| openstackgerrit | Fei Long Wang proposed openstack/nova: Fix nova backup for volume-backed instance https://review.openstack.org/164494 | 00:02 |
|---|---|---|
| *** pmcfadden has quit IRC | 00:04 | |
| *** jaypipes has quit IRC | 00:05 | |
| *** tpeoples has quit IRC | 00:06 | |
| *** sorrison_laptop has joined #openstack-nova | 00:06 | |
| *** xyang1 has quit IRC | 00:06 | |
| *** setuid has joined #openstack-nova | 00:08 | |
| *** pmcfadden has joined #openstack-nova | 00:14 | |
| *** jwcroppe has quit IRC | 00:15 | |
| *** pmcfadden has quit IRC | 00:17 | |
| *** Marga_ has joined #openstack-nova | 00:18 | |
| *** pmcfadden has joined #openstack-nova | 00:18 | |
| *** diogogmt has joined #openstack-nova | 00:19 | |
| openstackgerrit | Fei Long Wang proposed openstack/nova-specs: Add spec for volume backed backup https://review.openstack.org/190919 | 00:19 |
| *** IanGovett has quit IRC | 00:19 | |
| *** IanGovett has joined #openstack-nova | 00:20 | |
| *** Marga_ has quit IRC | 00:22 | |
| *** baoli has quit IRC | 00:24 | |
| *** alexpilotti has quit IRC | 00:26 | |
| *** hemna is now known as hemnafk | 00:27 | |
| *** ijw has quit IRC | 00:27 | |
| *** mtanino_ has joined #openstack-nova | 00:28 | |
| *** mtanino has quit IRC | 00:28 | |
| openstackgerrit | liaonanhai proposed openstack/nova-specs: Add user_id info: add 'user_id' to REST os-simple-tenant-usage output https://review.openstack.org/198937 | 00:30 |
| *** pmcfadden has quit IRC | 00:30 | |
| *** erlon has quit IRC | 00:30 | |
| *** yingjun has joined #openstack-nova | 00:32 | |
| *** ajayaa has quit IRC | 00:36 | |
| *** baoli has joined #openstack-nova | 00:39 | |
| *** thorst has joined #openstack-nova | 00:40 | |
| *** ijw has joined #openstack-nova | 00:40 | |
| *** apuimedo has quit IRC | 00:42 | |
| *** asettle has joined #openstack-nova | 00:42 | |
| *** ssurana has quit IRC | 00:43 | |
| openstackgerrit | liaonanhai proposed openstack/nova-specs: Add user_id info: add 'user_id' to REST os-simple-tenant-usage output https://review.openstack.org/198937 | 00:45 |
| *** baoli has quit IRC | 00:48 | |
| *** thedodd has quit IRC | 00:51 | |
| *** SourabhP has quit IRC | 00:51 | |
| *** armax has quit IRC | 00:51 | |
| *** SourabhP has joined #openstack-nova | 00:53 | |
| *** mwagner_lap has joined #openstack-nova | 00:54 | |
| *** Marga_ has joined #openstack-nova | 00:57 | |
| *** Marga_ has quit IRC | 00:58 | |
| *** Marga_ has joined #openstack-nova | 00:58 | |
| alex_xu | melwitt: hi, are you around? | 00:58 |
| melwitt | alex_xu: yes, hi | 00:59 |
| alex_xu | melwitt: not sure you see my email for a meeting about microversio client, is that time ok for you? what timezone you are in? | 01:00 |
| alex_xu | melwitt: appreciate if you can join the meeting, then we can get some suggestion from you :) | 01:00 |
| *** baoli has joined #openstack-nova | 01:01 | |
| *** marzif_ has joined #openstack-nova | 01:01 | |
| alex_xu | melwitt: we feel there are some communication issues around the spec and implementation, so hope setup a meeting to resolve those mess | 01:02 |
| melwitt | alex_xu: I'm in PST so it's 6am for me. I don't have a lot of knowledge about the microversions, so I'm mostly interested in hearing what johnthetubaguy and sdague have to say | 01:03 |
| melwitt | alex_xu: I was just surprised to see there is supposed to be a different behavior between Python API and CLI that I didn't understand | 01:03 |
| *** yingjun has quit IRC | 01:04 | |
| melwitt | I reviewed the patches based on the spec because that's the information I have about how microversions supposed to work from a client perspective | 01:04 |
| alex_xu | melwitt: yea, that is too early for you. | 01:05 |
| *** bradjones_ has joined #openstack-nova | 01:06 | |
| *** bradjones_ has quit IRC | 01:06 | |
| *** bradjones_ has joined #openstack-nova | 01:06 | |
| *** bradjones has quit IRC | 01:06 | |
| melwitt | alex_xu: I don't think I'm a critical piece for the meeting, so I will be fine to read about the outcome and go with the consensus you reach with everyone else | 01:06 |
| alex_xu | melwitt: melwitt yea, I didn't find out the spec describe that clearly also, but from the discussion, we said we agreement on API and CLi have different behavior | 01:06 |
| *** bradjones_ is now known as bradjones | 01:07 | |
| alex_xu | melwitt: from patchset4 https://review.openstack.org/#/c/188410/ there are comment about that, and I remember we agree on that in the nova api meeting...but finally that's not in the spec... | 01:07 |
| alex_xu | melwitt: ok, no problem, I will let you know the final decission | 01:08 |
| *** MentalRay has joined #openstack-nova | 01:08 | |
| melwitt | alex_xu: cool, thank you | 01:09 |
| alex_xu | melwitt: np | 01:09 |
| openstackgerrit | sajeesh cs proposed openstack/nova: Nested Quota Driver: Adding v2 and v3(v2.1) support https://review.openstack.org/149828 | 01:09 |
| openstackgerrit | sajeesh cs proposed openstack/nova: Nested Quota Driver: Adding v2 and v3(v2.1) support https://review.openstack.org/149828 | 01:09 |
| openstackgerrit | sajeesh cs proposed openstack/nova: Nested Quota Driver: Adding v2 and v3(v2.1) support https://review.openstack.org/149828 | 01:09 |
| openstackgerrit | sajeesh cs proposed openstack/nova: Nested Quota Driver: Adding v2 and v3(v2.1) support https://review.openstack.org/149828 | 01:10 |
| *** tonytan4ever has quit IRC | 01:10 | |
| *** andrewbogott is now known as andrewbogott_afk | 01:10 | |
| *** baoli has quit IRC | 01:10 | |
| *** vilobhmm has quit IRC | 01:11 | |
| openstackgerrit | sajeesh cs proposed openstack/nova: Nested Quota Driver: Adding v2 and v3(v2.1) support https://review.openstack.org/149828 | 01:11 |
| *** armax has joined #openstack-nova | 01:13 | |
| *** marzif_ has quit IRC | 01:14 | |
| *** IanGovett has quit IRC | 01:14 | |
| *** baoli has joined #openstack-nova | 01:14 | |
| *** thorst has quit IRC | 01:15 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge server list multi status tests between v2 and v2.1 https://review.openstack.org/201387 | 01:16 |
| *** SourabhP has quit IRC | 01:21 | |
| *** armax has quit IRC | 01:21 | |
| *** pixelbeat_ has quit IRC | 01:22 | |
| *** sorrison_laptop has quit IRC | 01:22 | |
| *** pixelbeat_ has joined #openstack-nova | 01:22 | |
| *** annashen has joined #openstack-nova | 01:25 | |
| *** ijw has quit IRC | 01:27 | |
| *** annashen has quit IRC | 01:30 | |
| *** obedmr has joined #openstack-nova | 01:33 | |
| *** dims__ has joined #openstack-nova | 01:34 | |
| obedmr | hi there, does anyone know if there's an example guide for adding a new driver in nova-compute? | 01:34 |
| *** karimb has quit IRC | 01:37 | |
| *** dannywilson has quit IRC | 01:37 | |
| *** dims_ has quit IRC | 01:38 | |
| *** otter768 has joined #openstack-nova | 01:40 | |
| *** SourabhP has joined #openstack-nova | 01:43 | |
| *** markvoelker has quit IRC | 01:44 | |
| *** otter768 has quit IRC | 01:45 | |
| *** baoli has quit IRC | 01:47 | |
| *** ishita has joined #openstack-nova | 01:47 | |
| *** baoli has joined #openstack-nova | 01:47 | |
| *** zzzeek has quit IRC | 01:48 | |
| *** SourabhP has quit IRC | 01:49 | |
| *** shuquan has joined #openstack-nova | 01:50 | |
| *** SourabhP has joined #openstack-nova | 01:50 | |
| *** jdandrea has quit IRC | 01:51 | |
| *** vilobhmm has joined #openstack-nova | 01:53 | |
| *** eliqiao has joined #openstack-nova | 01:53 | |
| *** dboik has joined #openstack-nova | 01:53 | |
| *** browne has quit IRC | 01:54 | |
| *** dboik_ has joined #openstack-nova | 01:54 | |
| *** dboik has quit IRC | 01:58 | |
| alex_xu | oomichi: good monrning, there two policy patches need another +2 https://review.openstack.org/#/c/199971/ https://review.openstack.org/#/c/199972/ , appreciate if you have time take a look at it | 01:59 |
| *** asettle is now known as asettle-afk | 01:59 | |
| *** mtanino_ has quit IRC | 01:59 | |
| *** tsekiyama has quit IRC | 02:01 | |
| *** artom_ has quit IRC | 02:03 | |
| *** ijw has joined #openstack-nova | 02:04 | |
| openstackgerrit | xing-yang proposed openstack/nova: Add os-brick based LibvirtVolumeDriver for ScaleIO https://review.openstack.org/194454 | 02:04 |
| *** lyanchih has joined #openstack-nova | 02:04 | |
| oomichi | alex_xu: hi, morning o/ | 02:05 |
| oomichi | alex_xu: ok, will check them soon :) | 02:05 |
| alex_xu | oomichi: thanks :) | 02:05 |
| oomichi | alex_xu: np | 02:05 |
| *** tpeoples has joined #openstack-nova | 02:08 | |
| *** liusheng has joined #openstack-nova | 02:08 | |
| *** jwcroppe has joined #openstack-nova | 02:12 | |
| *** dims_ has joined #openstack-nova | 02:12 | |
| *** dims__ has quit IRC | 02:13 | |
| oomichi | alex_xu: can you check a comment on https://review.openstack.org/#/c/199971/ ? | 02:14 |
| alex_xu | oomichi: sure | 02:14 |
| *** MentalRay has quit IRC | 02:15 | |
| alex_xu | oomichi: will update soon, thanks for the review! | 02:16 |
| *** jwcroppe has quit IRC | 02:16 | |
| oomichi | alex_xu: np :) | 02:17 |
| *** mutoulbj has joined #openstack-nova | 02:17 | |
| *** pixelbeat_ has quit IRC | 02:17 | |
| oomichi | alex_xu: if https://review.openstack.org/#/c/199972/ can be the first order on the patches, the patch can be merged soon. | 02:17 |
| *** liusheng has quit IRC | 02:18 | |
| *** dims__ has joined #openstack-nova | 02:18 | |
| *** liusheng has joined #openstack-nova | 02:18 | |
| alex_xu | oomichi: emm...good idea, let me adjust the order | 02:18 |
| *** dims_ has quit IRC | 02:19 | |
| oomichi | alex_xu: or enough to separate it(199972) from the patches. | 02:20 |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove useless db call instance_get_all_by_host_and_not_type https://review.openstack.org/199971 | 02:22 |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for provider_fw_rule_* https://review.openstack.org/200381 | 02:22 |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for archive_deleted_rows* https://review.openstack.org/200380 | 02:22 |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove the useless require_admin_context decorator https://review.openstack.org/200382 | 02:22 |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for reservation_expire https://review.openstack.org/199972 | 02:22 |
| alex_xu | oomichi: it's done, this https://review.openstack.org/#/c/199972/4 just need +w | 02:23 |
| oomichi | alex_xu: nice, +A | 02:24 |
| alex_xu | oomichi: oops, that is +A | 02:24 |
| alex_xu | oomichi: thanks | 02:24 |
| *** yamahata has quit IRC | 02:24 | |
| oomichi | alex_xu: https://review.openstack.org/#/c/199971/ is not updated yet, will you ? | 02:25 |
| alex_xu | oomichi: oh...my god...I forget | 02:25 |
| oomichi | alex_xu: no problem ;) | 02:26 |
| *** annashen has joined #openstack-nova | 02:26 | |
| *** bkopilov has quit IRC | 02:27 | |
| *** park_heijlong has joined #openstack-nova | 02:27 | |
| *** zhenguo has joined #openstack-nova | 02:27 | |
| *** SourabhP has quit IRC | 02:29 | |
| *** ishita has quit IRC | 02:30 | |
| *** annashen has quit IRC | 02:31 | |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove useless db call instance_get_all_by_host_and_not_type https://review.openstack.org/199971 | 02:31 |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove the useless require_admin_context decorator https://review.openstack.org/200382 | 02:31 |
| *** unicell has quit IRC | 02:31 | |
| *** obedmr has quit IRC | 02:33 | |
| *** tobe has joined #openstack-nova | 02:42 | |
| *** patrickeast has quit IRC | 02:42 | |
| *** dboik_ has quit IRC | 02:44 | |
| *** dboik has joined #openstack-nova | 02:44 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge server list multi status tests between v2 and v2.1 https://review.openstack.org/201387 | 02:47 |
| *** busterswt has joined #openstack-nova | 02:47 | |
| *** busterswt has quit IRC | 02:50 | |
| *** ijw has quit IRC | 02:51 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge block_device_mapping_v2_boot tests between v2 and v2.1 https://review.openstack.org/201401 | 02:51 |
| *** markvoelker has joined #openstack-nova | 02:54 | |
| *** markvoelker has quit IRC | 02:55 | |
| *** markvoelker has joined #openstack-nova | 02:55 | |
| *** browne has joined #openstack-nova | 02:55 | |
| *** ssurana has joined #openstack-nova | 03:00 | |
| *** shuquan has quit IRC | 03:05 | |
| openstackgerrit | wanghao proposed openstack/nova: Image/block device mapping combination is not valid https://review.openstack.org/171984 | 03:05 |
| *** wanghao has joined #openstack-nova | 03:07 | |
| *** jwcroppe has joined #openstack-nova | 03:13 | |
| *** jwcroppe has quit IRC | 03:21 | |
| *** yingjun has joined #openstack-nova | 03:23 | |
| *** dims__ has quit IRC | 03:27 | |
| *** lyanchih has quit IRC | 03:28 | |
| *** Jeffrey4l has quit IRC | 03:29 | |
| *** Jeffrey4l has joined #openstack-nova | 03:31 | |
| *** lyanchih has joined #openstack-nova | 03:34 | |
| *** Longgeek has joined #openstack-nova | 03:35 | |
| *** oomichi has quit IRC | 03:35 | |
| *** achanda has quit IRC | 03:38 | |
| *** otter768 has joined #openstack-nova | 03:41 | |
| *** otter768 has quit IRC | 03:45 | |
| *** baoli has quit IRC | 03:51 | |
| *** asettle-afk is now known as asettle | 03:52 | |
| *** yingjun has quit IRC | 03:53 | |
| *** annashen has joined #openstack-nova | 03:56 | |
| *** vilobhmm has quit IRC | 03:59 | |
| *** vilobhmm has joined #openstack-nova | 04:01 | |
| *** annashen has quit IRC | 04:01 | |
| *** asettle has quit IRC | 04:01 | |
| *** jwcroppe has joined #openstack-nova | 04:02 | |
| *** bkopilov has joined #openstack-nova | 04:07 | |
| *** deepthi has joined #openstack-nova | 04:09 | |
| *** armax has joined #openstack-nova | 04:10 | |
| *** arnaud___ has joined #openstack-nova | 04:17 | |
| *** tpeoples has quit IRC | 04:19 | |
| *** tpeoples has joined #openstack-nova | 04:21 | |
| *** boris-42 has quit IRC | 04:22 | |
| *** garyk has quit IRC | 04:24 | |
| *** oro has joined #openstack-nova | 04:24 | |
| *** dave-mccowan has quit IRC | 04:25 | |
| *** tjones1 has joined #openstack-nova | 04:25 | |
| *** tjones1 has left #openstack-nova | 04:27 | |
| *** asettle has joined #openstack-nova | 04:27 | |
| *** ssurana has quit IRC | 04:31 | |
| *** unicell has joined #openstack-nova | 04:33 | |
| *** fawadkhaliq has joined #openstack-nova | 04:34 | |
| *** sudipto has joined #openstack-nova | 04:35 | |
| *** fawadkhaliq has quit IRC | 04:38 | |
| *** whenry_ has quit IRC | 04:38 | |
| *** irenab has quit IRC | 04:38 | |
| *** shyu has quit IRC | 04:40 | |
| *** unicell has quit IRC | 04:41 | |
| *** shyu has joined #openstack-nova | 04:41 | |
| *** shyama has joined #openstack-nova | 04:43 | |
| *** VW_ has joined #openstack-nova | 04:43 | |
| *** unicell has joined #openstack-nova | 04:45 | |
| *** markvoelker_ has joined #openstack-nova | 04:48 | |
| *** markvoelker has quit IRC | 04:50 | |
| *** markvoelker_ has quit IRC | 04:52 | |
| *** achanda has joined #openstack-nova | 04:52 | |
| *** VW_ has quit IRC | 04:53 | |
| *** annashen has joined #openstack-nova | 04:57 | |
| *** sudipto has quit IRC | 04:57 | |
| *** markvoelker has joined #openstack-nova | 04:57 | |
| *** fawadkhaliq has joined #openstack-nova | 04:58 | |
| *** yingjun has joined #openstack-nova | 04:59 | |
| *** annashen has quit IRC | 05:02 | |
| *** markvoelker has quit IRC | 05:03 | |
| *** sharkJing has quit IRC | 05:07 | |
| *** sharkJing has joined #openstack-nova | 05:07 | |
| *** markvoelker has joined #openstack-nova | 05:08 | |
| *** moshele has joined #openstack-nova | 05:09 | |
| *** Nikolay_St has quit IRC | 05:13 | |
| *** markvoelker has quit IRC | 05:16 | |
| *** armax has quit IRC | 05:17 | |
| *** ildikov has quit IRC | 05:20 | |
| *** markvoelker has joined #openstack-nova | 05:22 | |
| *** garyk has joined #openstack-nova | 05:25 | |
| *** irenab has joined #openstack-nova | 05:26 | |
| *** markvoelker has quit IRC | 05:28 | |
| *** yingjun has quit IRC | 05:32 | |
| *** afazekas has joined #openstack-nova | 05:32 | |
| openstackgerrit | liaonanhai proposed openstack/nova-specs: Add user_id info: add 'user_id' to REST os-simple-tenant-usage output https://review.openstack.org/198937 | 05:33 |
| *** dheeraj-gupta-4 has joined #openstack-nova | 05:35 | |
| *** moshele has quit IRC | 05:38 | |
| *** kumar has joined #openstack-nova | 05:38 | |
| *** vilobhmm has quit IRC | 05:40 | |
| *** yingjun has joined #openstack-nova | 05:41 | |
| *** otter768 has joined #openstack-nova | 05:42 | |
| *** exploreshaifali has joined #openstack-nova | 05:42 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge block_device_mapping boot tests between v2 and v2.1 https://review.openstack.org/201401 | 05:46 |
| *** sharkJing has quit IRC | 05:47 | |
| *** otter768 has quit IRC | 05:47 | |
| *** sharkJing has joined #openstack-nova | 05:47 | |
| *** ssurana has joined #openstack-nova | 05:48 | |
| *** apuimedo has joined #openstack-nova | 05:52 | |
| *** arnaud___ has quit IRC | 05:53 | |
| *** lyanchih has quit IRC | 05:56 | |
| *** eliqiao has left #openstack-nova | 05:57 | |
| *** annashen has joined #openstack-nova | 05:58 | |
| *** coolsvap has joined #openstack-nova | 06:00 | |
| *** Murali has joined #openstack-nova | 06:01 | |
| *** Kennan has quit IRC | 06:02 | |
| *** Kennan has joined #openstack-nova | 06:03 | |
| *** annashen has quit IRC | 06:03 | |
| *** achanda has quit IRC | 06:07 | |
| *** pradipta has joined #openstack-nova | 06:07 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge block_device_mapping boot tests between v2 and v2.1 https://review.openstack.org/201401 | 06:08 |
| *** lyanchih has joined #openstack-nova | 06:09 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: CONF.allow_resize_to_same_host should check only once in controller https://review.openstack.org/186236 | 06:10 |
| *** lpetrut has joined #openstack-nova | 06:11 | |
| openstackgerrit | liaonanhai proposed openstack/nova-specs: Add user_id info: add 'user_id' to REST os-simple-tenant-usage output https://review.openstack.org/198937 | 06:12 |
| *** coolsvap has quit IRC | 06:13 | |
| *** coolsvap|away is now known as coolsvap | 06:13 | |
| *** dhvanan has joined #openstack-nova | 06:18 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex https://review.openstack.org/197213 | 06:18 |
| dhvanan | Hi, I am new to openstack and I needed help to know if we can make multiple concurrent requests through the nova-api? | 06:20 |
| dhvanan | Could someone help me out! | 06:21 |
| *** abhishekk has joined #openstack-nova | 06:22 | |
| *** achanda has joined #openstack-nova | 06:22 | |
| *** abhishekk has quit IRC | 06:22 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: network_request_obj: Clean up outdated code https://review.openstack.org/201425 | 06:22 |
| *** oro has quit IRC | 06:25 | |
| *** oro has joined #openstack-nova | 06:27 | |
| *** yingjun has quit IRC | 06:29 | |
| openstackgerrit | Rui Chen proposed openstack/nova: Fix incomplete error message of quota exceeded https://review.openstack.org/197460 | 06:29 |
| openstackgerrit | Rui Chen proposed openstack/nova: Set task_state=None when booting instance failed https://review.openstack.org/187145 | 06:29 |
| openstackgerrit | Rui Chen proposed openstack/nova: Improve functional test base for microversion https://review.openstack.org/190063 | 06:29 |
| openstackgerrit | Rui Chen proposed openstack/nova: Fix permission issue of server group API https://review.openstack.org/193117 | 06:29 |
| *** SourabhP has joined #openstack-nova | 06:32 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: libvirt: optimize multipath call to identify IQN https://review.openstack.org/163733 | 06:35 |
| *** eglynn has joined #openstack-nova | 06:35 | |
| *** SourabhP_ has joined #openstack-nova | 06:35 | |
| *** kiran-r has joined #openstack-nova | 06:36 | |
| *** arnaud___ has joined #openstack-nova | 06:36 | |
| *** belmoreira has joined #openstack-nova | 06:36 | |
| openstackgerrit | lyanchih proposed openstack/nova: libvirt - Flavor's quota instance resource not observed for non-ephemeral storage https://review.openstack.org/201019 | 06:36 |
| *** ankit_ag has joined #openstack-nova | 06:37 | |
| *** SourabhP has quit IRC | 06:37 | |
| *** SourabhP_ is now known as SourabhP | 06:37 | |
| *** scheuran has joined #openstack-nova | 06:43 | |
| *** _fortis has joined #openstack-nova | 06:44 | |
| *** tpeoples has quit IRC | 06:44 | |
| *** _fortis has quit IRC | 06:44 | |
| *** tpeoples has joined #openstack-nova | 06:44 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge flavor all_extensions tests between v2 and v2.1 https://review.openstack.org/201430 | 06:47 |
| *** kumar has quit IRC | 06:51 | |
| openstackgerrit | liaonanhai proposed openstack/nova: Add user_id info: add 'user_id' to REST os-simple-tenant-usage output https://review.openstack.org/151522 | 06:51 |
| *** ssurana has quit IRC | 06:52 | |
| *** _fortis has joined #openstack-nova | 06:53 | |
| *** lpetrut has quit IRC | 06:54 | |
| *** yingjun has joined #openstack-nova | 06:55 | |
| *** itzikb_ has joined #openstack-nova | 06:55 | |
| pkholkin | hi folks, could you please look at this patch https://review.openstack.org/#/c/173834/, I think unit-tests are broken without it | 06:55 |
| pkholkin | I wrote to Doug Hellmann for several times | 06:56 |
