*** thorst has quit IRC | 00:00 | |
*** achanda has quit IRC | 00:03 | |
*** flwang has quit IRC | 00:03 | |
*** achanda_ has quit IRC | 00:04 | |
*** flwang has joined #openstack-nova | 00:04 | |
*** zhangjn has quit IRC | 00:06 | |
*** EinstCrazy has quit IRC | 00:06 | |
*** tanee has quit IRC | 00:14 | |
*** tanee has joined #openstack-nova | 00:14 | |
*** rhedlind has quit IRC | 00:15 | |
*** su_zhang has quit IRC | 00:15 | |
*** rhedlind has joined #openstack-nova | 00:16 | |
openstackgerrit | Nicolas Simonds proposed openstack/nova: Make instance snapshots to respect the force_raw_images setting https://review.openstack.org/226069 | 00:17 |
---|---|---|
*** su_zhang has joined #openstack-nova | 00:17 | |
*** atuvenie__ has quit IRC | 00:21 | |
*** su_zhang has quit IRC | 00:23 | |
*** keekz has quit IRC | 00:24 | |
*** salv-orlando has quit IRC | 00:24 | |
*** keekz has joined #openstack-nova | 00:25 | |
*** angdraug has quit IRC | 00:26 | |
openstackgerrit | ZhuChunzhan proposed openstack/nova: Improve the performance of get instances https://review.openstack.org/214555 | 00:26 |
*** stevemar has joined #openstack-nova | 00:27 | |
*** Apoorva_ has joined #openstack-nova | 00:29 | |
*** Apoorva has quit IRC | 00:32 | |
*** nelsnelson has quit IRC | 00:33 | |
*** Apoorva_ has quit IRC | 00:33 | |
*** nelsnelson has joined #openstack-nova | 00:34 | |
*** ijw has quit IRC | 00:39 | |
*** dims_ has quit IRC | 00:42 | |
*** yuntongjin1 has joined #openstack-nova | 00:46 | |
*** SourabhP has joined #openstack-nova | 00:49 | |
*** EinstCrazy has joined #openstack-nova | 00:51 | |
*** yuntongjin1 has quit IRC | 00:53 | |
*** zhangjn has joined #openstack-nova | 00:54 | |
*** yuntongjin1 has joined #openstack-nova | 00:55 | |
*** ondergetekende has quit IRC | 00:56 | |
*** ondergetekende has joined #openstack-nova | 00:57 | |
*** ctrath has quit IRC | 01:02 | |
*** SourabhP has quit IRC | 01:07 | |
openstackgerrit | Sean McCully proposed openstack/nova: Creates disk.fallocate_test https://review.openstack.org/188249 | 01:10 |
*** ssurana has quit IRC | 01:10 | |
openstackgerrit | Claudiu Belu proposed openstack/nova: WIP: Hyper-V: adds os-win library https://review.openstack.org/212102 | 01:10 |
*** vilobhmm_11 has quit IRC | 01:12 | |
*** smatzek has joined #openstack-nova | 01:18 | |
*** claudiub has quit IRC | 01:19 | |
*** ijw has joined #openstack-nova | 01:20 | |
*** ijw has quit IRC | 01:21 | |
*** ijw has joined #openstack-nova | 01:22 | |
*** angdraug has joined #openstack-nova | 01:25 | |
*** dims has joined #openstack-nova | 01:26 | |
*** angdraug has quit IRC | 01:27 | |
*** artom has quit IRC | 01:30 | |
*** artom_ has joined #openstack-nova | 01:30 | |
*** browne has quit IRC | 01:30 | |
*** otter768 has joined #openstack-nova | 01:30 | |
*** nic has quit IRC | 01:33 | |
*** otter768 has quit IRC | 01:35 | |
*** VW has joined #openstack-nova | 01:43 | |
*** stevemar has quit IRC | 01:46 | |
*** zzzeek has quit IRC | 01:46 | |
*** wverdugo500 has quit IRC | 01:48 | |
*** thorst has joined #openstack-nova | 01:49 | |
openstackgerrit | Eli Qiao proposed openstack/nova: Give instance default hostname if hostname is empty https://review.openstack.org/224438 | 01:49 |
*** asettle has joined #openstack-nova | 01:50 | |
*** haomaiwa_ has joined #openstack-nova | 01:51 | |
*** mwagner_lap has quit IRC | 01:52 | |
*** thorst has quit IRC | 01:52 | |
*** VW has quit IRC | 01:53 | |
*** mwagner_lap has joined #openstack-nova | 01:53 | |
*** shyu has joined #openstack-nova | 01:55 | |
*** haomaiwa_ has quit IRC | 02:01 | |
*** 6A4AAH74W has joined #openstack-nova | 02:01 | |
*** smatzek has quit IRC | 02:01 | |
*** ferest has joined #openstack-nova | 02:03 | |
*** nelsnels_ has joined #openstack-nova | 02:07 | |
*** nelsnelson has quit IRC | 02:08 | |
*** ferest has quit IRC | 02:09 | |
openstackgerrit | Hiroyuki Eguchi proposed openstack/nova: Cannot rebuild a instance booted from volume https://review.openstack.org/201458 | 02:10 |
*** VW has joined #openstack-nova | 02:17 | |
*** otter768 has joined #openstack-nova | 02:17 | |
*** pixelb has quit IRC | 02:18 | |
*** browne has joined #openstack-nova | 02:19 | |
*** otter768 has quit IRC | 02:22 | |
*** otter768 has joined #openstack-nova | 02:23 | |
*** salv-orlando has joined #openstack-nova | 02:28 | |
*** diogogmt has joined #openstack-nova | 02:28 | |
*** mtreinish has quit IRC | 02:28 | |
*** salv-orlando has quit IRC | 02:32 | |
*** achanda has joined #openstack-nova | 02:33 | |
*** dims has quit IRC | 02:33 | |
*** yuntongjin1 has quit IRC | 02:35 | |
*** yuntongjin has joined #openstack-nova | 02:36 | |
*** rushil has joined #openstack-nova | 02:36 | |
*** unicell has quit IRC | 02:36 | |
*** rushil_ has joined #openstack-nova | 02:37 | |
*** david-ly_ has joined #openstack-nova | 02:39 | |
*** unicell has joined #openstack-nova | 02:40 | |
*** david-lyle has quit IRC | 02:40 | |
*** rushil has quit IRC | 02:40 | |
*** dhellmann has quit IRC | 02:40 | |
*** ccarmack1 has joined #openstack-nova | 02:40 | |
*** dhellmann has joined #openstack-nova | 02:41 | |
*** dansmith has quit IRC | 02:41 | |
*** dansmith has joined #openstack-nova | 02:42 | |
*** dansmith is now known as Guest24378 | 02:42 | |
*** unicell has quit IRC | 02:44 | |
*** tobe has joined #openstack-nova | 02:48 | |
*** tobe has quit IRC | 02:48 | |
*** VW has quit IRC | 02:52 | |
*** VW has joined #openstack-nova | 02:52 | |
*** yuner027 has quit IRC | 02:55 | |
*** yuner027 has joined #openstack-nova | 02:56 | |
*** stevemar has joined #openstack-nova | 02:56 | |
*** VW has quit IRC | 02:56 | |
*** mtreinish has joined #openstack-nova | 02:59 | |
*** 6A4AAH74W has quit IRC | 03:01 | |
*** haomaiwa_ has joined #openstack-nova | 03:01 | |
*** rushil_ has quit IRC | 03:01 | |
*** deepthi has joined #openstack-nova | 03:04 | |
*** RA_ has joined #openstack-nova | 03:04 | |
*** _RuiChen has joined #openstack-nova | 03:10 | |
*** RuiChen has quit IRC | 03:13 | |
*** david-ly_ is now known as david-lyle | 03:17 | |
*** Marga__ has joined #openstack-nova | 03:31 | |
*** dims has joined #openstack-nova | 03:34 | |
*** Marga_ has quit IRC | 03:34 | |
*** Marga__ has quit IRC | 03:35 | |
*** yuntongjin has quit IRC | 03:39 | |
*** yamahata has quit IRC | 03:41 | |
openstackgerrit | xing-yang proposed openstack/nova: Rollback is needed if initialize_connection times out https://review.openstack.org/225278 | 03:45 |
*** armax has quit IRC | 03:55 | |
*** haomaiwa_ has quit IRC | 04:01 | |
*** haomaiwa_ has joined #openstack-nova | 04:01 | |
*** dims has quit IRC | 04:09 | |
*** stevemar has quit IRC | 04:09 | |
*** stevemar has joined #openstack-nova | 04:10 | |
*** otter768 has quit IRC | 04:10 | |
openstackgerrit | Joshua Harlow proposed openstack/nova: Add timing logging around each filter being ran https://review.openstack.org/224355 | 04:14 |
*** darrenc is now known as darrenc_afk | 04:14 | |
openstackgerrit | Joshua Harlow proposed openstack/nova: Use the oslo.utils.reflection to extract the class name https://review.openstack.org/224843 | 04:16 |
*** mtreinish has quit IRC | 04:18 | |
openstackgerrit | Wei Jiangang proposed openstack/nova: lib/nova: make redirects happen in correct order https://review.openstack.org/226155 | 04:18 |
*** mtreinish has joined #openstack-nova | 04:23 | |
*** atuvenie__ has joined #openstack-nova | 04:27 | |
*** artom_ has quit IRC | 04:28 | |
*** deepthi has quit IRC | 04:29 | |
*** Sree has joined #openstack-nova | 04:29 | |
*** deepthi has joined #openstack-nova | 04:30 | |
*** vilobhmm11 has joined #openstack-nova | 04:32 | |
*** atuvenie__ has quit IRC | 04:34 | |
*** Sree_ has joined #openstack-nova | 04:37 | |
*** Sree_ is now known as Guest31757 | 04:38 | |
*** Sree has quit IRC | 04:39 | |
openstackgerrit | Wei Jiangang proposed openstack/nova: Make redirects happen in correct order https://review.openstack.org/226155 | 04:40 |
*** hitalia has joined #openstack-nova | 04:47 | |
*** Marga_ has joined #openstack-nova | 04:47 | |
*** Marga_ has quit IRC | 04:49 | |
*** Marga_ has joined #openstack-nova | 04:49 | |
*** yuntongjin has joined #openstack-nova | 04:52 | |
*** Sree has joined #openstack-nova | 04:53 | |
*** kumar has joined #openstack-nova | 04:54 | |
*** Sree has quit IRC | 04:55 | |
*** Guest31757 has quit IRC | 04:55 | |
*** Sree has joined #openstack-nova | 04:56 | |
*** haomaiwa_ has quit IRC | 05:01 | |
*** haomaiwang has joined #openstack-nova | 05:01 | |
*** deepakcs has joined #openstack-nova | 05:05 | |
*** Marga_ has quit IRC | 05:05 | |
*** yuntongjin1 has joined #openstack-nova | 05:07 | |
*** mwagner_lap has quit IRC | 05:08 | |
*** darrenc_afk is now known as darrenc | 05:09 | |
*** yuntongjin has quit IRC | 05:10 | |
*** mwagner_lap has joined #openstack-nova | 05:11 | |
*** changbl has quit IRC | 05:11 | |
*** smcginnis has quit IRC | 05:12 | |
*** changbl has joined #openstack-nova | 05:15 | |
*** ildikov has quit IRC | 05:18 | |
*** sudipto has joined #openstack-nova | 05:19 | |
*** salv-orlando has joined #openstack-nova | 05:26 | |
*** nkrinner has joined #openstack-nova | 05:33 | |
*** salv-orlando has quit IRC | 05:33 | |
*** ankit_ag has joined #openstack-nova | 05:35 | |
rajesht_ | ndipanov: could you please check https://review.openstack.org/#/c/223968/ I think matt has misunderstood it as code cleanup | 05:36 |
rajesht_ | ndipanov: thanks :) | 05:36 |
*** veena has joined #openstack-nova | 05:38 | |
*** shyu has quit IRC | 05:45 | |
*** sahid has joined #openstack-nova | 05:49 | |
*** stevemar has quit IRC | 05:53 | |
*** zhangjn_ has joined #openstack-nova | 05:55 | |
*** zhangjn has quit IRC | 05:55 | |
alex_xu | gmann: ping | 05:56 |
*** haomaiwang has quit IRC | 06:01 | |
*** haomaiwang has joined #openstack-nova | 06:01 | |
openstackgerrit | Feodor Tersin proposed openstack/nova: Do not specify default security group to Neutron by default https://review.openstack.org/173204 | 06:01 |
*** shyu has joined #openstack-nova | 06:02 | |
openstackgerrit | Jianghua Wang proposed openstack/nova: XenAPI: Expose labels for root and ephemeral disks https://review.openstack.org/213094 | 06:04 |
openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix quota update in init_instance on nova-compute restart https://review.openstack.org/226168 | 06:06 |
*** zhangjn has joined #openstack-nova | 06:07 | |
*** zhangjn_ has quit IRC | 06:11 | |
*** otter768 has joined #openstack-nova | 06:11 | |
*** zhangjn has quit IRC | 06:11 | |
*** otter768 has quit IRC | 06:16 | |
*** dave-mccowan has quit IRC | 06:19 | |
*** asettle has quit IRC | 06:19 | |
*** sahid has quit IRC | 06:25 | |
*** sabari has left #openstack-nova | 06:29 | |
*** cfriesen__ has quit IRC | 06:36 | |
*** browne has quit IRC | 06:43 | |
*** shuquan has joined #openstack-nova | 06:48 | |
openstackgerrit | Craige McWhirter proposed openstack/nova: WIP: Add drivers pulled in via stevedore.sphinxext https://review.openstack.org/226177 | 06:50 |
*** asettle has joined #openstack-nova | 06:50 | |
*** smcginnis has joined #openstack-nova | 06:53 | |
*** shuquan has quit IRC | 06:54 | |
*** Sree has quit IRC | 06:56 | |
*** mwagner_lap has quit IRC | 06:56 | |
*** Sree has joined #openstack-nova | 06:57 | |
*** mwagner_lap has joined #openstack-nova | 06:57 | |
*** shuquan has joined #openstack-nova | 06:58 | |
*** david-lyle has quit IRC | 06:58 | |
*** jlanoux has joined #openstack-nova | 07:00 | |
*** haomaiwang has quit IRC | 07:01 | |
*** haomaiwang has joined #openstack-nova | 07:01 | |
*** shuquan has quit IRC | 07:03 | |
*** subscope has quit IRC | 07:03 | |
*** salv-orlando has joined #openstack-nova | 07:03 | |
*** ildikov has joined #openstack-nova | 07:05 | |
*** kiran-r has joined #openstack-nova | 07:07 | |
*** eglynn_ has joined #openstack-nova | 07:07 | |
*** vilobhmm11 has quit IRC | 07:09 | |
*** matrohon has joined #openstack-nova | 07:12 | |
*** jichen has joined #openstack-nova | 07:13 | |
*** shyu has quit IRC | 07:14 | |
*** david-lyle has joined #openstack-nova | 07:14 | |
*** scheuran has joined #openstack-nova | 07:15 | |
*** jamielennox is now known as jamielennox|away | 07:15 | |
*** yuntongjin1 has quit IRC | 07:15 | |
*** achanda has quit IRC | 07:17 | |
*** eglynn_ has quit IRC | 07:17 | |
*** pbrooko has joined #openstack-nova | 07:18 | |
*** kiran-r has quit IRC | 07:18 | |
*** vilobhmm11 has joined #openstack-nova | 07:21 | |
*** kiran-r has joined #openstack-nova | 07:23 | |
*** yuntongjin has joined #openstack-nova | 07:25 | |
*** shyu has joined #openstack-nova | 07:26 | |
*** vilobhmm11 has quit IRC | 07:28 | |
*** takedakn has joined #openstack-nova | 07:28 | |
*** rdopiera has joined #openstack-nova | 07:30 | |
*** sahid has joined #openstack-nova | 07:31 | |
*** veena has quit IRC | 07:31 | |
*** su_zhang has joined #openstack-nova | 07:31 | |
*** zhangjn has joined #openstack-nova | 07:34 | |
*** zhangjn has quit IRC | 07:35 | |
*** zhangjn has joined #openstack-nova | 07:35 | |
*** pbrooko has quit IRC | 07:38 | |
*** yuntongjin has quit IRC | 07:38 | |
*** pbrooko has joined #openstack-nova | 07:40 | |
*** patrickeast has quit IRC | 07:41 | |
*** andrewbogott has quit IRC | 07:41 | |
*** DuncanT has quit IRC | 07:41 | |
*** dougwig has quit IRC | 07:41 | |
*** erlon has quit IRC | 07:41 | |
*** angular_mike_ has quit IRC | 07:41 | |
*** ramishra has quit IRC | 07:41 | |
*** serverascode has quit IRC | 07:41 | |
*** zhenguo has quit IRC | 07:41 | |
*** marcusvrn_ has quit IRC | 07:41 | |
*** scottda has quit IRC | 07:41 | |
*** rex_lee_ has quit IRC | 07:41 | |
*** boris-42 has quit IRC | 07:41 | |
*** wanghao has quit IRC | 07:41 | |
*** zhiyan has quit IRC | 07:41 | |
*** BadCub has quit IRC | 07:41 | |
*** ameade has quit IRC | 07:41 | |
*** Piet has quit IRC | 07:41 | |
*** aweeks has quit IRC | 07:41 | |
*** briancurtin has quit IRC | 07:41 | |
*** pbrooko has quit IRC | 07:42 | |
*** briancurtin has joined #openstack-nova | 07:43 | |
*** aweeks has joined #openstack-nova | 07:44 | |
*** zhenguo has joined #openstack-nova | 07:45 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Zanata https://review.openstack.org/225397 | 07:46 |
*** su_zhang has quit IRC | 07:47 | |
*** takedakn has quit IRC | 07:48 | |
*** ramishra has joined #openstack-nova | 07:49 | |
*** pbrooko has joined #openstack-nova | 07:50 | |
*** aix has joined #openstack-nova | 07:50 | |
*** hitalia has quit IRC | 07:51 | |
*** ndipanov has quit IRC | 07:52 | |
*** takedakn has joined #openstack-nova | 07:54 | |
*** yuntongjin has joined #openstack-nova | 07:55 | |
*** markus_z has joined #openstack-nova | 07:56 | |
*** nihilifer has quit IRC | 07:57 | |
*** nihilifer has joined #openstack-nova | 07:58 | |
*** haomaiwang has quit IRC | 08:01 | |
*** haomaiwang has joined #openstack-nova | 08:01 | |
*** pbrooko_ has joined #openstack-nova | 08:01 | |
*** rotbeard has joined #openstack-nova | 08:02 | |
*** pbrooko has quit IRC | 08:04 | |
*** sayali has joined #openstack-nova | 08:04 | |
*** pbrooko_ has quit IRC | 08:07 | |
*** derekh has joined #openstack-nova | 08:08 | |
openstackgerrit | John Garbutt proposed openstack/nova: docs: add the concept of Feature Classification https://review.openstack.org/215664 | 08:11 |
*** asettle has quit IRC | 08:11 | |
*** RA_ has quit IRC | 08:12 | |
openstackgerrit | John Garbutt proposed openstack/nova: devref: update the nova architecture doc https://review.openstack.org/177374 | 08:12 |
*** otter768 has joined #openstack-nova | 08:12 | |
*** angular_mike_ has joined #openstack-nova | 08:12 | |
openstackgerrit | John Garbutt proposed openstack/nova: devref: add a description of flavors https://review.openstack.org/178628 | 08:12 |
*** yassine_ has joined #openstack-nova | 08:13 | |
openstackgerrit | John Garbutt proposed openstack/nova: devref: discuss current state of error handling https://review.openstack.org/178635 | 08:13 |
eliqiao | hi jhonthetubaguy | 08:13 |
openstackgerrit | sahid proposed openstack/nova: libvirt: verify cpu bw policy capability for host https://review.openstack.org/224726 | 08:14 |
openstackgerrit | John Garbutt proposed openstack/nova: docs: add the scheduler evolution plans https://review.openstack.org/192260 | 08:14 |
openstackgerrit | John Garbutt proposed openstack/nova: docs: split up upgrade and upgrade futures https://review.openstack.org/201529 | 08:15 |
openstackgerrit | John Garbutt proposed openstack/nova: Add a summary of the architecture evolution work https://review.openstack.org/200001 | 08:16 |
