova | 17:41 | |
*** lucas-dinner has quit IRC | 17:43 | |
*** zhenguo has quit IRC | 17:43 | |
*** achanda has joined #openstack-nova | 17:46 | |
*** krtaylor has joined #openstack-nova | 17:46 | |
johnthetubaguy | vilobhmm: hi | 17:47 |
*** annashen has quit IRC | 17:47 | |
*** jcoufal has quit IRC | 17:48 | |
*** edmondsw has joined #openstack-nova | 17:49 | |
*** annashen has joined #openstack-nova | 17:49 | |
*** igordcard has joined #openstack-nova | 17:50 | |
*** edmondsw has quit IRC | 17:51 | |
*** edmondsw has joined #openstack-nova | 17:51 | |
openstackgerrit | Matt Riedemann proposed openstack/nova: Deprecate the osapi_v3.enabled option https://review.openstack.org/202724 | 17:52 |
*** moshele has quit IRC | 17:52 | |
*** dims_ has joined #openstack-nova | 17:52 | |
*** dims_ has quit IRC | 17:53 | |
vilobhmm | johnthetubaguy : regarding the code freeze for L release ; will WIP suffice to the creteria for a feature to be accepted for L release….because as per your update code needs to be up by 16th and needs to be merged by july 30th…i have one feature where i am missing unit tests (which I will try but might not be able to complete by 16th) and hence proposed them as WIP…so thought will check with you | 17:53 |
*** lucas-dinner has joined #openstack-nova | 17:56 | |
*** gyee has joined #openstack-nova | 17:57 | |
*** dims_ has joined #openstack-nova | 17:57 | |
*** paul-carlton has quit IRC | 17:58 | |
dguryanov|2 | Hello! | 17:59 |
*** y_sawai has quit IRC | 17:59 | |
*** tjones1 has quit IRC | 18:00 | |
dguryanov|2 | Could someone, please, look at https://review.openstack.org/#/c/202442/ . It a fix for a bug, introduced few days ago. Several cinder CI's are blocked by it. | 18:00 |
dims_ | mriedem: getting my pm? :) | 18:00 |
*** amotoki has quit IRC | 18:00 | |
*** rfolco has joined #openstack-nova | 18:01 | |
dansmith | dims_: earlier he told me he was busy | 18:02 |
dims_ | dansmith: thanks! | 18:02 |
*** Kennan has quit IRC | 18:02 | |
dansmith | but he might've just been avoiding me | 18:02 |
dims_ | haha | 18:02 |
dansmith | (for obvious reasons) | 18:02 |
*** dboik has quit IRC | 18:02 | |
*** redbeard has quit IRC | 18:03 | |
*** ssurana1 has joined #openstack-nova | 18:03 | |
*** Kennan has joined #openstack-nova | 18:03 | |
*** Sukhdev has joined #openstack-nova | 18:04 | |
*** rfolco has quit IRC | 18:05 | |
*** ssurana has quit IRC | 18:06 | |
*** Marga_ has quit IRC | 18:06 | |
*** davideagnello has joined #openstack-nova | 18:06 | |
*** Marga_ has joined #openstack-nova | 18:07 | |
*** pixelbeat_ has quit IRC | 18:07 | |
*** patrickeast has joined #openstack-nova | 18:07 | |
*** rfolco has joined #openstack-nova | 18:09 | |
*** _gryf has quit IRC | 18:10 | |
*** xek has quit IRC | 18:10 | |
*** isq has quit IRC | 18:10 | |
mgagne | I'm having an issue with Nova kilo (2015.1.0) (with cells) where the instance_extra.flavor field in the api cell is not properly updated after a resize. (the field in compute cell is properly updated) Is there any known bugs and/or fix for it? | 18:10 |
*** doude has quit IRC | 18:11 | |
*** xek has joined #openstack-nova | 18:12 | |
*** _gryf has joined #openstack-nova | 18:12 | |
*** isq has joined #openstack-nova | 18:13 | |
*** tonytan4ever has joined #openstack-nova | 18:14 | |
*** MentalRay__ has joined #openstack-nova | 18:15 | |
*** signed8bit is now known as signed8bit_ZZZzz | 18:16 | |
*** achanda has quit IRC | 18:16 | |
ndipanov | dansmith, https://review.openstack.org/202732 | 18:19 |
*** baoli has joined #openstack-nova | 18:19 | |
dansmith | ndipanov: I did it this morning | 18:19 |
dansmith | https://review.openstack.org/#/c/202641/ | 18:20 |
dansmith | sorry, I thought it was clear I would do it | 18:20 |
*** y_sawai has joined #openstack-nova | 18:21 | |
*** ildikov has joined #openstack-nova | 18:22 | |
*** Nikolay_St has quit IRC | 18:22 | |
ndipanov | dansmith, fwiw mine is cleaner :P | 18:22 |
*** y_sawai has quit IRC | 18:22 | |
*** mtanino has quit IRC | 18:22 | |
ndipanov | dansmith, can you +2 the kilo backport of that version fix if it pasess check today | 18:23 |
ndipanov | I'd need to drop off soon | 18:23 |
*** dboik has joined #openstack-nova | 18:23 | |
dansmith | did the master patch succeed? | 18:23 |
ndipanov | well the check is still running seems like it's pretty backed up :( | 18:23 |
ndipanov | it's running now | 18:24 |
*** achanda has joined #openstack-nova | 18:26 | |
*** agireud has joined #openstack-nova | 18:29 | |
*** MentalRay__ has quit IRC | 18:30 | |
*** promulo_ has joined #openstack-nova | 18:30 | |
*** erlon has quit IRC | 18:30 | |
*** jcoufal has joined #openstack-nova | 18:32 | |
*** jcoufal has quit IRC | 18:32 | |
*** yassine has quit IRC | 18:34 | |
*** mutoulbj has quit IRC | 18:35 | |
*** zsais has quit IRC | 18:35 | |
openstackgerrit | Dan Smith proposed openstack/nova: Add version counter to Service object https://review.openstack.org/201733 | 18:37 |