| *** moshele has joined #openstack-nova | 06:56 | |
| pkholkin | dhellmann: Hello, are you here? Could you please remove -2 from this patch https://review.openstack.org/#/c/173834/ | 06:56 |
| *** kiran-r has quit IRC | 06:58 | |
| *** PpPyg has quit IRC | 06:58 | |
| *** arnaud___ has quit IRC | 06:58 | |
| *** annashen has joined #openstack-nova | 06:59 | |
| *** asettle has quit IRC | 07:00 | |
| *** markus_z has joined #openstack-nova | 07:00 | |
| *** kiran-r has joined #openstack-nova | 07:01 | |
| openstackgerrit | Abhijeet Malawade proposed openstack/nova: POC: Extract cinder-client response to log request IDs https://review.openstack.org/201434 | 07:03 |
| *** annashen has quit IRC | 07:03 | |
| openstackgerrit | Abhijeet Malawade proposed openstack/nova: POC: Extract cinder-client response to log request IDs https://review.openstack.org/201434 | 07:05 |
| *** ssurana has joined #openstack-nova | 07:06 | |
| openstackgerrit | Abhishek Kekane proposed openstack/nova: Execute _poll_shelved_instances only if shelved_offload_time is > 0 https://review.openstack.org/201436 | 07:06 |
| *** SourabhP has quit IRC | 07:07 | |
| *** rdopiera has joined #openstack-nova | 07:08 | |
| itzikb_ | moshele: hi, you may want to attach https://bugs.launchpad.net/nova/+bug/1222990 to your patch | 07:08 |
| openstack | Launchpad bug 1222990 in OpenStack Compute (nova) "Cannot specify subset of PCI devices for PCI passthrough" [Wishlist,Confirmed] - Assigned to Yongli He (yongli-he) | 07:08 |
| *** kiran-r has quit IRC | 07:08 | |
| itzikb_ | moshele: https://review.openstack.org/#/c/199488/ | 07:09 |
| *** ildikov has joined #openstack-nova | 07:10 | |
| *** amotoki_ has joined #openstack-nova | 07:11 | |
| *** humble_ has joined #openstack-nova | 07:13 | |
| *** mutoulbj has quit IRC | 07:14 | |
| *** wanghua has quit IRC | 07:14 | |
| *** mutoulbj has joined #openstack-nova | 07:14 | |
| *** cfriesen has quit IRC | 07:14 | |
| *** _fortis has quit IRC | 07:17 | |
| *** hakimo has joined #openstack-nova | 07:18 | |
| *** rotbeard has joined #openstack-nova | 07:18 | |
| *** hakimo has quit IRC | 07:19 | |
| *** achanda has quit IRC | 07:19 | |
| *** hakimo has joined #openstack-nova | 07:20 | |
| *** Jeffrey4l has quit IRC | 07:21 | |
| *** _fortis has joined #openstack-nova | 07:21 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge "versions" tests between v2 and v2.1 https://review.openstack.org/201439 | 07:22 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge "versions" tests between v2 and v2.1 https://review.openstack.org/201439 | 07:24 |
| *** scheuran has quit IRC | 07:26 | |
| *** moshele has quit IRC | 07:29 | |
| *** jcoufal has joined #openstack-nova | 07:30 | |
| *** ndipanov has quit IRC | 07:34 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: network_request_obj: Clean up outdated code https://review.openstack.org/201425 | 07:36 |
| *** ndipanov has joined #openstack-nova | 07:38 | |
| *** nkrinner_afk is now known as nkrinner | 07:40 | |
| *** jlanoux has joined #openstack-nova | 07:40 | |
| *** browne has quit IRC | 07:41 | |
| *** jyuso1 has left #openstack-nova | 07:42 | |
| *** jistr has joined #openstack-nova | 07:42 | |
| *** jyuso1 has joined #openstack-nova | 07:42 | |
| *** otter768 has joined #openstack-nova | 07:43 | |
| openstackgerrit | Rodolfo Alonso proposed openstack/nova: Added new scheduler filter: AggregateTypeExtraSpecsAffinityFilter https://review.openstack.org/189279 | 07:45 |
| *** otter768 has quit IRC | 07:47 | |
| *** fawadkhaliq has quit IRC | 07:49 | |
| *** alex_klimov has joined #openstack-nova | 07:50 | |
| *** alex_xu has quit IRC | 07:52 | |
| *** Longgeek has quit IRC | 07:53 | |
| openstackgerrit | Andreas Scheuring proposed openstack/nova: Add macvtap as new virtual interface (vif) type to Nova's libvirt driver https://review.openstack.org/182283 | 07:54 |
| *** alex_xu has joined #openstack-nova | 07:55 | |
| *** shuquan has joined #openstack-nova | 07:58 | |
| *** annashen has joined #openstack-nova | 08:00 | |
| *** Longgeek has joined #openstack-nova | 08:00 | |
| *** pradipta has quit IRC | 08:01 | |
| *** deepthi has quit IRC | 08:02 | |
| *** annashen has quit IRC | 08:04 | |
| *** eliqiao has joined #openstack-nova | 08:05 | |
| *** lucasagomes has joined #openstack-nova | 08:06 | |
| *** fawadkhaliq has joined #openstack-nova | 08:11 | |
| *** yingjun has quit IRC | 08:13 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: Revert "Remove useless db call instance_get_all_hung_in_rebooting" https://review.openstack.org/201451 | 08:14 |
| ndipanov | johnthetubaguy, ^^ | 08:14 |
| ndipanov | the sooner this goes the better | 08:15 |
| ndipanov | who else is awake from the core team? | 08:15 |
| *** tpeoples has quit IRC | 08:15 | |
| garyk | ndipanov: why does that need to be reverted? | 08:16 |
| *** eliqiao1 has joined #openstack-nova | 08:17 | |
| ndipanov | because you can't drop methods without a major version bump | 08:17 |
| ndipanov | as always | 08:17 |
| garyk | ndipanov: butthat method is not used. | 08:17 |
| garyk | what is this breaking? | 08:18 |
| *** Jeffrey4l has joined #openstack-nova | 08:18 | |
| ndipanov | it was never ever used? | 08:18 |
| *** eliqiao has quit IRC | 08:18 | |
| garyk | not sure. | 08:18 |
| *** shuquan has quit IRC | 08:19 | |
| garyk | it may have been in a version prior to object :) | 08:19 |
| garyk | still do not think that warrants a revert | 08:19 |
| garyk | can you find code where it was an is used and then we can decide on whthee ot not to revert | 08:19 |
| ndipanov | yeah in that case no | 08:19 |
| garyk | whether (fat finger day) | 08:19 |
| *** Nikolay_St has joined #openstack-nova | 08:20 | |
| *** deepthi has joined #openstack-nova | 08:21 | |
| *** yingjun has joined #openstack-nova | 08:21 | |
| *** shyama has quit IRC | 08:21 | |
| *** amotoki_ has quit IRC | 08:22 | |
| *** gszasz has joined #openstack-nova | 08:23 | |
| *** ssurana has quit IRC | 08:23 | |
| *** ssurana has joined #openstack-nova | 08:26 | |
| ndipanov | so it seems it was actually never used | 08:26 |
| ndipanov | if only they had bothered to write that though | 08:26 |
| *** kiran-r has joined #openstack-nova | 08:29 | |
| *** ssurana has quit IRC | 08:29 | |
| *** kiran-r has quit IRC | 08:29 | |
| *** derekh has joined #openstack-nova | 08:30 | |
| *** pixelbeat_ has joined #openstack-nova | 08:30 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: libvirt: Add post_connection_terminated https://review.openstack.org/172341 | 08:30 |
| *** kiran-r has joined #openstack-nova | 08:31 | |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: libvirt: rename parallels driver to virtuozzo https://review.openstack.org/184311 | 08:32 |
| *** _kiran_ has joined #openstack-nova | 08:35 | |
| *** kiran-r has quit IRC | 08:35 | |
| johnthetubaguy | ndipanov: just taking a look | 08:36 |
| johnthetubaguy | ndipanov: I | 08:36 |
| ndipanov | johnthetubaguy, meh abandoned it | 08:36 |
| ndipanov | it seems that it was literally never used | 08:36 |
| johnthetubaguy | ndipanov: ah, fair enough | 08:36 |
| ndipanov | it still makes little sense to bump the version | 08:36 |
| johnthetubaguy | ndipanov: I kinda hope we got that right | 08:37 |
| ndipanov | they should have just removed it | 08:37 |
| ndipanov | version bump is meaningless there | 08:37 |
| johnthetubaguy | ndipanov: true, like RPC versions | 08:37 |
| ndipanov | right | 08:37 |
| openstackgerrit | Hiroyuki Eguchi proposed openstack/nova: Cannot rebuild a instance booted from volume https://review.openstack.org/201458 | 08:37 |
| ndipanov | garyk, ^ | 08:38 |
| garyk | ndipanov: gracias | 08:40 |
| *** _kiran_ has quit IRC | 08:40 | |
| johnthetubaguy | ndipanov: we might want to update the version bump comment in the test_objects.py thing with the object hashes in there | 08:42 |
| johnthetubaguy | a bump is always safer I guess | 08:42 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: libvirt: Add post_connection_terminated https://review.openstack.org/172341 | 08:48 |
| *** tobe has quit IRC | 08:48 | |
| *** deepthi has quit IRC | 08:49 | |
| garyk | johnthetubaguy: following the nova meeting last week i opened a specless bp as requested by you and matt - https://blueprints.launchpad.net/openstack/?searchtext=vmware-better-display-names | 08:50 |
| johnthetubaguy | garyk: cool, I can approve that, there is no way for me to see newly submitted blueprints, hence the request for the ping | 08:51 |
| *** aix has joined #openstack-nova | 08:51 | |
| garyk | johnthetubaguy: ok, thanks. | 08:52 |
| johnthetubaguy | garyk: that should be all approved now I think | 08:53 |
| *** lpetrut has joined #openstack-nova | 08:53 | |
| *** lpetrut has left #openstack-nova | 08:55 | |
| *** lpetrut has joined #openstack-nova | 08:55 | |
| garyk | johnthetubaguy: thanks! | 08:58 |
| *** tobe has joined #openstack-nova | 08:58 | |
| *** aix has quit IRC | 08:58 | |
| *** annashen has joined #openstack-nova | 09:00 | |
| *** aix has joined #openstack-nova | 09:03 | |
| *** annashen has quit IRC | 09:05 | |
| *** moshele has joined #openstack-nova | 09:05 | |
| moshele | itzikb_: I will look in to that | 09:09 |
| *** jcoufal_ has joined #openstack-nova | 09:09 | |
| *** xgsa has quit IRC | 09:09 | |
| *** xgsa has joined #openstack-nova | 09:10 | |
| *** jcoufal has quit IRC | 09:12 | |
| *** yingjun has quit IRC | 09:13 | |
| *** yingjun has joined #openstack-nova | 09:14 | |
| *** alexpilotti has joined #openstack-nova | 09:16 | |
| *** amotoki_ has joined #openstack-nova | 09:16 | |
| *** xgsa has quit IRC | 09:19 | |
| *** xgsa has joined #openstack-nova | 09:20 | |
| *** shyama has joined #openstack-nova | 09:21 | |
| *** dheeraj-gupta-4 has left #openstack-nova | 09:21 | |
| *** jang has quit IRC | 09:29 | |
| *** jang has joined #openstack-nova | 09:29 | |
| *** shuquan has joined #openstack-nova | 09:30 | |
| *** yingjun has quit IRC | 09:31 | |
| *** marzif_ has joined #openstack-nova | 09:33 | |
| *** shuquan has quit IRC | 09:34 | |
| *** shuquan has joined #openstack-nova | 09:36 | |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: libvirt: rename parallels driver to virtuozzo https://review.openstack.org/184311 | 09:38 |
| *** shuquan has quit IRC | 09:41 | |
| *** jichen has joined #openstack-nova | 09:41 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: add support for provider network portgroup https://review.openstack.org/199477 | 09:41 |
| *** shuquan has joined #openstack-nova | 09:42 | |
| *** shuquan has quit IRC | 09:42 | |
| *** otter768 has joined #openstack-nova | 09:44 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: Fix the incorect PciDeviceList version number https://review.openstack.org/201280 | 09:45 |
| *** shuquan has joined #openstack-nova | 09:45 | |
| *** shuquan has quit IRC | 09:46 | |
| *** shuquan has joined #openstack-nova | 09:48 | |
| *** otter768 has quit IRC | 09:48 | |
| *** dims_ has joined #openstack-nova | 09:48 | |
| *** shuquan has quit IRC | 09:49 | |
| *** shuquan has joined #openstack-nova | 09:50 | |
| *** Longgeek has quit IRC | 09:51 | |
| *** jcoufal has joined #openstack-nova | 09:53 | |
| *** shuquan has quit IRC | 09:54 | |
| *** shuquan has joined #openstack-nova | 09:55 | |
| *** shuquan has quit IRC | 09:56 | |
| *** jcoufal_ has quit IRC | 09:57 | |
| *** shuquan has joined #openstack-nova | 09:58 | |
| *** shyama has quit IRC | 09:58 | |
| *** shuquan has quit IRC | 09:58 | |
| *** shyama has joined #openstack-nova | 09:58 | |
| *** shyama_ has joined #openstack-nova | 10:00 | |
| *** annashen has joined #openstack-nova | 10:01 | |
| *** shyama has quit IRC | 10:03 | |
| *** shyama_ is now known as shyama | 10:03 | |
| *** Kennan has quit IRC | 10:03 | |
| *** Kennan has joined #openstack-nova | 10:03 | |
| *** kiran-r has joined #openstack-nova | 10:03 | |
| *** fawadkhaliq has quit IRC | 10:03 | |
| *** shuquan has joined #openstack-nova | 10:03 | |
| *** annashen has quit IRC | 10:06 | |
| *** Marga_ has quit IRC | 10:12 | |
| *** Marga_ has joined #openstack-nova | 10:13 | |
| *** shuquan has quit IRC | 10:14 | |
| *** shuquan has joined #openstack-nova | 10:16 | |
| *** ociuhandu has joined #openstack-nova | 10:16 | |
| *** Marga_ has quit IRC | 10:16 | |
| *** salv-orlando has quit IRC | 10:16 | |
| *** Marga_ has joined #openstack-nova | 10:16 | |
| *** moshele has quit IRC | 10:17 | |
| *** shuquan has quit IRC | 10:17 | |
| *** shuquan has joined #openstack-nova | 10:18 | |
| *** fawadkhaliq has joined #openstack-nova | 10:18 | |
| *** abhishekk has joined #openstack-nova | 10:21 | |
| *** aix has quit IRC | 10:21 | |
| *** shuquan has quit IRC | 10:22 | |
| *** Longgeek has joined #openstack-nova | 10:22 | |
| *** shuquan has joined #openstack-nova | 10:24 | |
| *** moshele has joined #openstack-nova | 10:24 | |
| *** shuquan has quit IRC | 10:24 | |
| *** shuquan has joined #openstack-nova | 10:26 | |
| *** shuquan has quit IRC | 10:27 | |
| *** jlanoux_ has joined #openstack-nova | 10:27 | |
| *** rdo has quit IRC | 10:27 | |
| *** oro has quit IRC | 10:27 | |
| *** shuquan has joined #openstack-nova | 10:27 | |
| *** exploreshaifali has quit IRC | 10:28 | |
| *** rdo has joined #openstack-nova | 10:29 | |
| *** lyanchih has quit IRC | 10:29 | |
| ndipanov | abhishekk, hey - so will you re-spin this one? https://review.openstack.org/#/c/192986/ | 10:29 |
| *** shuquan has quit IRC | 10:30 | |
| *** jlanoux has quit IRC | 10:30 | |
| abhishekk | ndipanov: yes, I am working on unit tests | 10:31 |
| *** moshele has quit IRC | 10:32 | |
| *** IanGovett has joined #openstack-nova | 10:32 | |
| *** moshele has joined #openstack-nova | 10:33 | |
| ndipanov | abhishekk, ok I will abandon my change here https://review.openstack.org/#/c/200621/ | 10:33 |
| ndipanov | ps - it has all the unittest done so you can steal ideas from there | 10:33 |
| *** mutoulbj has quit IRC | 10:34 | |
| *** aix has joined #openstack-nova | 10:34 | |
| abhishekk | ndipanov: sure, I will add your name in co-author | 10:35 |
| ndipanov | no no | 10:35 |
| ndipanov | you don't have to do that | 10:35 |
| ndipanov | you can't use those exact tests - you'd have to change them | 10:35 |
| ndipanov | but at least you can see where to do the tests | 10:35 |
| abhishekk | ndipanov: yes I need to change them, thank you for the suggestion | 10:36 |
| *** beagles has quit IRC | 10:36 | |
| *** deepthi has joined #openstack-nova | 10:40 | |
| alex_xu | ndipanov: just want to ask the generic case why we need bump major version when remove remote method. In normally we should keep the deprecate method in one release cycle, after use upgrade to new release cycle, then nobody use the old method, then we can remove them safely? so I didn't get the reason why major version bump needed | 10:40 |
| alex_xu | s/after use/after user/ | 10:40 |
| ndipanov | right | 10:40 |
| ndipanov | er | 10:41 |
| ndipanov | no | 10:41 |
| ndipanov | that's not correct | 10:41 |
| ndipanov | by keeping the version at 1.x | 10:41 |
| *** b3nt_pin has joined #openstack-nova | 10:41 | |
| *** b3nt_pin is now known as Guest83150 | 10:42 | |
| alex_xu | ndipanov: need bump to 2.x? so why? | 10:42 |
| ndipanov | we are saying that we are backwards compatible with any 1.x | 10:42 |
| ndipanov | if you remove the method | 10:42 |
| ndipanov | you can't really be backwards compatible as the new code will have no idea about that method anymore | 10:42 |
| ndipanov | it's just a coincidence that we bump the major version | 10:42 |
| garyk | is there any chance that one or more cores can please look at https://review.openstack.org/#/c/188789/ - this patch is blokcing a ton of other patches and it would open the flood gates... | 10:43 |
| ndipanov | very soon after we bump the realease number | 10:43 |
| ndipanov | alex_xu, ^ | 10:43 |
| ndipanov | in the case from today | 10:43 |
| ndipanov | alex_xu, does that make sense? | 10:43 |
| alex_xu | ndipanov: I think I didn't understand your word yet :( | 10:45 |
| alex_xu | ndipanov: I just ask the generic case, not for https://review.openstack.org/#/c/201451/1 | 10:45 |
| ndipanov | so bumping the major version means - we are no longer compatible with 1.x | 10:45 |
| *** Longgeek_ has joined #openstack-nova | 10:46 | |
| ndipanov | so if you want to make a backwards incompatible change | 10:46 |
| alex_xu | ndipanov: yes, but if just bump minor version, it won't break anything also? so bump the major version just for our verison policy? | 10:46 |
| ndipanov | and removing a method is | 10:46 |
| ndipanov | what do you mean it won't break anything | 10:46 |
| ndipanov | if 1.2 asks the conductor to do 'foo_method' which is remotable | 10:47 |
| ndipanov | and conductor is at 1.7 which does not have the method anymore | 10:47 |
| ndipanov | it breaks | 10:47 |
| ndipanov | the rule is: you can't do breaking changes | 10:47 |
| ndipanov | if you want to do breaking changes - wait for the major version bump and in the meantime add a comment saying it's deprecated | 10:48 |
| alex_xu | ndipanov: what happend if after 1.2 there isn't any code invoke foo_method, then if foo_method disappeared in 1.7, then that will safe | 10:48 |
| *** Longgeek has quit IRC | 10:48 | |
| *** jlanoux has joined #openstack-nova | 10:49 | |
| ndipanov | alex_xu, if it was invoked in any version of the code from 1.0 | 10:49 |
| sdague | anyone want to take a look at - https://review.openstack.org/#/c/174245/10 - it's really old review, got 2 reviews stacked on top of it that are ready to go, and hasn't got any -1s on it. | 10:49 |
| ndipanov | it needs to be kept | 10:49 |
| ndipanov | alex_xu, I don't understand what part is unclear | 10:50 |
| ndipanov | the only rule is - you can't make a change to the objects interface that will break any code that has the same major version | 10:52 |
| ndipanov | removing a method (if it was ever used) is one of those changes | 10:52 |
| *** jlanoux_ has quit IRC | 10:52 | |
| johnthetubaguy | alex_xu: you have to think about old and new deployments, and how upgrade works (old computes with new control plane), etc. | 10:57 |
| *** deepthi has quit IRC | 10:58 | |
| *** jichen has quit IRC | 10:58 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Change default compute api version https://review.openstack.org/184658 | 11:00 |
| alex_xu | ndipanov: johnthetubaguy, thanks for the explain, let me think about it~ | 11:01 |
| ndipanov | alex_xu, it's always good to remind yourself that in case of objects | 11:01 |
| johnthetubaguy | alex_xu: its not just because we are doing sem_ver, its how we keep upgrades working basically | 11:01 |
| *** afazekas has quit IRC | 11:02 | |
| ndipanov | the conductor knows all the latest versions (aka we need to upgrade conductor first | 11:02 |
| ndipanov | ) | 11:02 |
| johnthetubaguy | ndipanov: hoping we can get some more docs on this stuff asap, it easy for folks to loose the context here | 11:02 |
| *** annashen has joined #openstack-nova | 11:02 | |
| ndipanov | johnthetubaguy, well there are docs written by dansmith in the context of upgrades I think | 11:02 |
| ndipanov | maybe not posted in a well known place :/ | 11:03 |
| johnthetubaguy | ndipanov: yes, thats part of it, need to get this into devref | 11:03 |
| johnthetubaguy | its on my list, for the priority BP, but the list is getting silly long now | 11:03 |
| ndipanov | alex_xu, so that lets you have different versions of the code on compute nodes requesting stuff from the conductor | 11:03 |
| ndipanov | but for this to work - conductor needs to know all of their versions to be able to "backport" | 11:04 |
| ndipanov | so even if let's say no code in 1.7 uses "foo_method" | 11:04 |
| ndipanov | some code with 1.2 might | 11:04 |
| ndipanov | so the 1.7 conductor needs to know how to do it | 11:05 |
| sdague | gmann: on https://review.openstack.org/#/c/197394/2/nova/tests/functional/v3/test_fixed_ips.py,cm it would be really good to also negative test the new attributes not existing on v2.0 | 11:05 |
| *** sudipto has joined #openstack-nova | 11:07 | |
| alex_xu | ndipanov: yea, I think I just didn't get the major version why need bump | 11:07 |
| *** annashen has quit IRC | 11:07 | |
| *** openstackgerrit has quit IRC | 11:09 | |
| alex_xu | ndipanov: nvm..let me spend some time to think about it. thanks for your explain | 11:09 |
| *** jaypipes has joined #openstack-nova | 11:09 | |
| *** coolsvap is now known as coolsvap_ | 11:09 | |
| *** openstackgerrit has joined #openstack-nova | 11:10 | |
| openstackgerrit | John Garbutt proposed openstack/nova: docs: add the scheduler evolution plans https://review.openstack.org/192260 | 11:12 |
| alex_xu | ndipanov: one more similar patch https://review.openstack.org/#/c/199971/5/nova/objects/instance.py | 11:13 |
| *** mutoulbj has joined #openstack-nova | 11:15 | |
| ndipanov | alex_xu, well bumping the major version is kind of a coordinated thing | 11:15 |
| *** moshele has quit IRC | 11:16 | |
| ndipanov | so it means - we as a project now agree that we will not support anything older than X | 11:16 |
| ndipanov | and we can remove a bunch of code | 11:16 |
| ndipanov | so then your code that was marked deprecated can be removed | 11:16 |
| *** afazekas has joined #openstack-nova | 11:17 | |