*** otter768 has quit IRC | 08:17 | |
*** DuncanT has joined #openstack-nova | 08:19 | |
*** romainh has joined #openstack-nova | 08:19 | |
openstackgerrit | Andrea Rosa proposed openstack/nova-specs: Allow more instance operations during live migration https://review.openstack.org/226199 | 08:19 |
*** sahid has quit IRC | 08:21 | |
*** sahid has joined #openstack-nova | 08:21 | |
*** sahid has quit IRC | 08:22 | |
*** sahid has joined #openstack-nova | 08:22 | |
*** sayali has quit IRC | 08:23 | |
rajesht_ | johnthetubaguy: replied to your comment on https://review.openstack.org/#/c/215483/ | 08:25 |
rajesht_ | johnthetubaguy: thanks for review. :) | 08:26 |
*** ndipanov has joined #openstack-nova | 08:26 | |
*** sayali has joined #openstack-nova | 08:27 | |
*** takedakn has quit IRC | 08:27 | |
*** Marga_ has joined #openstack-nova | 08:27 | |
*** eglynn_ has joined #openstack-nova | 08:28 | |
*** pece has joined #openstack-nova | 08:28 | |
johnthetubaguy | rajesht_: your git commit message doesn't make your intent clear, if you are tying to make sure that look runs for live-migrate and migrate | 08:28 |
*** Marga_ has quit IRC | 08:28 | |
*** Marga_ has joined #openstack-nova | 08:29 | |
johnthetubaguy | rajesht_: when I read the git commit message it seems to just say "tidying up the code so its consistent", rather than, its fixing this bug for when XYZ happens | 08:29 |
johnthetubaguy | rajesht_: can you update the git commit message please? | 08:29 |
*** karimb has joined #openstack-nova | 08:29 | |
tdurakov | johnthetubaguy, hi, do you have some time to review https://review.openstack.org/#/c/225910/ | 08:30 |
eliqiao | hi johnthetubaguy: | 08:30 |
eliqiao | can you talk about sanitize_hostname? | 08:30 |
rajesht_ | johnthetubaguy: thanks for your review. It would be nice if you comment that on patch | 08:31 |
johnthetubaguy | tdurakov: sorry, don't have RC1 out yet, then I have to review the other ones that have been waiting | 08:31 |
johnthetubaguy | rajesht_: I can do if you want | 08:31 |
rajesht_ | johnthetubaguy: sure, it would be great | 08:32 |
rajesht_ | johnthetubaguy: thanks : | 08:32 |
*** lucasagomes has joined #openstack-nova | 08:32 | |
tdurakov | johnthetubaguy, ok, np. I hope we could merge that until spec freeze. | 08:32 |
*** alexpilotti has joined #openstack-nova | 08:33 | |
*** romainh has quit IRC | 08:34 | |
johnthetubaguy | tdurakov: I haven't started spec reviews yet, need to get liberty out the door right now, still got 45 bug pathes with no review yet :( | 08:34 |
johnthetubaguy | well, ish | 08:34 |
johnthetubaguy | eliqiao: sure | 08:34 |
*** pbrooko has joined #openstack-nova | 08:36 | |
*** pbrooko has quit IRC | 08:36 | |
eliqiao | johnthetubaguy: it's re https://review.openstack.org/#/c/224438/ , alex_xu feels we still need more test cases. | 08:37 |
eliqiao | such as " a b c " and "a" * 256 | 08:37 |
eliqiao | I wonder if api will restrict it already. | 08:38 |
openstackgerrit | John Garbutt proposed openstack/nova: rebuild: RPC sends additional args and claims are done https://review.openstack.org/200485 | 08:42 |
johnthetubaguy | eliqiao: v2.0 compat mode doesn't restrict that, for sure | 08:42 |
johnthetubaguy | ndipanov: fixed the pep8 gremlin in that patch for you ^ | 08:43 |
johnthetubaguy | just reading through it now | 08:43 |
eliqiao | johnthetubaguy: okay, for "a" * 256 , I am perfer to truncate it to 255, what do you think. | 08:44 |
johnthetubaguy | eliqiao: whats the alternative? | 08:44 |
johnthetubaguy | eliqiao: actually, you are right, the API will not let a * 256 in, oops | 08:45 |
johnthetubaguy | now I think thats consistent... but I would have to go double check the code | 08:45 |
*** romainh has joined #openstack-nova | 08:46 | |
openstackgerrit | Claudiu Belu proposed openstack/nova: WIP: Hyper-V: adds os-win library https://review.openstack.org/212102 | 08:46 |
*** aix has quit IRC | 08:46 | |
*** danpb has joined #openstack-nova | 08:46 | |
eliqiao | johnthetubaguy: alex_xu told raise a exception in that case. | 08:47 |
*** alexpilotti has quit IRC | 08:47 | |
ndipanov | johnthetubaguy, there was another issue | 08:47 |
ndipanov | I am uploading a new change now | 08:47 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: rebuild: RPC sends additional args and claims are done https://review.openstack.org/200485 | 08:48 |
ndipanov | johnthetubaguy, ^ | 08:48 |
ndipanov | that should be it | 08:49 |
ndipanov | johnthetubaguy, also note https://bugs.launchpad.net/bugs/1498126 that I reported yesterday after digging into the problem alex_xu spotted | 08:49 |
openstack | Launchpad bug 1498126 in OpenStack Compute (nova) "Inconsistencies with resource tracking in the case of resize operation." [Undecided,New] | 08:49 |
ndipanov | no need to block RC on it though | 08:50 |
johnthetubaguy | ndipanov: good point, that should be backportable I guess? | 08:50 |
johnthetubaguy | yeah | 08:50 |
*** lucas-dinner has joined #openstack-nova | 08:50 | |
ndipanov | most of it | 08:51 |
ndipanov | but it really requires careful analysis | 08:51 |
johnthetubaguy | yeah, +1 that | 08:51 |
ndipanov | I am currently trying to fugure out a way to write a test that catches these | 08:51 |
johnthetubaguy | ndipanov: looking at your patch, I think the self._error_out_instance_on_exception(context, instance) will save the instance and record the host to the DB | 08:51 |
ndipanov | but we never set it before the else | 08:52 |
ndipanov | now | 08:52 |
ndipanov | after the last update | 08:52 |
ndipanov | (I spotted that too) | 08:52 |
ndipanov | that's what I was fixing | 08:52 |
ndipanov | johnthetubaguy, ^ | 08:52 |
johnthetubaguy | ndipanov: yeah, I think the twisted folks have some cool tools for that, not so sure about for eventlet, might need to inject something horrible | 08:52 |
*** yuntongjin has quit IRC | 08:52 | |
johnthetubaguy | ndipanov: yeah, I think it works perfectly, really just checking to make sure thats intentional | 08:53 |
johnthetubaguy | the save on error bit | 08:53 |
johnthetubaguy | the delayed save makes sense, and I like the extra note | 08:53 |
*** xek has joined #openstack-nova | 08:54 | |
*** inara has quit IRC | 08:56 | |
*** haomaiwang has quit IRC | 09:01 | |
*** haomaiwang has joined #openstack-nova | 09:01 | |
*** inara has joined #openstack-nova | 09:02 | |
*** lpetrut has joined #openstack-nova | 09:03 | |
*** rdopiera has quit IRC | 09:05 | |
*** Marga_ has quit IRC | 09:05 | |
*** rdopiera has joined #openstack-nova | 09:06 | |
*** alex_klimov has joined #openstack-nova | 09:06 | |
bauzas | morning Nova | 09:06 |
*** abhishekk has joined #openstack-nova | 09:06 | |
johnthetubaguy | ndipanov: so I wonder if you latest one needs to set the host when there is an error, admins usually expect the host to be where the mess might be? | 09:09 |
*** alex_klimov has quit IRC | 09:09 | |
*** alex_klimov has joined #openstack-nova | 09:10 | |
ndipanov | johnthetubaguy, that's a good question... | 09:10 |
ndipanov | so in case of failing for resource reasons - that won't work definitely | 09:11 |
ndipanov | so that except: branch is fine | 09:11 |
ndipanov | and for general case... well the reason we introduced this whole migration thing is to give admin (monitoring tool) a way to know what's going on without guessing | 09:12 |
ndipanov | so a failed migration and instance on the old host made more sense to me | 09:12 |
ndipanov | the case we don't cover now is " I tried to evacuate to a host and it too died" but nova generally can't handle these well | 09:13 |
ndipanov | and dansmith has some nice ideas how to fix it maybe | 09:13 |
ndipanov | johnthetubaguy, ^ | 09:13 |
johnthetubaguy | ndipanov: yep, we sure have the migration record now, which does give them all the details | 09:15 |
openstackgerrit | Eli Qiao proposed openstack/nova: Give instance default hostname if hostname is empty https://review.openstack.org/224438 | 09:16 |
johnthetubaguy | ndipanov: just worried about a behaviour change for no reason, but screwing up the resource counting is a good reason to not do that | 09:16 |
*** aix has joined #openstack-nova | 09:16 | |
*** eliqiao has left #openstack-nova | 09:16 | |
* alex_xu busy on prepare api meeting...really hope to join the discussion about rebuild claim | 09:18 | |
*** xek_ has joined #openstack-nova | 09:18 | |
*** lalit has quit IRC | 09:18 | |
ndipanov | johnthetubaguy, well previously the update was made half way too so really one could not rely on it being set | 09:19 |
johnthetubaguy | ndipanov: sorry to keep bugging you, just looking at line 2708: https://review.openstack.org/#/c/200485/27/nova/compute/manager.py,cm I am thinking we don't want to do that update in the case where scheduled_node is None | 09:19 |
*** _gryf has joined #openstack-nova | 09:19 | |
johnthetubaguy | ndipanov: true, although it was before the stuff thats more likely to fail, but I like setting it later, its simpler | 09:19 |
johnthetubaguy | alex_xu: would love your eyes on the patch again, I think the issues you spotted have been fixed now | 09:20 |
*** shuquan has joined #openstack-nova | 09:21 | |
alex_xu | johnthetubaguy: I plan to review again after prepare api meeting | 09:21 |
*** ociuhandu has quit IRC | 09:21 | |
*** xek has quit IRC | 09:21 | |
*** wanghua has joined #openstack-nova | 09:24 | |
ndipanov | johnthetubaguy, responded to your comments | 09:24 |
ndipanov | (some of them at least | 09:24 |
*** boris-42 has joined #openstack-nova | 09:26 | |
ndipanov | alex_xu, thanks | 09:26 |
*** e0ne has joined #openstack-nova | 09:26 | |
alex_xu | ndipanov: you are welcome :) | 09:26 |
*** lucas-dinner has quit IRC | 09:41 | |
*** lucasagomes has quit IRC | 09:41 | |
*** lucasagomes_ has joined #openstack-nova | 09:41 | |
*** lucasagomes_ has quit IRC | 09:41 | |
*** lucasagomes has joined #openstack-nova | 09:41 | |
*** ociuhandu has joined #openstack-nova | 09:43 | |
alex_xu | johnthetubaguy: May I ask a stupid question, what is concept guide should looks like? Like this one http://docs.openstack.org/admin-guide-cloud/identity_concepts.html ? | 09:43 |
alex_xu | johnthetubaguy: if the concept guide, just a set of explain for some concept, I'm try to list all the concepts https://etherpad.openstack.org/p/nova-api-concept-guide then hope the api team can work on it | 09:44 |
*** xek_ is now known as xek | 09:45 | |
*** ociuhandu_ has joined #openstack-nova | 09:45 | |
alex_xu | sdague isn't here, not sure when he back...I'm thinking how to make progress on this | 09:45 |
johnthetubaguy | alex_xu: we used to have one for compute already | 09:45 |
*** lucasagomes_ has joined #openstack-nova | 09:45 | |
*** shuquan has quit IRC | 09:45 | |
*** kumar has quit IRC | 09:46 | |
alex_xu | johnthetubaguy: it isn't this one right? https://github.com/openstack/nova/blob/master/doc/source/v2/2.0_server_concepts.rst | 09:46 |
johnthetubaguy | alex_xu: I guess it is | 09:47 |
*** shuquan has joined #openstack-nova | 09:47 | |
*** dims has joined #openstack-nova | 09:47 | |
*** ociuhandu has quit IRC | 09:47 | |
*** ociuhandu_ is now known as ociuhandu | 09:47 | |
alex_xu | johnthetubaguy: sdague said it is old and wrong | 09:47 |
johnthetubaguy | alex_xu: yeah, it is | 09:47 |
johnthetubaguy | alex_xu: honestly, a guide that describes v2.1 micro versions, how they work, what they are for, how they should be used, is the top priority | 09:48 |
johnthetubaguy | alex_xu: and describing v2.0 vs v2.1, etc | 09:48 |
johnthetubaguy | alex_xu: after that we should probably describe all the top level resources in the API, like /servers /keypairs, etc | 09:49 |
*** marcusvrn_ has joined #openstack-nova | 09:49 | |
*** erlon has joined #openstack-nova | 09:49 | |
johnthetubaguy | alex_xu: I think thats the main bits | 09:49 |
johnthetubaguy | alex_xu: this is what I remember: https://github.com/openstack/nova/blob/master/doc/source/v2/2.0_general_info.rst#concepts | 09:50 |
johnthetubaguy | alex_xu: its about three years out of date, looking at it | 09:50 |
alex_xu | johnthetubaguy: we have devref for microversion https://github.com/openstack/nova/blob/master/doc/source/api_microversion_dev.rst | 09:50 |
*** njirap has joined #openstack-nova | 09:51 | |
alex_xu | johnthetubaguy: ok, thanks, I probably understand what is looks like now. let me try to structure the | 09:51 |
johnthetubaguy | alex_xu: yeah, its a reasonable starting point, but we need to change the audience from developers to API users | 09:51 |
alex_xu | johnthetubaguy: yea, got it | 09:52 |
johnthetubaguy | alex_xu: so maybe I should take a quick go at putting some structure in place, would that help? just do a WIP patch to rough out some ideas? | 09:52 |
*** lucasagomes has quit IRC | 09:52 | |
*** lucasagomes_ is now known as lucasagomes | 09:52 | |
johnthetubaguy | alex_xu: do we know what state this stuff is in? http://developer.openstack.org/api-ref-compute-v2.1.html | 09:52 |
alex_xu | johnthetubaguy: yea, that is helpful, or a patch wrote done the structure, then other contributes can work on it | 09:52 |
*** lucasagomes_ has joined #openstack-nova | 09:53 | |
*** ameade has joined #openstack-nova | 09:53 | |
johnthetubaguy | alex_xu: you had a list of missing extensions I think? | 09:53 |
alex_xu | johnthetubaguy: emm...not clear that, just know doc team contributer work on it | 09:53 |
alex_xu | johnthetubaguy: yea | 09:53 |
johnthetubaguy | alex_xu: ah, so we need to reach out to them | 09:53 |
*** dougwig has joined #openstack-nova | 09:53 | |
*** lucasagomes_ has quit IRC | 09:53 | |
*** IanGovett has joined #openstack-nova | 09:54 | |
*** serverascode has joined #openstack-nova | 09:54 | |
johnthetubaguy | alex_xu: so I think in that doc: http://developer.openstack.org/api-ref-compute-v2.1.html we need to add links to the microversion explanation | 09:54 |
*** zhiyan has joined #openstack-nova | 09:54 | |
johnthetubaguy | alex_xu: and ideally add lots of stub entries for the things that are missing, but the docs folks may not agree with that last bit! | 09:54 |
alex_xu | johnthetubaguy: I remember sdague ask some help to Anne about doc infrastruecture | 09:55 |
*** andrewbogott has joined #openstack-nova | 09:55 | |
*** scottda has joined #openstack-nova | 09:55 | |
johnthetubaguy | yeah, she will most likely drop into our meeting, if not I can catch up with her in a little bit | 09:55 |
*** wanghao has joined #openstack-nova | 09:56 | |
*** salv-orlando has quit IRC | 09:56 | |
*** pece has quit IRC | 09:56 | |
*** BadCub has joined #openstack-nova | 09:56 | |
*** rex_lee_ has joined #openstack-nova | 09:56 | |
johnthetubaguy | alex_xu: do you know if we publish the general compute API v2.0 stuff anywhere at the moment? | 09:56 |
*** pece has joined #openstack-nova | 09:57 | |
openstackgerrit | Vladyslav Drok proposed openstack/nova: Pass kernel id and ramdisk id to S3 image service https://review.openstack.org/226229 | 09:57 |
alex_xu | johnthetubaguy: no, I didn't find that | 09:57 |
johnthetubaguy | alex_xu: found it: http://docs.openstack.org/developer/nova/v2/index.html | 09:57 |
*** IanGovett1 has joined #openstack-nova | 09:57 | |
johnthetubaguy | although it claims to be a "local copy" | 09:57 |
*** markvoelker has quit IRC | 09:57 | |
bauzas | johnthetubaguy: btw. I discussed with mriedem_away about something and I added https://review.openstack.org/#/c/223802/ | 09:58 |
bauzas | (re: some devref for AZs) | 09:58 |
alex_xu | johnthetubaguy: looks like we have huge doc works :( | 09:58 |
johnthetubaguy | alex_xu: we sure do! | 09:58 |
*** IanGovett has quit IRC | 09:58 | |
johnthetubaguy | alex_xu: look on the brightside, lots of room for improvement :) | 09:58 |
alex_xu | johnthetubaguy: yea, too many debt | 09:59 |
johnthetubaguy | alex_xu: yeah, we are identifying and paying some of it down now, there is always that discovery curve, where you keep looking under a little rock and finding more broken stuff | 09:59 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: rebuild: RPC sends additional args and claims are done https://review.openstack.org/200485 | 09:59 |
*** sdague has joined #openstack-nova | 10:00 | |
alex_xu | johnthetubaguy: sorry, leave for a moment, family waiting for me dinner | 10:00 |
*** haomaiwang has quit IRC | 10:01 | |
alex_xu | johnthetubaguy: let's continue discussion that in api meeting later | 10:01 |