*** achanda has quit IRC | 18:38 | |
johnthetubaguy | vilobhmm: so the deadline was today really, and it has to be complete and ready for review not WIP by the end of today, if that happens do drop me an email | 18:38 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: objects: Don't import CellMapping from the objects module https://review.openstack.org/202737 | 18:38 |
*** jwang has joined #openstack-nova | 18:38 | |
johnthetubaguy | vilobhmm: sorry I have to sign off now, just got dragged into a meeting when you pinged me | 18:38 |
ndipanov | dansmith, a light objects one ^ | 18:39 |
* ndipanov & | 18:39 | |
vilobhmm | johnthetubaguy : np..will try and let you know | 18:39 |
*** xyang2 has quit IRC | 18:40 | |
vilobhmm | jaypipes : ping | 18:41 |
vilobhmm | dansmith : ping | 18:41 |
melwitt | mgagne: I don't think that's a known bug cc alaski | 18:42 |
*** otter768 has joined #openstack-nova | 18:42 | |
mgagne | melwitt: trying to backport the patch atm | 18:42 |
mgagne | melwitt: and see if it fixes anything | 18:43 |
alaski | melwitt: it was never reported before, because it was fixed by instance_udpate_at_top going to objects | 18:43 |
melwitt | alaski: ah, okay | 18:44 |
*** gszasz has quit IRC | 18:46 | |
*** signed8bit_ZZZzz is now known as signed8bit | 18:47 | |
alaski | melwitt: I was just chatting in #openstack-operators with mgagne so I already had context :) | 18:47 |
*** mtanino has joined #openstack-nova | 18:47 | |
*** mutoulbj has joined #openstack-nova | 18:48 | |
dansmith | vilobhmm: ack | 18:50 |
melwitt | alaski: nice! I was wondering, did I miss something? heh | 18:50 |
*** dims_ has quit IRC | 18:51 | |
*** dims_ has joined #openstack-nova | 18:52 | |
*** otter768 has quit IRC | 18:52 | |
*** mutoulbj has quit IRC | 18:53 | |
vilobhmm | dansmith : today saw a bug in resource tracker code and thought will check with you if you have some time | 18:54 |
dansmith | vilobhmm: I'm not the best to ask about RT stuff, but you can try | 18:55 |
vilobhmm | https://github.com/openstack/nova/blob/master/nova/compute/resource_tracker.py#L771 if reserved_host_disk_mb is never set and if used the default value ; "self.compute_node.local_gb_used" will always return 0 | 18:55 |
vilobhmm | which is bad ! IMHO | 18:55 |
*** setuid has quit IRC | 18:55 | |
vilobhmm | and can lead to -ve values in disk space calculations | 18:56 |
*** shyama has quit IRC | 18:56 | |
vilobhmm | its been the same since many releases now | 18:56 |
vilobhmm | dansmith : ^^ | 18:56 |
*** dims_ has quit IRC | 18:56 | |
vilobhmm | the fix is trivial and if needed can submit a patch for it… | 18:57 |
mriedem | vilobhmm: just report a bug and push up a change for review | 18:57 |
mriedem | if you think there is an issue | 18:58 |
dansmith | vilobhmm: I think that's well-understood actually | 18:58 |
dansmith | vilobhmm: I think jaypipes would be the one to ask | 18:58 |
vilobhmm | mriedem : agree…since i am new to resource tracker thought will check dansmith or jaypipes … | 18:59 |
mriedem | local_gb_used would be 0 so you're not reserving any disk gb for the host | 18:59 |
vilobhmm | mriedem : will do | 18:59 |
vilobhmm | mriedem : so does it mean that the code behaves as expected | 19:00 |
mriedem | it appears so | 19:00 |
vilobhmm | if so then when we used local_gb_used to calculate free_space it report inconsistent values | 19:00 |
*** annashen has quit IRC | 19:00 | |
vilobhmm | sometimes even -ve values | 19:01 |
*** jecarey_ has joined #openstack-nova | 19:01 | |
vilobhmm | for free_space | 19:01 |
mriedem | vilobhmm: maybe read up on the original change for reasons https://review.openstack.org/#/c/13182/ | 19:01 |
*** jecarey has quit IRC | 19:02 | |
vilobhmm | mriedem : alrite…lemme check | 19:02 |
*** setuid has joined #openstack-nova | 19:03 | |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtISCSIVolumeDriver into it's own module https://review.openstack.org/202301 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtISERVolumeDriver into it's own module https://review.openstack.org/197311 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move volume/remotefs/quobyte modules under volume subdir https://review.openstack.org/193223 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtNETVolumeDriver into it's own module https://review.openstack.org/197341 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtNFSVolumeDriver into it's own module https://review.openstack.org/196859 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move the LibvirtScalityVolumeDriver into it's own module https://review.openstack.org/193249 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move lvm/dmcrypt/rbd_utils modules under storage subdir https://review.openstack.org/193195 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtAOEVolumeDriver into it's own module https://review.openstack.org/196276 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move the LibvirtQuobyteVolumeDriver into the quobyte module https://review.openstack.org/193236 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtFibreChannelVolumeDriver into it's own module https://review.openstack.org/193278 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtGlusterfsVolumeDriver into it's own module https://review.openstack.org/193380 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move fibre channel specific utilities into fc module https://review.openstack.org/193317 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move the LibvirtGPFSVolumeDriver into it's own module https://review.openstack.org/193240 | 19:03 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtSMBFSVolumeDriver into it's own module https://review.openstack.org/196380 | 19:03 |
*** cpschult has quit IRC | 19:03 | |
*** READ10 has quit IRC | 19:04 | |
*** SourabhP has quit IRC | 19:06 | |
*** tjones1 has joined #openstack-nova | 19:08 | |
nagyz_ | can anybody enlighten me about keystone's authtoken's auth_uri vs identity_uri? I know this is not #openstack-keystone but maybe the folks here know :-) and I'm not getting an answer there. | 19:08 |
nagyz_ | I see that in the nova docs we're setting auth_uri AND auth_url | 19:09 |
nagyz_ | but not identity_uri | 19:09 |
mriedem | jamielennox: ^ | 19:09 |
dansmith | keystone midcycle is going right now | 19:09 |
dansmith | hence the lack of people on irc I bet | 19:09 |
nagyz_ | ah | 19:10 |
mriedem | so that's where bknudson has been all week... | 19:10 |
mriedem | i knew he didn't take vacations... | 19:10 |
nagyz_ | it was exactly jamielennox's code that tripped me over in heat actually which I spent like 3 hours tracking down yesterday | 19:10 |
nagyz_ | I've got back the (not-so-user-friendly) exception of NoneType not having a replace function from like 100 layers deep.. without the original exception's origin. :) | 19:10 |
* nagyz_ loves python | 19:11 | |
ndipanov | https://review.openstack.org/#/c/201280/ mriedem pls | 19:11 |
ndipanov | and the kilo backport | 19:11 |
*** belmoreira has joined #openstack-nova | 19:18 | |
*** lpetrut has joined #openstack-nova | 19:20 | |
*** agireud has quit IRC | 19:20 | |
*** annegentle has joined #openstack-nova | 19:21 | |
*** cpschult has joined #openstack-nova | 19:22 | |
*** agireud has joined #openstack-nova | 19:22 | |
*** baoli has quit IRC | 19:23 | |
*** agireud has quit IRC | 19:26 | |
*** baoli has joined #openstack-nova | 19:28 | |
*** Nikolay_St has joined #openstack-nova | 19:29 | |
*** agireud has joined #openstack-nova | 19:29 | |
*** annashen has joined #openstack-nova | 19:33 | |
mgagne | I would like a second opinion on this (yet to merge) review because I think we are clearly ending in the wrong direction: https://review.openstack.org/#/c/152703/5 | 19:36 |
sambol | alex_xu: I'm starting on the next part of the V3 code restructure, moving the V2.1 code to the correct directories. cc/ edleafe zsais | 19:37 |
tonyb | mgagne: It's +W and running through the check queue now, so I think it's too late for a second opinion. | 19:39 |
mgagne | tonyb: I think we need to reconsider it and revert it if necessary | 19:40 |
tonyb | mgagne: Sure, a revert is possible | 19:40 |
mgagne | the change got approved less than 2 hours after a supportive comment although serious concerns were raised previously. I didn't have time to raised my own concerns due to time constraint. | 19:42 |
jlearman | Running tox for the first time, stuff is missing such as pg_config and xslt-config, which I installed from libpq-dev and libxslt1.dev. Now, "pyconfig.h" is missing. Did I miss an important step when installing Ubuntu 14.04? | 19:48 |
sdague | it's not too late, the point is right now this is a legit problem for infra cloud coming online | 19:48 |
sdague | to give everyone additional resources to run tests with | 19:48 |
sdague | and it doesn't hurt anyone else | 19:49 |
*** harlowja has joined #openstack-nova | 19:49 | |
*** harlowja_ has quit IRC | 19:49 | |
*** tpatil has joined #openstack-nova | 19:51 | |
dansmith | it does affect our guests | 19:52 |
dansmith | and it's kindof undoable from that point of view | 19:52 |
dansmith | TBH, I didn't vote on this because it makes me uneasy as well | 19:52 |
*** jecarey_ has quit IRC | 19:52 | |
*** harlowja has quit IRC | 19:52 | |
*** jecarey has joined #openstack-nova | 19:52 | |
dansmith | (I also didn't -1 it either) | 19:52 |
*** harlowja has joined #openstack-nova | 19:53 | |
*** jecarey_ has joined #openstack-nova | 19:53 | |
sdague | well, if we want to hold till next week to talk it through, that's fine. I guess I didn't see the concern in exposing provider networks if that's the way you configured your network | 19:53 |