| sdague | alex_xu: can you rebase https://review.openstack.org/#/c/160206/16 ? | 11:18 |
| sdague | the rest of that stack is in merge conflict | 11:18 |
| ndipanov | alex_xu,, don't rebase unless you also respond to any pending comments | 11:19 |
| ndipanov | sdague, ^ | 11:19 |
| ndipanov | I've seen valid points get missed that way | 11:19 |
| sdague | ndipanov: there are no pending comments that I see | 11:20 |
| *** alex_klimov has quit IRC | 11:21 | |
| sdague | the last -1 in the logs is March, 13 revisions ago | 11:21 |
| ndipanov | sdague, cool just sayin' that it's something that happens | 11:21 |
| sdague | ndipanov: sure, it just seemed like you thought something specific was going on in this patch | 11:22 |
| *** thorst has joined #openstack-nova | 11:22 | |
| ndipanov | no ... | 11:22 |
| ndipanov | but people see the RED MERGE CONFLICT | 11:22 |
| ndipanov | and quickly rebase thinking that they will get back to the -1s... | 11:22 |
| sdague | ndipanov: sure, but I was reviewing the stack, and it was ready to go, except a conflict on patch 2 | 11:23 |
| ndipanov | yeah in that case all good | 11:23 |
| sdague | hey, so in https://review.openstack.org/#/c/189632/8 | 11:23 |
| ndipanov | I typed instead of clicking on the actual patch | 11:23 |
| ndipanov | yeeeees :) | 11:23 |
| sdague | I really don't like the idea of the fact that we have an API, and then we ignore params and don't tell the user | 11:23 |
| sdague | because a log warn doesn't get to the user | 11:24 |
| ndipanov | right | 11:24 |
| ndipanov | hence the upgradeImpact -> make this a releasenote kind of thing | 11:24 |
| ndipanov | release notes don't go to users necessarily either | 11:24 |
| ndipanov | but not sure how to resolve this any other way | 11:25 |
| sdague | sure, but if this is out in the field, how is anyone going to know | 11:25 |
| sdague | I'd like to figure out a way to bubble this issue back to the user | 11:25 |
| sdague | on the api call | 11:25 |
| jaypipes | *yawn* | 11:25 |
| johnthetubaguy | sdague: it does feel like we need virt driver specific validation on that API call, I guess? | 11:26 |
| ndipanov | johnthetubaguy, but that's not something that we can do really | 11:26 |
| sdague | johnthetubaguy: yeh, that's kind of what it's feeling like | 11:26 |
| sdague | ndipanov: why not? | 11:26 |
| johnthetubaguy | ndipanov: we kinda do already for live-migrate, but its not something we should copy, probably | 11:26 |
| ndipanov | johnthetubaguy, we do? | 11:27 |
| ndipanov | in the API? | 11:27 |
| sdague | so what API calls would trigger this? | 11:27 |
| ndipanov | boot | 11:27 |
| sdague | ndipanov: and attach? or only boot? | 11:27 |
| ndipanov | this particular patch boot, there is another one to fix it for attach | 11:27 |
| sudipto | garyk, Hi, regarding your comments here: https://review.openstack.org/#/c/200823/4/nova/tests/unit/api/openstack/compute/contrib/test_extended_virtual_interfaces_net.py | 11:27 |
| johnthetubaguy | ndipanov: yeah, after we pick the host, we check its compatible, by calling the specific compute node, and return errors the the user, but lets not go there | 11:27 |
| sudipto | garyk, I haven't dropped the uuid field? | 11:27 |
| ndipanov | johnthetubaguy, that's in the conductor | 11:27 |
| johnthetubaguy | ndipanov: sdague: in this case, it feels like libvirt failing builds that request a device, would be more consistent | 11:28 |
| ndipanov | after the API returns a 200 | 11:28 |
| johnthetubaguy | ndipanov: yeah, but its an RPC call into the conductor | 11:28 |
| sudipto | garyk, i just changed the net_uuid to be an int instead of a string, no? | 11:28 |
| ndipanov | johnthetubaguy, it's a call | 11:28 |
| ndipanov | hmmm | 11:28 |
| johnthetubaguy | ndipanov: its a horrible pattern, just saying we do that already, I would love a better pattern, Tasks was always pencilled in to fix that, but lets not go there | 11:29 |
| johnthetubaguy | so rather than silently fail, you could fail all builds that request a device id right/ | 11:29 |
| johnthetubaguy | ? | 11:29 |
| *** alex_klimov has joined #openstack-nova | 11:29 | |
| sdague | johnthetubaguy: yeh, I wonder if there is something sane we could do to return the error directly on the API | 11:29 |
| sdague | because, hypervisor_type is in the compute_nodes table | 11:29 |
| johnthetubaguy | sdague: in a multi virt driver cloud, its tricky right? I mean, should be possible, but its tricky as we don't have proper flavor modeling | 11:30 |
| ndipanov | so we could return an error for a device requested sure... | 11:30 |
| ndipanov | that would break everyone but they would know it | 11:31 |
| ndipanov | but the thing is | 11:31 |
| johnthetubaguy | ndipanov: yeah, its a trade off, but that feels more... honest and discoverable | 11:31 |
| ndipanov | libvirt guest still would not honor this device name in most OSs | 11:31 |
| sdague | johnthetubaguy: sure, but, honestly, if we solved the easy case, and said well in multi hypervisor, well things are complicated | 11:31 |
| johnthetubaguy | sdague: maybe | 11:31 |
| ndipanov | so from the perspective of the user this is less broken (maybe) since now at least nova gives you waht you actually see in the instance (most of the time) | 11:32 |
| ndipanov | but that's just imho | 11:32 |
| ndipanov | calling it a day and saying you can't specify device names anymore go away might be a solution too | 11:33 |
| ndipanov | this particular one seemed like it gets us where we want to be and the only thing it "breaks" is something that is already broken | 11:34 |
| johnthetubaguy | ndipanov: so we have drivers that work when you do that, well roughly work most of the time, and we have users using that, so its tricky to do that across the board, hence the discoverability point really | 11:34 |
| ndipanov | right | 11:34 |
| ndipanov | there are several ways to do this I guess | 11:35 |
| ndipanov | make it a filter | 11:35 |
| ndipanov | make it a config option that a libvirt only cloud can set | 11:35 |
| garyk | sudipto: srry, just saw your meesage now | 11:35 |
| sudipto | garyk, np :) | 11:35 |
| garyk | sudipto: the net_uuid should be a string (repesenting a uuid) as far as I understand | 11:36 |
| garyk | but then again mayne i am wrong :) | 11:36 |
| ndipanov | in defence of the proposed approach - it gets us three, by changing an already broken interaction so _might_ be worth it | 11:36 |
| ndipanov | sdague, ^ | 11:36 |
| sudipto | garyk, 'network_id': fields.IntegerField(), | 11:37 |
| sudipto | --the uuid is a separate field that is a UUIDField object, did you mean i change net_id to uuid? | 11:37 |
| sdague | ndipanov: yeh, I think the end state of the internals is mostly the right thing to do, I just want to take a couple of days and figure out if there is a useful way we can expose this back to the user | 11:37 |
| *** tobe has quit IRC | 11:37 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added InstanceTagExists exception https://review.openstack.org/201521 | 11:38 |
| *** tobe has joined #openstack-nova | 11:39 | |
| *** lyanchih has joined #openstack-nova | 11:39 | |
| sudipto | garyk, i think the older test file's nomenclature might have not matched with the nova objects nomenclature (?) | 11:39 |
| garyk | sudipto: ok, let me take a look agin | 11:41 |
| jaypipes | sdague: what about a microversion that removed the ability to request/set the device name? | 11:42 |
| *** afazekas has quit IRC | 11:42 | |
| *** Guest83150 is now known as beagles | 11:42 | |
| sudipto | garyk, i am wondering if get_vif_by_mac_address should actually be returning a nova object instead of a dict that it returns currently. | 11:42 |
| sudipto | garyk, in that test i mean. | 11:43 |
| *** exploreshaifali has joined #openstack-nova | 11:43 | |
| *** lucasagomes is now known as lucas-hungry | 11:44 | |
| ndipanov | sdague, one thing we want to keep in mind for the API disucssion - people writing an API for the services they run have way more data to go on than us - they know exactly what people are using and how, and can make informed decisions based on that | 11:44 |
| *** otter768 has joined #openstack-nova | 11:44 | |
| garyk | sudipto: it is kind of odd - the comment in the code indicate that the support should be dropped :) | 11:45 |
| ndipanov | we can't - and I fear that "default to making everything even a bugfix for something that never worked a version" will turn out to be too expensive | 11:45 |
| garyk | from what i understand it returns an object | 11:45 |
| *** tobe has quit IRC | 11:46 | |
| sudipto | garyk, yeah - i mean - i am not exactly sure of what all places the code is affected, so it's a bit of a hit and trial too - and as we speak i discovered that - this method is actually returning a dict instead of an object in the tests (which is not covered as a part of the review - but is an agreement for sure). :) | 11:46 |
| sdague | ndipanov: so, why is RPC, objects, and and database different? | 11:46 |
| sudipto | garyk, ok - let me have that method return a nova object...then. | 11:47 |
| garyk | sudipto: ok, sounds good | 11:47 |
| sdague | jaypipes: so, I thought some of the hypervisors supported the device setting | 11:47 |
| *** casusbelli has joined #openstack-nova | 11:48 | |
| ndipanov | sdague, I guess it's not but it's not really consumed by users... I think that we don't fully grasp the costs of "each commit is backwards compatible" | 11:48 |
| ndipanov | but it is what it is | 11:48 |
| jaypipes | sdague: I thought we were trying to get away from different behaviour depending on the underlying implementnation/driver? | 11:48 |
| ndipanov | jaypipes, in that case we should outright ban device names | 11:48 |
| *** dhvanan has quit IRC | 11:49 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: libvirt: Add post_connection_terminated https://review.openstack.org/172341 | 11:49 |
| *** otter768 has quit IRC | 11:49 | |
| sdague | jaypipes: maybe, though I don't know all the implications of throwing on this feature because libvirt can't do it correctly | 11:49 |
| sdague | i.e. don't take down a fence unless you know why it's there - http://www.chesterton.org/taking-a-fence-down/ | 11:49 |
| ndipanov | sdague, ec2 api needs it | 11:50 |
| ndipanov | well | 11:50 |
| ndipanov | they could maybe in theory work around it by storing it on their side | 11:50 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added InstanceTagExists exception https://review.openstack.org/201521 | 11:50 |
| *** moshele has joined #openstack-nova | 11:52 | |
| *** lyanchih has quit IRC | 11:52 | |
| sdague | ndipanov: so, my point is that we added versioning into all our internal data structures because it was impossible to build a complex system like Nova without it. Nova isn't the end all for users. We expect them to build complex applications on top of it. How can we expect them to do that without the same tools we gave ourselves. | 11:53 |
| *** jichen has joined #openstack-nova | 11:54 | |
| *** sudipto has quit IRC | 11:56 | |
| jaypipes | sdague: I guess I'm not quite seeing how a microversion doesn't solve this problem. If we add a microversion that removes the ability to request a device name, then the EC2 API folks can block on that microversion (i.e. return MIN_VERSION = N - 1 where N == microversion introducing no device name specified) until they store the device name themselves and work around that... | 11:56 |
| ndipanov | sdague, well nova rest API has widely different semantics than any RPC API - if we continue down that road what we'll find is that we need to fully redesign the API | 11:57 |
| sdague | jaypipes: is it only ec2? | 11:57 |
| ndipanov | jaypipes, that still means being backwards compatible | 11:57 |
| ndipanov | that's a massive cost | 11:57 |
| sdague | jaypipes: mostly I just don't know all the implications of getting rid of this | 11:57 |
| ndipanov | that makes sense only if we have 3.x on the horizon | 11:57 |
| ndipanov | which we don't | 11:57 |
| ndipanov | nova rest api has no way to tell the user that it has no idea really about these things it just returned | 11:58 |
| ndipanov | if it was a task we could query later - this problem would not exist | 11:58 |
| ndipanov | but it's not | 11:58 |
| ndipanov | nova API is a blatant lie | 11:59 |
| ndipanov | :) | 11:59 |
| ndipanov | the solution posted is pragmatic acknowledigng the current limitations | 12:00 |
| ndipanov | acknowledging* | 12:02 |
| jaypipes | ndipanov: the microversion would be backwards-compatible, no? the behaviour would not change for clients requesting an earlier microversion. | 12:03 |
| ndipanov | jaypipes, but that means keeping backwards compatibility in nova code | 12:03 |
| sdague | well, we never acknowledged it to the user. That's my concern. We changed behavior on them (in potentially a significant way) and didn't tell them. | 12:03 |
| openstackgerrit | John Garbutt proposed openstack/nova: docs: split up upgrade and upgrade futures https://review.openstack.org/201529 | 12:03 |
| *** annashen has joined #openstack-nova | 12:03 | |
| *** shyu has quit IRC | 12:03 | |
| jaypipes | ndipanov: yes, for a time, it does indeed. that's what microversions are all about. | 12:03 |
| *** exploreshaifali has quit IRC | 12:03 | |
| sdague | johnthetubaguy, jaypipes, ndipanov : anyone want to take a look at this - https://review.openstack.org/#/c/174245/10 - seemed pretty straight forward | 12:04 |
| jaypipes | sdague: not quite following you... | 12:04 |
| jaypipes | sdague: looking... | 12:04 |
| ndipanov | yes I have it open but jaypipes will probably do it before me | 12:05 |
| ndipanov | but now that I have you here gentlemen https://review.openstack.org/#/c/201280/ | 12:05 |
| *** mwagner_lap has quit IRC | 12:05 | |
| ndipanov | the kilo backport of this will look differently as discussed with dansmith yesterday | 12:06 |
| openstackgerrit | John Garbutt proposed openstack/nova: docs: split up upgrade and upgrade futures https://review.openstack.org/201529 | 12:06 |
| *** raildo has joined #openstack-nova | 12:07 | |
| *** salv-orlando has joined #openstack-nova | 12:08 | |
| *** annashen has quit IRC | 12:08 | |
| ndipanov | hmmm actually | 12:09 |
| *** edmondsw has joined #openstack-nova | 12:09 | |
| sdague | we seem to end up with a lot of these rpc version fixes after the fact. Is there a way to test and prevent this in the future? | 12:10 |
| *** afazekas has joined #openstack-nova | 12:10 | |
| *** openstack has joined #openstack-nova | 12:12 | |
| jaypipes | sdague: I | 12:13 |
| *** edmondsw has quit IRC | 12:14 | |
| jaypipes | 'm a little annoyed with it because it's furthering a flawed database model. | 12:14 |
| *** gmann_ has joined #openstack-nova | 12:14 | |
| sdague | jaypipes: ok, well how would you rather be doing this? Can you get a discussion going on it? | 12:16 |
| sdague | and, it's already in, so is all this work that uses it blocked until this gets redone, or can this continue? | 12:16 |
| *** Longgeek_ has quit IRC | 12:17 | |
| ndipanov | sdague, well | 12:17 |
| *** dprince has joined #openstack-nova | 12:18 | |
| jaypipes | sdague: how long has this task log database model been around... | 12:18 |
| ndipanov | sdague, basically we should be building a hash tree in our tests | 12:18 |
| *** Longgeek has joined #openstack-nova | 12:19 | |
| *** edmondsw has joined #openstack-nova | 12:19 | |
| *** fawadkhaliq has quit IRC | 12:20 | |
| ndipanov | so instead of having a version that's a number - we should have a fingerprint that is a hash of all the things that are relevant to the current version including hashes of related objects | 12:20 |
| ndipanov | (we could then alias that to a human readable version) | 12:21 |
| ndipanov | where have I seen something like that before... hmmm | 12:21 |
| ndipanov | :) | 12:21 |
| sdague | jaypipes: tasklog the object showed up here - https://review.openstack.org/#/c/174244/10 | 12:22 |
| *** aysyd has joined #openstack-nova | 12:23 | |
| *** ig0r_ has joined #openstack-nova | 12:23 | |
| jaypipes | sdague: gimme a sec.. on a phone call :) | 12:24 |
| sdague | task_log in the db goes back.... to - https://review.openstack.org/#/c/9455/7/nova/db/sqlalchemy/migrate_repo/versions/108_task_log.py,cm | 12:24 |
| sdague | Jul 10, 2012 6:28 PM | 12:24 |
| *** annegent_ has quit IRC | 12:24 | |
| *** annegentle has joined #openstack-nova | 12:25 | |
| *** READ10 has joined #openstack-nova | 12:25 | |
| *** garyk has quit IRC | 12:27 | |
| *** gmann_ has quit IRC | 12:29 | |
| jaypipes | sdague: k, sorry, off call now. | 12:31 |
| jaypipes | sdague: yes, I see the db model is an old one... | 12:31 |
| jaypipes | sdague: I guess I'm just a bit disappointed we couldn't have a strong object model that didn't pin us to the broken (denormalized) db model. | 12:32 |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: libvirt:Rsync remote FS driver was added https://review.openstack.org/146887 | 12:32 |
| sdague | jaypipes: ok, well that's changable over time right? | 12:32 |
| sdague | I guess, does this make things worse? | 12:32 |
| sdague | or just not improve them the way you want them improved | 12:32 |
| *** bkopilov has quit IRC | 12:32 | |
| sdague | honestly, I don't know, I'm trying to figure out if hans' work should be nixed or merged | 12:33 |
| *** yingjun has joined #openstack-nova | 12:33 | |
| jaypipes | sdague: well, yes it's changeable in some ways. not in others... for instance, you can't change the name of the object. however I suppose we could just essentially deprecate the object and replace with new Task and TaskItem objects | 12:33 |
| *** moshele has quit IRC | 12:34 | |
| openstackgerrit | ZhaoBo proposed openstack/nova: Fixed raise SecurityGroupCannotBeApplied as network port_security off https://review.openstack.org/188260 | 12:34 |
| *** moshele has joined #openstack-nova | 12:35 | |
| *** Piet has quit IRC | 12:36 | |
| *** moshele has quit IRC | 12:37 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge block_device_mapping boot tests between v2 and v2.1 https://review.openstack.org/201401 | 12:37 |
| *** dave-mccowan has joined #openstack-nova | 12:38 | |
| *** coolsvap_ is now known as coolsvap|away | 12:39 | |
| *** lucas-hungry is now known as lucasagomes | 12:40 | |
| johnthetubaguy | sdague: task log object, oh dear, thats the third version of the same thing... whats the context of the patch chain we have here? | 12:42 |
| *** busterswt has joined #openstack-nova | 12:42 | |
| *** moshele has joined #openstack-nova | 12:42 | |
| *** bnemec has joined #openstack-nova | 12:42 | |
| *** markvoelker has joined #openstack-nova | 12:43 | |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for quota_class_create/update https://review.openstack.org/160206 | 12:43 |
| openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for quota_class_get_all_by_name https://review.openstack.org/160205 | 12:43 |
| openstackgerrit | Alex Xu proposed openstack/nova: Cleanup quota_class unittest with appropriate request context https://review.openstack.org/150351 | 12:43 |
| alex_xu | sdague: ^ just rebased | 12:43 |
| johnthetubaguy | jaypipes: sdague: this sounds like good stuff, but do we have the plan written down anywhere for this bit? | 12:44 |
| jaypipes | johnthetubaguy: no. I didn't even know about it until just this morning when I saw sdague | 12:46 |
| jaypipes | 's link. | 12:46 |
| johnthetubaguy | jaypipes: ok, so I feel a bit better that its not just me | 12:46 |
| jaypipes | johnthetubaguy: I don't think it's terrible, though... we can evolve to a Task and TaskItem model later if need be. | 12:46 |
| jaypipes | johnthetubaguy: at least this gets rid of the conductor calls. | 12:46 |
| jaypipes | direct conductor calls... | 12:47 |
| jaypipes | sdague: ^^ ok, I am cool with this for the time being. | 12:47 |
| jaypipes | sdague: will review fully shortly. | 12:47 |
| jaypipes | sorry for the fire drill. | 12:47 |
| *** markvoelker has quit IRC | 12:47 | |
| johnthetubaguy | jaypipes: sdague: yeah, at a quick glance it looks cool, just feels like we should add a retrospective blueprint (maybe spec) to report the progress here | 12:48 |
| jaypipes | johnthetubaguy: sure, good idea. | 12:48 |
| jaypipes | johnthetubaguy: I'll try to remember to do that sometime this week.. | 12:48 |
| johnthetubaguy | jaypipes: ah, that would be awesome, thank you | 12:48 |
| johnthetubaguy | jaypipes: my main worry is what it means for instance actions and instance faults, and the cells API redirection work, as it seems to conflict a little, I would love alaski to take a look | 12:50 |
| *** lyanchih has joined #openstack-nova | 12:51 | |
| *** ivasev has joined #openstack-nova | 12:55 | |
| kashyap | cburgess: Heya, I unassigned this bug, as there is no word on it since you said "look into it" about 6 months ago. If you have more concreten analysis/patches, feel free to undo my action -- https://bugs.launchpad.net/nova/+bug/1334398. | 12:56 |
| openstack | Launchpad bug 1334398 in OpenStack Compute (nova) "libvirt live_snapshot periodically explodes on libvirt 1.2.2 in the gate" [High,In progress] | 12:56 |
| *** markvoelker has joined #openstack-nova | 12:57 | |
| *** eglynn has quit IRC | 12:59 | |