*** haomaiwang has joined #openstack-nova | 10:01 | |
johnthetubaguy | alex_xu: you should totally join them, and +1 to that | 10:01 |
bauzas | johnthetubaguy: https://review.openstack.org/#/c/178635/ is having a merge issue, weirdo | 10:03 |
*** shuquan has quit IRC | 10:04 | |
*** patrickeast has joined #openstack-nova | 10:07 | |
*** RA_ has joined #openstack-nova | 10:08 | |
*** ptm_ has joined #openstack-nova | 10:09 | |
*** romainh has left #openstack-nova | 10:10 | |
*** Sree has quit IRC | 10:10 | |
*** shuquan has joined #openstack-nova | 10:11 | |
*** otter768 has joined #openstack-nova | 10:13 | |
johnthetubaguy | bauzas: yeah, needs a tone of re-work, just leaving that for now, more worried about the API docs | 10:14 |
*** shyu has quit IRC | 10:14 | |
bauzas | johnthetubaguy: no worries, just wondered why it got a merge issue | 10:14 |
bauzas | johnthetubaguy: I'm just reworking the Tasks manifesto, it sounds quite related to your error-handling devref | 10:15 |
johnthetubaguy | bauzas: usually the index page | 10:15 |
openstackgerrit | Lucas Alvares Gomes proposed openstack/nova: Ironic: Workaround to mitigate bug #1341420 https://review.openstack.org/226235 | 10:15 |
openstack | bug 1341420 in OpenStack Compute (nova) "gap between scheduler selection and claim causes spurious failures when the instance is the last one to fit" [High,Confirmed] https://launchpad.net/bugs/1341420 | 10:15 |
johnthetubaguy | bauzas: it is yeah | 10:15 |
johnthetubaguy | bauzas: feel free to take over my patch, if thats helpful :) | 10:15 |
johnthetubaguy | bauzas: also feel free to create a new one, if thats easier! | 10:16 |
bauzas | johnthetubaguy: I first need to draw another iteration for the tasks manifesto, but that's doable | 10:16 |
*** otter768 has quit IRC | 10:17 | |
*** gszasz has joined #openstack-nova | 10:18 | |
openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: libvirt: make live_migration_uri flag dependent on virt_type https://review.openstack.org/175780 | 10:19 |
*** mpaolino has joined #openstack-nova | 10:21 | |
*** shuquan has quit IRC | 10:22 | |
bauzas | johnthetubaguy: meh, we broke a weigher when merging https://review.openstack.org/#/c/213317/15/nova/scheduler/host_manager.py,cm because of https://bugs.launchpad.net/nova/+bug/1493680 | 10:24 |
openstack | Launchpad bug 1493680 in OpenStack Compute (nova) "HostState.metrics is no longer a dict" [Medium,In progress] - Assigned to Yingxin (cyx1231st) | 10:24 |
bauzas | johnthetubaguy: marked the bug as Medium, but I wonder if it would be good for a RC candidate | 10:24 |
bauzas | not exactly for RC1, since we could backport it | 10:25 |
bauzas | johnthetubaguy: if not, metrics are non-functional with Liberty codebase :/ | 10:25 |
johnthetubaguy | bauzas: as a regression, thats a great canditate | 10:25 |
bauzas | johnthetubaguy: adding the tag, and reviewing the change | 10:26 |
johnthetubaguy | bauzas: yep, already added it | 10:26 |
bauzas | johnthetubaguy: okay I did on my own | 10:26 |
bauzas | johnthetubaguy: raising the bar up to High ? | 10:27 |
BobBall | johnthetubaguy: Regarding https://review.openstack.org/#/c/203513/ - I assume the change in behaviour with soft reboot falling back to hard reboot in XenAPI won't cause an unexpected change in behaviour in RAX? | 10:27 |
*** mleroy has left #openstack-nova | 10:28 | |
*** e0ne has quit IRC | 10:30 | |
*** RA_ has quit IRC | 10:31 | |
*** e0ne has joined #openstack-nova | 10:35 | |
*** smatzek has joined #openstack-nova | 10:37 | |
openstackgerrit | Lucas Alvares Gomes proposed openstack/nova: Ironic: Add support for microversion https://review.openstack.org/221848 | 10:43 |
*** salv-orlando has joined #openstack-nova | 10:47 | |
*** romainh has joined #openstack-nova | 10:49 | |
*** aix has quit IRC | 10:53 | |
johnthetubaguy | bauzas: high probably makes sense | 10:54 |
johnthetubaguy | BobBall: nope, makes perfect sense to me, I guess it takes a long with some XenServer versions | 10:55 |
johnthetubaguy | BobBall: the UI folks gave up I think, and now only have a hard shutdown | 10:55 |
BobBall | I see; OK - well as long as it won't cause you guys a hassle then that's fine | 10:55 |
*** sahid has quit IRC | 10:56 | |
*** rdopiera has quit IRC | 10:57 | |
*** rdopiera has joined #openstack-nova | 10:57 | |
*** haomaiwang has quit IRC | 11:01 | |
*** haomaiwang has joined #openstack-nova | 11:01 | |
*** sahid has joined #openstack-nova | 11:03 | |
*** jlanoux has quit IRC | 11:04 | |
*** aix has joined #openstack-nova | 11:05 | |
openstackgerrit | John Garbutt proposed openstack/nova: docs: rerganize the API concept docs https://review.openstack.org/226253 | 11:08 |
alex_xu | ndipanov: just commented on https://review.openstack.org/#/c/200485/, I feel the _move_claim doesn't work for evacuate, but please correct me if I'm wrong. | 11:10 |
*** mpaolino has quit IRC | 11:13 | |
*** _pece has joined #openstack-nova | 11:13 | |
*** pece has quit IRC | 11:14 | |
openstackgerrit | venkatesh proposed openstack/nova: make the driver.Scheduler as abstract class https://review.openstack.org/212016 | 11:15 |
*** baoli has joined #openstack-nova | 11:17 | |
*** baoli_ has joined #openstack-nova | 11:19 | |
sdague | alex_xu: I'm a bit confused about this - https://review.openstack.org/#/c/222032 | 11:21 |
sdague | it's not really clear to me why a magic additional parameter changes things | 11:21 |
alex_xu | sdague: welcome back :) | 11:21 |
sdague | alex_xu: thanks :) | 11:21 |
*** pixelb has joined #openstack-nova | 11:21 | |
*** baoli has quit IRC | 11:22 | |
alex_xu | sdague: it is for line 149 in https://review.openstack.org/#/c/220791/16/nova/api/openstack/compute/availability_zone.py | 11:22 |
*** thorst has joined #openstack-nova | 11:23 | |
sdague | ok, so it's kind of weird to have that in a separate patch, because it's hard to know if those are needed without cross referencing it with the follow on patch | 11:24 |
alex_xu | sdague: yea, but if we put them together, there will be a huge patch, it is hard for review | 11:24 |
sdague | honestly, I think it's harder to review this separately because you loose all the context | 11:25 |
alex_xu | sdague: or we should separate this one https://review.openstack.org/#/c/222032/3 | 11:25 |
alex_xu | sdague: oops, i mean this one https://review.openstack.org/#/c/220791/16 | 11:25 |
sdague | especially because it looks like access ips is changed in the first patch, but not in the second | 11:26 |
*** njirap has quit IRC | 11:27 | |
*** njirap has joined #openstack-nova | 11:27 | |
alex_xu | sdague: ok, got it | 11:28 |
alex_xu | sdague: do you need me merge them? | 11:28 |
*** markvoelker has joined #openstack-nova | 11:28 | |
openstackgerrit | jichenjc proposed openstack/nova: Remove duplicate VALID_NAME_REGEX https://review.openstack.org/218489 | 11:29 |
*** Murali has joined #openstack-nova | 11:30 | |
sdague | well, I'd also like to figure out what oomichi's concerns are with the top patch. Because I'm really surprised by a -2 there | 11:30 |
sdague | johnthetubaguy: you have any more insights there? | 11:30 |
bauzas | markus_z: aloha | 11:31 |
johnthetubaguy | sdague: running to lunch right now, but I think he is worried about v2.1 code getting messy and/or not quite being the same as v2.0, but honestly a bit confused by it too | 11:31 |
johnthetubaguy | alex_xu: sdague: attempted this some doc work on the API concept guide, to try and make it flow better: https://review.openstack.org/226253 | 11:31 |
bauzas | markus_z: I just discovered https://bugs.launchpad.net/nova/+bug/1493680 which was not checked, nor prioritized | 11:31 |
openstack | Launchpad bug 1493680 in OpenStack Compute (nova) "HostState.metrics is no longer a dict" [High,In progress] - Assigned to Yingxin (cyx1231st) | 11:31 |
johnthetubaguy | (and make the gaps more obvious) | 11:31 |
alex_xu | johnthetubaguy: thanks a lot | 11:32 |
sdague | johnthetubaguy: ok, cool, I'll add it to my list. Still processing email from being out. | 11:32 |
bauzas | markus_z: I wonder if we need to review the NEW and IN_PROGRESS bugs to see which ones are regressions | 11:32 |
*** deepthi has quit IRC | 11:32 | |
*** markvoelker has quit IRC | 11:33 | |
*** lucasagomes is now known as lucas-hungry | 11:34 | |
openstackgerrit | jichenjc proposed openstack/nova: devref: change error messages no need microversion https://review.openstack.org/218491 | 11:40 |
raorn | greetings | 11:46 |
raorn | fix for bug https://bugs.launchpad.net/nova/+bug/1160309 is incomplete | 11:46 |
openstack | Launchpad bug 1160309 in tempest "Nova API floating IP error code inconsistent between Nova-Net and Quantum" [High,Fix released] - Assigned to Ala Rezmerita (arezmerita) | 11:46 |
raorn | I have a patch, how do I submit it? | 11:46 |
raorn | I mean, can I make topic "bug/1160309" ? | 11:47 |
raorn | or do I need to submit another bug? | 11:48 |
raorn | s/make/reuse/ | 11:48 |
bauzas | raorn: you should open another bug | 11:48 |
bauzas | raorn: and mention that one | 11:48 |
raorn | OK | 11:48 |
*** _RuiChen has quit IRC | 11:50 | |
*** _RuiChen has joined #openstack-nova | 11:50 | |
*** manas has joined #openstack-nova | 11:52 | |
*** claudiub has joined #openstack-nova | 11:53 | |
*** dave-mccowan has joined #openstack-nova | 11:54 | |
*** haomaiwang has quit IRC | 11:55 | |
openstackgerrit | Alexey I. Froloff proposed openstack/nova: Fix more inconsistency between Nova-Net and Neutron https://review.openstack.org/226274 | 11:55 |
*** manas has quit IRC | 11:55 | |
alex_xu | nova api meeing in 5 mins~ | 11:55 |
*** ociuhandu has quit IRC | 11:58 | |
openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix order of arguments in assertEqual https://review.openstack.org/220142 | 11:59 |
*** dosaboy has quit IRC | 12:00 | |
*** jlanoux has joined #openstack-nova | 12:00 | |
*** kodokuu has joined #openstack-nova | 12:01 | |
*** claudiub|2 has joined #openstack-nova | 12:01 | |
*** tojuvone has joined #openstack-nova | 12:03 | |
*** claudiub has quit IRC | 12:03 | |
*** e0ne has quit IRC | 12:03 | |
*** ociuhandu has joined #openstack-nova | 12:03 | |
*** Sree has joined #openstack-nova | 12:04 | |
*** yuner027 has quit IRC | 12:04 | |
*** sudipto has quit IRC | 12:06 | |
*** atuvenie has joined #openstack-nova | 12:06 | |
*** yuner027 has joined #openstack-nova | 12:06 | |
*** dosaboy has joined #openstack-nova | 12:07 | |
*** alexpilotti has joined #openstack-nova | 12:09 | |
ndipanov | alex_xu, responded to your comment - you are right about _update_from_migration | 12:12 |
ankit_ag | alaski: hi ... | 12:13 |
*** otter768 has joined #openstack-nova | 12:14 | |
*** markvoelker has joined #openstack-nova | 12:18 | |
*** raildo-afk is now known as raildo | 12:19 | |
*** otter768 has quit IRC | 12:19 | |
*** yuner027 has quit IRC | 12:21 | |
*** yuner027 has joined #openstack-nova | 12:21 | |
*** rfolco has joined #openstack-nova | 12:21 | |
*** e0ne has joined #openstack-nova | 12:23 | |
lxsli | tojuvone: ping | 12:24 |
*** bswartz has quit IRC | 12:26 | |
*** lucas-hungry is now known as lucasagomes | 12:27 | |
*** irenab has left #openstack-nova | 12:30 | |
*** burgerk has joined #openstack-nova | 12:31 | |
*** edmondsw has joined #openstack-nova | 12:33 | |
*** RA_ has joined #openstack-nova | 12:37 | |
*** sahid has quit IRC | 12:39 | |
*** leitan has joined #openstack-nova | 12:40 | |
*** achanda has joined #openstack-nova | 12:41 | |
*** mehdi has joined #openstack-nova | 12:41 | |
*** nelsnels_ has quit IRC | 12:41 | |
*** artom has joined #openstack-nova | 12:42 | |
mehdi | hi all dears | 12:42 |
*** achanda has quit IRC | 12:42 | |
*** nelsnelson has joined #openstack-nova | 12:42 | |
*** njirap has quit IRC | 12:42 | |
mehdi | i have a question about hypervisors, actually code on nova/nova/virt/* | 12:43 |
sdague | alex_xu: ok, you going to respin that patch now ? | 12:44 |
alex_xu | sdague: yea, will do | 12:44 |
*** kiran-r has quit IRC | 12:45 | |
*** _kiran_ has joined #openstack-nova | 12:45 | |
*** _kiran_ is now known as kiran-r | 12:45 | |
*** aysyd has joined #openstack-nova | 12:45 | |
*** xek has quit IRC | 12:46 | |
alex_xu | ndipanov: sorry, just back from api meeting | 12:46 |
ndipanov | alex_xu, np - check out my answers on that patch | 12:46 |
ndipanov | fixing the migration issue now | 12:46 |
alex_xu | ndipanov: will do, thanks the fixing | 12:46 |
*** xek has joined #openstack-nova | 12:47 | |
*** mehdi has quit IRC | 12:47 | |
*** alejandrito has joined #openstack-nova | 12:47 | |
*** vladikr has joined #openstack-nova | 12:47 | |
*** mehdi__ has joined #openstack-nova | 12:47 | |
openstackgerrit | jichenjc proposed openstack/nova: Remove duplicate VALID_NAME_REGEX https://review.openstack.org/218489 | 12:48 |
*** boris-42 has quit IRC | 12:49 | |
*** romainh has quit IRC | 12:51 | |
alex_xu | ndipanov: just replied | 12:52 |
*** xyang1 has joined #openstack-nova | 12:53 | |
*** eharney has joined #openstack-nova | 12:54 | |
*** jichen has quit IRC | 12:54 | |
*** romainh has joined #openstack-nova | 12:54 | |
ndipanov | alex_xu, responded back - the context manager takes care of it | 12:55 |
alex_xu | ndipanov: but the code didn't use the context manager, otherwise the code should look like "with rebuild_claim() as claim" at line 2674 | 12:57 |
*** rlrossit has joined #openstack-nova | 12:58 | |
alex_xu | ndipanov: and the 'drop_move_claim()' will drop the migration record, right? it shouldn't drop for failure case | 12:58 |
*** tongli has joined #openstack-nova | 12:58 | |
*** RA_ has quit IRC | 12:58 | |
ndipanov | alex_xu, actually it won't | 12:58 |
ndipanov | it will only drop the usage and the migratio_context record | 12:59 |
ndipanov | we don't really delete migration records at all we just set them to a final state | 12:59 |
openstackgerrit | xing-yang proposed openstack/nova: Rollback is needed if initialize_connection times out https://review.openstack.org/225278 | 12:59 |
alex_xu | ndipanov: oops, you are right, checked the code | 13:00 |
*** haomaiwang has joined #openstack-nova | 13:00 | |
alex_xu | ndipanov: but the code didn't use the context manager, is it right? | 13:00 |
ndipanov | it does now | 13:00 |
alex_xu | ndipanov: ok, cool | 13:00 |
ndipanov | that patch adds it | 13:00 |
*** haomaiwang has quit IRC | 13:01 | |
*** romainh has quit IRC | 13:01 | |
*** haomaiwang has joined #openstack-nova | 13:01 | |
*** ijuwang has joined #openstack-nova | 13:01 | |
alex_xu | ndipanov: you mean next patchset? | 13:01 |
*** mdrabe has joined #openstack-nova | 13:02 | |
ndipanov | no no | 13:03 |
ndipanov | this one | 13:03 |
ndipanov | alex_xu, self._do_rebuild_instance_with_claim acutally does the rebuild in a with block | 13:03 |
ndipanov | so if anything blows up the claim is rolled back as expected | 13:04 |
*** romainh has joined #openstack-nova | 13:04 | |
alex_xu | ndipanov: oh, I got it, thanks a lot | 13:04 |
*** mehdi__ has quit IRC | 13:07 | |
openstackgerrit | Alex Xu proposed openstack/nova: Filter leading/trailing spaces for name field in v2.1 compat mode https://review.openstack.org/222032 | 13:08 |
*** mehdi__ has joined #openstack-nova | 13:08 | |
alex_xu | sdague: ^ just merged the patch | 13:08 |
alaski | ankit_ag: hi | 13:10 |
*** dims has quit IRC | 13:14 | |
mehdi__ | i have just started nova openstack, i 'll actually work on hypervisor features. i intend to find a way that add missing hypervisors features to openstack | 13:14 |
mehdi__ | If I could find any article about architecture of the code specially /nova/nova/virt/* | 13:14 |
*** dims has joined #openstack-nova | 13:15 | |
*** dustins has joined #openstack-nova | 13:15 | |
*** baoli_ has quit IRC | 13:16 | |
*** baoli has joined #openstack-nova | 13:16 | |
*** _kiran_ has joined #openstack-nova | 13:16 | |
bauzas | mehdi__: I guess you know http://docs.openstack.org/developer/nova/support-matrix.html ? | 13:17 |
*** kiran-r has quit IRC | 13:17 | |
*** _kiran_ is now known as kiran-r | 13:17 | |
*** tongli has quit IRC | 13:17 | |
raorn | tdurakov: ayt? | 13:17 |
tdurakov | raorn, hi | 13:18 |
*** salv-orlando has quit IRC | 13:18 | |
raorn | tdurakov: review 226274, os-floating-ips | 13:19 |
tdurakov | raorn, let me check | 13:20 |
*** mehdi_1 has joined #openstack-nova | 13:20 | |
*** mehdi_1 has left #openstack-nova | 13:20 | |
raorn | I think "GET /os-floating-ips" should show empty list even if Neutron's L3 suport is disabled | 13:21 |