jroll | sdague: mgagne makes a valid point there, I think | 19:54 |
sdague | jenkins rando failed it anyway | 19:54 |
lifeless | jlearman: sadly yes, thats commonly tricky. You might find https://rbtcollins.wordpress.com/2015/07/12/bootstrapping-developer-environments-for-openstack/ helpful | 19:54 |
mgagne | jroll: we already implemented all the blueprint locally a couple of months ago. I know the baremetal use case is very valid. but we have to think more about it then just exposing everything and hope cloud-init will know what to do with it. | 19:55 |
jroll | mgagne: yeah, I think I agree with you | 19:55 |
mgagne | jroll: we made the exposition of vlan info conditional based on the compute driver | 19:55 |
jroll | my colleague wrote the patch, I primarily work on ironic as well, I'll be around next week if we want to talk about it | 19:56 |
sdague | ok, but this is coming off the provider networks only from neutron, which means you are only going to get this information if you are on provider networks right? | 19:56 |
dansmith | yeah, that pretty much sucks | 19:56 |
dansmith | conditional networkinfo based on the compute driver | 19:56 |
sdague | which means you were given a port | 19:56 |
mgagne | sdague: we use provider networks in our virtual cloud | 19:56 |
dansmith | even if it is "ironic is special" | 19:56 |
mgagne | sdague: and virtual instance aren't using vlan interfaces. it is transparent | 19:57 |
*** jecarey has quit IRC | 19:57 | |
jroll | so, like I said, I'll be at the sprint next week. if y'all want to set aside an hour to hash this out I'm very +1 on that | 19:57 |
mgagne | dansmith: it really depends on how you wish to expose the network to the baremetal: access or trunk mode | 19:57 |
jroll | there's valid points here. | 19:57 |
jlearman | Thanks, lifeless! | 19:58 |
dansmith | mgagne: yeah, but that doesn't need to be virt-driver-specific | 19:58 |
sdague | so, it would probably be worth getting mordred and jeblair into the conversation as well, because this is basically their needed feature | 19:58 |
mgagne | dansmith: virt driver specific is ONE possible implementation which I happen to use and fit my need | 19:59 |
jroll | it might also depend on the neutron backend | 19:59 |
dansmith | mgagne: okay, sure | 19:59 |
mgagne | dansmith: the case is virtual instance shouldn't be exposed to this information because they don't need it | 19:59 |
dansmith | having a -1 from a neutron core is kindof bad form for us to merge, IMHO | 19:59 |
mgagne | dansmith: and will only cause confusion | 19:59 |
*** Mike_D_laptop has quit IRC | 20:02 | |
*** achanda has joined #openstack-nova | 20:03 | |
*** Sukhdev has quit IRC | 20:04 | |
*** exploreshaifali has quit IRC | 20:04 | |
sdague | ok, I'll hold. mgagne can you get this out on the mailing list for discussion, because all the parties involved are not going to be in any one location | 20:05 |
openstackgerrit | Dan Smith proposed openstack/nova: Add version counter to Service object https://review.openstack.org/201733 | 20:07 |
*** achanda has quit IRC | 20:09 | |
*** xyang1 has joined #openstack-nova | 20:09 | |
*** cpschult has quit IRC | 20:10 | |
openstackgerrit | Giridhar Jayavelu proposed openstack/nova: Derive resource quota from Flavor's extra_spec when attaching interface to an instance https://review.openstack.org/196171 | 20:10 |
openstackgerrit | Giridhar Jayavelu proposed openstack/nova: Create a new dictionary for type_data in VMwareAPIVMTestCase class https://review.openstack.org/202768 | 20:10 |
*** belmoreira has quit IRC | 20:11 | |
*** salv-orlando has quit IRC | 20:11 | |
*** agireud has quit IRC | 20:13 | |
melwitt | dansmith: you need to create your blueprint on that one :) | 20:14 |
dansmith | melwitt: what? | 20:14 |
*** Sukhdev has joined #openstack-nova | 20:15 | |
dansmith | oh | 20:15 |
dansmith | service-version-nuymber? | 20:15 |
*** agireud has joined #openstack-nova | 20:17 | |
dansmith | melwitt: done.. I'll let johnthetubaguy do the approval bits so it doesn't look like I'm approving my own stuff :) | 20:17 |
openstackgerrit | Anthony Lee proposed openstack/nova: Fix live-migrations usage of the wrong connector information https://review.openstack.org/202770 | 20:18 |
*** achanda has joined #openstack-nova | 20:19 | |
*** cpschult has joined #openstack-nova | 20:23 | |
*** ZZelle_ has joined #openstack-nova | 20:23 | |
mordred | sdague: what did I do? | 20:25 |
*** igordcard has quit IRC | 20:25 | |
sdague | mordred: there was some discussion of the patch to get the provider network info into config drive | 20:25 |
sdague | and concerns | 20:26 |
mordred | ya- there was a discussion with jroll and devananda a while ago in #openstack-ironic too | 20:26 |
sdague | ok | 20:26 |
sdague | how about we get this out to the mailing list | 20:26 |
sdague | because there are a lot of pockets of people | 20:26 |