| alex_xu | johnthetubaguy: sdague, andreykurilin let's talk about microversion client? | 13:00 |
| andreykurilin | alex_xu: hi | 13:00 |
| *** whenry_ has joined #openstack-nova | 13:01 | |
| alex_xu | andreykurilin: hi | 13:01 |
| johnthetubaguy | alex_xu: ah, yes, can you PM me the link for that? | 13:01 |
| *** markvoelker has quit IRC | 13:01 | |
| alex_xu | johnthetubaguy: https://review.openstack.org/152569 | 13:01 |
| *** markvoelker has joined #openstack-nova | 13:01 | |
| johnthetubaguy | alex_xu: ah, gotcha | 13:01 |
| *** oro has joined #openstack-nova | 13:01 | |
| alex_xu | johnthetubaguy: sdague, the current implementation adds version negotiation before each API call. | 13:02 |
| *** fawadkhaliq has joined #openstack-nova | 13:02 | |
| *** ZZelle has quit IRC | 13:02 | |
| alex_xu | andreykurilin: ^ is it right? | 13:02 |
| *** ZZelle has joined #openstack-nova | 13:03 | |
| andreykurilin | alex_xu: it's true for "latest" versions | 13:03 |
| *** annashen has joined #openstack-nova | 13:04 | |
| alex_xu | so two questions, 1. we said we only implement version negotiation for CLI, 2. The microversion is global version, whether we should discover version for each API? | 13:04 |
| *** mdrabe has joined #openstack-nova | 13:06 | |
| *** markvoelker has quit IRC | 13:06 | |
| *** whenry_ has quit IRC | 13:06 | |
| *** rlrossit has joined #openstack-nova | 13:06 | |
| *** shuquan has joined #openstack-nova | 13:07 | |
| *** shuquan has quit IRC | 13:07 | |
| sdague | I think on point 1. yes, for the API we shouldn't support implicit version negotiation, we should just provide the CLI as an example of how a client might want to do that themselves. | 13:08 |
| *** whenry_ has joined #openstack-nova | 13:08 | |
| *** shuquan has joined #openstack-nova | 13:08 | |
| *** dims_ has quit IRC | 13:09 | |
| *** annashen has quit IRC | 13:09 | |
| *** krtaylor has quit IRC | 13:09 | |
| alex_xu | sdague: so what mean for explicit version negotiation? user use version discover version manually, or there is parameter enable API bit discover version? | 13:10 |
| *** markvoelker has joined #openstack-nova | 13:10 | |
| *** dims_ has joined #openstack-nova | 13:10 | |
| *** krtaylor has joined #openstack-nova | 13:10 | |
| sdague | alex_xu: honestly, I just assumed the application writer would put that into their own code | 13:11 |
| alex_xu | johnthetubaguy: sdague andreykurilin, I create etherpad, want to write down how to use our client lib, https://etherpad.openstack.org/p/microversion_client, maybe this is helpful our communicate | 13:11 |
| alex_xu | sdague: yea, that I remember we agreement on | 13:12 |
| *** dims_ has quit IRC | 13:13 | |
| *** dims_ has joined #openstack-nova | 13:13 | |
| andreykurilin | alex_xu: https://etherpad.openstack.org/p/microversion_client looks like users will write a lot of lines of code to use our client. | 13:14 |
| andreykurilin | *will need to write | 13:14 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: Adding user_id handling to keypair index, show and create api calls https://review.openstack.org/70485 | 13:15 |
| johnthetubaguy | andreykurilin: I think I had assumed most users will just code to one version of the API, and stick with that until they need something more | 13:15 |
| andreykurilin | alex_xu: so what the negative side of using putting discovering latest version to API? | 13:15 |
| sdague | johnthetubaguy: ++ | 13:15 |
| sdague | andreykurilin: because the point of the API here was to be explicit | 13:16 |
| *** shuquan has quit IRC | 13:16 | |
| alex_xu | johnthetubaguy: ++ | 13:16 |
| johnthetubaguy | andreykurilin: we want people to code to the specific versions really, rather than just "try get the latest stuff" | 13:16 |
| johnthetubaguy | now the issues is when you work across multiple clouds | 13:16 |
| johnthetubaguy | but the CLI is a very easy case to fix, but the solution we pick for the CLI probably isn't want we want most folks to do | 13:17 |
| *** vladikr has joined #openstack-nova | 13:17 | |
| *** shuquan has joined #openstack-nova | 13:17 | |
| sdague | I think it's fine to add a utility method here that lives above Client for getting the latest client you can get. It would be used by the cli code, and a user could choose to use it. But Client(...) down should be explicit version I think | 13:17 |
| andreykurilin | ok | 13:17 |
| *** abhishekk has quit IRC | 13:17 | |
| sdague | and that's the only real hang up at this point right | 13:17 |
| johnthetubaguy | sdague: +1 I think | 13:17 |
| sdague | whether we discover / negotiate above or below Client() | 13:17 |
| johnthetubaguy | some of this is probably more of a docs effort too | 13:18 |
| johnthetubaguy | well, once we get past this initial bit | 13:18 |
| *** changbl has quit IRC | 13:18 | |
| sdague | right, but I mean that's why the patches are blocked | 13:18 |
| andreykurilin | In this case, the question: "latest" version should be global or per API? | 13:19 |
| alex_xu | andreykurilin: global, and we should let user use latest for Client() I think | 13:19 |
| alex_xu | if use latest, the Client() may break | 13:20 |
| alex_xu | when server latest version > client latest version | 13:20 |
| sdague | alex_xu: I thought we just said Client() would always be explicit | 13:20 |
| sdague | which means we shouldn't put latest into it | 13:20 |
| alex_xu | sdague: yea, ..I forget type 'not' again... | 13:21 |
| sdague | andreykurilin: yes, and it's global | 13:21 |
| andreykurilin | So if some API will be deprecated in N-10 version, client should discover N version(when both client and server side supports N version). right? | 13:21 |
| rlrossit | dansmith: ping | 13:21 |
| alex_xu | andreykurilin: the Client() with explicit version, if that API didn't existed in requested version, we can just raise exception, like in the ethepad line 12 | 13:23 |
| *** RuiChen has quit IRC | 13:23 | |
| andreykurilin | alex_xu: yes, Client would always be explicit, but it seems to me we agreed that we can provide some utils to find latest version. so users will be able to call something like client.Client(find_latest()) | 13:25 |
| raildo | sdague: ping, do you have some free minutes? I think that you can help me with a problem :) | 13:25 |
| andreykurilin | also such method will be used in cli | 13:26 |
| alex_xu | andreykurilin: I think ti should be find_latest(client.Client()) | 13:26 |
| *** eharney has joined #openstack-nova | 13:26 | |
| *** mriedem_away is now known as mriedem | 13:27 | |
| *** Piet has joined #openstack-nova | 13:27 | |
| andreykurilin | ok | 13:28 |
| andreykurilin | sdague, alex_xu: there is a one more question | 13:28 |
| *** artom has joined #openstack-nova | 13:28 | |
| *** jdandrea has joined #openstack-nova | 13:29 | |
| alex_xu | andreykurilin: that find_latest may codes like in line 16-30 at https://etherpad.openstack.org/p/microversion_client | 13:29 |
| andreykurilin | alex_xu: it's a little bit harder:) `client.version.get()` is invalid method. | 13:30 |
| andreykurilin | client.versions have only list method | 13:31 |
| dansmith | rlrossit: ack | 13:31 |
| alex_xu | andreykurilin: yes, I'm think we add one, version.list() is request to '/', version.get is request '/2.1' or '/2', then you can the version info the endpoint directly | 13:31 |
| *** eglynn has joined #openstack-nova | 13:31 | |
| alex_xu | version.get request to the endpoint version API directly | 13:32 |
| andreykurilin | alex_xu: the simplest way is list all versions and check url with management_url of httpclient | 13:32 |
| *** yingjun has quit IRC | 13:32 | |
| *** marzif_ has quit IRC | 13:33 | |
| *** shhfeng has joined #openstack-nova | 13:33 | |
| *** yingjun has joined #openstack-nova | 13:33 | |
| *** marzif_ has joined #openstack-nova | 13:33 | |
| alex_xu | andreykurilin: you still need check url with management_url with list. if with get, then you needn't, you can get version info directly | 13:33 |
| andreykurilin | ok | 13:34 |
| rlrossit | dansmith: looking through ndipanov's version bumps, and noticed some of the objects don't have hash updates... looked at the fingerprint code and saw Lists have child_versions included in the relevant_data, but regular objects don't have obj_relationships in there (https://github.com/openstack/nova/blob/master/nova/tests/unit/objects/test_objects.py#L1242-L1246) don't we want to pay attention to obj_relationships so people don't change | 13:35 |
| alex_xu | so everybody are on the same page :) | 13:35 |
| andreykurilin | alex_xu: I'll change a spec today and sen a link to review | 13:35 |
| alex_xu | andreykurilin: we said in the meeting, just focus on code, needn't update spec | 13:35 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: virt: introduce libosinfo library to set hardware policy https://review.openstack.org/149625 | 13:35 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring network model https://review.openstack.org/149627 | 13:35 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adding a class to retrieve hardware properties https://review.openstack.org/149626 | 13:35 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring the disk bus https://review.openstack.org/149628 | 13:35 |
| andreykurilin | alex_xu: great! | 13:36 |
| *** thangp has joined #openstack-nova | 13:36 | |
| ndipanov | rlrossit, we do | 13:36 |
| ndipanov | it's a bug | 13:36 |
| alex_xu | sdague: johnthetubaguy, andreykurilin is it worth you can review this etherpad https://etherpad.openstack.org/p/microversion_client then we nail down how client lib looks like? | 13:36 |
| dansmith | ndipanov: rlrossit hang on | 13:36 |
| dansmith | rlrossit: we added child_versions to the hash just so we'd catch it because list objects were always getting skipped for bumps | 13:37 |
| dansmith | rlrossit: the problem with that is that it doesn't tell you what you need to change, just that something changed | 13:37 |
| andreykurilin | alex_xu: Should I fix first two patches in chain or I can add new patch with restrict latest for Client() ? | 13:37 |
| ndipanov | ok anyway | 13:37 |
| dansmith | rlrossit: if the relationships test is working, then that is much more detailed information than just "hash don't match yp" | 13:37 |
| dansmith | rlrossit: so now that that's working, I'd say make the fingerprint just cover the things we don't have explicit correctness checks for | 13:38 |
| rlrossit | dansmith: ok so when I get rid of child_versions that will compress that down a little to just do methods and fields basically? | 13:39 |
| dansmith | rlrossit: when you bump the child object, that'll trigger a hash change, but the version will also trigger the parent's relationship test to fail, and when you change that version, it will change the hash, and make some other object that depends on it fail the relationship test, and so on | 13:39 |
| dansmith | yeah | 13:39 |
| alex_xu | andreykurilin: we should fix first two patches I think, we can't introduce two wrong thing first, then fix them in later patch | 13:39 |
| ndipanov | dansmith, almost like a hash tree | 13:39 |
| ndipanov | where have I seen that before... hmmm | 13:39 |
| ndipanov | anyway: https://bugs.launchpad.net/nova/+bug/1474074 | 13:40 |
| openstack | Launchpad bug 1474074 in OpenStack Compute (nova) "PciDeviceList is not versioned properly in liberty and kilo" [High,In progress] - Assigned to Nikola Đipanov (ndipanov) | 13:40 |
| ndipanov | dansmith, ^ | 13:40 |
| ndipanov | in order for this to work - we need to make liberty (master) aware of the .z versions | 13:41 |
| ndipanov | present in kilo stable | 13:41 |
| shhfeng | johnthetubaguy, ping | 13:41 |
| ndipanov | so that it won't backport relations by mistake | 13:41 |
| johnthetubaguy | shhfeng: hi | 13:41 |
| *** bradjones has quit IRC | 13:41 | |
| shhfeng | hello | 13:41 |
| dansmith | ndipanov: I'm not sure I know what you mean | 13:41 |
| ndipanov | well it's like this | 13:42 |
| shhfeng | I'd like to talk about 139910 | 13:42 |
| shhfeng | https://review.openstack.org/#/c/139910/ | 13:42 |
| ndipanov | if kilo with instance version 1.19 asks for it from a lemming conductor | 13:42 |
| shhfeng | johnthetubaguy ^ | 13:42 |
| shhfeng | johnthetubaguy, I seen your comment on it. | 13:42 |
| ndipanov | conductor looks at it and says - in my book - 1.19 means pcilist 1.1 | 13:42 |
| ndipanov | because everything before 1.22 (Instance) means 1.1 (PciDeviceList) | 13:43 |
| ndipanov | and the same thing happens for 1.19.1 | 13:43 |
| ndipanov | but that's wrong | 13:43 |
| ndipanov | conductor in lemming needs to know that | 13:43 |
| ndipanov | so it's not enough to only insert .z versions in kilo | 13:44 |
| ndipanov | we need to make master know about those for all the stuff that ultimately uses the PciDeviceList | 13:44 |
| *** Piet has quit IRC | 13:44 | |
| ndipanov | which is sadly Instance :( | 13:44 |
| rlrossit | ndipanov: so the .z is a half update that lets Liberty know about new kilo things without backporting the entire version tree? | 13:44 |
| ndipanov | yes | 13:44 |
| ndipanov | it's basically a branch | 13:44 |
| ndipanov | of the tree | 13:45 |
| ndipanov | well a parallel tree if you will :) | 13:45 |
| *** mwagner_lap has joined #openstack-nova | 13:45 | |
| *** otter768 has joined #openstack-nova | 13:45 | |
| dansmith | which instance version is in kilo? | 13:45 |
| dansmith | okay, 1.19 | 13:46 |
| dansmith | so, I'm not sure why we can't just correct this in place | 13:47 |
| dansmith | bump pcidevicelist to 1.3 | 13:47 |
| dansmith | insert a relationship for instance 1.19 to be dependent on pci device list 1.3 | 13:47 |
| *** Jeffrey4l has quit IRC | 13:47 | |
| dansmith | it wont be correct for people that deployed mid-kilo, | 13:47 |
| dansmith | but those people aren't able to work properly anyway | 13:47 |
| *** bradjones has joined #openstack-nova | 13:48 | |
| *** bradjones has quit IRC | 13:48 | |
| *** bradjones has joined #openstack-nova | 13:48 | |
| *** Jeffrey4l has joined #openstack-nova | 13:48 | |
| *** jecarey has joined #openstack-nova | 13:48 | |
| rlrossit | ndipanov: dansmith: so, possibly stupid question, but what are we saving 2.x objects for, or are we always going to be on 1.x? | 13:48 |
| dansmith | rlrossit: that's a really sad question | 13:49 |
| dansmith | rlrossit: because I had a set to bump everything to 2.0 so we could drop a bunch of compatibility | 13:49 |
| dansmith | rlrossit: but SOMEone didn't review it | 13:49 |
| dansmith | rlrossit: either way, that wouldn't help us here | 13:49 |
| *** ctrath has joined #openstack-nova | 13:50 | |
| rlrossit | well I wasn't sure if we could back up the .z to be .y and then just bump all liberty everything to 2.x so when we get a 1.x we know it's from "before" | 13:50 |
| *** otter768 has quit IRC | 13:50 | |
| rlrossit | I keep mixing up my x y and z | 13:50 |
| rlrossit | but I hope that was understandable | 13:50 |
| dansmith | I don't think it helps | 13:50 |
| dansmith | I also don't think this is that complicated to fix | 13:51 |
| dansmith | we know that kilo people are on instance 1.19, and that they need to depend on a pcidevicelist of 1.2 | 13:52 |
| *** krtaylor has quit IRC | 13:52 | |
| ndipanov | dansmith, yar | 13:52 |
| dansmith | and that right now they're broken anyway | 13:52 |
| ndipanov | yes | 13:52 |
| dansmith | so kilo people are going to have to deploy code to both sides to get back to working | 13:52 |
| *** dboik has quit IRC | 13:52 | |
| dansmith | so we might as well just insert the version, the relationship and get then back on the right track | 13:52 |
| ndipanov | so just actually add a tuple to obj_relationships | 13:53 |
| rlrossit | dansmith: well with what I discovered, the hash test won't kill you on that :) | 13:53 |
| ndipanov | and not do any of the version bumps | 13:53 |
| ndipanov | other than PciDevice list | 13:53 |
| dansmith | ndipanov: you have to bump the pcidevicelist to 1.2 right? but the relationship is already there | 13:53 |
| ndipanov | that actually makes sense | 13:53 |
| dansmith | right | 13:53 |
| dansmith | exactly | 13:53 |
| ndipanov | cool | 13:53 |
| ndipanov | and then we do the same thing in kilo and we should be good | 13:54 |
| * dansmith gets coffee | 13:54 | |
| rlrossit | good thing I keep putting off rebasing that change | 13:54 |
| dansmith | yes | 13:54 |
| ndipanov | well you will still have to only the change will be much smaller | 13:54 |
| ndipanov | that will break pre - numa_node kilo folks but not really since it's a purely additive change | 13:55 |
| ndipanov | ok | 13:55 |
| ndipanov | hmm or would it | 13:56 |
| dansmith | they're broken anyway | 13:56 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: libvirt:Rsync remote FS driver was added https://review.openstack.org/146887 | 13:56 |
| dansmith | well, depends on where they are I guess | 13:56 |
| *** Longgeek has quit IRC | 13:56 | |
| dansmith | if they are before the numa node change, then this fixes it for them rolling over that actually | 13:57 |
| dansmith | because all you're doing is adding the actual accounting for what happened | 13:57 |
| dansmith | if they had a post-numa-node conductor and a pre compute, | 13:57 |
| dansmith | then they'd be sending back unbackported pcidevice objects | 13:57 |
| ndipanov | they were broken anyway | 13:57 |
| dansmith | now they will properly get backported with this change | 13:57 |
| dansmith | the kinds of changes where we get stuck are when a version number isn't allocated for a change that is made | 13:58 |
| dansmith | but this did allocate one, | 13:58 |
| dansmith | it just didn't actually set it :) | 13:58 |
| ndipanov | right | 13:58 |
| *** yamahata has joined #openstack-nova | 13:58 | |
| *** annegentle has quit IRC | 14:00 | |
| *** cfriesen has joined #openstack-nova | 14:00 | |
| *** markvoelker_ has joined #openstack-nova | 14:00 | |
| dansmith | rlrossit: in the meantime, if you could think about whether the tests we're going to end up with would have caught this case, that'd be awesome | 14:00 |
| *** irenab has quit IRC | 14:00 | |
| dansmith | rlrossit: because we might want a test to make sure that we don't have a version in relationships that is newer than ourselves | 14:00 |
| dansmith | rlrossit: to catch this kind of thing | 14:00 |
| *** Longgeek has joined #openstack-nova | 14:01 | |
| dansmith | rlrossit: you can see that the impact to missing these things can be large, so improvements in that area of testing are very welcome | 14:01 |
| *** VW_ has joined #openstack-nova | 14:02 | |
| rlrossit | dansmith: what do you mean by testing so we don't have a version newer than ourselves? | 14:03 |
| rlrossit | like if you version is 1.1, and we have obj_relationships with ('1.2', '1.3') or something? | 14:03 |
| dansmith | yes, which is what happened | 14:03 |
| dansmith | hang on | 14:03 |
| openstackgerrit | Alex Xu proposed openstack/nova: Cleanup quota_class unittest with appropriate request context https://review.openstack.org/150351 | 14:03 |
| dansmith | rlrossit: https://review.openstack.org/#/c/140289/4/nova/objects/pci_device.py | 14:03 |
| *** VW__ has joined #openstack-nova | 14:03 | |
| dansmith | rlrossit: see they forgot to bump the version, but did put in the relationship to satisfy the test | 14:03 |
| *** markvoelker has quit IRC | 14:04 | |
| *** krtaylor has joined #openstack-nova | 14:04 | |
| *** VW__ has quit IRC | 14:05 | |
| *** annashen has joined #openstack-nova | 14:05 | |
| rlrossit | dansmith: gotcha. I thought you were talking about a test for catching what the original bug was for, which is what all this list stuff is | 14:06 |
| ndipanov | dansmith, actually the code that gets screwed is code between having the numa_node in pci device and the next chronological Instance bump - you can't fix those | 14:06 |
| *** agireud has joined #openstack-nova | 14:06 | |
| ndipanov | since you can't insert an instance version | 14:06 |
| ndipanov | even on stable kilo | 14:06 |
| ndipanov | will post this patch soon | 14:06 |
| dansmith | um | 14:06 |
| mriedem | rlrossit: dansmith: seems like a relatively easy test, just lookup relationships that i'm in and see if a newer version is listed than what i have | 14:06 |
| dansmith | ndipanov: oh right, because of the other list bugs he's fixing the instance thing didn't get caught | 14:07 |
| dansmith | I see | 14:07 |
| *** VW_ has quit IRC | 14:07 | |
| dansmith | mriedem: yep | 14:07 |
| rlrossit | mriedem: yep, but I'm kind of curious why the hash test didn't check this, because it should've seen child_versions was updated, so the hash should be different. Unless the hash was updated, but not the version part of the fingerprint... | 14:07 |
| mriedem | rlrossit: that was probably it | 14:08 |
| mriedem | they updated the hash but not the version, | 14:08 |
| mriedem | and the test assumes you know what you're doing with the version | 14:08 |
| *** diogogmt has quit IRC | 14:08 | |
| mriedem | and the test output says 'make sure you bump the version' but some people don't read | 14:08 |
| rlrossit | so this will just make things more strict then. which is good | 14:08 |
| dansmith | yep | 14:08 |
| alaski | johnthetubaguy: the TaskLog object doesn't appear to be related to "tasks" as we've been talking about them. it's related to instance usage audits, so just confusingly named | 14:09 |