*** baoli_ has joined #openstack-nova | 13:21 | |
*** Boopalan has joined #openstack-nova | 13:21 | |
raorn | index() method doesn't even expect 404 error | 13:21 |
*** Boopalan is now known as boopalan | 13:21 | |
raorn | (and I know, I screwed up the tests) | 13:22 |
bauzas | ndipanov: johnthetubaguy: I see https://bugs.launchpad.net/nova/+bug/1417667 being RC potential with https://review.openstack.org/#/c/200485/28 bumping a RPC version, I guess it's why it would be needed for RC1 and not backportable ? | 13:22 |
openstack | Launchpad bug 1417667 in OpenStack Compute (nova) "migration/evacuation/rebuild/resize of instance with dedicated cpus needs to recalculate cpus on destination" [Medium,In progress] - Assigned to Nikola Đipanov (ndipanov) | 13:22 |
*** abhishekk has quit IRC | 13:22 | |
ndipanov | bauzas, I guess so | 13:22 |
ndipanov | I am still working on some updates to the patch that will probably result in one more patch | 13:23 |
bauzas | ndipanov: okay, that makes the change really prioritized | 13:23 |
johnthetubaguy | bauzas: yes, its the RPC bump | 13:23 |
bauzas | ndipanov: okay, fair enough, the change is good but a bit hairy so I'm taking time to review it correctly | 13:24 |
*** ctrath has joined #openstack-nova | 13:24 | |
*** baoli has quit IRC | 13:24 | |
bauzas | johnthetubaguy: when do you plan to call it a RC1 ? :) | 13:24 |
bauzas | today? | 13:24 |
*** mehdi__ has quit IRC | 13:25 | |
bauzas | johnthetubaguy: are you already providing the setup.cfg change for bumping the nova version ? or is it something now done by the releases team ? | 13:25 |
dims | bauzas, release team needs a SHA | 13:26 |
dims | they won't touch our repo | 13:26 |
johnthetubaguy | yeah | 13:26 |
johnthetubaguy | the current plan is wait till Thursday to see if we can get the above bug sorted | 13:26 |
tdurakov | raorn, prior to your patch client.list_floatingips could raise exception. your patch suspend it by returning empty list instead, could it cause any side effects? | 13:26 |
bauzas | dims: okay, so someone needs stiill to provide a change for bumping setup.cfg version ? | 13:26 |
johnthetubaguy | if thats not what we want, then, we can change our plans | 13:26 |
bauzas | johnthetubaguy: okay, ack | 13:27 |
*** VW has joined #openstack-nova | 13:27 | |
johnthetubaguy | yeah, its usually the commit before that becomes the branch, from memory | 13:27 |
*** beagles has quit IRC | 13:27 | |
ankit_ag | alaski: I have replied to your comment on https://review.openstack.org/#/c/194063/ | 13:27 |
*** meh_di has joined #openstack-nova | 13:27 | |
*** b3nt_pin has joined #openstack-nova | 13:28 | |
*** b3nt_pin is now known as Guest13372 | 13:28 | |
*** sudipto has joined #openstack-nova | 13:28 | |
ankit_ag | alaski: Unfortunately its not possible to fix that issue without removing the code you mentioned | 13:28 |
*** jecarey has joined #openstack-nova | 13:29 | |
meh_di | dear all | 13:29 |
*** Guest13372 is now known as beagles | 13:29 | |
alaski | ankit_ag: can you put that code in a patch, and then make the other changes in that review depend on that? | 13:29 |
meh_di | is there anybody how can help me on nova codes | 13:29 |
ankit_ag | alaski: sure, thank you ! | 13:29 |
*** zzzeek has joined #openstack-nova | 13:30 | |
openstackgerrit | John Garbutt proposed openstack/nova: docs: rerganize the API concept docs https://review.openstack.org/226253 | 13:30 |
bauzas | meh_di: (15:17:09) bauzas: mehdi__: I guess you know http://docs.openstack.org/developer/nova/support-matrix.html ? | 13:30 |
*** burgerk has quit IRC | 13:31 | |
*** mriedem_away is now known as mriedem | 13:31 | |
meh_di | i have just studied /nova/nova/virt/libvirt/driver.py but more than 7k lines of codes is quiet complicated to undrestand how it works | 13:32 |
*** karimb has quit IRC | 13:32 | |
meh_di | help | 13:32 |
openstackgerrit | Feodor Tersin proposed openstack/nova: Give names to volumes during block device mapping https://review.openstack.org/213433 | 13:32 |
*** karimb has joined #openstack-nova | 13:33 | |
openstackgerrit | John Garbutt proposed openstack/nova: docs: rerganize the API concept docs https://review.openstack.org/226253 | 13:34 |
*** smatzek has quit IRC | 13:34 | |
mriedem | meh_di: the nova developer docs are a good place to start http://docs.openstack.org/developer/nova/ | 13:35 |
ndipanov | johnthetubaguy, thanks! will update the patch today | 13:36 |
*** e0ne has quit IRC | 13:37 | |
*** boopalan has quit IRC | 13:37 | |
mriedem | meh_di: we can always use help with bugs too: https://bugs.launchpad.net/nova/+bugs | 13:38 |
*** RA_ has joined #openstack-nova | 13:39 | |
*** e0ne has joined #openstack-nova | 13:39 | |
*** e0ne has quit IRC | 13:39 | |
*** e0ne has joined #openstack-nova | 13:40 | |
*** kiran-r has quit IRC | 13:40 | |
mriedem | meh_di: also https://wiki.openstack.org/wiki/Nova/Mentoring | 13:40 |
raorn | tdurakov: nova would return 500 Error on /os-floating-ips. I don't think this is a correct behavior. index() method does not expect any exception to happen\ | 13:41 |
*** ankit_ag has quit IRC | 13:41 | |
*** burt has joined #openstack-nova | 13:41 | |
openstackgerrit | Andrey Kurilin proposed openstack/nova: [Py34] Enable test_legacy_v2_compatible_wrapper https://review.openstack.org/226326 | 13:42 |
*** Daisy has joined #openstack-nova | 13:42 | |
*** claudiub|2 has quit IRC | 13:43 | |
*** deepakcs has quit IRC | 13:43 | |
*** fawadkhaliq has joined #openstack-nova | 13:44 | |
*** atuvenie has quit IRC | 13:44 | |
*** takedakn has joined #openstack-nova | 13:45 | |
*** jaypipes has joined #openstack-nova | 13:46 | |
*** obedmr has joined #openstack-nova | 13:47 | |
*** jaypipes has quit IRC | 13:49 | |
*** jaypipes has joined #openstack-nova | 13:50 | |
*** Piet has joined #openstack-nova | 13:52 | |
*** emagana has joined #openstack-nova | 13:52 | |
*** mriedem is now known as mriedem_meeting | 13:53 | |
*** sudipto has quit IRC | 13:53 | |
*** amotoki has joined #openstack-nova | 13:53 | |
raorn | tdurakov: _get_floating_ips_by_fixed_and_port does exactly same thing | 13:55 |
openstackgerrit | Alexey I. Froloff proposed openstack/nova: Fix more inconsistency between Nova-Net and Neutron https://review.openstack.org/226274 | 13:56 |
*** romainh has quit IRC | 13:58 | |
*** fawadkhaliq has quit IRC | 13:59 | |
*** julien_ZTE has joined #openstack-nova | 13:59 | |
*** karimb has quit IRC | 14:00 | |
*** haomaiwang has quit IRC | 14:01 | |
*** smatzek has joined #openstack-nova | 14:01 | |
*** fawadkhaliq has joined #openstack-nova | 14:01 | |
openstackgerrit | Vladyslav Drok proposed openstack/nova: Pass kernel id and ramdisk id to S3 image service https://review.openstack.org/226229 | 14:01 |
*** haomaiwa_ has joined #openstack-nova | 14:01 | |
*** fawadk has joined #openstack-nova | 14:02 | |
*** EinstCrazy has quit IRC | 14:03 | |
*** zhangjn has quit IRC | 14:03 | |
*** stevemar has joined #openstack-nova | 14:03 | |
*** burgerk has joined #openstack-nova | 14:05 | |
*** fawadkhaliq has quit IRC | 14:05 | |
*** Murali has quit IRC | 14:06 | |
Daisy | johnthetubaguy: are you going to make RC1 today ? | 14:06 |
johnthetubaguy | Daisy: hi, I meant to go find you, thanks for reaching out | 14:06 |
johnthetubaguy | Daisy: basically no | 14:07 |
johnthetubaguy | Daisy: its looking more like Thursday right now :( | 14:07 |
Daisy | ok. So I will wait for another day to create new versions in the translation website. | 14:07 |
johnthetubaguy | Daisy: OK, thank you | 14:08 |
Daisy | When you make stable/liberty, I need to create the corresponding version in the translation website. I don't want to miss the time. | 14:08 |
Daisy | So when you are going to do, let me know. | 14:08 |
*** irenab_ has joined #openstack-nova | 14:08 | |
*** fawadkhaliq has joined #openstack-nova | 14:09 | |
*** breitz has quit IRC | 14:09 | |
*** breitz has joined #openstack-nova | 14:10 | |
*** irenab_ is now known as irenab | 14:11 | |
*** fawadk has quit IRC | 14:11 | |
openstackgerrit | John Garbutt proposed openstack/nova: Open Mitaka development https://review.openstack.org/226340 | 14:12 |
*** thangp has joined #openstack-nova | 14:12 | |
*** inara has quit IRC | 14:12 | |
*** browne has joined #openstack-nova | 14:13 | |
johnthetubaguy | Daisy: will do, thank you | 14:13 |
Daisy | thank you | 14:13 |
*** exploreshaifali has joined #openstack-nova | 14:14 | |
*** otter768 has joined #openstack-nova | 14:15 | |
*** inara has joined #openstack-nova | 14:18 | |
*** otter768 has quit IRC | 14:20 | |
openstackgerrit | Vladyslav Drok proposed openstack/nova: Pass kernel id and ramdisk id to S3 image service https://review.openstack.org/226229 | 14:21 |
*** jhesketh has quit IRC | 14:25 | |
openstackgerrit | Andrey Kurilin proposed openstack/nova: [Py34] Enable nova.tests.unit.api.test_auth https://review.openstack.org/226346 | 14:27 |
*** jhesketh has joined #openstack-nova | 14:28 | |
*** alexpilotti has quit IRC | 14:30 | |
*** boris-42 has joined #openstack-nova | 14:30 | |
*** zhangjn has joined #openstack-nova | 14:30 | |
*** zhangjn has quit IRC | 14:30 | |
*** atuvenie has joined #openstack-nova | 14:30 | |
*** VW_ has joined #openstack-nova | 14:31 | |
*** zhangjn has joined #openstack-nova | 14:31 | |
*** zhangjn has quit IRC | 14:31 | |
*** zhangjn has joined #openstack-nova | 14:32 | |
*** zhangjn has quit IRC | 14:32 | |
*** sahid has joined #openstack-nova | 14:34 | |
*** VW__ has joined #openstack-nova | 14:34 | |
*** VW_ has quit IRC | 14:34 | |
*** VW__ has quit IRC | 14:34 | |
*** VW_ has joined #openstack-nova | 14:34 | |
*** VW has quit IRC | 14:35 | |
*** romainh has joined #openstack-nova | 14:35 | |
*** electrocucaracha has joined #openstack-nova | 14:36 | |
*** Marga_ has joined #openstack-nova | 14:37 | |
*** zhangjn has joined #openstack-nova | 14:38 | |
*** zhangjn has quit IRC | 14:38 | |
*** zhangjn has joined #openstack-nova | 14:39 | |
*** zhangjn has quit IRC | 14:39 | |
*** fawadkhaliq has quit IRC | 14:39 | |
*** zhangjn has joined #openstack-nova | 14:41 | |
*** zhangjn has quit IRC | 14:41 | |
*** mtanino has joined #openstack-nova | 14:41 | |
*** zhangjn has joined #openstack-nova | 14:42 | |
*** zhangjn has quit IRC | 14:42 | |
*** cfriesen__ has joined #openstack-nova | 14:42 | |
*** zhangjn has joined #openstack-nova | 14:43 | |
*** exploreshaifali has quit IRC | 14:43 | |
*** zhangjn has quit IRC | 14:43 | |
*** zhangjn has joined #openstack-nova | 14:44 | |
*** ssurana has joined #openstack-nova | 14:44 | |
openstackgerrit | Alexis Lee proposed openstack/nova-specs: Add list-server-invalid-option-as-error.rst https://review.openstack.org/219755 | 14:45 |
*** oomichi has joined #openstack-nova | 14:46 | |
oomichi | alex_xu: Nihao | 14:47 |
oomichi | alex_xu: sorry for not-joining today meeting | 14:47 |
*** x3k has joined #openstack-nova | 14:48 | |
*** mdbooth has quit IRC | 14:49 | |
johnthetubaguy | oomichi: hey, I promised to ask about the -2 | 14:50 |
johnthetubaguy | you probably saw from the meeting notes | 14:50 |
*** tsekiyama has joined #openstack-nova | 14:51 | |
oomichi | johnthetubaguy: good timing, I'd like to know the situation. | 14:51 |
*** tsekiyam_ has joined #openstack-nova | 14:52 | |
oomichi | johnthetubaguy: https://review.openstack.org/#/c/222032/ and https://review.openstack.org/#/c/220791/ seems the same, right? | 14:52 |
*** emagana has quit IRC | 14:52 | |
oomichi | johnthetubaguy: and my -2's https://review.openstack.org/#/c/220791/ gets old, because https://review.openstack.org/#/c/222032/ seems newer | 14:52 |
johnthetubaguy | oomichi: basically squashed the two reviews together, so its easier to review | 14:52 |
sdague | oomichi we didn't understand your -2, and it seemed like it was contingent on an alternate approach, which doesn't exist | 14:53 |
oomichi | johnthetubaguy: humm, almost the same | 14:54 |
*** kodokuu has quit IRC | 14:54 | |
sdague | oomichi: yes, it's the same code | 14:54 |
oomichi | sdague: -2 because the patch adds different api behaviors also from legacy v2 code, even if that is consistent behavior in v2 compatible mode | 14:55 |
*** mdbooth has joined #openstack-nova | 14:55 | |
*** tsekiyama has quit IRC | 14:56 | |
oomichi | sdague: I have updated alex_xu's patch and I'd like to ask alex_xu that can I update it | 14:56 |
Guest24378 | ndipanov: I'm a little lost in the comments with alex_xu because he still has a -1 on your patch | 14:56 |
*** Guest24378 is now known as dansmith | 14:56 | |
*** angdraug has joined #openstack-nova | 14:56 | |
oomichi | sdague: for avoiding different api behaviors | 14:56 |
dansmith | ndipanov: is that still valid? | 14:56 |
sdague | oomichi: ok, let's get that pushed then and look | 14:57 |
oomichi | sdague: thanks | 14:57 |
dansmith | ndipanov: I got a little distracted yesterday while trying to figure out the new save ordering you changed, and thinking about how unfortunately repetitive it is with rpc calls | 14:57 |
ndipanov | dansmith, yep he's right on one of the points | 14:57 |
dansmith | ndipanov: but not sure there's anything we can do about it right now | 14:57 |
dansmith | ndipanov: okay | 14:57 |
ndipanov | it's an easy fix but I am not trying to make sure it's properly tested | 14:57 |
dansmith | not trying? | 14:58 |
ndipanov | because the thing he noticed should have been caught by tests | 14:58 |
ndipanov | now | 14:58 |
ndipanov | lol | 14:58 |
dansmith | ah, better :D | 14:58 |
ndipanov | what do you mean repetitive | 14:58 |
*** emagana has joined #openstack-nova | 14:58 | |
ndipanov | well | 14:58 |
ndipanov | it's 2 different objects | 14:58 |
*** ctrath has quit IRC | 14:59 | |
*** eharney has quit IRC | 14:59 | |
ndipanov | unless we make a bulk save rpc call to the conductor which is basically going back full circle :) | 14:59 |
ndipanov | it has to be at least 2 saves | 14:59 |
ndipanov | dansmith, ^ | 14:59 |
dansmith | ndipanov: if you did the migration context destroy as part of the save | 14:59 |
ndipanov | ah that | 14:59 |
dansmith | instance save, that is | 14:59 |
ndipanov | yeah that one | 14:59 |
ndipanov | I am thinking now that I had the reason to do it this way... | 15:00 |
ndipanov | can't remember | 15:00 |
dansmith | tbh, I thought it was "because time is tight, we can fix it later" | 15:00 |
*** haomaiwa_ has quit IRC | 15:01 | |
ndipanov | might have been | 15:01 |
bauzas | ndipanov: what do you mean by removing the lookup on https://review.openstack.org/#/c/200485/28/nova/conductor/manager.py,cm ? | 15:01 |
openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Allow leading/trailing spaces for name field in v2.1 compat mode https://review.openstack.org/222032 | 15:01 |
bauzas | ndipanov: I guess we still need to get the migration object, right? | 15:01 |
ndipanov | bauzas, yes | 15:02 |
*** nkrinner has quit IRC | 15:02 | |
*** haomaiwa_ has joined #openstack-nova | 15:02 | |
bauzas | ndipanov: okay, but then what do you want to remove, sorry if I'm seeing confused | 15:02 |
ndipanov | actually I don't think that cmment is clear | 15:02 |
ndipanov | so provided that we add migration to the conductor rpc api for rebuild | 15:03 |
ndipanov | in 3.0 | 15:03 |
ndipanov | then we can remove that check in 4.0 | 15:03 |
ndipanov | maybe I should remove that comment | 15:03 |
ndipanov | I am respinning this anyway so | 15:03 |
bauzas | ndipanov: yes please then | 15:03 |
*** eglynn_ has quit IRC | 15:03 | |
bauzas | ndipanov: because you wrote the correct compatibility check in the RPC API | 15:04 |
bauzas | ah, the conductor RPC API... | 15:04 |
*** takedakn has quit IRC | 15:05 | |
*** thorst has quit IRC | 15:06 | |
openstackgerrit | Andrew Laski proposed openstack/nova: Cells: Limit instances pulled in _heal_instances https://review.openstack.org/176925 | 15:06 |
*** mriedem_meeting is now known as mriedem | 15:06 | |
ndipanov | bauzas, yes so once we are passing in the migration AND there is no API node that is not sending it - we can drop the lookup | 15:06 |
ndipanov | as it will always be in the args | 15:07 |
*** armax has joined #openstack-nova | 15:07 | |
*** kiranr has joined #openstack-nova | 15:07 | |
johnthetubaguy | oomichi: is the man thing you dropped the stripping of the trailing whitespace? | 15:08 |
bauzas | ndipanov: you mean provided you pass the migration object from the API ? | 15:08 |
ndipanov | y es which we are not doing yet so the comment is confusing | 15:09 |
bauzas | ndipanov: gotcha, I see now | 15:09 |
bauzas | ndipanov: okay, you want to move that lookup to the compute API, gotcha | 15:09 |