mordred | so - I agree that exposing vlan info to a guest that does not need vlan info is silly | 20:26 |
jroll | sdague: +1 | 20:26 |
mordred | but that's the nice thing about the format of the metadata | 20:26 |
sdague | and I don't think they are talking to each other | 20:26 |
sdague | but, I'm running away for beer o'clock, because I just got grenade multinode d-g actually working, doing the right thing, and passing all the tests | 20:27 |
mordred | ugh. :( That means I'll have to read a mailing list thread - but ok | 20:27 |
mordred | I honestly have zero ideas why this is problematic - but maybe the mailing list thread will illuminate them for me | 20:28 |
sdague | right, so someone start that thread who is vested in this. | 20:28 |
sdague | ok, beer | 20:28 |
mordred | I'll send an email | 20:28 |
mordred | shoudl I tag it with nova, neutron and ironic | 20:28 |
mordred | ? | 20:28 |
*** annegentle has quit IRC | 20:28 | |
dansmith | come on | 20:29 |
dansmith | you can do better than that | 20:29 |
dansmith | and your tags better have parameters | 20:29 |
clarkb | [tags] | 20:29 |
jroll | [all] [vlan] [tags] [neutron] [nova] [ironic] [magnum] | 20:29 |
jlearman | Trying to run openstack for the first time, getting "ERROR: openstack No user with a name or ID of '4268ea30f4154909844eaa6a788dc376' exists." Any clues? | 20:29 |
*** sambol has quit IRC | 20:30 | |
dansmith | [nova:network=neutron:relevance=50%:priority=12][neutron:driver=foo:priority=27][ironic:isweird=True] | 20:30 |
jroll | lol | 20:30 |
mordred | oh god | 20:30 |
jroll | nice. | 20:30 |
*** jasondotstar has joined #openstack-nova | 20:30 | |
*** SourabhP has joined #openstack-nova | 20:30 | |
*** agireud has quit IRC | 20:31 | |
jlearman | the emotes really improve it | 20:31 |
*** agireud has joined #openstack-nova | 20:32 | |
mordred | mgagne: oh! I think I see from your comment where we may be missing each other | 20:32 |
mordred | mgagne: and it's possible that I'm looking at the consumption not the implementation side | 20:32 |
mordred | mgagne: I do not believe that vlan information should be included in the instance metadata if the hypervisor has already decapsulated the traffic for you | 20:33 |
mordred | I agree - that woudl be crazypants | 20:33 |
mgagne | mordred: yep | 20:33 |
mgagne | mordred: I already implemented a similar patch into our installation and I do understand the baremetal use case because we did address it (in our way) in our implementation. | 20:33 |
*** doude has joined #openstack-nova | 20:33 | |
*** doude has quit IRC | 20:33 | |
*** doude has joined #openstack-nova | 20:33 | |
mgagne | mordred: addressed in the way you describe | 20:33 |
mordred | mgagne: great - so is it just the implementation of this patch you have issue with? | 20:34 |
*** doude has quit IRC | 20:34 | |
mordred | and not the concept of being able to express vlan information in the metadata if it's needed? | 20:34 |
mgagne | mordred: well yea because it breaks the world | 20:34 |
mordred | awesome! | 20:34 |
*** doude has joined #openstack-nova | 20:34 | |
*** doude has quit IRC | 20:34 | |
*** doude has joined #openstack-nova | 20:34 | |
mgagne | mordred: vlan is part of the spec | 20:34 |
mordred | vlan should be an optional part of the spec - and it shoudl be there if and only if the compute target needs to know taht vlan info | 20:34 |
mgagne | yep | 20:35 |
mordred | ok. if that is not what's happening here, then I agree with you | 20:35 |
openstackgerrit | Anthony Lee proposed openstack/nova: Fix live-migrations usage of the wrong connector information https://review.openstack.org/202770 | 20:35 |
mordred | mgagne: I want glean or cloud-init to be able to look at the metadata and say "oh, look, it says I need a vlan, I should configure it" | 20:35 |
mgagne | mordred: sure | 20:36 |
mordred | and have that always work and always be true | 20:36 |
mgagne | mordred: have fun: https://github.com/jayofdoom/cloud-init-debian-pkg/blob/master/debian/patches/rackspace-onmetal-configdrive.patch | 20:36 |
mgagne | mordred: if I install a similar cloud-init in a virtual image, everything falls apart | 20:36 |
mordred | mgagne: oh - I don't use cloud-init :) | 20:36 |
*** krtaylor has quit IRC | 20:36 | |
*** dims_ has joined #openstack-nova | 20:37 | |
mordred | mgagne: I wrote a minimal replacement that only does network and ssh keys | 20:37 |
mgagne | mordred: well, we have to considerate other sane people too =) | 20:37 |
mordred | mgagne: looking at the spec, I expect it will take me approx 20 minutes to add the vlan support to it once we agree on an input format | 20:37 |
mordred | mgagne: :) | 20:37 |
*** alex_klimov has joined #openstack-nova | 20:37 | |
mordred | mgagne: you shoudl start using glean!!! | 20:37 |
jroll | mgagne: sane people don't use cloud-init :P | 20:37 |