| johnthetubaguy | alaski: oh, I should go read that more closely | 14:09 |
| *** annashen has quit IRC | 14:09 | |
| *** tpeoples has joined #openstack-nova | 14:09 | |
| *** Kennan2 has joined #openstack-nova | 14:10 | |
| *** Kennan has quit IRC | 14:10 | |
| *** bkopilov has joined #openstack-nova | 14:12 | |
| *** dboik has joined #openstack-nova | 14:12 | |
| *** sudipto has joined #openstack-nova | 14:15 | |
| ankit_ag | mriedem: hi | 14:16 |
| mriedem | ankit_ag: hi | 14:17 |
| ankit_ag | mriedem: Could you please take a look on a pending review https://review.openstack.org/#/c/194063/ | 14:17 |
| ankit_ag | whenever you get some time | 14:18 |
| *** browne has joined #openstack-nova | 14:18 | |
| mriedem | looks scary | 14:18 |
| mriedem | i remember the revert | 14:18 |
| ankit_ag | mriedem: yeah, I have fixed that race condition issue | 14:18 |
| ankit_ag | mriedem: in the same patch | 14:18 |
| ankit_ag | mriedem: It was because I had removed an exception from the master code, which I have added back | 14:19 |
| ankit_ag | in the current patch | 14:19 |
| *** shhfeng has quit IRC | 14:23 | |
| *** exploreshaifali has joined #openstack-nova | 14:23 | |
| *** kbyrne has quit IRC | 14:23 | |
| *** tonytan4ever has joined #openstack-nova | 14:23 | |
| *** RuiChen has joined #openstack-nova | 14:24 | |
| *** MentalRay has joined #openstack-nova | 14:25 | |
| *** samueldmq has quit IRC | 14:26 | |
| *** ankit_ag has quit IRC | 14:26 | |
| *** gszasz has quit IRC | 14:28 | |
| *** agentle_ has joined #openstack-nova | 14:28 | |
| *** samueldmq has joined #openstack-nova | 14:28 | |
| jaypipes | dansmith: heya. want to chat about sudipto's memory bandwidth monitor metric object field changes? | 14:29 |
| dansmith | jaypipes: remind me with a link? | 14:29 |
| sudipto | dansmith, https://review.openstack.org/#/c/198972/ | 14:29 |
| *** andrewbogott_afk is now known as andrewbogott | 14:29 | |
| dansmith | jaypipes: hmm | 14:30 |
| jaypipes | dansmith: I wanted to explain what my original thoughts were on how to seamlessly make the value field polymorphic, to accomodate the existing integer values as well as the new NUMA-node-specific memory bandwitdh values. | 14:30 |
| gibi | Hi nova cores! I have this small patch regarding making on_shared_storage optional with a +2 already and I'm eager to get another one ;) https://review.openstack.org/#/c/197951/ | 14:31 |
| *** gszasz has joined #openstack-nova | 14:31 | |
| dansmith | jaypipes: okay... | 14:31 |
| *** _kiran_ has joined #openstack-nova | 14:31 | |
| jaypipes | dansmith: so, my idea was to use a @property on the MonitorMetric value field and use the from_dict() method to just populate either the IntegerField or the NUMANodeMemoryBandwidths ObjectField, depending on the metric_type. | 14:33 |
| mriedem | gibi: since that's related to evacuate code you probably want dansmith on your review since he's playing in the guts of evacuate | 14:33 |
| thomasem | Anyone have a few minutes to look at https://review.openstack.org/#/c/187012/ today? | 14:34 |
| dansmith | jaypipes: value is already a property | 14:34 |
| thomasem | Been open for a while, gate was blocking it previously, finally got things passing again. | 14:34 |
| dansmith | jaypipes: I'm missing where there is an objectfield in here | 14:35 |
| jaypipes | dansmith: right, but it's easy enough to change the name of that field attribute to something like "integer_value". | 14:35 |
| *** kiran-r has quit IRC | 14:35 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 1 https://review.openstack.org/152569 | 14:35 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 2 https://review.openstack.org/167408 | 14:35 |
| jaypipes | dansmith: sorry, it's not an object field... it's a DictOfIntegersField. | 14:35 |
| gibi | mriedem: Thanks. He was involved in the spec review and already agreed that his is independent from mine. | 14:35 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Adds support for x509 certificates as keypairs https://review.openstack.org/136458 | 14:35 |
| dansmith | jaypipes: not really :) Aside from breaking our semantic rules, if you change that and then send an older version to the newer version it'll get dropped. But, that's a cosmetic thing I guess | 14:36 |
| rlrossit | dansmith: we need to get my tests in... they missed updating MonitorMetricList :P | 14:36 |
| dansmith | sudipto: ^ | 14:36 |
| dansmith | rlrossit: aye :) | 14:36 |
| jaypipes | dansmith: right. it can all be handled in from_dict() and obj_make_compatible() and to_dict(). | 14:36 |
| dansmith | jaypipes: so you want the .value (whatever we call it) to return either an integer or a dict depending on the type? | 14:36 |
| pkholkin | dhellmann: Are you here? Could you remove -2 from the patch https://review.openstack.org/#/c/173834/ | 14:36 |
| jaypipes | dansmith: this never actually goes over the RPC wire... only currently to the notification API. | 14:36 |
| jaypipes | dansmith: yes, that's correct. | 14:37 |
| *** itzikb_ has quit IRC | 14:37 | |
| *** afazekas has quit IRC | 14:37 | |
| jaypipes | dansmith: FTR, this is precisely how I'm building the resource objects serialization, which also needs to be polymorphic in the same way. | 14:37 |
| dansmith | jaypipes: you mean it doesn't go over the wire in object format yet because we convert to dict first? These do get sent back to the scheduler or db or something so that we can schedule on them, right? | 14:38 |
| *** unicell1 has joined #openstack-nova | 14:38 | |
| openstackgerrit | lyanchih proposed openstack/nova: gate-nova-python34 sometimes test failed on test_save_updates_numa_topology https://review.openstack.org/201579 | 14:38 |
| *** zsais has joined #openstack-nova | 14:38 | |
| jaypipes | dansmith: unfortunately, that has yet to be "corrected". right now, the resource tracker just does jsonutils.dumps directly on the metrics dict. | 14:38 |
| jaypipes | :( | 14:38 |
| jaypipes | dansmith: I was working to correct that. | 14:38 |
| dansmith | okay | 14:39 |
| mriedem | gibi: i approved, but, | 14:39 |
| jaypipes | dansmith: it's the next in my patch series that unf**ks the monitor system. | 14:39 |
| mriedem | gibi: new tests should be using mock now | 14:39 |
| *** unicell has quit IRC | 14:39 | |
| mriedem | rather than mox | 14:39 |
| *** MentalRay has quit IRC | 14:39 | |
| dansmith | either way, I guess I'm not sure why you'd ever want to return either an integer or a dict | 14:39 |
| *** Murali has quit IRC | 14:39 | |
| dansmith | why not just store both and access the one you think you're going to want? | 14:39 |
| dansmith | i.e. it seems confusing to me to have int(foo.value) in one place and foo.value[3] in another place | 14:40 |
| *** shuquan has quit IRC | 14:40 | |
| gibi | mriedem: Thanks. I wanted to follow the patter in the test file this is why the test was written with mox. | 14:40 |
| *** jcoufal has quit IRC | 14:41 | |
| rlrossit | dansmith: jaypipes: may be missing something, but are you basically trying to mux the value and numa_nodes_values in MonitorMetric? | 14:41 |
| mriedem | gibi: yeah, we also don't want more new tests in nova.tests.unit.compute.test_compute since that uses a db backend, | 14:41 |
| mriedem | gibi: new compute manager tests should use mock and live in nova.tests.unit.compute.test_compute_mgr | 14:41 |
| mriedem | gibi: but i didn't want to block your change on that | 14:41 |
| dansmith | jaypipes: I think I'm derailing this conversation.. what did you specifically want to chat about? the polymorphism? | 14:41 |
| gibi | mriedem: thanks, I was nat aware of that. | 14:41 |
| *** jichen has quit IRC | 14:41 | |
| dansmith | rlrossit: yeah, sounds like it | 14:41 |
| sudipto | rlrossit, yeah sorta. | 14:41 |
| *** iamjarvo has joined #openstack-nova | 14:42 | |
| *** jichen has joined #openstack-nova | 14:42 | |
| *** shuquan has joined #openstack-nova | 14:42 | |
| jaypipes | dansmith: sec | 14:42 |
| jaypipes | dansmith: think of it like a union struct in C. | 14:43 |
| *** kbyrne has joined #openstack-nova | 14:43 | |
| dansmith | jaypipes: right, but in a union you reference the element in the union that has the data type you expect... :D | 14:43 |
| dansmith | you don't say int c = (int)&myunion | 14:44 |
| dansmith | you say int c = myunion.intval | 14:44 |
| jaypipes | dansmith: sure, but what I'm saying is that the union doesn't store multiple things at the same time inside itself. | 14:44 |
| rlrossit | dansmith: so you're saying we need to make a UnionField? | 14:44 |
| jaypipes | rlrossit: no... hold on, mate. | 14:45 |
| dansmith | jaypipes: right, but the object already doesn't store things that aren't set, so there's no duplication or re-use or anything | 14:45 |
| dansmith | jaypipes: sorry if I'm being dense | 14:46 |
| dansmith | jaypipes: we can do this high bw if you want to make hand gestures and such | 14:46 |
| jaypipes | dansmith: I'm not following you... In the original patchset: https://review.openstack.org/#/c/198972/1/nova/objects/monitor_metric.py, the value @property on MonitorMetricValue is doing the work of "selecting" the union'd field by simply returning the first non-NULL value. | 14:47 |
| jaypipes | dansmith: yeah, let's hop on a hangout. I hve 13 minutes. | 14:47 |
| dansmith | is that enough? want to do it after your next thing? I'm free the rest of the day | 14:47 |
| jaypipes | dansmith: https://plus.google.com/hangouts/_/gwhyffus4mhpvx7y32dpb23b4ua?hl=en | 14:47 |
| jaypipes | dansmith: I'm not, umnfortunately. | 14:48 |
| dansmith | okay | 14:48 |
| *** MentalRay has joined #openstack-nova | 14:49 | |
| *** markvoelker_ has quit IRC | 14:49 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added server tags support in nova-api https://review.openstack.org/128940 | 14:49 |
| *** markvoelker has joined #openstack-nova | 14:50 | |
| sudipto | rlrossit, can you help with the changes you mentioned that missed the metriclist? | 14:51 |
| openstackgerrit | Belmiro Moreira proposed openstack/nova: Add flavor* tables to API database https://review.openstack.org/201606 | 14:51 |
| *** tsekiyama has joined #openstack-nova | 14:52 | |
| rlrossit | sudipto: sure. what do you need help with? | 14:52 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/200708 | 14:53 |
| *** pmcfadden has joined #openstack-nova | 14:53 | |
| sudipto | rlrossit, oh, i was referring to your comment: "we need to get my tests in... they missed updating MonitorMetricList" | 14:54 |
| *** shuquan has quit IRC | 14:54 | |
| *** markvoelker has quit IRC | 14:55 | |
| rlrossit | sudipto: I left a comment in the review too, but basically you need to update the child_versions in MonitorMetricList to have '1.1': '1.1' after '1.0': '1.0' and then also bump the version of the list to 1.1 (which will then break unit tests because you'll need to update hashes and the object tree) | 14:56 |
| *** armax has joined #openstack-nova | 14:56 | |
| *** rwsu has joined #openstack-nova | 14:56 | |
| *** zzzeek has joined #openstack-nova | 14:57 | |
| *** rwsu has quit IRC | 14:58 | |
| dansmith | sudipto: rlrossit you guys could have joined, FWIW, but I think we're good so I'll leave comments on the review | 14:58 |
| *** rwsu has joined #openstack-nova | 14:58 | |
| *** achanda has joined #openstack-nova | 14:58 | |
| sudipto | rlrossit, ah i get it now. So there are two things, let jaypipes and dansmith decide on the best approach for the field (and if that would eventually lead to a version bump), secondly, i guess you are saying we need to bump the version of the metriclist because even though it's accepting the same 'type' of object, the hashes would be different? | 14:59 |
| sdague | andreykurilin: https://review.openstack.org/#/c/200908/ does not look like a race, it's got to be actually resolved | 14:59 |
| sdague | please look at the test logs before just running recheck | 14:59 |
| rlrossit | dansmith: didn't want to crash a party I wasn't invited to :) but good to see you guys reached an agreement | 14:59 |
| sudipto | since effectively the version of metricvalue would have changed with the new field. | 14:59 |
| openstackgerrit | Belmiro Moreira proposed openstack/nova: Add flavor* tables to API database https://review.openstack.org/201606 | 14:59 |
| sudipto | dansmith, sure... :) | 15:00 |
| rlrossit | sudipto: correct | 15:00 |
| *** ddaskal has joined #openstack-nova | 15:00 | |
| dansmith | rlrossit: pasting the url in public channel means it's not a private party :) | 15:00 |
| sudipto | rlrossit, alright, good catch :) | 15:00 |
| andreykurilin | sdague: It doesn't repeat each time, so I prefer to fix it in separate patch, since novaclient gates are fully blocked now | 15:00 |
| sdague | andreykurilin: did you look at the logs? | 15:01 |
| andreykurilin | sdague: yes | 15:01 |
| sdague | it's very clearly a different set of parameters coming back | 15:01 |
| *** pmcfadden has quit IRC | 15:01 | |
| *** shuquan has joined #openstack-nova | 15:01 | |
| andreykurilin | sdague: and it can be reproduced locally. 3 of 4 tests launches are failed. | 15:02 |
| sdague | andreykurilin: if the failure rate is that high, it's pointless to try to recheck grind through that race | 15:02 |
| sdague | the root cause needs to be addressed | 15:02 |
| *** busterswt has quit IRC | 15:02 | |
| dansmith | sudipto: rlrossit: left comments, see if you agree | 15:02 |
| *** burt has joined #openstack-nova | 15:03 | |
| *** tpeoples has quit IRC | 15:04 | |
| *** tpeoples has joined #openstack-nova | 15:04 | |
| *** rdopiera has quit IRC | 15:04 | |
| *** coolsvap|away is now known as coolsvap | 15:04 | |
| *** xyang2 has joined #openstack-nova | 15:05 | |
| *** mdrabe has quit IRC | 15:05 | |
| *** _kiran_ is now known as kiran-r | 15:05 | |
| andreykurilin | sdague: I suppose the root issue in this mock - https://github.com/openstack/python-novaclient/blob/master/novaclient/tests/unit/v2/test_shell.py#L2370-L2371 . Since the global object is mocked and this test is launched several times via different testcases, it can produce collision. | 15:05 |
| *** annashen has joined #openstack-nova | 15:06 | |
| *** achanda has quit IRC | 15:06 | |
| *** shuquan has quit IRC | 15:06 | |
| andreykurilin | sdague: can we replace mock_open with real file? | 15:06 |
| sudipto | dansmith, agree with all the comments. So you are suggesting that we report only a single 'value' field based on the attribute that's set (either numa one or the normal int one) right? | 15:06 |
| dansmith | sudipto: since it's the legacy dict, that's up to you I guess | 15:07 |
| dansmith | sudipto: probably better to set it to None in the dict if it's not set on the object I guess | 15:07 |
| dansmith | if you're going to report them as separate names in the dict | 15:07 |
| *** pmcfadden has joined #openstack-nova | 15:07 | |
| sudipto | dansmith, ok, i tried to take a middle path between your and jaypipes 's approach - and preserve the legacy dict by just having the 'value' field. | 15:08 |
| sdague | andreykurilin: we could | 15:09 |
| sdague | andreykurilin: I'm trying to figure out where else it's called though | 15:09 |
| sdague | like why would it be a race | 15:09 |
| *** alex_xu has quit IRC | 15:09 | |
| sdague | or if this is another bug in mock 1.1 | 15:09 |
| dansmith | sudipto: okay, I'm not sure about the consumer of that blob so I'm not sure if that's the best approach or not | 15:10 |
| *** alex_xu has joined #openstack-nova | 15:10 | |
| *** ig0r__ has joined #openstack-nova | 15:10 | |
| *** diogogmt has joined #openstack-nova | 15:10 | |
| *** annashen has quit IRC | 15:10 | |
| *** tdurakov has joined #openstack-nova | 15:10 | |
| dansmith | sudipto: however, I'd leave it the way you have it (but with my comments addressed) and then let jaypipes look at it.. easy to change the dict again if need be | 15:10 |
| tdurakov | Hi, got problems with py27 on master: http://paste.openstack.org/show/374595/ | 15:10 |
| sudipto | dansmith, sounds good! thx... | 15:10 |
| *** ig0r_ has quit IRC | 15:10 | |
| tdurakov | anyone else has same problems? | 15:11 |
| sdague | andreykurilin: oh, you know what, I bet it's because this is decorated | 15:12 |
| *** Piet has joined #openstack-nova | 15:12 | |
| andreykurilin | sdague: "where else it's called though". there two classes are inheritors of ShellTest: ShellTestV11 and ShellWithSessionClientTest. When all these classes are launched in parallel, it produces some collision. But I suppose this is a new bug of mock, because previous releases doesn't have any problems | 15:14 |
| sdague | yeh, I just noticed that | 15:14 |
| sdague | andreykurilin: can you file an upstream mock bug, I'm going to see if we can just turn it into a context manager instead | 15:15 |
| andreykurilin | sdague: ok, I'll do | 15:16 |
| *** ig0r_ has joined #openstack-nova | 15:17 | |
| *** ferest has joined #openstack-nova | 15:17 | |
| *** ig0r__ has quit IRC | 15:19 | |
| openstackgerrit | lyanchih proposed openstack/nova: NUMATopologyFilter raise exception and not continue filter next node https://review.openstack.org/200630 | 15:19 |
| *** mdrabe has joined #openstack-nova | 15:20 | |
| openstackgerrit | Mark Goddard proposed openstack/nova: Map physical ports to networks. https://review.openstack.org/153230 | 15:20 |
| sdague | andreykurilin: I'm testing my theory, if it's true, I'll have a patch in a minute | 15:20 |
| sdague | yeh, so you can get a 100% reproduce with: tox -e py27 -- --concurrency=1 | 15:20 |
| *** gszasz has quit IRC | 15:21 | |
| sdague | the reason it sometimes works is there are 3 tests | 15:21 |
| sdague | and 8 workers | 15:21 |
| sdague | so it only fails if 2 tests end up in the same worker | 15:21 |
| *** ferest has quit IRC | 15:21 | |
| *** IanGovett has quit IRC | 15:21 | |
| *** dannywilson has joined #openstack-nova | 15:22 | |
| *** IanGovett has joined #openstack-nova | 15:22 | |
| *** READ10 has quit IRC | 15:23 | |
| openstackgerrit | Sean Dague proposed openstack/python-novaclient: patch open in a context manager instead of decorator https://review.openstack.org/201629 | 15:24 |
| sdague | andreykurilin: ^^^ | 15:24 |
| *** dannywil_ has joined #openstack-nova | 15:25 | |
| sdague | so, I guess because we're blocking on py26 on the other patch, we recheck grind it, then get this one in to fix the mock 1.1 issue | 15:25 |
| sdague | lifeless fyi, another mock 1.1 issue exposed in https://review.openstack.org/201629 | 15:25 |
| *** dannywilson has quit IRC | 15:25 | |
| tdurakov | jaypipes: Hi Jay! | 15:25 |
| andreykurilin | sdague: maybe it would be better to squash two patches in one? | 15:26 |
| *** belmoreira has quit IRC | 15:26 | |
| sdague | andreykurilin: maybe, lets see if you get lucking on the rechecks | 15:26 |
| sdague | if it fails, we can merge them | 15:26 |
| andreykurilin | :) | 15:26 |
| sdague | I at least feel better in understanding it, and having a reproduce | 15:26 |
| *** markvoelker has joined #openstack-nova | 15:26 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: Fix the incorect PciDeviceList version number https://review.openstack.org/201280 | 15:27 |
| ndipanov | dansmith, ^^ | 15:27 |
| *** krtaylor has quit IRC | 15:27 | |
| jaypipes | tdurakov: hello! | 15:27 |
| *** ildikov_ has joined #openstack-nova | 15:28 | |
| *** Mike_D_laptop has joined #openstack-nova | 15:29 | |
| tdurakov | @jaypipes, it seems that nova master is broken now | 15:29 |
| tdurakov | could you verify that | 15:29 |
| tdurakov | http://paste.openstack.org/show/374595/ | 15:30 |
| dansmith | ndipanov: replied | 15:30 |
| tdurakov | jaypipes:^ | 15:30 |
| dansmith | ndipanov: assuming the obvious answer to my question, code looks good I think | 15:30 |
| *** ildikov has quit IRC | 15:30 | |
| openstackgerrit | Bence Romsics proposed openstack/nova: Do not block instance boot on net without subnet https://review.openstack.org/200046 | 15:31 |
| RuiChen | dims_, can you take a look this patch https://review.openstack.org/#/c/193117/ ? I add some reply. | 15:31 |
| jaypipes | tdurakov: :( not sure... | 15:31 |
| dims_ | RuiChen: ack | 15:31 |
| dansmith | tdurakov: yes, same | 15:32 |
| dansmith | tdurakov: try rebuilding your tox | 15:32 |
| openstackgerrit | John Garbutt proposed openstack/nova: docs: split up upgrade and upgrade futures https://review.openstack.org/201529 | 15:32 |
| tdurakov | dansmith: ok, I'll check that | 15:32 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: Fix the incorrect PciDeviceList version number https://review.openstack.org/201280 | 15:33 |
| johnthetubaguy | dansmith: ndipanov: spoke to you both about more docs around upgrade, I have started moving the existing ones forward a bit here: https://review.openstack.org/201529 | 15:34 |
| dansmith | johnthetubaguy: okay, I was also going to start a "what to look for in reviews for upgrades" | 15:35 |
| dansmith | maybe that's covered in your changes? | 15:35 |
| * dansmith pushes that on the growing stack | 15:35 | |
| *** wuhao has joined #openstack-nova | 15:35 | |
| johnthetubaguy | dansmith: it adds a section that should cover that | 15:35 |
| dansmith | okay | 15:35 |
| *** READ10 has joined #openstack-nova | 15:36 | |
| johnthetubaguy | dansmith: I am trying to focus on what we want to happen, and why, and the more general stuff first | 15:36 |