*** ZZelle has quit IRC | 15:09 | |
bauzas | I mean, even creating it and passing it | 15:09 |
*** ctrath has joined #openstack-nova | 15:09 | |
oomichi | johnthetubaguy: I dropped strip() calls from the APIs which didn't strip() on legacy v2 API | 15:09 |
ndipanov | bauzas, no - it's created there | 15:09 |
ndipanov | there will be no lookup | 15:10 |
*** ZZelle has joined #openstack-nova | 15:10 | |
bauzas | ndipanov: yeah, sorry, my last sentence :) | 15:10 |
johnthetubaguy | oomichi: could we not just add the strip call to both legacy and v2.1? | 15:10 |
oomichi | johnthetubaguy: if we get a consensus on API-WG or some places as a rule, we can do it based on concrete manner. | 15:11 |
bauzas | ndipanov: thanks for your explanations | 15:11 |
tdurakov | raorn, left comment in gerrit | 15:12 |
*** dave-mccowan has quit IRC | 15:12 | |
johnthetubaguy | oomichi: the consensus should be to not allow them, as we do in v2.1 | 15:12 |
*** kiran-r has joined #openstack-nova | 15:12 | |
tdurakov | markus_z, hey) | 15:12 |
johnthetubaguy | oomichi: this is a messier case really, I thought we had consensus in nova to do the strip everywhere? | 15:12 |
markus_z | tdurakov: Hey ho | 15:12 |
tdurakov | seen your comment for spec, thanks, need some time to answer them | 15:13 |
*** schwicke has joined #openstack-nova | 15:13 | |
oomichi | johnthetubaguy: I just want to avoid pendulum when unexpected users complain it | 15:13 |
markus_z | tdurakov: sure, take your time. | 15:13 |
markus_z | tdurakov: Thanks for considering. | 15:13 |
ndipanov | bauzas, np thanks for looking into it | 15:14 |
tdurakov | markus_z, :) | 15:14 |
johnthetubaguy | oomichi: this just feels like a bug to me, I guess | 15:14 |
*** EinstCrazy has joined #openstack-nova | 15:14 | |
*** annegentle has joined #openstack-nova | 15:14 | |
*** schwicke_ has joined #openstack-nova | 15:14 | |
*** schwicke__ has joined #openstack-nova | 15:14 | |
*** schwicke__ has quit IRC | 15:14 | |
*** schwicke_ has quit IRC | 15:14 | |
*** schwicke has quit IRC | 15:14 | |
*** meh_di has left #openstack-nova | 15:16 | |
oomichi | johnthetubaguy: yeah, maybe you are right. but we faced backwards incompatible issues when I did think the change make Nova better. | 15:17 |
oomichi | johnthetubaguy: so I just wanted to avoid doing it again :( | 15:17 |
johnthetubaguy | oomichi: the key thing in my head is about existing API calls succeeding | 15:17 |
johnthetubaguy | oomichi: I mean API calls that previous succeed, continue to succeed | 15:18 |
raorn | tdurakov: replied | 15:18 |
oomichi | johnthetubaguy: let me pick up serious example | 15:19 |
johnthetubaguy | oomichi: the edge case we break is if the user adds a space at the end or the beginning, and keeps using that name to do other look up operations, but that seems like a case we can ignore to gain quite a lot more sanity | 15:19 |
oomichi | johnthetubaguy: $ nova keypair-add "def foo()" | 15:19 |
oomichi | johnthetubaguy: $ nova keypair-add " print(‘foo’)” | 15:19 |
tdurakov | raorn, can't find any | 15:19 |
oomichi | johnthetubaguy: $ nova keypair-add “foo()” | 15:19 |
oomichi | johnthetubaguy: after adding strip(), the above fails due to pep8 check ;) | 15:20 |
*** dims has quit IRC | 15:20 | |
raorn | tdurakov: I've updated patch, can't post reply to patset 1. | 15:21 |
*** pixelb has quit IRC | 15:21 | |
ndipanov | so dansmith I remember why I did the drop_mig_context that way | 15:21 |
raorn | tdurakov: look at _get_floating_ips_by_fixed_and_port, it returns empty list for 404 error | 15:21 |
tdurakov | raorn, just select patch-set 1 and click review button for p-s 1 | 15:21 |
ndipanov | so we call it from drop_resize_claim | 15:21 |
raorn | posted | 15:22 |
ndipanov | which gets called from __exit__ of the claim and in random places | 15:22 |
ndipanov | and it's really not expected for it to mutate the instance | 15:22 |
*** tsekiyam_ has quit IRC | 15:22 | |
ndipanov | that needs to be saved later | 15:22 |
*** tsekiyama has joined #openstack-nova | 15:23 | |
ndipanov | so I did this as a shortcut | 15:23 |
ndipanov | dansmith, ^ | 15:23 |
ndipanov | not saying that was an awesome idea - just there there was some glimpse of thought behind it not only rush :) | 15:23 |
dansmith | okay | 15:24 |
dansmith | well, nothing we can reasonably do about it right now | 15:24 |
tdurakov | raorn, yes, saw that | 15:25 |
johnthetubaguy | oomichi: it just feels like the wrong trade off | 15:25 |
*** diogogmt has quit IRC | 15:26 | |
*** eharney has joined #openstack-nova | 15:26 | |
ndipanov | dansmith, and forcing a save there would also not be excellent I think | 15:26 |
johnthetubaguy | oomichi: having said that, I would rather skip the strip than not fix this for liberty | 15:26 |
bauzas | ndipanov: I may again have a stupid question, but since you removed saving the new destination for the instance in https://review.openstack.org/#/c/200485/28/nova/compute/manager.py,cm L2692 (old version), I guess it's set elsewhere ? | 15:26 |
*** Sree has quit IRC | 15:26 | |
bauzas | ndipanov: or is it a miss ? | 15:27 |
tdurakov | raorn, let's give a chance for other reviewers | 15:27 |
*** annegent_ has joined #openstack-nova | 15:27 | |
bauzas | ndipanov: nevermind my silly question... | 15:28 |
*** e0ne has quit IRC | 15:28 | |
*** dims has joined #openstack-nova | 15:28 | |
ndipanov | bauzas, so I'd like to see forcing a host either die completely or become sane | 15:29 |
raorn | tdurakov: agree. | 15:30 |
*** e0ne has joined #openstack-nova | 15:30 | |
raorn | and call it a day! | 15:30 |
bauzas | ndipanov: yeah, you moved the save up | 15:30 |
bauzas | ndipanov: and I agree with you | 15:30 |
ndipanov | as you can see at this point I just give up if your trying to force a host- you _have_ to know what you're doing | 15:30 |
ndipanov | which save? | 15:30 |
ndipanov | I moved it down | 15:30 |
ndipanov | it has to move down otherwise there's a race | 15:30 |
*** jdurgin1 has joined #openstack-nova | 15:30 | |
*** annegentle has quit IRC | 15:30 | |
ndipanov | there is a bug like that already in resize | 15:30 |
bauzas | ndipanov: I call it 'up' because that's in the calling method | 15:31 |
bauzas | once the called method is called | 15:31 |
bauzas | ndipanov: anyway | 15:31 |
tdurakov | jaypipes, hi Jay, could you look at https://review.openstack.org/#/c/225910/ ? | 15:31 |
bauzas | ndipanov: so, back to your point about forcing, that's all the rationale for http://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/check-destination-on-migrations.html | 15:32 |
bauzas | ndipanov: ie. never even force a destination unless you explicitely know what you do | 15:32 |
*** atuvenie has quit IRC | 15:34 | |
*** thorst has joined #openstack-nova | 15:34 | |
*** annegent_ has quit IRC | 15:36 | |
ndipanov | bauzas, that text is not easy to read | 15:36 |
*** annegentle has joined #openstack-nova | 15:36 | |
*** diogogmt has joined #openstack-nova | 15:36 | |
oomichi | johnthetubaguy: you are saying that it is better to skip strip() calls on servers API also if we don't apply strip() calls to all APIs, right? | 15:37 |
oomichi | johnthetubaguy: if so, that also is good option for me | 15:37 |
bauzas | ndipanov: that's understandable, I should propose the spec back for Mitaka, so you should stay tuned for commenting it out if you wish | 15:38 |
johnthetubaguy | oomichi: I just mean we urgently need to allow the trailing spaces in the v2.0 compat mode, I am less fussed about stripping them all, although my preference is to strip them all | 15:38 |
*** dave-mccowan has joined #openstack-nova | 15:38 | |
ndipanov | so bauzas I read that spec | 15:39 |
ndipanov | and I remember asking before | 15:39 |
ndipanov | this seems to be the same thing as claims | 15:39 |
bauzas | ndipanov: that's just an UX change | 15:39 |
ndipanov | but in a place where we are making it super racy | 15:39 |
johnthetubaguy | dansmith: I promised to reach out about setting priorities on the specs, do you think this set of spec categories would work? https://etherpad.openstack.org/p/mitaka-nova-spec-review-tracking | 15:39 |
ndipanov | I have to say that I don't get it | 15:39 |
bauzas | ndipanov: I mean, we're changing how the user is asking for an evacuate | 15:40 |
bauzas | ndipanov: not how the evacuate works | 15:40 |
bauzas | ndipanov: for the moment, it's just super confusing and error-prone to consider that the scheduler is called while it's not | 15:40 |
ndipanov | ok so that part is fine | 15:40 |
ndipanov | I don't get the | 15:40 |
ndipanov | conductor runs the scheduler and then recrods the fact that it failed | 15:41 |
ndipanov | this info is useless | 15:41 |
ndipanov | imho | 15:41 |
*** e0ne has quit IRC | 15:41 | |
ndipanov | what you want to do is maybe make the claim do that | 15:41 |
ndipanov | but doing it in the conductor makes no sense to me really | 15:42 |
*** romainh has left #openstack-nova | 15:42 | |
ndipanov | I feel we had this conversation though | 15:42 |
bauzas | ndipanov: tbh, by reading again the spec, I feel something is wrong | 15:44 |
bauzas | ndipanov: when someone is providing a destination and not using the force flag, then we're verifying the scheduler | 15:44 |
bauzas | ndipanov: in the spec, it gives a 409 when the user is calling, but that would mean a blocking call to the scheduler than I doesn't want | 15:45 |
bauzas | ndipanov: so, instead, we should still accept a 202 and change the instance status later when the scheduler returns | 15:45 |
ndipanov | bauzas, I am really confused | 15:46 |
ndipanov | once you're executing conductor | 15:46 |
ndipanov | request is already gone | 15:46 |
ndipanov | no idea what that has to do with this | 15:46 |
ndipanov | and I find it really hard to understand that text at all | 15:46 |
ndipanov | maybe try to rewrite it | 15:47 |
ndipanov | I have no idea what that thing is supposed to do | 15:47 |
bauzas | ndipanov: that's my point, the spec shouldn't have been accepted | 15:47 |
ndipanov | well it' | 15:47 |
ndipanov | s gone now so | 15:47 |
ndipanov | :) | 15:48 |
bauzas | ndipanov: I mean | 15:48 |
bauzas | ndipanov: my spec is badly written on that, we need to get another consensus for that specific case | 15:49 |
bauzas | lxsli: ^ | 15:49 |
bauzas | lxsli: tl;dr: we can't serve a Conflict to the user if we want to verify the scheduler | 15:49 |
bauzas | lxsli: http://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/check-destination-on-migrations.html#rest-api-impact (bullet #1) | 15:50 |
*** julien_ZTE has quit IRC | 15:50 | |
ndipanov | so I am a very strong -1 on even trying to do that | 15:50 |
ndipanov | we are kidding ourselves if we think we can | 15:50 |
bauzas | ndipanov: because we would need the 2PC that is the claim ? | 15:51 |
ndipanov | the only way forward that will not make us insane is that we accpet that user facing API is a very poor abstraction | 15:51 |
ndipanov | if your instance is in error - tough shit | 15:51 |
ndipanov | run a new one | 15:51 |
ndipanov | only thing we can do is 200 json looks ok | 15:52 |
ndipanov | "who knows what's gonna happen to the instance though I have like 40 services and a galera cluster" | 15:52 |
bauzas | ndipanov: so that would strongly reduce the scope of it, but I see your point | 15:52 |
ndipanov | "gut good luck with that" | 15:52 |
bauzas | ndipanov: I mean the scope of the spec | 15:53 |
bauzas | ndipanov: what I want to avoid is some very poor user experience where he doesn't know why it should be scheduled or not | 15:53 |
ndipanov | yeah exposing scheduling in the API is just not something we can reasonably do without shooting our both feet off and then proceeding to knees ... | 15:53 |
*** oomichi has quit IRC | 15:53 | |
ndipanov | bauzas, so that's not poor user exp | 15:54 |
ndipanov | that's poor admin exp | 15:54 |
bauzas | ndipanov: yeah, the rule of thumb is never ever do RPC blocking calls | 15:54 |
ndipanov | IMHO admin should not even be using the same API | 15:54 |
bauzas | ndipanov: sec, verifying the rules | 15:54 |
ndipanov | but that's a different discussion | 15:54 |
*** cfriesen__ has quit IRC | 15:54 | |
*** cfriesen__ has joined #openstack-nova | 15:54 | |
*** f13o has joined #openstack-nova | 15:55 | |
ndipanov | all things that are admin only should belong to a different endpoint and have a different set of resources verbs and a different version | 15:55 |
*** dustins is now known as dustins|afl | 15:55 | |
*** vladikr has quit IRC | 15:55 | |
*** dustins|afl is now known as dustins|afk | 15:55 | |
bauzas | ndipanov: so, yeah, live-migrate and evacuate are admin-only, agreed then on the bad admin UX | 15:56 |
ndipanov | anyway that's just like my opinion | 15:56 |
bauzas | ndipanov: anyway, it needs another spec | 15:57 |
*** RA_ has quit IRC | 15:57 | |
bauzas | ndipanov: so I'll provide a change by modifying http://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/check-destination-on-migrations.html#rest-api-impact | 15:57 |
bauzas | ndipanov: and we could discuss that on the spec and see with all others how to find a gentleman's agreement | 15:57 |
bauzas | ndipanov: but I got your idea anyway | 15:58 |
*** RA_ has joined #openstack-nova | 15:58 | |
*** Apoorva has joined #openstack-nova | 16:00 | |
*** zhangjn has quit IRC | 16:00 | |
*** mdrabe has quit IRC | 16:00 | |
*** burt has quit IRC | 16:00 | |
*** haomaiwa_ has quit IRC | 16:01 | |
*** burt has joined #openstack-nova | 16:01 | |
*** haomaiwa_ has joined #openstack-nova | 16:01 | |
*** alex_klimov has quit IRC | 16:03 | |
*** whydidyoustealmy is now known as shakamunyi | 16:03 | |
*** zhangjn has joined #openstack-nova | 16:04 | |
*** RA_ has quit IRC | 16:04 | |
*** jlanoux has quit IRC | 16:04 | |
lxsli | bauzas: thanks for the ping, I'll await the new spec with interest | 16:04 |
*** VW_ has quit IRC | 16:05 | |
*** Sree has joined #openstack-nova | 16:06 | |
*** annegent_ has joined #openstack-nova | 16:06 | |
*** VW has joined #openstack-nova | 16:06 | |
*** annegentle has quit IRC | 16:07 | |
*** vladikr has joined #openstack-nova | 16:09 | |
dansmith | johnthetubaguy: this is like spec review priorities, somewhat distinct from what we decide on as actual project priorities at summit? | 16:10 |
johnthetubaguy | dansmith: yes | 16:10 |
dansmith | johnthetubaguy: okay, seems good I guess.. I'll reserve final judgment until we put things in these buckets and see how well they fit | 16:11 |
*** VW has quit IRC | 16:11 | |
*** scheuran has quit IRC | 16:11 | |
*** exploreshaifali has joined #openstack-nova | 16:12 | |
johnthetubaguy | dansmith: yeah, we can always add some new categories, etc, as we go through | 16:13 |
dansmith | yeah | 16:13 |
*** yamahata has joined #openstack-nova | 16:13 | |
*** Marga_ has quit IRC | 16:13 | |
*** leeantho has joined #openstack-nova | 16:13 | |
*** tonytan4ever has joined #openstack-nova | 16:14 | |
*** ildikov has quit IRC | 16:14 | |
*** zzzeek has quit IRC | 16:14 | |
*** tjones has joined #openstack-nova | 16:15 | |
*** otter768 has joined #openstack-nova | 16:16 | |
*** VW has joined #openstack-nova | 16:16 | |
openstackgerrit | Chuck Carmack proposed openstack/nova: WIP: Disable security group tests in tempest-dsvm-cells-rc using a config option https://review.openstack.org/226043 | 16:16 |
*** VW_ has joined #openstack-nova | 16:17 | |
*** EinstCrazy has quit IRC | 16:19 | |
*** baoli_ has quit IRC | 16:19 | |
*** otter768 has quit IRC | 16:20 | |
*** VW has quit IRC | 16:20 | |
*** matrohon has quit IRC | 16:21 | |
*** tjones has left #openstack-nova | 16:22 | |
dansmith | johnthetubaguy: hit that string wrap one | 16:23 |
*** zzzeek has joined #openstack-nova | 16:23 | |
*** mdrabe has joined #openstack-nova | 16:23 | |
johnthetubaguy | dansmith: :) | 16:23 |
*** kiranr has quit IRC | 16:24 | |
*** rdopiera has quit IRC | 16:25 | |
*** dosaboy_ has joined #openstack-nova | 16:25 | |
*** dosaboy_ has quit IRC | 16:26 | |
ndipanov | johnthetubaguy, posting that patch soon - hopefully tomorrow we can merge it in the AM | 16:31 |
johnthetubaguy | ndipanov: sounds like a good aim | 16:32 |
*** dosaboy has quit IRC | 16:32 | |
*** nic has joined #openstack-nova | 16:33 | |
ndipanov | it turned into 2 patches but only for clarity - they could be squashed although it willl end up being 500+ line diff | 16:33 |