mgagne | mordred: I provide solutions for common people =) | 20:38 |
lifeless | I really wish I understood the hate-on that folk have for cloud-init | 20:38 |
lifeless | I realise its become crufty | 20:38 |
jroll | source: I use cloud-init | 20:38 |
mordred | lifeless: too complex | 20:38 |
mgagne | works for me | 20:38 |
lifeless | but it still solves the core use cases | 20:38 |
mordred | lifeless: does completely unnecessary things | 20:38 |
mordred | lifeless: we're having to work aroudn like 8 things that cloud-init breaks in infra right now | 20:38 |
mgagne | only problem I have is that it's hard as hell to get stuff merged upstream | 20:38 |
lifeless | mordred: which can all be turned off | 20:38 |
lifeless | mordred: (AIUI) | 20:39 |
mordred | lifeless: OR - I could just not have a thing installed on my system which consistently breaks it in the first place | 20:39 |
* mgagne opened a can of worms | 20:39 | |
mordred | rather than installing a complex piece of software which I think include complex configs for disabling much of | 20:39 |
*** thangp has quit IRC | 20:39 | |
*** xyang1 has quit IRC | 20:39 | |
lifeless | mgagne: indeed | 20:39 |
mordred | s/think/then can/ | 20:39 |
*** xyang1 has joined #openstack-nova | 20:39 | |
mgagne | ok, let me reword that: please consider insane people who didn't discover the light yet =) | 20:40 |
mordred | lifeless: I do not need a remote shell execution environment that is in addition to ssh. I have ssh. ssh is consistent and debuggable. | 20:40 |
mordred | mgagne: hahaha. nice :) | 20:40 |
mgagne | mordred: so the way I addressed it is by making the ironic driver includes extra data in the instance so I can determine if network info should include vlan info or not. | 20:41 |
*** dave-mccowan has joined #openstack-nova | 20:41 | |
mordred | mgagne: how did you get the extra data into the instance? | 20:41 |
mordred | mgagne: (if there is an eas(y|ier) way to solve this, I'm in) | 20:41 |
mgagne | mordred: by patching ironic driver and updating instance_system_metadata | 20:41 |
mordred | mgagne: sorry - I asked a vague question | 20:42 |
mordred | mgagne: how does that cause that information to be communicated into the inside of the instance? | 20:42 |
dansmith | mgagne: :( | 20:42 |
mgagne | mordred: a big if/else | 20:42 |
mordred | mgagne: where? | 20:42 |
mgagne | mordred: in network info implementation | 20:42 |
mgagne | mordred: we have our own because upstream wasn't ready yet | 20:42 |
mordred | ok - so you're still using teh network_info.json as the communication vehicle | 20:42 |
mgagne | mordred: sure | 20:42 |
mordred | AWESOME | 20:42 |
mordred | that's literally all I care about | 20:43 |
mgagne | mordred: you just have to provide the right info | 20:43 |
mordred | if it takes 2 more years to figure out the right way to get the information into network_info.json that's fine with me | 20:43 |
mordred | I just want to know that the network_info.json interface is a thing | 20:43 |
mordred | that I can use | 20:43 |
mordred | inside of the instance | 20:43 |
*** oro has quit IRC | 20:43 | |
mordred | from config-drive | 20:43 |
mgagne | mordred: well I already waited 2 years :-/ | 20:43 |
jroll | mgagne: I'd love to see your patch | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: Introduce LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202777 | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: convert NFS volume driver to LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202778 | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: convert SMBFS vol driver to LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202779 | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: convert GlusterFS driver to LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202780 | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: convert GPFS volume driver to LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202781 | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: convert Quotbyte driver to LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202782 | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: convert Scality vol driver to LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202783 | 20:44 |
*** kmARC has quit IRC | 20:44 | |
mgagne | jroll: nothing special tbh. if baremetal: else: | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: convert Scality vol driver to LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202783 | 20:44 |
openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: convert Quobyte driver to LibvirtBaseFileSystemVolumeDriver https://review.openstack.org/202782 | 20:44 |
*** salv-orlando has joined #openstack-nova | 20:45 | |
*** tonytan4ever has quit IRC | 20:48 | |
*** sambol has joined #openstack-nova | 20:48 | |
tonyb | c | 20:49 |
tonyb | stupid macos and no focus follows mouse! | 20:50 |