| *** exploreshaifali has quit IRC | 15:36 | |
| johnthetubaguy | dansmith: but yeah, that review check list would be super handy | 15:36 |
| RuiChen | dims_, thank you :-) | 15:37 |
| dansmith | the review checklist is important because it's hard to keep track of all the banned things for folks that don't live and breathe it | 15:37 |
| dansmith | johnthetubaguy: anyway, yeah, I'll look at that in a bit | 15:37 |
| *** oro_ has joined #openstack-nova | 15:38 | |
| *** ig0r__ has joined #openstack-nova | 15:40 | |
| *** krtaylor has joined #openstack-nova | 15:40 | |
| *** jogo has quit IRC | 15:40 | |
| *** jogo has joined #openstack-nova | 15:41 | |
| tdurakov | dansmith, thanks tox rebuild helped | 15:42 |
| ndipanov | sdague, johnthetubaguy https://review.openstack.org/#/c/189632/ | 15:43 |
| ndipanov | I mean whatever but at least let's make a plan to talk about this | 15:43 |
| dansmith | tdurakov: cool | 15:43 |
| ndipanov | device names have been broken since forever | 15:43 |
| *** ig0r_ has quit IRC | 15:43 | |
| ndipanov | and this is one way to fix it | 15:43 |
| ndipanov | if no one else picks it up (and I don't think they will) | 15:44 |
| ndipanov | I don't see why we should keep it broken indefinitely | 15:44 |
| johnthetubaguy | ndipanov: sdague: feels like it needs an ML thread, with someone summarising both sides of the story, to unstick the review? | 15:45 |
| mriedem | ndipanov: maybe bring this to the mailing list | 15:45 |
| mriedem | ha | 15:45 |
| mriedem | jinx | 15:45 |
| *** tjones1 has joined #openstack-nova | 15:45 | |
| andreykurilin | sdague: sorry for delay | 15:46 |
| andreykurilin | sdague, lifeless: https://github.com/testing-cabal/mock/issues/280 | 15:46 |
| ndipanov | yeah the question is with whom does the burden of ML post sit | 15:46 |
| johnthetubaguy | ndipanov: I am tempted to say we should error out builds, so the lack of support is discoverable, but that is was more disruptive | 15:46 |
| lyanchih | guys please help me to review this https://review.openstack.org/#/c/197514/ | 15:46 |
| ndipanov | johnthetubaguy, yes | 15:46 |
| ndipanov | why I decided against it | 15:46 |
| *** otter768 has joined #openstack-nova | 15:46 | |
| mriedem | ndipanov: i think the ML is just a place to make this known and see if there are any strong reactions against this | 15:46 |
| ndipanov | I'll post something | 15:46 |
| mriedem | there isn't a -2 from sdague, just a -1 that we need to make sure everyone is aware and ok with this since it is throwing away an api parameter, which is why i said it was maybe an APIImpact | 15:47 |
| johnthetubaguy | ndipanov: that would be great, thank you | 15:47 |
| *** oro_ has quit IRC | 15:47 | |
| johnthetubaguy | mriedem: thats true, it is maybe is worth APIImpact | 15:48 |
| *** oro has quit IRC | 15:48 | |
| johnthetubaguy | but an ML thread should help with all that | 15:48 |
| andreykurilin | sdague: I'm not lucky and patch with requirements updates is failed - https://jenkins01.openstack.org/job/gate-python-novaclient-python27/37/ :( | 15:48 |
| sdague | andreykurilin: ok, you want to merge the patches together? | 15:48 |
| *** dprince has quit IRC | 15:48 | |
| *** yingjun has quit IRC | 15:48 | |
| andreykurilin | sdague: yes | 15:49 |
| sdague | andreykurilin: thanks | 15:50 |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: Allow NUMA based reporting for Monitors https://review.openstack.org/198972 | 15:50 |
| sudipto | rlrossit, dansmith ^ whenever you can... | 15:51 |
| *** changbl has joined #openstack-nova | 15:51 | |
| *** otter768 has quit IRC | 15:51 | |
| *** marzif_ has quit IRC | 15:52 | |
| *** marzif_ has joined #openstack-nova | 15:52 | |
| sdague | so I think what we need to understand is "is this useful feature anywhere?" if not, I'd say we make a version change, get rid of it. And figure out what the fallback behavior would be. I think in the libvirt case, it doesn't work anyway, so it's just a warning like we have. | 15:53 |
| *** tpeoples has quit IRC | 15:53 | |
| dansmith | last I checked, only xen could honor that flag, | 15:53 |
| johnthetubaguy | sdague: to the XenAPI driver works with this, I think, at least in the most common cases, and folks seemed to want to use it, now calling it "useful" is harder... | 15:54 |
| dansmith | and maybe not even recent xenserver | 15:54 |
| *** tpeoples has joined #openstack-nova | 15:54 | |
| *** wuhao has quit IRC | 15:54 | |
| johnthetubaguy | dansmith: so it seems to work, although I haven't tried really hard to break it recently | 15:54 |
| sdague | anyway, it seems like we should try to get that detail out on the table. | 15:54 |
| *** wuhao has joined #openstack-nova | 15:54 | |
| dansmith | johnthetubaguy: I thought you said recently that there was something about it that wasn't right, like you could only control ordering but not actually the device name it got in the guest | 15:54 |
| dansmith | it used to be with blkfront that you could just say "this should be sdzzfoo" and it would do it | 15:55 |
| dansmith | but I thought nowadays it just orders them by index | 15:55 |
| johnthetubaguy | dansmith: we just control the index not the name, yeah | 15:55 |
| ndipanov | dansmith, missed your comment about explaining the version change - will do | 15:55 |
| dansmith | johnthetubaguy: right, then this is broken | 15:55 |
| *** jichen has quit IRC | 15:55 | |
| johnthetubaguy | dansmith: the API does a nasty covert | 15:55 |
| johnthetubaguy | dansmith: yeah, thats true | 15:55 |
| dansmith | right, but it can't actually honor what is asked for | 15:55 |
| dansmith | so the API is broken | 15:55 |
| dansmith | and not useful anywhere, AFAICT | 15:55 |
| dansmith | ordering is different and we can do that properly | 15:55 |
| ndipanov | nova kinda already does that | 15:56 |
| *** exploreshaifali has joined #openstack-nova | 15:56 | |
| dansmith | right | 15:56 |
| ndipanov | and there's hooks to make xen driver replicate what the hypervisor would do | 15:56 |
| johnthetubaguy | dansmith: well if you request the correct sort of names, it does work, but yes, in general is a pile of seaming something nasty | 15:56 |
| ndipanov | i.e. the methods I added with this and the next patch | 15:56 |
| ndipanov | not actually with this... it was there since havana | 15:56 |
| dansmith | johnthetubaguy: but I can't request sdz and have it give me sdz if the next one in line is sdb right? | 15:57 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Updated from global requirements https://review.openstack.org/200908 | 15:57 |
| *** hemnafk is now known as hemna | 15:57 | |
| andreykurilin | sdague: ^ | 15:58 |
| johnthetubaguy | dansmith: oh, good question, I think it actually tries to give you /dev/xvdz or something like that, but I haven't tried it enough to be sure that still works everywere (PV vs PVops vs HVM+PV... sigh) | 15:58 |
| dansmith | johnthetubaguy: okay, I didn't think even that worked everywhere anymore | 15:58 |
| dansmith | johnthetubaguy: certainly not fully HVM | 15:59 |
| johnthetubaguy | dansmith: I have a feeling it works in linux and breaks in windows, but yeah | 15:59 |
| dansmith | johnthetubaguy: either way, I'd rather we have an index (and order if we prefer) and drop the "you can specify a device name" thing if we can, since it's soooo broken | 15:59 |
| sdague | so, can we get another person to sign off on this - https://review.openstack.org/#/c/200908/2 as I had a hand in the fix | 15:59 |
| sdague | but all novaclient patches are blocked until it lands | 15:59 |
| *** marzif_ has quit IRC | 15:59 | |
| johnthetubaguy | dansmith: ah... so it is possibly because xenserver still uses the xen QEMU rather than upstream | 16:00 |
| *** tpeoples has quit IRC | 16:00 | |
| *** tpeoples has joined #openstack-nova | 16:01 | |
| *** amotoki_ has quit IRC | 16:01 | |
| johnthetubaguy | dansmith: so yeah, I need to get someone to dig into all this, its confusing me now | 16:03 |
| dansmith | well, regardless, | 16:04 |
| dansmith | unless I can specify random device names, I think this is problematic | 16:04 |
| dansmith | it's extremely problematic for libvirt users | 16:04 |
| dansmith | because it breaks us and we can never actually honor it | 16:04 |
| *** tjones1 has quit IRC | 16:04 | |
| dansmith | so since we can't ever, we should at least fix it for libvirt if we're going to wait on the larger question of removing it from the api | 16:05 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: objects: add missing enum values to DiskBus field https://review.openstack.org/200280 | 16:05 |
| johnthetubaguy | dansmith: agreed with that | 16:05 |
| dansmith | sdague: are you opposed to making libvirt fully ignore it separately from the API decision? | 16:05 |
| dansmith | the latter will take some time either for consensus and/or deprecation/removal | 16:05 |
| sdague | dansmith: no, I'm fine with libvirt ignoring it, as it doesn't work anyway | 16:06 |
| *** jlanoux has quit IRC | 16:06 | |
| sdague | I'd just like to make it clearer to the user that's what we are doing | 16:06 |
| sdague | ideally with removing that feature from the API entirely if we can | 16:06 |
| *** annashen has joined #openstack-nova | 16:06 | |
| dansmith | sdague: yep, agreed with all that | 16:07 |
| sdague | my -1 was mostly a "can we hold and figure out the API story first before merge" | 16:07 |
| sdague | which I think I said in the message | 16:08 |
| *** dprince has joined #openstack-nova | 16:08 | |
| mriedem | sdague: oh i remember it quite differently.... http://fakeplus.com/pictures/jpg/hateful-cat-is-hateful_20120425122425.jpg | 16:08 |
| *** moshele has quit IRC | 16:09 | |
| sdague | mriedem: lol | 16:09 |
| dansmith | sdague: okay, well, I'm not sure we need to hold this up for that.. do you just want something that marks the decision point? like another review to deprecate it from the API that we can hold until we decide or something? | 16:09 |
| sdague | well can we take a day and figure out our direction here, get the ML thread going? | 16:10 |
| sdague | is there a reason this is a must merge today? it's been broken for a while. | 16:10 |
| ndipanov | I'm cool with taking a day it's been up for months | 16:10 |
| *** annashen has quit IRC | 16:11 | |
| ndipanov | well maybe 2 months :) | 16:11 |
| mriedem | heh, ~35 days :) | 16:11 |
| mriedem | yeah | 16:11 |
| mriedem | ha | 16:11 |
| mriedem | FOREVER! | 16:11 |
| dansmith | sure, I just don't want to hold the patch on $potentially_neverending_convo | 16:11 |
| dansmith | holding it on starting the convo is fine | 16:11 |
| ndipanov | exactly what irked me about this | 16:11 |
| *** signed8bit has joined #openstack-nova | 16:11 | |
| *** RuiChen has quit IRC | 16:11 | |
| *** ijw has joined #openstack-nova | 16:12 | |
| sdague | ndipanov: well, don't be irked, my intent was once we figured out the plan to +2 it myself, I just didn' | 16:12 |
| sdague | t want it to accidentally merge while we were still talking about stuff | 16:12 |
| *** unicell1 has quit IRC | 16:13 | |
| ndipanov | sure - will follow up with a ML post | 16:13 |
| *** zhenguo has quit IRC | 16:13 | |
| sdague | ndipanov: great | 16:13 |
| sdague | thank you | 16:14 |
| *** ishita has joined #openstack-nova | 16:14 | |
| *** boris-42 has joined #openstack-nova | 16:14 | |
| ndipanov | just hope we'll decide something within the thread... | 16:15 |
| *** kjelly has joined #openstack-nova | 16:15 | |
| ndipanov | tor soonish | 16:15 |
| ndipanov | or* | 16:15 |
| sdague | ndipanov: yep, agreed. I feel like we should get this merged by end of week one way or another. | 16:16 |
| *** alex_klimov has quit IRC | 16:16 | |
| *** Sukhdev has joined #openstack-nova | 16:18 | |
| *** kiran-r has quit IRC | 16:21 | |
| *** gmurphy has left #openstack-nova | 16:23 | |
| *** iamjarvo has quit IRC | 16:24 | |
| sdague | mriedem: https://review.openstack.org/#/c/197769/4 - parallel tests run in different address spaces, so there should be no issues there | 16:27 |
| sdague | was that your only objection to that patch? | 16:27 |
| mriedem | sdague: i just think it's unnecessary to have the hard-coded uuid in there | 16:28 |
| mriedem | sdague: also these aren't nodbtestcase tests so that worried me | 16:29 |
| *** nic has joined #openstack-nova | 16:30 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: Fix the incorrect PciDeviceList version number https://review.openstack.org/201280 | 16:32 |
| ndipanov | dansmith, ^ | 16:32 |
| *** apuimedo has quit IRC | 16:32 | |
| *** fawadkhaliq has quit IRC | 16:33 | |
| *** derekh has quit IRC | 16:34 | |
| *** tjones1 has joined #openstack-nova | 16:34 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Split instance resize operation into 2 phases https://review.openstack.org/195088 | 16:35 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Create class hierarchy for tasks in conductor https://review.openstack.org/184495 | 16:35 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP: Base NovaObject on VersionedObject https://review.openstack.org/201655 | 16:37 |
| sdague | mriedem: the db is reset on every test | 16:38 |
| sdague | and it's in memory | 16:38 |
| *** ijw has quit IRC | 16:38 | |
| dansmith | mriedem: wanna hit this so we can get it on the backport queue? https://review.openstack.org/201280 | 16:39 |
| dansmith | or sdague | 16:39 |
| *** annashen has joined #openstack-nova | 16:39 | |
| sdague | dansmith: done | 16:39 |
| dansmith | thanks | 16:39 |
| *** ijw has joined #openstack-nova | 16:39 | |
| sdague | dansmith: oh, so the thing I asked ndipanov this morning, we seem to have to land one of these oops patches every month. Is there any way we can get ahead of that with testing? | 16:40 |
| dansmith | sdague: we have tests for it | 16:40 |
| dansmith | sdague: there were some critical missteps in there, which rlrossit has patches up to fix | 16:40 |
| sdague | dansmith: ok | 16:41 |
| dansmith | sdague: he had already fixed this one in that set, but we wanted to pull this out into a backportable thing | 16:41 |
| sdague | ok, cool. link to those fixes? | 16:41 |
| dansmith | and I don't think it's every month :) | 16:41 |
| rlrossit | sdague: https://review.openstack.org/#/c/198730/ https://review.openstack.org/#/c/198904/ | 16:41 |
| dansmith | yeah that | 16:42 |
| rlrossit | they're kinda stale though because I'm waiting for the dust to settle before going back in and rebasing | 16:42 |
| dansmith | but it needs rebasing on top of this still I think | 16:42 |
| dansmith | yeah | 16:42 |
| dansmith | the critical missing link was that we weren't following links through list objects in the test | 16:43 |
| rlrossit | dansmith: should I add the checking for incorrect future versioning in one of these changes? or make that a separate change? | 16:43 |
| dansmith | which was an exception put in place by yjiang5 | 16:43 |
| dansmith | rlrossit: either is fine with me | 16:43 |
| dansmith | but the exception kinda missed the point that if we check half the structure in one test and half in the other, | 16:44 |
| dansmith | we miss the intersection points between them | 16:44 |
| dansmith | which is what rlrossit has fixed | 16:44 |
| *** nkrinner is now known as nkrinner_afk | 16:44 | |
| dansmith | so, we thought we were doing well, but in reality we were missing a few things | 16:44 |
| dansmith | that, and we still have reviewers that don't notice when someone changes the hash for an object and not the version :( | 16:44 |
| *** yamahata has quit IRC | 16:45 | |
| *** oro_ has joined #openstack-nova | 16:46 | |
| *** fawadkhaliq has joined #openstack-nova | 16:46 | |
| *** oro has joined #openstack-nova | 16:47 | |
| ndipanov | dansmith, can you at least add a +1 on the kilo backport so that I can bug stable maint to merge it? | 16:47 |
| dansmith | ndipanov: you already have a backport? | 16:47 |
| mriedem | ndipanov: i can look at the backport | 16:48 |
| ndipanov | well yes | 16:48 |
| mriedem | https://review.openstack.org/#/c/201654/ | 16:48 |
| *** unicell has joined #openstack-nova | 16:48 | |
| dansmith | yeah, I see | 16:48 |
| *** lyanchih has quit IRC | 16:48 | |
| dansmith | ndipanov: if you just wait until master lands, then mriedem and I can merge it | 16:48 |
| *** unicell has quit IRC | 16:48 | |
| *** unicell has joined #openstack-nova | 16:48 | |
| ndipanov | cools! | 16:48 |
| mriedem | ndipanov: the commit message is different | 16:49 |
| ndipanov | yes | 16:49 |
| *** lyanchih has joined #openstack-nova | 16:49 | |
| ndipanov | so as to make sense | 16:49 |
| mriedem | sure yeah i guess | 16:49 |
| *** Jeffrey4l has quit IRC | 16:49 | |
| mriedem | jwcroppe: 2nd bullet under ibm specifics https://wiki.openstack.org/wiki/Sprints/NovaLibertySprint#IBM_Specifics | 16:52 |
| jwcroppe | mriedem: thanks | 16:54 |
| *** xyang2 has quit IRC | 16:55 | |
| raildo | Hi guys, I'm working to provide nested quotas here in nova. With this feature, a parent project will be able to set the quotas for their children. | 16:56 |
| raildo | but during the implementation, we found a project_id check here: https://github.com/openstack/nova/blob/8fd7b44d1ae4fd919dfdab4aab358a3137708ff4/nova/api/openstack/wsgi.py#L805-L812 | 16:56 |
| raildo | that prohibit make some action a project which isn't the project in the context. | 16:56 |
| raildo | Anyone know a way to don't make this check only for the quota update? | 16:57 |
| *** lpetrut has quit IRC | 16:58 | |
| *** ijw has quit IRC | 16:59 | |
| *** ijw has joined #openstack-nova | 16:59 | |
| *** aix has quit IRC | 16:59 | |
| *** mwagner_lap has quit IRC | 17:00 | |
| *** mgarza_ has joined #openstack-nova | 17:00 | |
| *** gmurphy has joined #openstack-nova | 17:01 | |
| *** tjones1 has quit IRC | 17:01 | |
| *** Jeffrey4l has joined #openstack-nova | 17:01 | |
| *** lucasagomes is now known as lucas-dinner | 17:01 | |
| *** browne has quit IRC | 17:04 | |
| *** kjelly has quit IRC | 17:05 | |
| *** davideagnello has quit IRC | 17:06 | |
| *** ildikov_ has quit IRC | 17:07 | |
| *** yamahata has joined #openstack-nova | 17:11 | |
| *** veena has joined #openstack-nova | 17:12 | |
| *** vilobhmm has joined #openstack-nova | 17:12 | |
| *** jistr has quit IRC | 17:14 | |
| *** davideagnello has joined #openstack-nova | 17:14 | |
| sdague | mriedem: as I consider you the oslo.i18n expert, is this right - https://review.openstack.org/#/c/193860/1/nova/virt/libvirt/volume.py,cm ? | 17:15 |
| *** jcoufal has joined #openstack-nova | 17:15 | |
| *** ssurana has joined #openstack-nova | 17:15 | |
| *** kylek3h has joined #openstack-nova | 17:15 | |
| mriedem | looking | 17:15 |
| mriedem | jecarey is the expert | 17:15 |
| *** tonytan4ever has quit IRC | 17:15 | |
| sdague | mriedem: sure, but I can blame you | 17:15 |
| *** MentalRay has quit IRC | 17:16 | |
| mriedem | http://docs.openstack.org/developer/oslo.i18n/guidelines.html#choosing-a-marker-function | 17:16 |
| *** fawadkhaliq has quit IRC | 17:16 | |
| sdague | mriedem: yeh | 17:16 |
| mriedem | +2 | 17:16 |
| sdague | it's all just confusing with "do this" "except" | 17:16 |
| mriedem | don't do what donny don't does | 17:17 |
| *** signed8bit is now known as signed8bit_ZZZzz | 17:17 | |
| mriedem | http://i.imgur.com/VwKnQ5H.jpg | 17:17 |
| sdague | are the _L* functions only used to build priority lists? | 17:17 |
| mriedem | i guess? "The purpose of the different marker functions is to separate the translatable messages into different catalogs, which the translation teams can prioritize translating. It is important to choose the right marker function, to ensure that strings the user sees will be translated and to help the translation team manage their work load." | 17:18 |
| openstackgerrit | Qin Zhao proposed openstack/nova: VMware: Use virtual disk size instead of image size https://review.openstack.org/168013 | 17:18 |
| mriedem | so i'd prioritize _LE | 17:18 |
| *** patrickeast has joined #openstack-nova | 17:18 | |
| *** jcoufal has quit IRC | 17:19 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: libvirt:Rsync remote FS driver was added https://review.openstack.org/146887 | 17:20 |
| *** coolsvap is now known as coolsvap|afk | 17:20 | |
| openstackgerrit | Qin Zhao proposed openstack/nova: VMware: Use virtual disk size instead of image size https://review.openstack.org/168013 | 17:22 |
| *** annashen has quit IRC | 17:23 | |
| *** kiran-r has joined #openstack-nova | 17:28 | |
| *** krtaylor has quit IRC | 17:29 | |
| *** mtanino has joined #openstack-nova | 17:29 | |
| *** achanda has joined #openstack-nova | 17:29 | |
| openstackgerrit | Mike Dorman proposed openstack/nova: Add console allowed origins setting https://review.openstack.org/201677 | 17:32 |
| *** Sukhdev has quit IRC | 17:33 | |
| *** fawadkhaliq has joined #openstack-nova | 17:34 | |
| *** markus_z has quit IRC | 17:35 | |
| *** lpetrut has joined #openstack-nova | 17:36 | |
| *** mpaolino has joined #openstack-nova | 17:37 | |
| *** ijw has quit IRC | 17:40 | |
| *** ijw has joined #openstack-nova | 17:40 | |
| *** achanda has quit IRC | 17:41 | |
| *** andrewbogott is now known as andrewbogott_afk | 17:41 | |
| *** achanda has joined #openstack-nova | 17:41 | |
| *** otter768 has joined #openstack-nova | 17:47 | |