*** dosaboy has joined #openstack-nova | 16:33 | |
*** Marga_ has joined #openstack-nova | 16:33 | |
*** jdurgin1 has quit IRC | 16:34 | |
*** kiran-r is now known as _kiran_ | 16:35 | |
*** annegent_ has quit IRC | 16:40 | |
*** _kiran_ has quit IRC | 16:40 | |
*** kiran-r has joined #openstack-nova | 16:40 | |
*** kiran-r has quit IRC | 16:40 | |
*** kiranr_ has joined #openstack-nova | 16:40 | |
*** kiranr_ has quit IRC | 16:40 | |
openstackgerrit | Artom Lifshitz proposed openstack/nova: Virtual device role tagging https://review.openstack.org/226409 | 16:40 |
*** tdurakov has quit IRC | 16:40 | |
openstackgerrit | Nikola Dipanov proposed openstack/nova: rebuild: RPC sends additional args and claims are done https://review.openstack.org/200485 | 16:41 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: RT: track evacuation migrations https://review.openstack.org/226411 | 16:41 |
ndipanov | johnthetubaguy, ^ | 16:42 |
*** annegent_ has joined #openstack-nova | 16:42 | |
openstackgerrit | Artom Lifshitz proposed openstack/nova: WIP: Virtual device role tagging https://review.openstack.org/226409 | 16:43 |
dansmith | ndipanov: do we need to do the %sflavor thing for migrations? Those should always be the same flavor | 16:45 |
*** sahid has quit IRC | 16:46 | |
*** david-lyle has quit IRC | 16:46 | |
*** achanda has joined #openstack-nova | 16:47 | |
ndipanov | dansmith, hmmm so do we still stash it even if it is the same? | 16:47 |
ndipanov | because if not then there is a bug in the code anyway | 16:48 |
dansmith | ndipanov: I think we do for migration, but not for any good reason, and we could change that later, so seems like using instance.flavor for that case too is just better in the long run | 16:48 |
*** annegent_ has quit IRC | 16:48 | |
ndipanov | if yes then - yar - but we can improve it | 16:48 |
ndipanov | yeah | 16:48 |
dansmith | johnthetubaguy: so are we good to land this now, or are we racing with any branching or something? I saw the "open mitaka" patch is up | 16:49 |
openstackgerrit | Bartosz Fic proposed openstack/nova: Live migration failure in API doesn't revert state of vm https://review.openstack.org/168916 | 16:50 |
*** yamahata has quit IRC | 16:50 | |
*** julien_ZTE has joined #openstack-nova | 16:51 | |
*** markus_z has quit IRC | 16:54 | |
*** baoli has joined #openstack-nova | 16:55 | |
*** danpb has quit IRC | 16:55 | |
*** fawadkhaliq has joined #openstack-nova | 16:56 | |
*** xyang1 has quit IRC | 16:57 | |
*** julien_ZTE has quit IRC | 16:57 | |
*** xyang1 has joined #openstack-nova | 16:57 | |
*** rushil has joined #openstack-nova | 16:58 | |
*** Marga_ has quit IRC | 16:59 | |
johnthetubaguy | dansmith: so the patch is up really just so we can throw the hammer down fast | 17:00 |
*** ctrath has quit IRC | 17:00 | |
*** vilobhmm11 has joined #openstack-nova | 17:00 | |
*** Marga_ has joined #openstack-nova | 17:00 | |
*** rushil has quit IRC | 17:00 | |
*** haomaiwa_ has quit IRC | 17:01 | |
johnthetubaguy | dansmith: I am still planning on waiting for that claims patch | 17:01 |
johnthetubaguy | seems worth having | 17:01 |
*** haomaiwang has joined #openstack-nova | 17:01 | |
dansmith | yep, cool | 17:01 |
*** dustins|afk is now known as dustins | 17:01 | |
*** Marga_ has quit IRC | 17:02 | |
*** Marga_ has joined #openstack-nova | 17:03 | |
*** kiranr has joined #openstack-nova | 17:03 | |
*** ZZelle_ has joined #openstack-nova | 17:03 | |
*** ctrath has joined #openstack-nova | 17:03 | |
*** Marga_ has quit IRC | 17:05 | |
*** su_zhang has joined #openstack-nova | 17:06 | |
*** Marga_ has joined #openstack-nova | 17:06 | |
*** derekh has quit IRC | 17:06 | |
*** hemnafk is now known as hemna | 17:06 | |
*** davideagnello has quit IRC | 17:08 | |
*** su_zhang has quit IRC | 17:08 | |
*** su_zhang has joined #openstack-nova | 17:08 | |
*** davideagnello has joined #openstack-nova | 17:10 | |
*** penick has joined #openstack-nova | 17:10 | |
*** ildikov has joined #openstack-nova | 17:11 | |
*** lucasagomes is now known as lucas-dinner | 17:12 | |
*** gszasz has quit IRC | 17:12 | |
openstackgerrit | John Garbutt proposed openstack/nova-specs: Blueprint for Websockify security proxy framework https://review.openstack.org/207006 | 17:12 |
*** electrocucaracha has quit IRC | 17:14 | |
*** ptm_ has quit IRC | 17:14 | |
*** unicell has joined #openstack-nova | 17:14 | |
*** tongli has joined #openstack-nova | 17:17 | |
*** achanda_ has joined #openstack-nova | 17:17 | |
*** achanda has quit IRC | 17:17 | |
*** Marga_ has quit IRC | 17:20 | |
*** Marga_ has joined #openstack-nova | 17:20 | |
*** Marga_ has quit IRC | 17:21 | |
*** BobBall is now known as BobBall_AWOL | 17:22 | |
*** baoli has quit IRC | 17:27 | |
*** acc_ has joined #openstack-nova | 17:27 | |
*** hitalia has joined #openstack-nova | 17:29 | |
raildo | johnthetubaguy, hey :) thanks for you comments about nested quotas. I totally agree that we have to discuss more about this. Unfortunately, I'm not going to the next summit, but another guy from my team, that implemented nested quota on Cinder are going. | 17:29 |
raildo | johnthetubaguy, maybe we can propose a design session for this subject? or just a piece of time | 17:30 |
*** yuner027 has quit IRC | 17:31 | |
*** dims has quit IRC | 17:31 | |
*** Daisy has quit IRC | 17:31 | |
*** dims has joined #openstack-nova | 17:32 | |
*** lpetrut has quit IRC | 17:32 | |
*** Daisy has joined #openstack-nova | 17:32 | |
*** kiran-r has joined #openstack-nova | 17:32 | |
*** Marga_ has joined #openstack-nova | 17:33 | |
*** Marga_ has quit IRC | 17:34 | |
*** Marga_ has joined #openstack-nova | 17:34 | |
johnthetubaguy | raildo: ah, so we should look at what was done for cinder, for user | 17:36 |
johnthetubaguy | for sure | 17:36 |
*** claudiub has joined #openstack-nova | 17:36 | |
*** Daisy has quit IRC | 17:36 | |
*** rotbeard has quit IRC | 17:37 | |
*** garyk has quit IRC | 17:37 | |
raildo | johnthetubaguy, you mean quota for users? Cinder doesn't have this, just quota for projects | 17:38 |
openstackgerrit | Joshua Harlow proposed openstack/nova: Avoid needless list copy in 'scheduler_host_subset_size' case https://review.openstack.org/226443 | 17:38 |
johnthetubaguy | raildo: sorry, thats just my bad typing, I meant s/for user/for sure/ | 17:38 |
raildo | ah, ok :) | 17:38 |
*** haomaiwang has quit IRC | 17:39 | |
raildo | I'll keep the spec as "work in progress" and I'll re-write after the summit :) | 17:39 |
raildo | johnthetubaguy, thanks | 17:39 |
johnthetubaguy | raildo: ideally I would like to merge all the specs before the summit where possible | 17:40 |
johnthetubaguy | raildo: that way its just the ones that need in person discuss that need to take up summit sessions | 17:40 |
johnthetubaguy | raildo: we should be open for mitaka code in a few days | 17:40 |
*** Sree has quit IRC | 17:41 | |
raildo | johnthetubaguy, got it... so I'll just send the spec (as it is written) for mitaka, and we can discuss on the review, or maybe put a topic in the nova meeting (if we have a big discussion on this subject) | 17:43 |
openstackgerrit | John Garbutt proposed openstack/nova-specs: Libvirt hardware policy from libosinfo https://review.openstack.org/218962 | 17:45 |
vilobhmm11 | johnthetubaguy : ping | 17:46 |
johnthetubaguy | vilobhmm11: I am about to run away I am afraid | 17:46 |
vilobhmm11 | johnthetubaguy : sure later then | 17:47 |
*** jichen has joined #openstack-nova | 17:47 | |
vilobhmm11 | :) | 17:47 |
johnthetubaguy | vilobhmm11: do drop me an email, if thats easier | 17:47 |
vilobhmm11 | johnthetubaguy : sure thanks! | 17:47 |
*** lpetrut has joined #openstack-nova | 17:48 | |
*** emagana has quit IRC | 17:51 | |
*** emagana has joined #openstack-nova | 17:52 | |
*** mriedem has quit IRC | 17:52 | |
*** mriedem1 has joined #openstack-nova | 17:53 | |
*** obedmr has quit IRC | 17:55 | |
*** rushil has joined #openstack-nova | 17:57 | |
*** electrocucaracha has joined #openstack-nova | 17:57 | |
*** rlrossit1 has joined #openstack-nova | 17:57 | |
*** exploreshaifali has quit IRC | 17:57 | |
*** fawadkhaliq has quit IRC | 17:58 | |
*** rlrossit has quit IRC | 17:58 | |
*** electroc_ has joined #openstack-nova | 17:58 | |
*** obedmr has joined #openstack-nova | 18:00 | |
*** electrocucaracha has quit IRC | 18:02 | |
*** VW_ has quit IRC | 18:02 | |
*** VW has joined #openstack-nova | 18:03 | |
*** penick has quit IRC | 18:04 | |
*** julien_ZTE has joined #openstack-nova | 18:05 | |
*** rushil has quit IRC | 18:06 | |
*** VW has quit IRC | 18:07 | |
*** julien_ZTE has quit IRC | 18:10 | |
openstackgerrit | Joshua Harlow proposed openstack/nova: Add dynamic filter reordering, based on runtime/run count (WIP) https://review.openstack.org/226478 | 18:11 |
sdague | mriedem1: so on things like this - https://review.openstack.org/#/c/222573/1 my feeling is voting check is sufficient. | 18:12 |
*** ociuhandu has quit IRC | 18:16 | |
*** krtaylor has quit IRC | 18:16 | |
harlowja | let me know what u guys think of 226478 if u get anytime, basically certain filters are really slow in a larger cloud (ours), and shouldn't be done up front, so reordering them based on prior runs automaticaally seems useful (or maybe it doesn't, ha, u decide) | 18:16 |
*** otter768 has joined #openstack-nova | 18:16 | |
mriedem1 | sdague: because it's only running on nova changes? | 18:19 |
sdague | mriedem1: yeh | 18:19 |
bauzas | harlowja: I'm very -1 for that | 18:19 |
harlowja | bauzas i'd be interested in knowing why | 18:20 |
*** lpetrut has quit IRC | 18:20 | |
mriedem1 | sdague: i guess you have to convince andreas | 18:20 |
bauzas | harlowja: sure, just because it really depends on many situations | 18:20 |
mriedem1 | sdague: and probably call that out in the commit message | 18:20 |
bauzas | harlowja: and that's something that can be done by operators | 18:20 |
sdague | well, other nova people giving it a +1 would be good | 18:20 |
mriedem1 | sdague: i might +1 after a commit message justifying check queue only :) | 18:21 |
mriedem1 | i can't speak for other "nova people", most of them are...well, you know | 18:21 |
bauzas | harlowja: and for example, the ComputeFilter could be very quick to be called, but it would be still the first to be run | 18:21 |
harlowja | bauzas i get that, and yes it can, although i'm sorta of the opinion that a system should correct itself if it can | 18:21 |
sdague | mriedem1: ok, I feel like this is a common pattern we've used, and it was agreed in the nova-api meeting. But I can further explain | 18:21 |
bauzas | harlowja: no, that's not the FilterScheduler design | 18:22 |
*** otter768 has quit IRC | 18:22 | |
mriedem1 | could use some other eyes on this https://review.openstack.org/#/c/225278/ - i'm not really crazy about explicitly handling 500/504 in there | 18:22 |
bauzas | harlowja: the design is that you have something you can use and modify if you want | 18:22 |
mriedem1 | sdague: i wasn't in the nova-api meeting | 18:22 |
mriedem1 | so context is helpful in a review | 18:22 |
bauzas | harlowja: for example, putting a new filter if you want | 18:22 |
harlowja | bauzas sure, modify if want, or let it auto-tune itself | 18:22 |
bauzas | harlowja: again, I'm not sure that having the filter run by the time it takes is a good thing | 18:23 |
harlowja | bauzas ok what other metric would u prefer? | 18:23 |
sdague | mriedem1: ok sure | 18:23 |
*** achanda has joined #openstack-nova | 18:23 | |
mriedem1 | sdague: btw, welcome back :P | 18:23 |
harlowja | operators have metrics in there brain, sure, i get that | 18:23 |
bauzas | harlowja: just an operator metric, ie. leaving him changing the order if he wants :) | 18:23 |
sdague | :) | 18:23 |
*** achanda_ has quit IRC | 18:24 | |
harlowja | bauzas i get that to, but then operators could also deploy vms manually for people to :-P | 18:24 |
harlowja | except for this cloud thing, lol | 18:24 |
harlowja | where we try to do it for them | 18:24 |
harlowja | bauzas we can do more smartness like a decay function as well, so that filters eventually reset | 18:25 |
harlowja | just know that we got hit by slow filters at yahoo, and yes, operators can go in and tweak things, except i feel somewhat that software should do most of the tweaking (thus the reason we have software in the first place) | 18:26 |
*** penick has joined #openstack-nova | 18:26 | |
harlowja | vs more wetware, lol | 18:26 |
harlowja | if u don't start somewhere with software though, u never get away from wetware making all the decisions, and therefore https://review.openstack.org/#/c/226478/ could be something useful to some people (maybe not all) | 18:28 |
bauzas | harlowja: okay, so saying for example that you run TrustedFilter which makes an external call (okay, I hate that one but that's still in-tree), and you have CoreFilter that compares the left vcpus with the flavor | 18:29 |
harlowja | k | 18:29 |
bauzas | harlowja: so, using your change would have the TrustedFilter run first, right? | 18:29 |
*** jwcroppe has joined #openstack-nova | 18:29 | |
*** kiran-r has quit IRC | 18:29 | |
harlowja | unlikely, i would expect TrustedFilter to be slower, due to external call, and therefore shift to running later | 18:30 |
*** krtaylor has joined #openstack-nova | 18:30 | |
harlowja | CoreFilter afaik is just running on some in-memory stuff, so should be quicker | 18:30 |
bauzas | okay | 18:30 |
jwcroppe | jaypipes, bauzas, dansmith: anyone available to check out the metrics review we had for that performance-related case - https://review.openstack.org/#/c/219153/ ? No rush, just wanted to get some eyes on this. | 18:30 |
bauzas | harlowja: so, now, say you want to run ComputeFilter which makes a call to the SG API | 18:31 |
harlowja | k, that one by default also uses an in-memory stuff | 18:31 |
bauzas | then the CoreFilter could possibly be run before ComputeFilter, right? | 18:31 |
bauzas | harlowja: it doesn't | 18:31 |
bauzas | harlowja: it calls the sg.is_up() API | 18:31 |
harlowja | right, the default of that is just in-memory lookups | 18:32 |
*** vilobhmm11 has quit IRC | 18:32 | |
bauzas | harlowja: for checking a disabled state, it calls the sg API anyway, so it's not a local lookup | 18:32 |
*** ctrath1 has joined #openstack-nova | 18:32 | |
jwcroppe | bauzas: btw, I am in Rennes right now - interesting place :) | 18:32 |
*** vilobhmm11 has joined #openstack-nova | 18:32 | |
harlowja | bauzas right, https://github.com/openstack/nova/blob/master/nova/servicegroup/drivers/db.py#L57 lookup into passed in service_ref | 18:33 |
bauzas | harlowja: but you can't assume that the DB driver is the selected one | 18:33 |
harlowja | sure, most of the other ones are broken currently, so i'm assuming that, lol | 18:33 |
bauzas | harlowja: at least, you don't have that information in the scheduler | 18:33 |
bauzas | harlowja: but what if tooz becomes the main API ? | 18:34 |
bauzas | s/API/driver | 18:34 |
harlowja | someday | 18:34 |
*** Piet has quit IRC | 18:34 | |
bauzas | agreed, but we're discussing now about something which would become default tomorrow | 18:34 |
bauzas | I mean your change | 18:34 |
harlowja | k | 18:34 |
harlowja | so maybe we need a way to lock certain filters out of the auto-tuning then | 18:35 |
bauzas | so, there could possibly some way that the CoreFilter would be considered faster than the ComputeFilter | 18:35 |
harlowja | sure | 18:35 |
bauzas | given the complexity of that is linear | 18:35 |
harlowja | right | 18:35 |
bauzas | (for each filter) | 18:35 |
doffm | mriedem1: https://review.openstack.org/#/c/225289/ | 18:36 |
*** atuvenie has joined #openstack-nova | 18:36 | |
*** ctrath has quit IRC | 18:36 | |
bauzas | it would mean that you would take more time at the end, because you would call more | 18:36 |
bauzas | eg. say that I have 1000 nodes, but only one active | 18:36 |
bauzas | then, ordering the other way would be : | 18:36 |
bauzas | 1000 calls + 1 call | 18:37 |
ndipanov | dansmith, what do you want me to do with this? https://review.openstack.org/#/c/226411/ | 18:37 |
bauzas | instead of 1 + 1 call | 18:37 |
harlowja | bauzas sure, i get what u are saying | 18:37 |
dansmith | ndipanov: thought you were just going to change it to "if migration.type == resize: instance.%sflavor else instance.flavor" ? | 18:37 |
harlowja | and yes, so it might be useful to lock certain filters to a certain order | 18:37 |
harlowja | and let other ones be autotuned | 18:37 |
bauzas | harlowja: so that's exactly why we leave the possibiliy to the operator to order the list the way he wants | 18:38 |