*** tpatil has quit IRC | 20:51 | |
mriedem | nova meeting in 8 minutes | 20:52 |
*** tpatil has joined #openstack-nova | 20:52 | |
*** tonytan4ever has joined #openstack-nova | 20:52 | |
tonyb | Hey I can do that one ;P How very novel | 20:52 |
*** otter768 has joined #openstack-nova | 20:53 | |
*** tpatil has left #openstack-nova | 20:54 | |
*** otter768 has quit IRC | 20:58 | |
*** ociuhandu has joined #openstack-nova | 21:02 | |
*** gyee has quit IRC | 21:02 | |
*** armax has joined #openstack-nova | 21:03 | |
*** ijw has joined #openstack-nova | 21:06 | |
*** signed8bit is now known as signed8bit_ZZZzz | 21:11 | |
*** signed8bit_ZZZzz is now known as signed8bit | 21:12 | |
*** dane-fichter has joined #openstack-nova | 21:17 | |
*** doude has quit IRC | 21:20 | |
*** dane-fichter has quit IRC | 21:22 | |
*** chuckcarmack has left #openstack-nova | 21:29 | |
jroll | johnthetubaguy: I went looking for a blueprint and noticed this should have been approved as it was landed in kilo https://blueprints.launchpad.net/nova/+spec/use-configdrive-with-ironic | 21:29 |
johnthetubaguy | jroll: ah, oops, good heads up | 21:30 |
johnthetubaguy | jroll: I marked that as implemented in kilo, I think | 21:31 |
openstackgerrit | Jim Rollenhagen proposed openstack/nova-specs: Ironic: Multiple compute host support https://review.openstack.org/194453 | 21:31 |
jroll | johnthetubaguy: cool, just noticed it was pending approval so thought I'd point it out | 21:31 |
johnthetubaguy | jroll: we unapprove when it gets bumped from a release | 21:32 |
jroll | johnthetubaguy: right, it didn't get bumped, it fully landed in kilo. | 21:32 |
johnthetubaguy | jroll: I guess it got merged anyways | 21:32 |
johnthetubaguy | jroll: it probably missed feature freeze, but merged afterwards, or something | 21:33 |
jroll | gotcha | 21:33 |
johnthetubaguy | mriedem: I think I got those all approved now | 21:35 |
*** krtaylor has joined #openstack-nova | 21:36 | |
*** mriedem has quit IRC | 21:38 | |
gabriel-bezerra | johnthetubaguy: that code has not been queued into Zuul yet. Is it normal? | 21:42 |
*** boris-42 has joined #openstack-nova | 21:42 | |
gabriel-bezerra | https://review.openstack.org/190675 | 21:42 |
gabriel-bezerra | dansmith: ^ | 21:43 |
*** edmondsw has quit IRC | 21:45 | |
*** bnemec has quit IRC | 21:46 | |
*** alex_klimov has quit IRC | 21:47 | |
*** angdraug has quit IRC | 21:48 | |
*** agireud has quit IRC | 21:49 | |
*** mutoulbj has joined #openstack-nova | 21:51 | |
gabriel-bezerra | johnthetubaguy: I've just talked to -infra people. Can you remove the -2 from the patch, please? Zuul won't enqueue while the -2 be there: https://review.openstack.org/190675 | 21:55 |
*** mutoulbj has quit IRC | 21:56 | |
*** annashen has quit IRC | 21:57 | |
*** lucas-dinner has quit IRC | 21:59 | |
*** vilobhmm has quit IRC | 21:59 | |
*** vilobhmm has joined #openstack-nova | 22:00 | |
*** jecarey_ has quit IRC | 22:01 | |
*** dboik has quit IRC | 22:02 | |
*** davideagnello has quit IRC | 22:04 | |
*** cpschult has quit IRC | 22:05 | |
*** burt has quit IRC | 22:05 | |
*** ociuhandu has quit IRC | 22:06 | |
*** ijw_ has joined #openstack-nova | 22:06 | |
openstackgerrit | Jim Rollenhagen proposed openstack/nova: Fix deleted markers with changed-since https://review.openstack.org/202812 | 22:08 |
*** ijw has quit IRC | 22:09 | |
*** achanda has quit IRC | 22:10 | |
*** ddaskal has quit IRC | 22:10 | |
*** mdrabe has quit IRC | 22:11 | |
*** lucas-dinner has joined #openstack-nova | 22:11 | |
*** Sukhdev has quit IRC | 22:14 | |
*** jasondotstar has quit IRC | 22:15 | |
*** Sukhdev has joined #openstack-nova | 22:16 | |
*** pixelbeat_ has joined #openstack-nova | 22:17 | |
*** signed8bit is now known as signed8bit_ZZZzz | 22:19 | |
andreykurilin__ | johnthetubaguy: hi | 22:19 |
andreykurilin__ | Can you look at https://review.openstack.org/#/c/202656/ ? This a fix for stable/kilo branch of novaclient, which is broken now. | 22:19 |
andreykurilin__ | johnthetubaguy: ^ | 22:20 |
*** thorst has quit IRC | 22:21 | |
tonyb | andreykurilin__: I think you've missed johnthetubaguy it's 11pm for him. | 22:22 |
andreykurilin__ | tonyb: oh...got it. thanks:( | 22:23 |
edleafe | tonyb: saw that you pinged me yesterday. Wassup? | 22:24 |
tonyb | edleafe: You're in San Antonio right? | 22:24 |
edleafe | tonyb: yep | 22:25 |
edleafe | tonyb: you headed to mother ship before the midcycle? | 22:25 |
tonyb | edleafe: wanna grab lunch or a beer? | 22:25 |
tonyb | edleafe: I'm presently in the mother ship | 22:25 |
edleafe | tonyb: can't tonight - daughter's birthday | 22:26 |
edleafe | tonyb: how long are you around? | 22:26 |
tonyb | edleafe: okay np | 22:26 |
tonyb | I fly to MN on Monday. | 22:26 |
edleafe | tonyb: me too | 22:26 |
*** vipul is now known as vipoul | 22:27 | |
*** vipoul is now known as vipul | 22:27 | |
edleafe | tonyb: you got plans for tomorrow? | 22:27 |