| *** armax has quit IRC | 17:48 | |
| *** MentalRay has joined #openstack-nova | 17:48 | |
| *** pradipta has joined #openstack-nova | 17:51 | |
| *** ildikov has joined #openstack-nova | 17:51 | |
| *** otter768 has quit IRC | 17:52 | |
| *** ociuhandu has quit IRC | 17:52 | |
| *** browne has joined #openstack-nova | 17:53 | |
| *** mwagner_lap has joined #openstack-nova | 17:53 | |
| *** ijw has quit IRC | 17:55 | |
| *** openstackgerrit has quit IRC | 17:56 | |
| sdague | so, honestly, it feels like we should be able to build a tool to do that instead of having to annotate every message | 17:56 |
| *** ijw_ has joined #openstack-nova | 17:57 | |
| *** openstackgerrit has joined #openstack-nova | 17:57 | |
| sdague | hmm... numa datastructure isn't stable ordered - https://jenkins02.openstack.org/job/gate-nova-python34/581/console ? | 17:57 |
| *** mpaolino has quit IRC | 17:58 | |
| ndipanov | sdague why would you assume that | 18:00 |
| *** lpetrut has quit IRC | 18:00 | |
| sdague | ndipanov: because of how that test result fails | 18:00 |
| sdague | it's because the order of a highly nested array is different than expected | 18:00 |
| sdague | same content though | 18:01 |
| ndipanov | it's a dict | 18:01 |
| sdague | ndipanov: did you look at the failure? | 18:01 |
| *** agentle_ has quit IRC | 18:02 | |
| ndipanov | yes but those are dicts - not objects | 18:02 |
| *** agentle_ has joined #openstack-nova | 18:02 | |
| *** apuimedo has joined #openstack-nova | 18:03 | |
| sdague | ndipanov: "nova_object.changes": ["memory", "cpuset", "id"]}] vs. "nova_object.changes": ["cpuset", "memory", "id"]}] | 18:03 |
| sdague | ndipanov: so maybe, if you have another theory about what's going on, would be great | 18:03 |
| ndipanov | ah I see what you mean by stable ordered... | 18:04 |
| ndipanov | so the solution to that problem is to not store those changes because they are useless | 18:04 |
| ndipanov | I think there was a patch to do this | 18:04 |
| sdague | ok, if you have a pointer, that would be great, because this is rando failing unrelated patches | 18:05 |
| ndipanov | and apparently never merged | 18:05 |
| ndipanov | so the thing is - that's a set in the object | 18:05 |
| ndipanov | but json has no set | 18:05 |
| ndipanov | so it gets turned into a list | 18:05 |
| *** Nikolay_St has quit IRC | 18:05 | |
| ndipanov | I assume py27 has a deterministic ordering | 18:05 |
| ndipanov | and 3.4 doesn't | 18:05 |
| ndipanov | sdague, ^ | 18:05 |
| ndipanov | would that make sense? | 18:05 |
| ndipanov | hmm no json there actually | 18:06 |
| ndipanov | let me look into it more | 18:06 |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: libvirt: rename parallels driver to virtuozzo https://review.openstack.org/184311 | 18:06 |
| *** chuckcarmack has joined #openstack-nova | 18:07 | |
| ndipanov | ah no there is json there actually sdague | 18:07 |
| ndipanov | yeah sdague there's your problem | 18:08 |
| *** annashen has joined #openstack-nova | 18:09 | |
| *** agireud has quit IRC | 18:09 | |
| sdague | ndipanov: well, if you are relying on natural ordering of python data structures, that changes between releases | 18:09 |
| ndipanov | yes | 18:09 |
| ndipanov | but we rely on it here only for this test to work really | 18:10 |
| ndipanov | so the real solution is - drop those changes from the DB | 18:10 |
| ndipanov | (I will dig up the patch I know one was posted) | 18:10 |
| ndipanov | and if that's too uncertain to wait | 18:10 |
| ndipanov | just add sorting to that list there | 18:11 |
| ndipanov | I vote for 1) | 18:11 |
| *** signed8bit_ZZZzz is now known as signed8bit | 18:11 | |
| sdague | ndipanov: sure, though it seems like fake_obj_numa_topology._to_json() is the thing that needs the stable ordering, right? | 18:11 |
| ndipanov | well that stable ordering comes from obj_to_primitive | 18:12 |
| *** veena has quit IRC | 18:12 | |
| ndipanov | which just calls list() | 18:12 |
| ndipanov | a super quick hack would be to do sorted() as well | 18:12 |
| dansmith | edleafe wrote some things for stable object comparisons, IIRC | 18:12 |
| *** iamjarvo has joined #openstack-nova | 18:13 | |
| *** tpeoples has quit IRC | 18:13 | |
| *** tdp has joined #openstack-nova | 18:13 | |
| ndipanov | if self.obj_what_changed(): | 18:13 |
| ndipanov | obj['nova_object.changes'] = list(self.obj_what_changed()) | 18:13 |
| ndipanov | sdague, ^ | 18:13 |
| ndipanov | this is what causes it | 18:13 |
| ndipanov | (I think) | 18:13 |
| sdague | yeh, that could be. dansmith you want to weigh in on this? | 18:14 |
| ndipanov | sorting that list would be the quickest way for that to go away | 18:14 |
| sdague | ndipanov: sure | 18:14 |
| ndipanov | but in reality there is no need to sort that list | 18:14 |
| dansmith | there is no need to sort that list | 18:14 |
| dansmith | did you see what I said about edleafe's object comparison stuff? | 18:14 |
| ndipanov | but this is not object comparison | 18:15 |
| ndipanov | this is only for numa | 18:15 |
| ndipanov | which serializes this into the db | 18:15 |
| ndipanov | and then fails tests on python 3 | 18:15 |
| ndipanov | we would still need to compare json strings in that test | 18:15 |
| ndipanov | not objects | 18:15 |
| *** xyang2 has joined #openstack-nova | 18:15 | |
| dansmith | yeah, I know | 18:16 |
| dansmith | I think he had a stable json comparator for that reason | 18:16 |
| ndipanov | in that case awesome | 18:17 |
| sdague | dansmith: except, that wouldn't help here | 18:17 |
| ndipanov | a slightly less horrible way to fix this is to sort only in _to_json() method of InstanceNumaTopology | 18:17 |
| sdague | because it's mock doing the comparison | 18:17 |
| ndipanov | that way we would have deterministic ordering always | 18:18 |
| *** agireud has joined #openstack-nova | 18:18 | |
| *** krtaylor has joined #openstack-nova | 18:18 | |
| ndipanov | but the real fix is to not save those changes | 18:18 |
| ndipanov | dansmith, do you remember if there was an actual fix for that? | 18:18 |
| sdague | http://logs.openstack.org/05/160205/13/gate/gate-nova-python34/e789e1c/console.html#_2015-07-14_17_33_29_557 | 18:18 |
| dansmith | sdague: ah, using mox you mean? | 18:18 |
| *** lpetrut has joined #openstack-nova | 18:19 | |
| dansmith | ah, I see | 18:19 |
| dansmith | well, you could just inspect the call instead of using assert_called_with | 18:20 |
| sdague | dansmith: sure | 18:20 |
| dansmith | we don't need to persist the changes, which is what ndipanov is saying, but that's just sidestepping the issue | 18:20 |
| sdague | is the stable json comparitor in? | 18:20 |
| *** ijw_ has quit IRC | 18:20 | |
| dansmith | yeah, it's been for ages | 18:20 |
| dansmith | I need to find it though :) | 18:20 |
| ndipanov | dansmith, well imho there is no issue - changes don't need to be ordered | 18:21 |
| sdague | it does seem like forcing stable order on the json we generate would be a long term good thing | 18:21 |
| *** ijw has joined #openstack-nova | 18:21 | |
| *** tdp is now known as tpeoples | 18:21 | |
| dansmith | sdague: why? | 18:22 |
| dansmith | that's just burning CPU every time we serialize an object | 18:22 |
| dansmith | it's an unordered thing, it doesn't matter at all | 18:22 |
| *** MentalRay has quit IRC | 18:22 | |
| sdague | I'd be pretty surprised if it was a measurable amount of CPU | 18:23 |
| dansmith | it's just a silly thing to do to make our unit tests pass | 18:23 |
| sdague | maybe, it feels like stable structures are more useful should they get logs or compared in other contexts, but that's fine | 18:24 |
| *** wasmum- has joined #openstack-nova | 18:24 | |
| *** hakimo_ has joined #openstack-nova | 18:24 | |
| dansmith | so in the actual code, this is a set not a list | 18:25 |
| dansmith | it's only a list for the trip over the wire because json has no set structure | 18:25 |
| *** iamjarvo has quit IRC | 18:25 | |
| dansmith | sorting it before we serialize it and then converting it to a set on the other end is just waste | 18:25 |
| sdague | ok | 18:25 |
| *** raorn_ has joined #openstack-nova | 18:25 | |
| sdague | if it's really a set, then I'm fine with that answer | 18:25 |
| dansmith | https://github.com/openstack/nova/blob/master/nova/tests/unit/objects/test_objects.py#L218-253 | 18:26 |
| dansmith | so that's the top level thing | 18:26 |
| dansmith | which takes comparators | 18:26 |
| *** rdo_ has joined #openstack-nova | 18:26 | |
| dansmith | and there is a json one I think which I think we might be able to use | 18:26 |
| dansmith | ah, we have an assertJsonEqual somewhere | 18:27 |
| *** armax has joined #openstack-nova | 18:27 | |
| dansmith | so if we just use that and pull the call out ourselves | 18:27 |
| dansmith | that makes more sense I think | 18:27 |
| sdague | sure | 18:27 |
| dansmith | but, we also need not persist changes in the database like this, | 18:28 |
| sdague | how do you extract the calling params for a mock call? I've never done that before | 18:28 |
| dansmith | so just calling obj_reset_changes() on the thing before we save it is also saving us bytes in the db | 18:28 |
| dansmith | sdague: mock_thingy.call_args_list[0][0][0] is the first argument | 18:28 |
| dansmith | of the first call | 18:28 |
| sdague | ok | 18:28 |
| sdague | 3 [0]? | 18:28 |
| sdague | what's the 3rd one for | 18:29 |
| *** changbl has quit IRC | 18:29 | |
| *** shyama has quit IRC | 18:29 | |
| *** tlbr_ has joined #openstack-nova | 18:29 | |
| dansmith | https://github.com/openstack/nova/blob/master/nova/tests/unit/objects/test_instance.py#L456-456 | 18:29 |
| *** adam_g` has joined #openstack-nova | 18:29 | |
| *** mgagne_ has joined #openstack-nova | 18:29 | |
| *** __maurosr__ has joined #openstack-nova | 18:29 | |
| dansmith | actually, nevermind that link | 18:30 |
| *** kcaj_ has joined #openstack-nova | 18:30 | |
| dansmith | sdague: it's mock being silly | 18:30 |
| *** edleafe- has joined #openstack-nova | 18:30 | |
| dansmith | sdague: one call of foo(123, bar=456) looks likethis: | 18:30 |
| *** fawadkhaliq has quit IRC | 18:30 | |
| *** vilobhmm has quit IRC | 18:30 | |
| *** jogo has quit IRC | 18:30 | |
| *** READ10 has quit IRC | 18:30 | |
| *** ddaskal has quit IRC | 18:30 | |
| *** zsais has quit IRC | 18:30 | |
| *** jdandrea has quit IRC | 18:30 | |
| *** rlrossit has quit IRC | 18:30 | |
| *** ivasev has quit IRC | 18:30 | |
| *** rdo has quit IRC | 18:30 | |
| *** Marga_ has quit IRC | 18:30 | |
| *** hakimo has quit IRC | 18:30 | |
| *** angular_mike_ has quit IRC | 18:30 | |
| *** edleafe has quit IRC | 18:30 | |
| *** tlbr has quit IRC | 18:30 | |
| *** isq has quit IRC | 18:30 | |
| *** ccard has quit IRC | 18:30 | |
| *** PaulMurray has quit IRC | 18:30 | |
| *** kcaj has quit IRC | 18:30 | |
| *** zhiyan has quit IRC | 18:30 | |
| *** raorn has quit IRC | 18:30 | |
| *** rex_lee has quit IRC | 18:30 | |
| *** maurosr has quit IRC | 18:30 | |
| *** scottda has quit IRC | 18:30 | |
| *** wasmum has quit IRC | 18:30 | |
| *** mgagne has quit IRC | 18:30 | |
| *** adam_g has quit IRC | 18:30 | |
| *** sileht has quit IRC | 18:30 | |
| *** stefanb has quit IRC | 18:30 | |
| *** mrda-away has quit IRC | 18:30 | |
| *** __maurosr__ is now known as maurosr | 18:30 | |
| dansmith | call_args_list[([123],{'bar':456})] | 18:30 |
| *** MentalRay has joined #openstack-nova | 18:31 | |
| dansmith | sdague: it's a list of calls, each entry is a tuple of args,kwargs | 18:31 |
| *** Guest10549 is now known as dan | 18:31 | |
| *** changbl has joined #openstack-nova | 18:32 | |
| *** iamjarvo has joined #openstack-nova | 18:32 | |
| *** armax has quit IRC | 18:32 | |
| *** patrickeast has quit IRC | 18:32 | |
| *** patrickeast_ has joined #openstack-nova | 18:32 | |
| *** mgagne_ has left #openstack-nova | 18:33 | |
| *** armax has joined #openstack-nova | 18:33 | |
| *** oro has quit IRC | 18:34 | |
| *** oro_ has quit IRC | 18:34 | |
| *** Longgeek has quit IRC | 18:36 | |
| edleafe- | dansmith: you sure that was me? | 18:36 |
| edleafe- | doesn't ring a bell | 18:36 |
| dansmith | well, I thought it was | 18:36 |
| * dansmith blames | 18:36 | |
| sdague | edleafe-: just take the credit and move on | 18:37 |
| edleafe- | sdague: heh | 18:37 |
| dansmith | I dunno that stuff has changed enough that it's not easy to tell | 18:37 |
| dansmith | whatever | 18:37 |
| *** erlon has joined #openstack-nova | 18:37 | |
| edleafe- | dansmith: I remember that bauzas did some comparison stuff | 18:38 |
| sdague | ok, so I have the change, man is this the ugs :) | 18:38 |
| edleafe- | dansmith: maybe that's what you had in mind | 18:38 |
| dansmith | I don't think that was this | 18:38 |
| *** krotscheck has joined #openstack-nova | 18:39 | |
| *** armax has quit IRC | 18:39 | |
| *** ijw has quit IRC | 18:39 | |
| sdague | because assertJsonEqual uses actual json, not python structures | 18:40 |
| sdague | so we have to turn it back into strings | 18:40 |
| *** mgagne_ has joined #openstack-nova | 18:40 | |
| mriedem | mtreinish: ha, see the section on building docs here https://github.com/openstack/nova/blob/master/HACKING.rst | 18:42 |
| *** sudipto has quit IRC | 18:43 | |
| sdague | mtreinish: hey, what's up with the random pdf build script? | 18:44 |
| mriedem | mtreinish: yeah, apparently we already have "tox -evenv -- python setup.py build_sphinx_latex" | 18:45 |
| *** kcaj_ has quit IRC | 18:47 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Split instance resize operation into 2 phases https://review.openstack.org/195088 | 18:47 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Create class hierarchy for tasks in conductor https://review.openstack.org/184495 | 18:47 |
| lifeless | sdague: thanks | 18:48 |
| lifeless | sdague: the exception thing is fixed btw, can make those clean again | 18:48 |
| lifeless | (in 1.1.3) | 18:48 |
| sdague | lifeless: cool | 18:48 |
| sdague | the autospec issue still out there? | 18:48 |
| *** kcaj has joined #openstack-nova | 18:49 | |
| lifeless | sdague: it was exposed via autospec but not an autospec issue | 18:49 |
| lifeless | sdague: see http://bugs.python.org/issue23661 | 18:50 |
| *** gmurphy has left #openstack-nova | 18:51 | |
| *** Sukhdev has joined #openstack-nova | 18:51 | |
| openstackgerrit | Sean Dague proposed openstack/nova: make test_save_updates_numa_topology stable accross python versions https://review.openstack.org/201719 | 18:53 |
| sdague | ndipanov / dansmith ^^^ | 18:53 |
| *** sgordon has quit IRC | 18:55 | |
| *** openstackgerrit has quit IRC | 18:56 | |
| *** openstackgerrit has joined #openstack-nova | 18:56 | |
| *** sgordon has joined #openstack-nova | 18:57 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Update HACKING.rst for running tests and building docs https://review.openstack.org/201722 | 18:58 |
| mtreinish | sdague: sphinx doesn't generate valid latex so if you try to use it it'll fail | 19:01 |
| mtreinish | so I wrote a helper script to work around it | 19:01 |
| mtreinish | I still need to report the bugs upstream though | 19:01 |
| sdague | yeh, I'd feel more ok about that after the bugs were filed upstream | 19:02 |
| sdague | I also think doing some error handling is probably appropriate | 19:02 |
| lifeless | andreykurilin: is that issue you filed the one sdague was mentioning ? | 19:02 |
| mtreinish | sdague: well, tbh I'm not sure how much of them will be considered sphinx bugs vs nova docs bugs | 19:03 |
| mtreinish | like the 2 issues were the unicode check and x mark usage and the svg arch diagram | 19:03 |
| *** tpatil has joined #openstack-nova | 19:03 | |
| openstackgerrit | Ming Yang proposed openstack/nova: Refactor get_filtered_hosts in filter host manager https://review.openstack.org/192878 | 19:03 |
| openstackgerrit | Ming Yang proposed openstack/nova: Case-insensitive filters https://review.openstack.org/201725 | 19:03 |
| mtreinish | sdague: well found one open for the svg issue: https://github.com/sphinx-doc/sphinx/issues/1907 | 19:04 |
| sdague | lifeless: yes, andreykurilin filed the bug | 19:05 |
| sdague | we were working through the patch together | 19:05 |
| *** mutoulbj has quit IRC | 19:07 | |
| *** ijw has joined #openstack-nova | 19:10 | |
| lifeless | cool | 19:11 |
| *** achanda has quit IRC | 19:17 | |
| *** achanda has joined #openstack-nova | 19:17 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova-specs: Spec: add nova-manage db purge-deleted-instances https://review.openstack.org/200224 | 19:18 |
| *** xyang2 has quit IRC | 19:19 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Add version counter to Service object https://review.openstack.org/201733 | 19:20 |
| *** agentle__ has joined #openstack-nova | 19:20 | |
| *** ZZelle_ has joined #openstack-nova | 19:21 | |
| openstackgerrit | Sean Dague proposed openstack/nova: make test_save_updates_numa_topology stable across python versions https://review.openstack.org/201719 | 19:21 |
| sdague | ok, dansmith how is that language? | 19:21 |
| mriedem | sdague: is this a thing in stable/kilo also? | 19:22 |
| mriedem | if so, we should have a bug for this for backporting | 19:22 |
| openstackgerrit | lyanchih proposed openstack/nova: Creating instance fail when inject ssh key in cells mode https://review.openstack.org/197514 | 19:22 |
| dansmith | sdague: still sounds like it's a NUMA-specific thing, which it's not, but it's much better than it was | 19:22 |
| sdague | mriedem: no idea, I caught a gate failure | 19:23 |
| dansmith | mriedem: it's just a unit test thing | 19:23 |
| dansmith | mriedem: not sure it's worth a backport | 19:23 |
| mriedem | oh and it's py34 right? | 19:23 |
| mriedem | yeah, nvm | 19:23 |
| sdague | I saw it expose on py34 | 19:23 |
| sdague | I have no idea if that's the only place it will or not | 19:23 |
| sdague | but, yeh, backport seems unneeded | 19:23 |
| *** MentalRay has quit IRC | 19:24 | |
| *** agentle_ has quit IRC | 19:24 | |
| dims_ | sdague: there's a bug as well - https://bugs.launchpad.net/nova/+bug/1473949 | 19:24 |
| openstack | Launchpad bug 1473949 in OpenStack Compute (nova) "gate-nova-python34 some times test failed on test_save_updates_numa_topology" [Undecided,In progress] - Assigned to lyanchih (lyanchih) | 19:24 |
| sdague | dims_: oh, I didn't even know | 19:25 |
| sdague | I can close bug that | 19:25 |
| dims_ | ++ sdague | 19:25 |
| openstackgerrit | Sean Dague proposed openstack/nova: make test_save_updates_numa_topology stable across python versions https://review.openstack.org/201719 | 19:25 |
| sdague | mriedem can you reapprove, now with updated commit message, which I know you like | 19:25 |
| mriedem | sdague: man i already did like 50 seconds ago | 19:27 |
| *** MentalRay has joined #openstack-nova | 19:27 | |
| mriedem | btw | 19:27 |
| mriedem | is it odd that i was thinking about what happened to the intel pci ci in the middle of the night last night? | 19:28 |
| dansmith | yes | 19:28 |
| mriedem | has anyone noticed it's gone? | 19:28 |
| mriedem | n0ano: do you know anything about that? ^ | 19:29 |
| dansmith | gone for over 30 days it looks like | 19:29 |
| mriedem | right, hence my cold sweats at 3am | 19:29 |
| mriedem | :) | 19:29 |
| mriedem | it's like 'home alone: 4' | 19:29 |
| mriedem | in other ci news, jhesketh's fixes for t-h in kilo have been in the gate for over 3 hours now, really like 2 days | 19:30 |
| mriedem | sucks | 19:30 |
| * sdague has a vision of mriedem putting "LOST: Intel CI" fliers up in his neighboorhood | 19:30 | |
| n0ano | mriedem, no, hadn't heard but I can ask around, I think the ones running the CI are in China so I won't find out until this evening | 19:31 |
| mriedem | ok, i think i got thinking about it b/c of this pci device list version snafu | 19:32 |
| *** lyanchih has quit IRC | 19:32 | |
| *** pradipta has quit IRC | 19:32 | |
| *** Nikolay_St has joined #openstack-nova | 19:32 | |
| *** tonytan4ever has joined #openstack-nova | 19:32 | |
| *** Piet has quit IRC | 19:33 | |
| *** Piet has joined #openstack-nova | 19:33 | |
| *** achanda has quit IRC | 19:34 | |
| dansmith | mriedem: that's a good point, I would have expected it to start catching that pretty much as soon as it landed | 19:34 |
| dansmith | and it was running back then, which makes me wonder what it was doing... | 19:34 |
| *** achanda has joined #openstack-nova | 19:35 | |
| *** eglynn has quit IRC | 19:35 | |
| dansmith | never ran on that patch | 19:35 |
| sdague | it's got a different test payload, that can be looked at | 19:36 |
| *** mgarza_ has quit IRC | 19:37 | |
| *** mgarza_ has joined #openstack-nova | 19:38 | |
| *** Nikolay_St has quit IRC | 19:39 | |
| *** Nikolay_St has joined #openstack-nova | 19:40 | |
| *** Mike_D_laptop has quit IRC | 19:41 | |
| *** iamjarvo has quit IRC | 19:43 | |
| openstackgerrit | Matthew Treinish proposed openstack/nova: Add tool to build a doc latex pdf https://review.openstack.org/199332 | 19:45 |
| mtreinish | sdague, mriedem, dansmith: ^^^ | 19:45 |
| *** achanda has quit IRC | 19:47 | |
| *** mgarza_ has quit IRC | 19:49 | |
| *** vilobhmm has joined #openstack-nova | 19:49 | |
| *** isq has joined #openstack-nova | 19:49 | |
| *** PaulMurray has joined #openstack-nova | 19:49 | |