harlowja | or maybe the whole idea sucks, and we should just stick to wetware | 18:38 |
harlowja | idk, the whole idea that we can't do better in software always makes me wonder about this | 18:38 |
bauzas | harlowja: then, you would need to find some way to say "as an operator, I don't want to monitor the time for that specific filter" | 18:38 |
harlowja | right, thats not so hard | 18:38 |
ndipanov | dansmith, ok I can do that - I kind of think this is better since we do stash it | 18:38 |
bauzas | harlowja: IMHO, all of that could be considered as a potential separate scheduler driver | 18:39 |
dansmith | ndipanov: you think what you have there is better? my point was just that we probably shouldn't be using %sflavor, and if we changed it later, we'd need to change this for no reason | 18:39 |
bauzas | harlowja: I mean, like the CachingScheduler | 18:39 |
ndipanov | anyway maybe no need to hold up over that since I though there is still a window to get this in for rC1 | 18:39 |
ndipanov | dansmith, ok | 18:39 |
harlowja | bauzas or a config that says 'auto-tune' these, lol | 18:39 |
harlowja | and only auto-tune those | 18:39 |
dansmith | ndipanov: yeah, I thought based on the earlier convo you were going to change it that minute, so I didn't think it'd hold anything up | 18:39 |
dansmith | ndipanov: if you don't want to then that's fine I guess | 18:40 |
bauzas | harlowja: I'd rather consider that orthogonal | 18:40 |
jroll | harlowja: TPainScheduler() | 18:40 |
harlowja | to the pain? | 18:40 |
harlowja | i didn't find the 6 fingered guy yet | 18:40 |
harlowja | lol | 18:40 |
jroll | harlowja: "The use of Auto-Tune as a musical effect was bolstered in the late-2000s by R&B singer T-Pain," https://en.wikipedia.org/wiki/Auto-Tune | 18:40 |
* jroll now realizes maybe that question wasn't for him and sneaks away | 18:41 | |
bauzas | jroll: I love that | 18:41 |
bauzas | jroll: that's pretty ironic | 18:41 |
bauzas | :p | 18:41 |
jroll | heh | 18:42 |
penick | bauzas: I see what you did there | 18:42 |
*** VW has joined #openstack-nova | 18:42 | |
*** achanda has quit IRC | 18:42 | |
bauzas | penick: MHO is that we could possibly consider that optional for the operator, but I'd prefer to see that landed in a specific driver | 18:42 |
bauzas | penick: but again, that's MHO like I said to harlowja | 18:43 |
ndipanov | dansmith, re-running tests now | 18:43 |
dansmith | ndipanov: polishing my +2 hammer now | 18:44 |
bauzas | harlowja: given your change, I'd wonder if you could just provide an out-of-tree driver for your own internal needs | 18:44 |
ndipanov | I thought there was a -2 hammer and a +2 fairy-dust sack | 18:44 |
bauzas | harlowja: that's easily doable thanks to a conf flag | 18:44 |
harlowja | bauzas anythings possible | 18:44 |
bauzas | harlowja: sure, but we need to be very careful when adding features to something like the scheduler | 18:45 |
bauzas | because we have lots of usecases | 18:45 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: RT: track evacuation migrations https://review.openstack.org/226411 | 18:45 |
bauzas | harlowja: again, see for example the CachingScheduler and why it's only considered experimental | 18:45 |
harlowja | ya, i wonder about that one honestly | 18:46 |
*** rushil has joined #openstack-nova | 18:46 | |
*** rushil has quit IRC | 18:46 | |
harlowja | in our usage recently, it seems like the filters runtime/ordering is the slow parts, not the extraction of hosts | 18:47 |
harlowja | and i thought the caching scheduler just caches the extraction of hosts info | 18:47 |
*** achanda has joined #openstack-nova | 18:48 | |
*** obedmr has quit IRC | 18:48 | |
*** electroc_ has quit IRC | 18:48 | |
*** aix has quit IRC | 18:49 | |
harlowja | which a database should (likely) be caching anyway (since its pretty frequently used)? | 18:49 |
bauzas | harlowja: so, like alaski said, I think it's worth writing a spec for your thoughts | 18:50 |
harlowja | bauzas sure, i think penick and melwitt might be up for that, further expand that idea and take it from there, i was just scribbling down a POC i thought of this morning :-P | 18:50 |
dansmith | ndipanov: lol | 18:50 |
harlowja | and/or that was mentioned as an idear yesterday | 18:50 |
penick | Don’t look at me :P melwitt’s idea, she’s got something she’s cooking up | 18:51 |
*** lpetrut has joined #openstack-nova | 18:51 | |
bauzas | harlowja: sure, thanks for rising up the idea, that's always a good thing | 18:51 |
harlowja | penick fair enough, no looking at u, lol | 18:51 |
*** rushil has joined #openstack-nova | 18:52 | |
harlowja | but it should be called the auto-tune scheduler, ha | 18:52 |
harlowja | thats my input, lol | 18:52 |
bauzas | :) | 18:52 |
penick | hahahaha “I’m on a boat” | 18:53 |
*** su_zhang has quit IRC | 18:53 | |
*** ociuhandu has joined #openstack-nova | 18:53 | |
mriedem1 | weird, large ops timed out waiting for an instance to go active http://logs.openstack.org/33/165233/31/check/gate-tempest-dsvm-large-ops/e7a691a/console.html | 18:53 |
mriedem1 | looks like networking gets setup and then, kaput | 18:54 |
*** obedmr has joined #openstack-nova | 18:54 | |
bauzas | jwcroppe: re: https://review.openstack.org/#/c/219153/ there is rather more a high prio bug which makes the metrics regressed | 18:54 |
bauzas | jwcroppe: https://bugs.launchpad.net/nova/+bug/1493680 | 18:55 |
openstack | Launchpad bug 1493680 in OpenStack Compute (nova) "HostState.metrics is no longer a dict" [High,In progress] - Assigned to Yingxin (cyx1231st) | 18:55 |
*** lpetrut has quit IRC | 18:55 | |
ccarmack1 | mriedem: question - I did "check experimental" on a nova change, but gate-tempest-dsvm-cells-rc didn't run | 18:56 |
bauzas | jwcroppe: so I'd tend to hold your change until Mitaka opens | 18:56 |
bauzas | which should be very soon | 18:56 |
*** rushil has quit IRC | 18:56 | |
*** whenry_ has quit IRC | 18:56 | |
mriedem1 | ccarmack1: gate-tempest-dsvm-cells is in the check queue, you don't need to 'check experimental' on the nova change | 18:56 |
bauzas | - and for Rennes, I'm not a super expert of that city, by far away from my spot:) | 18:56 |
mriedem1 | ccarmack1: the cells job is experimental in the devstack and tempest queues i think | 18:57 |
ccarmack1 | mriedem: thanks | 18:57 |
*** obedmr has quit IRC | 19:00 | |
_diana_ | zzzeek: any tips wrt database debugging using devstack (turning query logging on, not truncating long query lines etc)? thanks! | 19:00 |
melwitt | bauzas: yeah... I had been thinking of a scheduler that reorders filters based on capacity. for example as cores are consumed, CoreFilter moves earlier in the order, so that more restrictive filters are first. a scheduler that uses capacity and other available metrics to dynamically order filters to pare down the list of candidates as early in the order as possible | 19:01 |
zzzeek | _diana_: i would turn on query logging, yes. you can also refine logging using a totally separate log config of your own, though some devstack features might get tripped up by it | 19:01 |
*** atuvenie has quit IRC | 19:01 | |
zzzeek | _diana_: its very tough to work with, devstack, etc. | 19:01 |
zzzeek | _diana_: im sort of looking forward to see what you come up with b.c. I dont’ have enough pateince to get good at it :) | 19:02 |
zzzeek | _diana_: if you want to see how the world of openstack logging works you can read the oslo.log source, which also interacts iwth oslo.config | 19:02 |
bauzas | melwitt: so, I think looking at how long it takes is a bad metric, we should rather see the percentage of falses returned by each filter IMHO | 19:02 |
melwitt | bauzas: makes sense | 19:03 |
harlowja | bauzas melwitt perhaps a combined set of metrics, time, capabity... | 19:04 |
harlowja | pretty sure this is how JIT compilers work as well (to figure out which functions to jit...) | 19:04 |
bauzas | harlowja: perhaps... | 19:04 |
*** electrocucaracha has joined #openstack-nova | 19:04 | |
melwitt | bauzas: I wasn't thinking of doing it by how long it takes, but by figuring capacity based on resource tracker. cores used % for example | 19:04 |
_diana_ | zzzeek: Thanks! if you have any example to config to start with, it would me much appreciated. For now, I just set DATABASE_QUERY_LOGGING=True in the devstack local.conf, but long queries appear to be truncated. | 19:04 |
bauzas | melwitt: I see your thoughts | 19:04 |
bauzas | melwitt: like I said to harlowja, considering an experimental driver for that could makes sense | 19:05 |
_diana_ | zzzeek: I'll go look at oslo.log now too, thanks! | 19:05 |
bauzas | make even | 19:05 |
harlowja | melwitt perhaps a filter should try to guess_a_mate its position based on its own knowledge, idk | 19:05 |
harlowja | guesstimate | 19:05 |
harlowja | or whatver, lol | 19:05 |
bauzas | melwitt: but I guess operators would love to take control from the automatic pilot when they want | 19:06 |
bauzas | melwitt: and I also think about one specific problem | 19:06 |
penick | bauzas: If they want to, they’re welcome to. Just need to remove it from the filter chain. | 19:07 |
bauzas | melwitt: say that you have identified all your filters | 19:07 |
bauzas | penick: that requires a restart | 19:07 |
melwitt | bauzas: yeah. if it were a different driver then it would be a config change + reload (to go manual) | 19:07 |
penick | True. Takes only a second to restart the scheduler | 19:07 |
bauzas | melwitt: so, for some specific requests, you could have filters behaviour very different from the median | 19:08 |
bauzas | melwitt: say for example the NUMAFilter | 19:08 |
bauzas | melwitt: in general, that filter would be returning 100% | 19:08 |
bauzas | melwitt: because 99% of the requests wouldn't ask for NUMA requirements | 19:08 |
*** obedmr has joined #openstack-nova | 19:08 | |
*** earlephilhower has joined #openstack-nova | 19:09 | |
bauzas | melwitt: but, say that you have one guy now asking for cpu-pinning using a very specific flavor | 19:09 |
bauzas | melwitt: then there would some chance that 95% of the hosts would return False with that filter | 19:09 |
melwitt | bauzas: I see your point | 19:09 |
*** electrocucaracha has quit IRC | 19:10 | |
*** e0ne has joined #openstack-nova | 19:10 | |
bauzas | melwitt: so, that really depends on your cloud, and while I think it could be a nice improvement, I'm just a bit concerned by any corner case that could come up | 19:11 |
bauzas | hence my idea to consider that experimental in a separate driver | 19:12 |
bauzas | we could refine the weighting logic based on the request type, but that could become a very huge implementation | 19:13 |
melwitt | bauzas: yeah. I agree it would be a separate driver | 19:13 |
bauzas | for a very little gain, given that operators can still set their own order if they wish | 19:13 |
_diana_ | zzzeek: this is a good read: https://wiki.openstack.org/wiki/OpenStack_and_SQLAlchemy thanks! | 19:13 |
bauzas | melwitt: happy with writing a spec ? | 19:13 |
zzzeek | _diana_: there’s an option log_config_append you can put in nova.conf for eaxmple, give it a filename with a standard Python logging config | 19:14 |
harlowja | have there been any thoughts on making the whole scheduler filters a parallel map then join (just wondering) | 19:14 |
*** exploreshaifali has joined #openstack-nova | 19:14 | |
harlowja | vs a linear map/reduce like thing | 19:14 |
bauzas | harlowja: we have some thoughts on that, but not exactly speaking of the same | 19:15 |
harlowja | k | 19:15 |
melwitt | bauzas: yes, sure. I need to think about it more before materializing the spec. I didn't mean to discuss it yet but I had mentioned it in some talks we had working on some scheduler filter ordering recently | 19:15 |
bauzas | harlowja: what I would like to address is how to scale-out the scheduler | 19:15 |
bauzas | but that's food for thoughts atm | 19:15 |
bauzas | mostly because we're priotizing our interfaces before working on that | 19:16 |
harlowja | understood | 19:16 |
harlowja | isn't primarily the interface select_destinations ? | 19:19 |
harlowja | given some requirements | 19:19 |
*** ijw has quit IRC | 19:19 | |
*** su_zhang has joined #openstack-nova | 19:19 | |
mriedem1 | so, | 19:22 |
*** rlrossit1 has quit IRC | 19:23 | |
mriedem1 | jgriffith: hemna: didn't cinder change the detach API in liberty such that if you try to detach a volume that isn't attached, it's just a noop? | 19:23 |
mriedem1 | rather than a 400 | 19:23 |
*** whenry_ has joined #openstack-nova | 19:24 | |
*** angdraug has quit IRC | 19:24 | |
*** rlrossit has joined #openstack-nova | 19:25 | |
*** su_zhang has quit IRC | 19:29 | |
*** ijw has joined #openstack-nova | 19:29 | |
openstackgerrit | jichenjc proposed openstack/nova: Pep8 didn't check api/openstack/common.py https://review.openstack.org/226516 | 19:35 |
*** rlrossit has quit IRC | 19:36 | |
*** achanda has quit IRC | 19:37 | |
*** su_zhang has joined #openstack-nova | 19:38 | |
*** rlrossit has joined #openstack-nova | 19:38 | |
*** yassine_ has quit IRC | 19:39 | |
openstackgerrit | Nicolas Simonds proposed openstack/nova-specs: RBD instance snapshots https://review.openstack.org/188244 | 19:40 |
nic | If M is now opened for development, when are the procedural -2s on specs getting uncorked? | 19:42 |
bauzas | nic: not yey | 19:43 |
bauzas | yet | 19:43 |
nic | Clearly. | 19:43 |
bauzas | nic: https://review.openstack.org/#/c/226340/ | 19:43 |
nic | A ha. I thought I saw it merge, my mistake. | 19:43 |
*** electrocucaracha has joined #openstack-nova | 19:44 | |
*** bradjones has joined #openstack-nova | 19:46 | |
*** bradjones has quit IRC | 19:46 | |
*** bradjones has joined #openstack-nova | 19:46 | |
*** _pece has quit IRC | 19:48 | |
*** penick has quit IRC | 19:51 | |
edleafe | harlowja: what about having the filtering process export metrics that operators could use to tweak things? | 19:54 |
harlowja | edleafe thats possible to | 19:55 |
harlowja | although i have a feeling operators really want the thing to work auto-pilot like | 19:55 |
harlowja | and not have to micromanage it | 19:55 |
harlowja | *at least some operators i know probably don't want to micromanage, just have it work | 19:55 |
*** hitalia has quit IRC | 19:55 | |
edleafe | harlowja: do you think that these timings would vary from day to day? | 19:56 |
harlowja | depends on how much hardware is imported per day? | 19:56 |
harlowja | and/or who is scheduling want | 19:56 |
harlowja | and requesting what kind of filters to be activated | 19:56 |
edleafe | harlowja: it would seem that these adjustments would always be trailing reality | 19:56 |
*** penick has joined #openstack-nova | 19:57 | |
harlowja | perhaps | 19:57 |
harlowja | but the same could be said about any jit compiler :-P | 19:57 |
edleafe | and I'd really like to see how much adding a new rack of servers changes the relative performance of filters | 19:57 |
harlowja | k | 19:58 |
harlowja | we (y!) can probably get that info somehow | 19:58 |
harlowja | cause i'm pretty sure its happening weekly (?) | 19:58 |
edleafe | I can sure see how more servers == longer time to filter | 19:58 |
harlowja | yup, that def happens :-P | 19:58 |
edleafe | but I can't see how more servers would make one filter perform better/worse | 19:59 |
openstackgerrit | jichenjc proposed openstack/nova: Pep8 didn't check api/openstack/common.py https://review.openstack.org/226516 | 19:59 |
harlowja | true, some kind of blog post or something seems in order :-P | 19:59 |
harlowja | with pretty graphs, ha | 19:59 |
edleafe | harlowja: heh | 19:59 |
edleafe | harlowja: if something like that showed up, I'd probably suspect the filter :) | 20:00 |
harlowja | an idear, it'd be interesting to track metrics per-tenant and reorder per-tenant (optimize ordering per-tenant filters they typically apply?) | 20:00 |
harlowja | edleafe ya, some filters def call to db i think, so thats not good | 20:00 |
edleafe | harlowja: they shouldn't any more | 20:01 |
harlowja | k | 20:01 |
edleafe | at least for in-tree | 20:01 |
*** annegentle has joined #openstack-nova | 20:02 | |
harlowja | cool, i might be behind in that reality | 20:02 |
*** jamielennox|away is now known as jamielennox | 20:03 | |
*** julien_ZTE has joined #openstack-nova | 20:06 | |
*** lucas-dinner has quit IRC | 20:08 | |
hemna | mriedem1, yes I believe that's the case. | 20:09 |