*** achanda has joined #openstack-nova | 22:27 | |
tonyb | edleafe: Just work, but if you're free I can ditch that ;P | 22:27 |
edleafe | tonyb: heh | 22:28 |
edleafe | tonyb: ibm is a little stricter :) | 22:29 |
edleafe | tonyb: how about later in the day? Where are you staying? | 22:29 |
tonyb | Aloft near the airport | 22:29 |
tonyb | edleafe: but I'll probably pick up a car tomorrow at some stage. | 22:29 |
* tonyb shoudl book that ;P | 22:30 | |
edleafe | tonyb: wanna meet at Lion & Rose next door? | 22:30 |
tonyb | edleafe: Sure. That's pretty easy and I'm told that do whiskey samplers ;P | 22:31 |
*** sambol has quit IRC | 22:31 | |
*** mriedem has joined #openstack-nova | 22:32 | |
*** xyang1 has quit IRC | 22:33 | |
*** lpetrut has quit IRC | 22:37 | |
*** IanGovett1 has quit IRC | 22:38 | |
*** mmedvede has quit IRC | 22:39 | |
*** achanda has quit IRC | 22:41 | |
*** ijw_ has quit IRC | 22:44 | |
*** lpetrut has joined #openstack-nova | 22:45 | |
*** signed8bit_ZZZzz is now known as signed8bit | 22:45 | |
*** foexle has quit IRC | 22:46 | |
*** lucas-dinner has quit IRC | 22:46 | |
*** mmedvede has joined #openstack-nova | 22:47 | |
*** ctrath has quit IRC | 22:50 | |
*** otter768 has joined #openstack-nova | 22:54 | |
*** thorst has joined #openstack-nova | 22:54 | |
*** signed8bit is now known as signed8bit_ZZZzz | 22:55 | |
*** thorst has quit IRC | 22:56 | |
*** vilobhmm has quit IRC | 22:57 | |
*** vilobhmm has joined #openstack-nova | 22:57 | |
openstackgerrit | Dan Smith proposed openstack/nova: Fix the incorrect PciDeviceList version number https://review.openstack.org/201280 | 22:58 |
*** otter768 has quit IRC | 22:58 | |
*** pixelbeat_ has quit IRC | 22:59 | |
*** ijuwang has quit IRC | 23:01 | |
*** hogepodge has quit IRC | 23:02 | |
*** kylek3h has quit IRC | 23:05 | |
*** tpeoples has quit IRC | 23:05 | |
*** zzzeek has quit IRC | 23:08 | |
*** achanda has joined #openstack-nova | 23:09 | |
*** salv-orlando has quit IRC | 23:16 | |
kfox1111 | can you tag flavors with metadata that isn't used by the schedular? | 23:18 |
kfox1111 | (something magnum can look at?) | 23:18 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/200708 | 23:19 |
*** zigo has quit IRC | 23:21 | |
*** zigo has joined #openstack-nova | 23:22 | |
*** jasondotstar has joined #openstack-nova | 23:25 | |
*** hemna is now known as hemnafk | 23:25 | |
*** openstackstatus has joined #openstack-nova | 23:34 | |
*** ChanServ sets mode: +v openstackstatus | 23:34 | |
*** lpetrut has quit IRC | 23:34 | |
mriedem | kfox1111: i'm not aware of validation on extra_specs having specific keys | 23:36 |
kfox1111 | looking at the docs, it looks like the schedular uses al the values to filter, and if I throw some magnum something key in there, it may cause the vm to fail to launch? | 23:37 |
*** ijw has joined #openstack-nova | 23:38 | |
*** aysyd has quit IRC | 23:39 | |
*** erlon has joined #openstack-nova | 23:39 | |
kfox1111 | can a filter remove stuff from extra_specs before it goes to the next filter? | 23:42 |
*** wanghua has joined #openstack-nova | 23:46 | |
melwitt | kfox1111: I don't think so. I think if a filter doesn't care about a key, it will ignore it. but I don't know for certain | 23:49 |
*** jaypipes has quit IRC | 23:50 | |
kfox1111 | I'm thinking I've tagged some flavors with a piece of metadata, and then put the same key/value in a host aggregate to pin flavors to aggregates. | 23:50 |
kfox1111 | I think that use case might break if arbitrary extra metadata's added to extra specs. | 23:50 |
kfox1111 | nova/filter.py line 89, looks like it does not allow a class to remove filter_properties? | 23:52 |
kfox1111 | except.... it looks like its passing in a hash.. that should be by reference. so changes to it should propigate... | 23:54 |
*** barra204 has joined #openstack-nova | 23:54 | |
melwitt | yes, that is a thing you can do to pin flavors to host aggregates | 23:55 |
*** mriedem1 has joined #openstack-nova | 23:55 | |
*** mriedem has quit IRC | 23:55 | |
kfox1111 | so it would break without a filter of some kind to remove the extra_spec before it hit the host aggrigate filter. | 23:56 |
kfox1111 | its looking like it would be possible to write such a filter... not sure how nova would feal about such a thing though. thoughts? | 23:56 |
mriedem1 | kfox1111: out of tree filters aren't a new thing | 23:57 |
kfox1111 | yeah, but the api kind of is for filtering out hosts that dont match the extra spec info. its a bit of a strange use to have the filter just leave the list untouched and tweak the extra spec data itself. | 23:58 |
melwitt | kfox1111: why do you think you need to remove an arbitrary extra_spec? I don't see what you mean in nova/filters.py | 23:59 |
kfox1111 | ok. example. | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!