| *** rlrossit has joined #openstack-nova | 19:49 | |
| *** ccard has joined #openstack-nova | 19:49 | |
| *** mgarza_ has joined #openstack-nova | 19:49 | |
| *** ddaskal has joined #openstack-nova | 19:49 | |
| *** Marga_ has joined #openstack-nova | 19:49 | |
| *** wverdugo500 has joined #openstack-nova | 19:49 | |
| *** sileht has joined #openstack-nova | 19:49 | |
| sdague | mtreinish: I still don't understand why you do the backup | 19:50 |
| *** iamjarvo has joined #openstack-nova | 19:51 | |
| sdague | it's also not idempotent with your sed line | 19:51 |
| mtreinish | oh, I guess we don't need to. I was doing that to test the sed stuff and do a diff | 19:52 |
| mtreinish | I can drop that bit | 19:52 |
| *** tjones1 has joined #openstack-nova | 19:52 | |
| *** akscram has joined #openstack-nova | 19:52 | |
| *** angular_mike_ has joined #openstack-nova | 19:52 | |
| *** stefanb has joined #openstack-nova | 19:52 | |
| *** mrda-away has joined #openstack-nova | 19:52 | |
| *** angular_mike_ has quit IRC | 19:53 | |
| *** scottda has joined #openstack-nova | 19:56 | |
| sdague | the non idempotent part concerns me a bit, what happens if it gets run more than once | 19:56 |
| *** mgarza_ has quit IRC | 19:57 | |
| *** angular_mike_ has joined #openstack-nova | 19:57 | |
| *** mgarza_ has joined #openstack-nova | 19:57 | |
| *** kiran-r has quit IRC | 19:57 | |
| *** ijw has quit IRC | 19:58 | |
| cfriesen | ndipanov: you around? | 19:59 |
| *** zhiyan has joined #openstack-nova | 20:02 | |
| *** agentle__ is now known as annegentle | 20:02 | |
| *** rex_lee has joined #openstack-nova | 20:03 | |
| mriedem | ctrath: were you going to follow up in the mailing list on the options for this? https://review.openstack.org/#/c/199604/ | 20:03 |
| mriedem | or want me to? | 20:03 |
| mriedem | i see that alex_xu put in his 2 cents | 20:04 |
| lifeless | sdague: did https://github.com/testing-cabal/mock/issues/280 previously work ? | 20:04 |
| mtreinish | sdague: oh if it's run more than once the sphinx will regenerate first | 20:05 |
| mtreinish | which will restore the original file | 20:05 |
| *** iamjarvo has quit IRC | 20:08 | |
| mriedem | alaski: melwitt: have you seen this type of cells failure in the gate before? http://logs.openstack.org/85/197185/10/gate/gate-tempest-dsvm-cells/7ef1949/logs/screen-n-cpu.txt.gz#_2015-07-14_16_39_06_862 | 20:09 |
| lifeless | sdague: nvm, reproduced. It did. | 20:09 |
| mtreinish | sdague: so the script is idempotent but the actual sed call isn't | 20:09 |
| sdague | mtreinish: and the latex doesn't care with it being in there a million times? | 20:10 |
| sdague | lifeless: cool | 20:10 |
| mtreinish | sdague: care that what's there a million times? | 20:10 |
| sdague | if you call the script a second time | 20:10 |
| sdague | you'll double up on your imports | 20:11 |
| sdague | that will keep growing | 20:11 |
| sdague | -e 's/\\usepackage{multirow}/\\usepackage{multirow}\n\\usepackage{amsmath,amssymb,latexsym}\n\\usepackage{pifont}/g' | 20:11 |
| mtreinish | sdague: I doesn't because the sphinx-build call at the start of the script recreates the .tex file | 20:11 |
| mtreinish | which will undo the changes that sed is making | 20:12 |
| sdague | ah, gotcha | 20:12 |
| alaski | mriedem: that's new to me. Is this only occurring in cells jobs? | 20:12 |
| sdague | ok | 20:12 |
| mriedem | checking logstash now | 20:12 |
| sdague | mtreinish: ok, I'm +2 if you get rid of the backup file | 20:13 |
| sdague | because that's just cruft in your git tree for no use | 20:13 |
| mtreinish | sdague: ok, one sec | 20:13 |
| mriedem | mtreinish: did you check to see if there was a way to exclude doc/source/api from the latex build? | 20:13 |
| mriedem | alaski: not just cells | 20:13 |
| mriedem | looks like it just started in the last 24 hours | 20:14 |
| lifeless | sdague: and found the upstream bug | 20:14 |
| sdague | mtreinish: and... if you file sphinx bugs | 20:14 |
| sdague | lifeless: cool | 20:14 |
| openstackgerrit | Matthew Treinish proposed openstack/nova: Add tool to build a doc latex pdf https://review.openstack.org/199332 | 20:14 |
| lifeless | sdague: http://bugs.python.org/issue21750 | 20:14 |
| mtreinish | sdague: ^^^ | 20:14 |
| mtreinish | sdague: see the commit msg | 20:14 |
| alaski | mriedem: okay. I'll take a look, but glad it's not cells related | 20:14 |
| mtreinish | mriedem: err, not yet I'll have to look into that | 20:14 |
| sdague | mtreinish: oh, win | 20:14 |
| mtreinish | mriedem: how do you do it for the html and man page builds? | 20:14 |
| sdague | lifeless: so... am I the only one concerned that the upstream version apparently has so many regressions ? | 20:15 |
| ctrath | mriedem: regarding https://review.openstack.org/#/c/199604/, I have no idea which option should be chosen. I would reply, but I don't know the downstream impacts | 20:15 |
| mriedem | alaski: i do see this in the cell-region log http://logs.openstack.org/85/197185/10/gate/gate-tempest-dsvm-cells/7ef1949/logs/screen-n-cell-region.txt.gz?level=TRACE | 20:15 |
| mriedem | ctrath: that's the reason for taking it to the ML | 20:15 |
| lifeless | sdague: it was inevitable | 20:15 |
| lifeless | sdague: most upstream libs have few resources maintaining them | 20:15 |
| *** annashen has quit IRC | 20:16 | |
| lifeless | sdague: and few users (literally, only the cpython developers) | 20:16 |
| lifeless | sdague: ... until releases happen | 20:16 |
| sdague | right, well also, python3 usage being much lower I'm sure helps let stuff slip through | 20:16 |
| lifeless | sdague: and mock (like mox in this regard) is very clever code | 20:16 |
| lifeless | so what we need to do is get on this, get the fixes together and into cPython - which I can help with | 20:17 |
| lifeless | sdague: we'd have hit all of these ourselvs *anyway* with the strategy of 'use mock for 2.7, unittest.mock for 3.4' - except we wouldn't have been able to consume the fixes | 20:17 |
| sdague | lifeless: yeh, sure | 20:18 |
| lifeless | e.g. mock_open was broken by 3.4 | 20:18 |
| lifeless | (or possibly between the import and 3.3. releasing) | 20:18 |
| sdague | I'm just kind of surprised we're still uncovering things. It would also be useful to realize that the nova code base makes a pretty good testbed for a wide range of mock use | 20:18 |
| sdague | so it could be used as a test set for future versions | 20:19 |
| mriedem | mtreinish: i need help with an e-r query | 20:20 |
| mriedem | mtreinish: given http://logs.openstack.org/85/197185/10/gate/gate-tempest-dsvm-cells/7ef1949/logs/screen-n-cpu.txt.gz?level=TRACE#_2015-07-14_16_39_06_862 | 20:20 |
| mriedem | why won't this work: | 20:20 |
| mriedem | message:"_build_networks_for_instance" AND message:"InstanceUpdateConflict_Remote" work? | 20:20 |
| alaski | mriedem: the UnexpectedVMStateError is not totally uncommon, it's a messaging race | 20:20 |
| mriedem | alaski: yeah, and that's a delete test too | 20:21 |
| mtreinish | mriedem: hmm so looking at logstash for the individual methods it treating those as separate lines now | 20:23 |
| mriedem | now? | 20:23 |
| mtreinish | so you won't be able to find a message that contains both | 20:23 |
| mriedem | is that new? | 20:23 |
| mtreinish | mriedem: well I though tracebacks used to be treated as a single big line in logstash right? | 20:23 |
| mriedem | yeah | 20:23 |
| mriedem | that's what i thought too | 20:23 |
| mriedem | and why i was using this | 20:23 |
| *** iamjarvo has joined #openstack-nova | 20:24 | |
| *** rotbeard has quit IRC | 20:24 | |
| *** annashen has joined #openstack-nova | 20:24 | |
| *** coolsvap|afk is now known as coolsvap|away | 20:25 | |
| mtreinish | mriedem: maybe a recent oslo_log release changed something | 20:25 |
| mriedem | gdi | 20:25 |
| mriedem | i can't just use InstanceUpdateConflict | 20:25 |
| lifeless | sdague: so, basically the stdlib is a bad idea | 20:25 |
| mriedem | too generic | 20:25 |
| mtreinish | not sure, but that's a guess | 20:25 |
| lifeless | sdague: but there's huge inertia around it | 20:25 |
| sdague | lifeless: them sound like drinking words | 20:25 |
| lifeless | sdague: batteries included is a great idea. But that isn't the same thing as the stdlib :) | 20:25 |
| mriedem | dims_: was there a change to oslo.log such that all lines in a traceback are now separate lines? | 20:25 |
| lifeless | sdague: the expectation in cPython is that during alpha/beta periods one finds out about bugs and issues in the changes from many months of development | 20:26 |
| lifeless | sdague: and we know how well that works .... | 20:26 |
| dims_ | mriedem: does not ring a bell | 20:26 |
| mtreinish | mriedem: well it'd be more subtle than that, it would have changed the format so the parsing rules treat it as multiple lines | 20:26 |
| *** dave-mccowan has quit IRC | 20:27 | |
| mriedem | well gdi | 20:28 |
| mriedem | don't tell me there are subtle bugs | 20:28 |
| mtreinish | mriedem, dims_: it might have been: http://git.openstack.org/cgit/openstack/oslo.log/commit/?id=27f7fe52fefcb81c5fe01736841d7a20149caf02 | 20:28 |
| *** lucas-dinner has quit IRC | 20:28 | |
| mtreinish | let me dig up the oslo filters for logstash | 20:28 |
| *** dave-mccowan has joined #openstack-nova | 20:29 | |
| *** dprince has quit IRC | 20:30 | |
| *** jdandrea has joined #openstack-nova | 20:32 | |
| mriedem | mtreinish: pattern => "^%{TIMESTAMP_ISO8601}%{SPACE}%{NUMBER}?%{SPACE}?TRACE" | 20:32 |
| mriedem | for multiline | 20:32 |
| mtreinish | mriedem: yep | 20:33 |
| mriedem | ok, let me update a thing | 20:33 |
| mriedem | but | 20:33 |
| *** jogo has joined #openstack-nova | 20:33 | |
| mriedem | doesn't seem like i could just change TRACE to ERROR there | 20:33 |
| mriedem | since not all errors are multiline | 20:34 |
| mtreinish | mriedem: yeah, that might not be the best choice | 20:34 |
| mtreinish | mriedem: I don't think we have a way to differentiate a traceback from a normal error now | 20:35 |
| mtreinish | mriedem: it's all dims_ fault :) | 20:35 |
| mtreinish | mriedem: this also means a bunch of are existing queries won't be matching anything anymore... | 20:36 |
| mriedem | yes | 20:36 |
| mriedem | i know | 20:36 |
| mriedem | revert is up | 20:37 |
| dims_ | haha | 20:38 |
| *** mgarza_ has quit IRC | 20:38 | |
| *** achanda has joined #openstack-nova | 20:38 | |
| *** vladikr has quit IRC | 20:38 | |
| *** mgarza has joined #openstack-nova | 20:39 | |
| alaski | mriedem: https://review.openstack.org/#/c/141115/ is the cause of that failure | 20:40 |
| alaski | there's a new error raised, but no code to deal with it | 20:41 |
| *** mgarza has quit IRC | 20:42 | |
| mriedem | alaski: ok, yeah, i hadn't checked git history yet | 20:43 |
| mriedem | but that's spot on | 20:43 |
| *** oro_ has joined #openstack-nova | 20:43 | |
| *** oro has joined #openstack-nova | 20:43 | |
| mriedem | alaski: let me get a bug writen | 20:43 |
| mriedem | *written | 20:43 |
| mriedem | unf'ing logstash logs atm | 20:43 |
| ctrath | mriedem: I sent that email to the ML. Hopefully I didn't murder it too badly | 20:44 |
| *** Alexei_987 has joined #openstack-nova | 20:45 | |
| *** Alexei_987 has left #openstack-nova | 20:45 | |
| mriedem | alaski: https://bugs.launchpad.net/nova/+bug/1474550 | 20:45 |
| openstack | Launchpad bug 1474550 in OpenStack Compute (nova) "network allocation randomly failing with InstanceUpdateConflict after compare and swap was merged" [Undecided,New] | 20:45 |
| mriedem | i see it drops off a bit in logstash, do you see anything merged which fixed it? | 20:45 |
| alaski | not yet | 20:46 |
| mriedem | revert? | 20:46 |
| mriedem | mdbooth: around? | 20:46 |
| mriedem | ^ | 20:46 |
| alaski | considering a revert | 20:46 |
| mriedem | it would crush him :) | 20:47 |
| alaski | the other option would be to have InstanceUpdateConflict inherit from UnexpectedVMStateError | 20:47 |
| alaski | heh | 20:47 |
| alaski | since that's an expected error from that method | 20:47 |
| alaski | and then separate handling could be added in later | 20:47 |
| *** ijw has joined #openstack-nova | 20:48 | |
| mriedem | yeah, or UnexpectedVMStateError | 20:48 |
| mriedem | those are both coming out of that | 20:48 |
| mriedem | the comment in the code says we know something conflicted but we don't know what | 20:49 |
| flwang | mriedem: dansmith: could you please review the nova backup spec https://review.openstack.org/190919 ? | 20:49 |
| *** mgagne_ is now known as mgagne | 20:50 | |
| mriedem | alaski: i'm leaning toward revert | 20:50 |
| alaski | mriedem: yeah, that would be better | 20:51 |
| *** Sukhdev has quit IRC | 20:51 | |
| alaski | the commit message indicates this is expected to be an exceptionally rare case, which obviously isn't true | 20:52 |
| alaski | so better handling is needed before going this route | 20:52 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Revert "Implement compare-and-swap for instance update" https://review.openstack.org/201771 | 20:52 |
| mriedem | there we go | 20:53 |
| *** raildo has quit IRC | 20:53 | |
| alaski | mriedem: +2 | 20:54 |
| mriedem | sdague: jaypipes: dansmith: ^ | 20:55 |
| sdague | mriedem: +A | 20:55 |
| sdague | mriedem: do we need to promote that? | 20:55 |
| mriedem | sdague: no | 20:56 |
| sdague | ok | 20:56 |
| mriedem | flwang: sorry, in the middle of some things atm | 20:56 |
| jaypipes | *sigh* | 20:56 |
| jaypipes | why do I feel like every reverted patch has my fingerprints on it... | 20:57 |
| alaski | jaypipes: it's great that it exposed the race. it just needs some handling for it | 20:57 |
| *** dannywilson has joined #openstack-nova | 20:59 | |
| *** dannywi__ has joined #openstack-nova | 20:59 | |
| *** dannywilson has quit IRC | 20:59 | |
| *** Sukhdev has joined #openstack-nova | 21:00 | |
| *** edleafe- is now known as edleafe | 21:02 | |
| *** jaypipes has quit IRC | 21:02 | |
| *** aysyd has quit IRC | 21:02 | |
| *** dannywil_ has quit IRC | 21:02 | |
| *** changbl has quit IRC | 21:03 | |
| *** rlrossit has left #openstack-nova | 21:03 | |
| *** chuckcarmack has left #openstack-nova | 21:05 | |
| *** thangp has quit IRC | 21:06 | |
| mriedem | ctrath: heh, well, i thought you'd at least put something in the commit message for context | 21:09 |
| ctrath | something from the commit message in the note? | 21:10 |
| mriedem | i'm replying and will updtae the subject | 21:10 |
| ctrath | thanks | 21:11 |
| tjones1 | hi - i have a wierd thing. I ran tox -egenconfig on my kilo env and the sample is missing things like log_dir, debug, and verbose. so i looked at upstream master and there is etc/nova/nova-config-generator.conf which is missing from stable/kilo. was that done after kilo and if so anyone know why the sample would be missing key items like verbose? | 21:11 |
| flwang | mriedem: no worries, it would be appreciated if you can revisit the patch at your most convenience, thanks | 21:14 |
| mriedem | flwang: is feodor +1 on it? | 21:14 |
| flwang | mriedem: not yet, but i have addressed his comments | 21:15 |
| mriedem | ok, i'd like to see him agree with it since he's done a lot of the review work on it and the code already | 21:15 |
| flwang | mriedem: ok, sounds good | 21:15 |
| *** thorst has quit IRC | 21:16 | |
| *** iamjarvo has quit IRC | 21:18 | |
| *** eharney has quit IRC | 21:19 | |
| *** ishita has quit IRC | 21:19 | |
| *** mriedem is now known as mriedem_away | 21:19 | |
| *** jwcroppe_ has joined #openstack-nova | 21:22 | |
| *** ishita has joined #openstack-nova | 21:22 | |
| *** jwcroppe has quit IRC | 21:23 | |
| *** jwcroppe has joined #openstack-nova | 21:26 | |
| *** jwcroppe_ has quit IRC | 21:27 | |
| *** Piet has quit IRC | 21:30 | |
| *** iamjarvo has joined #openstack-nova | 21:30 | |
| *** iamjarvo has quit IRC | 21:31 | |
| *** coolsvap has joined #openstack-nova | 21:31 | |
| *** iamjarvo has joined #openstack-nova | 21:32 | |
| *** iamjarvo has quit IRC | 21:32 | |
| *** ndipanov has quit IRC | 21:37 | |
| *** dboik has quit IRC | 21:37 | |
| *** iamjarvo has joined #openstack-nova | 21:38 | |
| *** iamjarvo has quit IRC | 21:39 | |
| *** lpetrut has quit IRC | 21:40 | |
| *** krtaylor has quit IRC | 21:45 | |
| *** iamjarvo has joined #openstack-nova | 21:48 | |
| *** iamjarvo has quit IRC | 21:49 | |
| *** otter768 has joined #openstack-nova | 21:49 | |
| *** Piet has joined #openstack-nova | 21:51 | |
| *** oro has quit IRC | 21:53 | |
| *** otter768 has quit IRC | 21:53 | |
| *** oro_ has quit IRC | 21:54 | |
| *** cloudrancher has joined #openstack-nova | 21:55 | |
| *** cloudrancher has left #openstack-nova | 21:56 | |
| *** iamjarvo has joined #openstack-nova | 21:56 | |
| *** iamjarvo has quit IRC | 21:56 | |
| *** asettle has joined #openstack-nova | 21:56 | |
| *** jwcroppe has quit IRC | 21:57 | |
| *** shakamunyi has joined #openstack-nova | 21:58 | |
| *** signed8bit has quit IRC | 21:59 | |
| *** tjones1 has quit IRC | 22:01 | |
| *** burt has quit IRC | 22:02 | |
| *** ddaskal has quit IRC | 22:02 | |
| kfox1111 | johnthetubaguy: Thanks. What can we do to keep the ball rolling besides bringing it up in the last meeing? It hasn't seen a comment in almost two weeks? :/ | 22:03 |
| *** pkholkin has quit IRC | 22:04 | |
| *** pkholkin has joined #openstack-nova | 22:04 | |
| *** dannywi__ has quit IRC | 22:06 | |
| *** dannywilson has joined #openstack-nova | 22:07 | |
| *** karimb has joined #openstack-nova | 22:07 | |
| *** edmondsw has quit IRC | 22:08 | |
| *** mdrabe has quit IRC | 22:08 | |
| *** ishita has quit IRC | 22:15 | |
| *** pmcfadden has quit IRC | 22:15 | |
| *** kylek3h has quit IRC | 22:17 | |
| *** annegentle has quit IRC | 22:18 | |
| *** jecarey has quit IRC | 22:19 | |
| *** pmcfadden has joined #openstack-nova | 22:19 | |
| *** tjones1 has joined #openstack-nova | 22:21 | |
| *** ishita has joined #openstack-nova | 22:22 | |
| kfox1111 | having a very strange issue. nova list shows nothing in ip address section on some vm's. nova show doesn't either. but it shows up correctly in the horizon ui. | 22:24 |
| kfox1111 | but horizon just uses the same client the cli does, doesn't it? | 22:25 |
| *** pmcfadden has quit IRC | 22:29 | |
| *** pmcfadden has joined #openstack-nova | 22:31 | |
| *** signed8bit has joined #openstack-nova | 22:31 | |
| *** signed8bit is now known as signed8bit_ZZZzz | 22:31 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/200708 | 22:34 |
| *** signed8bit_ZZZzz has quit IRC | 22:34 | |
| *** zzzeek has quit IRC | 22:41 | |
| *** zzzeek has joined #openstack-nova | 22:41 | |
| *** ijuwang has quit IRC | 22:41 | |
| *** karimb has quit IRC | 22:41 | |
| *** agireud has quit IRC | 22:42 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/200708 | 22:42 |
| *** ijw has quit IRC | 22:42 | |
| *** MentalRay has quit IRC | 22:46 | |
| *** ijw has joined #openstack-nova | 22:47 | |
| *** thorst has joined #openstack-nova | 22:49 | |
| *** xyang1 has joined #openstack-nova | 22:49 | |
| *** zzzeek has quit IRC | 22:50 | |
| *** ctrath has quit IRC | 22:51 | |
| *** artom has quit IRC | 22:51 | |
| openstackgerrit | Earle F. Philhower, III proposed openstack/nova: Add os_brick-based VolumeDriver for HGST connector https://review.openstack.org/194412 | 22:52 |
| *** slong has joined #openstack-nova | 22:56 | |
| *** bnemec has quit IRC | 22:59 | |
| *** Sukhdev has quit IRC | 22:59 | |
| *** annashen has quit IRC | 23:09 | |
| *** dims__ has joined #openstack-nova | 23:17 | |
| *** vilobhmm has quit IRC | 23:17 | |
| *** vilobhmm has joined #openstack-nova | 23:18 | |
| *** diogogmt has quit IRC | 23:18 | |
| *** tpeoples is now known as tpeoples_away | 23:19 | |
| *** dims_ has quit IRC | 23:19 | |
| *** signed8b_ has joined #openstack-nova | 23:20 | |
| *** dprince has joined #openstack-nova | 23:24 | |
| *** ijw_ has joined #openstack-nova | 23:26 | |
| *** signed8b_ is now known as signed8bit_ZZZzz | 23:27 | |
| *** signed8bit_ZZZzz has quit IRC | 23:27 | |
| *** thorst has quit IRC | 23:28 | |
| *** oomichi has joined #openstack-nova | 23:28 | |
| *** ijw has quit IRC | 23:29 | |
| *** ZZelle_ has quit IRC | 23:30 | |
| *** erlon has quit IRC | 23:30 | |
| *** bitblt has joined #openstack-nova | 23:31 | |
| openstackgerrit | ZhaoBo proposed openstack/nova: Fixed raise SecurityGroupCannotBeApplied as network port_security off https://review.openstack.org/188260 | 23:33 |
| *** barra204 has joined #openstack-nova | 23:35 | |
| *** shakamunyi has quit IRC | 23:36 | |
| *** exploreshaifali has quit IRC | 23:38 | |
| *** IanGovett has quit IRC | 23:39 | |
| *** tonytan4ever has quit IRC | 23:40 | |
| *** alexpilotti has quit IRC | 23:40 | |
| *** apuimedo has quit IRC | 23:41 | |
| *** ijw_ has quit IRC | 23:42 | |
| *** pmcfadden has quit IRC | 23:48 | |
| *** asettle has quit IRC | 23:50 | |
| *** otter768 has joined #openstack-nova | 23:50 | |
| *** signed8b_ has joined #openstack-nova | 23:53 | |
| *** otter768 has quit IRC | 23:54 | |
| *** signed8b_ is now known as signed8bit_ZZZzz | 23:55 | |
| *** signed8bit_ZZZzz has quit IRC | 23:55 | |
| *** dannywilson has quit IRC | 23:57 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Add version counter to Service object https://review.openstack.org/201733 | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