*** jecarey has left #openstack-nova | 20:10 | |
*** julien_ZTE has quit IRC | 20:10 | |
mriedem1 | hemna: ok, i was looking at some wonky code in the libvirt driver's pre_live_migration method where it connects volumes and then in the next line checks if doing block migration and fails, saying it can't live block migration with volumes, | 20:11 |
hemna | is this related to the bug sseverson filed ? | 20:11 |
mriedem1 | but i wasn't sure if i could move that check up b/c the rollback code assumes the volume connection happened and just blindly tries to disconnect | 20:11 |
mriedem1 | hemna: i found it while looking through that code | 20:12 |
mriedem1 | i wouldn't say it's directly related | 20:12 |
hemna | ah ok | 20:12 |
*** dustins is now known as dustins|afk | 20:12 | |
mriedem1 | problem is here https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L6214 | 20:12 |
mriedem1 | right before that check we've connected volumes | 20:12 |
mriedem1 | which it seems like we shouldn't do if we know we're going to fail | 20:12 |
mriedem1 | the compute manager calls a rollback method which does the volume disconnect, but i'm not sure how graceful that is about handling 'i don't need to disconnect b/c i'm not connected' things | 20:13 |
hemna | it seems that if we know we have a cinder volume, we shouldn't be copying those volumes, and simply attach at the other side. | 20:13 |
*** babababa has joined #openstack-nova | 20:13 | |
hemna | avoiding the copy a volume onto itself | 20:13 |
hemna | I see | 20:14 |
hemna | couldn't we move that check in front of the for loop connecting volumes? | 20:15 |
hemna | thus avoiding the self._connect_volume calls if we know we are going to raise | 20:15 |
*** achanda has joined #openstack-nova | 20:15 | |
*** otter768 has joined #openstack-nova | 20:17 | |
*** claudiub has quit IRC | 20:19 | |
*** lucas-dinner has joined #openstack-nova | 20:20 | |
mriedem1 | hemna: yeah, that's what i was going to do | 20:20 |
mriedem1 | but | 20:20 |
mriedem1 | if pre_live_migratoin fails, we go into this rollback https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L5342 | 20:21 |
mriedem1 | which gets to https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L4791 | 20:22 |
*** otter768 has quit IRC | 20:22 | |
*** smatzek has quit IRC | 20:22 | |
mriedem1 | and i'm not entirely sure how the driver detach and volume_api.terminate_connection handle the case when we're not actually connected | 20:22 |
hemna | hrmm | 20:23 |
hemna | yah if the volume paths aren't actually there | 20:23 |
hemna | then it depends on each os-brick connector's disconnect_volume to deal w/ that in a nice way | 20:24 |
hemna | as well as cinder's terminate_connection | 20:24 |
mriedem1 | if we get a 404 then we handle that | 20:24 |
hemna | I think iSCSI is safe in this respect | 20:25 |
hemna | looks like it tries to find any paths for the volume before doing anything to it | 20:25 |
hemna | as long as we aren't calling cinder's initialize_connection to fetch the connection_info prior to the detach then it's probably ok | 20:27 |
hemna | that would induce an attachment | 20:28 |
hemna | would have to test it out man. | 20:28 |
openstackgerrit | Brian Rosmaita proposed openstack/nova: Change parameter name in utility function https://review.openstack.org/226539 | 20:28 |
*** su_zhang has quit IRC | 20:30 | |
mriedem1 | hemna: yeah, i'm going to punt instead | 20:30 |
mriedem1 | o-) | 20:30 |
hemna | :) | 20:30 |
*** kiranr has quit IRC | 20:32 | |
mriedem1 | onto more weird things | 20:32 |
openstackgerrit | Sylvain Bauza proposed openstack/nova: Add some devref for AZs https://review.openstack.org/223802 | 20:34 |
mriedem1 | alaski: dansmith: question about a weird state thing with rescue if you have a sec, | 20:34 |
mriedem1 | you can rescue an instance that is active/stopped/error state | 20:34 |
*** baoli has joined #openstack-nova | 20:35 | |
mriedem1 | the compute manager powers off the instance (or tries to) regardless of vm_state - if it's not running the virt driver should ignore that | 20:35 |
mriedem1 | but if driver.rescue fails, we don't do anything about the vm_state, we assume it should just stay as-is; but if you came in with an active instance and tried to rescue it, and we stopped it, then active is wrong | 20:35 |
*** jichen has quit IRC | 20:37 | |
mriedem1 | i'm trying to salvage a patch that's been up since february to set the instance to error state if driver.rescue fails | 20:37 |
*** gyee has joined #openstack-nova | 20:39 | |
dansmith | mriedem1: I dunno, if you need to rescue the instance and that fails, having it restart is probably not important | 20:39 |
*** su_zhang has joined #openstack-nova | 20:40 | |
*** tonytan4ever has quit IRC | 20:40 | |
mriedem1 | yeah, i wasn't sure if punting and just setting vm_state to ERROR was valid, but if rescue fails and we don't really know why, it's not valid to just leave the vm_state as active either | 20:40 |
dansmith | yeah | 20:41 |
*** arithx has joined #openstack-nova | 20:41 | |
openstackgerrit | xing-yang proposed openstack/nova: Rollback is needed if initialize_connection times out https://review.openstack.org/225278 | 20:42 |
*** ctrath1 has quit IRC | 20:43 | |
*** claudiub has joined #openstack-nova | 20:43 | |
*** tdurakov has joined #openstack-nova | 20:44 | |
tdurakov | hi, who could help with nova-network? | 20:44 |
*** whenry_ has quit IRC | 20:46 | |
tdurakov | tr3buchet, are you around? | 20:48 |
mriedem1 | it's a good thing we don't have the auto-abandon bot anymore i guess... https://review.openstack.org/#/c/155120/ | 20:48 |
*** e0ne has quit IRC | 20:51 | |
tr3buchet | tdurakov: yes hello | 20:51 |
tdurakov | tr3buchet, what do you think about async version of setup_networks_on_host? | 20:52 |
*** thangp has quit IRC | 20:52 | |
*** diogogmt has quit IRC | 20:52 | |
*** andrewbogott has quit IRC | 20:52 | |
*** andrewbogott has joined #openstack-nova | 20:52 | |
*** cfarquhar has joined #openstack-nova | 20:53 | |
*** cfarquhar has quit IRC | 20:53 | |
*** cfarquhar has joined #openstack-nova | 20:53 | |
*** annegentle has quit IRC | 20:54 | |
*** diogogmt has joined #openstack-nova | 20:58 | |
alaski | mriedem1: dansmith agreed. if rescue fails and we go back to active users have no idea there was a failure. ERROR seems valid to me | 20:59 |
mriedem1 | yar, just running tests locally and then i'll have that change up | 20:59 |
mriedem1 | in all it's co-authored-by glory | 21:00 |
*** eharney has quit IRC | 21:00 | |
tr3buchet | tdurakov: i think it would fine depending on the situation | 21:00 |
*** ctrath has joined #openstack-nova | 21:00 | |
*** baoli has quit IRC | 21:01 | |
tdurakov | tr3buchet, rollback_live_migration_on destination, if network teardown fails, driver rollback not even attempted to execute | 21:01 |
*** baoli has joined #openstack-nova | 21:02 | |
tr3buchet | tdurakov: there is also the race condition situation | 21:02 |
tdurakov | tr3buchet, where? | 21:02 |
tr3buchet | tdurakov: i'd say agreement on state and potential race conditions could be arguments against | 21:02 |
tdurakov | ok, ack | 21:03 |
tr3buchet | tdurakov: nowhere specific and only for setup network | 21:03 |
tr3buchet | what is your use case? | 21:03 |
tdurakov | see above, live-migration rollback on destination host | 21:04 |
openstackgerrit | Matt Riedemann proposed openstack/nova: If rescue failed set instance to ERROR https://review.openstack.org/155120 | 21:04 |
*** hitalia has joined #openstack-nova | 21:05 | |
*** su_zhang has quit IRC | 21:05 | |
mriedem1 | ctrath: https://github.com/lxc/nova-compute-lxd | 21:05 |
openstackgerrit | Chuck Carmack proposed openstack/nova: WIP: Disable security group tests in tempest-dsvm-cells-rc using a config option https://review.openstack.org/226043 | 21:08 |
*** aysyd has quit IRC | 21:08 | |
*** claudiub has quit IRC | 21:09 | |
*** thorst has quit IRC | 21:11 | |
*** baoli has quit IRC | 21:12 | |
*** baoli has joined #openstack-nova | 21:12 | |
*** raildo is now known as raildo-afk | 21:12 | |
openstackgerrit | Rafael Folco proposed openstack/nova: libvirt: check min required qemu/libvirt versions on ppc64/ppc64le https://review.openstack.org/226555 | 21:15 |
*** IanGovett1 has quit IRC | 21:16 | |
*** ccarmack1 has quit IRC | 21:17 | |
*** tdurakov has quit IRC | 21:19 | |
*** jaypipes has quit IRC | 21:19 | |
*** ccarmack has joined #openstack-nova | 21:20 | |
mriedem1 | doffm: http://logs.openstack.org/48/219448/4/check/gate-tempest-dsvm-postgres-full/c8a5c07/logs/screen-n-cpu.txt.gz?level=TRACE#_2015-09-21_19_44_28_025 | 21:20 |
*** rlrossit has left #openstack-nova | 21:20 | |
*** Marga_ has quit IRC | 21:21 | |
*** achanda has quit IRC | 21:21 | |
*** Marga__ has joined #openstack-nova | 21:22 | |
*** annegentle has joined #openstack-nova | 21:24 | |
*** artom has quit IRC | 21:26 | |
*** artom has joined #openstack-nova | 21:28 | |
*** alejandrito has quit IRC | 21:32 | |
*** alejandrito has joined #openstack-nova | 21:33 | |
*** julien_ZTE has joined #openstack-nova | 21:37 | |
jwcroppe | bauzas: you still awake? | 21:37 |
*** julien_ZTE has quit IRC | 21:42 | |
mriedem1 | doffm: debug logging added here http://logs.openstack.org/48/219448/4/check/gate-tempest-dsvm-postgres-full/c8a5c07/logs/screen-n-cpu.txt.gz#_2015-09-21_19_44_15_361 | 21:42 |
mriedem1 | doffm: that's called right before clean_lxc_namespace | 21:42 |
openstackgerrit | Mark Doffman proposed openstack/nova: DO NOT MERGE: Extra logging for lxc teardown fail. https://review.openstack.org/223125 | 21:42 |
*** burgerk has quit IRC | 21:42 | |
mriedem1 | doffm: looks like the umount happens here http://logs.openstack.org/48/219448/4/check/gate-tempest-dsvm-postgres-full/c8a5c07/logs/screen-n-cpu.txt.gz#_2015-09-21_19_44_15_526 | 21:45 |
*** su_zhang has joined #openstack-nova | 21:46 | |
mriedem1 | doffm: oh nvm, api.Mount init calls reset_dev() and if self.device is set then it sets self.mounted = True | 21:46 |
mriedem1 | so the unmount happens | 21:46 |
*** babababa has quit IRC | 21:46 | |
*** mriedem1 has quit IRC | 21:48 | |
*** salv-orlando has joined #openstack-nova | 21:48 | |
*** arithx has quit IRC | 21:50 | |
*** julien_ZTE has joined #openstack-nova | 21:53 | |
*** mriedem has joined #openstack-nova | 21:53 | |
*** annegentle has quit IRC | 21:54 | |
*** annegentle has joined #openstack-nova | 21:54 | |
*** ctrath has quit IRC | 21:57 | |
*** atuvenie has joined #openstack-nova | 21:58 | |
*** asettle has joined #openstack-nova | 21:58 | |
*** su_zhang has quit IRC | 21:58 | |
*** baoli has quit IRC | 22:01 | |
*** baoli has joined #openstack-nova | 22:01 | |
*** atuvenie has quit IRC | 22:02 | |
*** burt has quit IRC | 22:02 | |
*** baoli has quit IRC | 22:02 | |
*** ctrath has joined #openstack-nova | 22:03 | |
*** asettle has quit IRC | 22:03 | |
*** baoli has joined #openstack-nova | 22:04 | |
*** asettle has joined #openstack-nova | 22:04 | |
*** VW_ has joined #openstack-nova | 22:04 | |
openstackgerrit | Claudiu Belu proposed openstack/nova: Hyper-V: adds os-win library https://review.openstack.org/212102 | 22:05 |
openstackgerrit | Claudiu Belu proposed openstack/nova: Hyper-V: removes *Utils modules and unit tests https://review.openstack.org/215917 | 22:05 |
mikal | Morning | 22:05 |
*** asettle has quit IRC | 22:06 | |
*** asettle has joined #openstack-nova | 22:06 | |
*** VW has quit IRC | 22:07 | |
*** jerrygb has joined #openstack-nova | 22:08 | |
*** BadCub has quit IRC | 22:08 | |
*** ndipanov has quit IRC | 22:08 | |
*** davidlenwell has quit IRC | 22:08 | |
*** markmcclain has quit IRC | 22:08 | |
*** VW_ has quit IRC | 22:08 | |
*** ndipanov has joined #openstack-nova | 22:09 | |
*** davidlenwell has joined #openstack-nova | 22:09 | |
*** BadCub has joined #openstack-nova | 22:10 | |
*** edmondsw has quit IRC | 22:10 | |
*** mwagner_lap has quit IRC | 22:10 | |
*** inara has quit IRC | 22:10 | |
*** Guest61243 is now known as mfisch | 22:11 | |
*** zhenguo has quit IRC | 22:12 | |
*** mfisch is now known as Guest50715 | 22:12 | |
*** gus has quit IRC | 22:12 | |
*** zhenguo has joined #openstack-nova | 22:13 | |
*** dustins|afk has quit IRC | 22:13 | |
*** inara has joined #openstack-nova | 22:13 | |
*** Guest50715 is now known as mfisch | 22:13 | |
*** mdrabe has quit IRC | 22:13 | |
*** mfisch has quit IRC | 22:13 | |
*** mfisch has joined #openstack-nova | 22:13 | |
*** gus has joined #openstack-nova | 22:14 | |
*** markmcclain has joined #openstack-nova | 22:16 | |
*** Piet has joined #openstack-nova | 22:16 | |
*** tanee has quit IRC | 22:17 | |
*** baoli has quit IRC | 22:18 | |
*** otter768 has joined #openstack-nova | 22:18 | |
*** tanee has joined #openstack-nova | 22:20 | |
*** baoli has joined #openstack-nova | 22:20 | |
*** babababa has joined #openstack-nova | 22:20 | |
*** penick has quit IRC | 22:21 | |
*** boris-42 has quit IRC | 22:22 | |
*** achanda has joined #openstack-nova | 22:22 | |
*** StevenK has quit IRC | 22:22 | |
*** StevenK has joined #openstack-nova | 22:22 | |
*** babababa_ has joined #openstack-nova | 22:22 | |
*** otter768 has quit IRC | 22:23 | |
*** gridinv has quit IRC | 22:24 | |
*** boris-42 has joined #openstack-nova | 22:24 | |
*** gridinv has joined #openstack-nova | 22:25 | |
*** babababa has quit IRC | 22:25 | |
*** babababa_ is now known as babababa | 22:25 | |
*** jwcroppe has quit IRC | 22:25 | |
*** zhangjn has quit IRC | 22:26 | |
*** jwcroppe has joined #openstack-nova | 22:26 | |
*** ccarmack has quit IRC | 22:26 | |
*** pkholkin has quit IRC | 22:26 | |
*** ccarmack has joined #openstack-nova | 22:26 | |
*** pkholkin has joined #openstack-nova | 22:26 | |
*** ndipanov has quit IRC | 22:26 | |
*** briancurtin has quit IRC | 22:26 | |
*** ndipanov has joined #openstack-nova | 22:27 | |
*** darrenc has quit IRC | 22:27 | |
*** achanda has quit IRC | 22:27 | |
*** briancurtin has joined #openstack-nova | 22:28 | |
*** babababa has quit IRC | 22:28 | |
*** frickler has quit IRC | 22:29 | |
*** frickler has joined #openstack-nova | 22:29 | |
*** babababa has joined #openstack-nova | 22:30 | |
*** darrenc has joined #openstack-nova | 22:30 | |
*** jwcroppe has quit IRC | 22:30 | |
*** babababa has quit IRC | 22:33 | |
*** vilobhmm11 has quit IRC | 22:36 | |
*** ctrath has quit IRC | 22:36 | |
*** babababa has joined #openstack-nova | 22:36 | |
*** vilobhmm11 has joined #openstack-nova | 22:36 | |
*** babababa has quit IRC | 22:39 | |
*** _diana_ has quit IRC | 22:43 | |
*** zhangjn has joined #openstack-nova | 22:43 | |
*** zhangjn has quit IRC | 22:44 | |
*** zhangjn has joined #openstack-nova | 22:45 | |
*** baoli has quit IRC | 22:47 | |
*** su_zhang has joined #openstack-nova | 22:51 | |
*** ijuwang has quit IRC | 22:51 | |
*** obedmr has quit IRC | 22:54 | |
*** alejandrito has quit IRC | 22:55 | |
*** _diana_ has joined #openstack-nova | 22:56 | |
*** annegentle has quit IRC | 22:56 | |
*** tongli has quit IRC | 22:58 | |
*** wanghua has quit IRC | 23:00 | |
*** leitan has quit IRC | 23:00 | |
*** angdraug has joined #openstack-nova | 23:02 | |
*** mcarden has quit IRC | 23:02 | |
*** mcarden has joined #openstack-nova | 23:03 | |
*** zzzeek has quit IRC | 23:08 | |
*** asettle has quit IRC | 23:10 | |
*** ZZelle_ has quit IRC | 23:10 | |
*** achanda has joined #openstack-nova | 23:12 | |
*** vilobhmm11 has quit IRC | 23:17 | |
*** vilobhmm11 has joined #openstack-nova | 23:18 | |
*** dims has quit IRC | 23:18 | |
*** vilobhmm111 has joined #openstack-nova | 23:19 | |
*** zzzeek has joined #openstack-nova | 23:22 | |
*** vilobhmm11 has quit IRC | 23:22 | |
*** stevemar has quit IRC | 23:23 | |
*** salv-orlando has quit IRC | 23:25 | |
*** artom has quit IRC | 23:30 | |
*** darrenc is now known as darrenc_afk | 23:30 | |
*** stevemar has joined #openstack-nova | 23:32 | |
*** emagana has quit IRC | 23:33 | |
*** stevemar has quit IRC | 23:36 | |
*** markvoelker has quit IRC | 23:40 | |
*** su_zhang has quit IRC | 23:43 | |
*** darrenc_afk is now known as darrenc | 23:46 | |
*** otter768 has joined #openstack-nova | 23:47 | |
*** electrocucaracha has quit IRC | 23:50 | |
*** rushil has joined #openstack-nova | 23:54 | |
*** dims has joined #openstack-nova | 23:54 | |
*** dims has quit IRC | 23:55 | |
*** Guest18166 has joined #openstack-nova | 23:55 | |
*** julien_ZTE has quit IRC | 23:55 | |
*** tjones has joined #openstack-nova | 23:56 | |
*** tjones has left #openstack-nova | 23:56 | |
*** tonytan4ever has joined #openstack-nova | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!