| *** haomaiwa_ has quit IRC | 00:01 | |
| *** haomaiwang has joined #openstack-nova | 00:01 | |
| *** davideagnello has joined #openstack-nova | 00:03 | |
| *** zhangjn has joined #openstack-nova | 00:03 | |
| *** mc_nair has joined #openstack-nova | 00:07 | |
| *** heyongli has quit IRC | 00:09 | |
| *** heyongli has joined #openstack-nova | 00:09 | |
| *** pratikmallya has joined #openstack-nova | 00:11 | |
| *** su_zhang has quit IRC | 00:11 | |
| *** Marga__ has quit IRC | 00:12 | |
| *** Marga_ has joined #openstack-nova | 00:13 | |
| *** Marga_ has quit IRC | 00:17 | |
| *** Marga_ has joined #openstack-nova | 00:17 | |
| *** takedakn has joined #openstack-nova | 00:19 | |
| *** heyongli has quit IRC | 00:19 | |
| *** heyongli has joined #openstack-nova | 00:20 | |
| *** ijw has joined #openstack-nova | 00:21 | |
| *** vilobhmm11 has quit IRC | 00:21 | |
| alex_xu_ | gmann: thanks :) | 00:21 |
|---|---|---|
| *** zhangjn has quit IRC | 00:22 | |
| alex_xu_ | jaypipes: will review today | 00:22 |
| *** EinstCrazy has quit IRC | 00:22 | |
| *** tonytan4ever has joined #openstack-nova | 00:26 | |
| *** Jeffrey4l has joined #openstack-nova | 00:27 | |
| *** diogogmt has joined #openstack-nova | 00:29 | |
| *** heyongli has quit IRC | 00:29 | |
| *** heyongli has joined #openstack-nova | 00:30 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: Add project-id and user-id when list server-groups https://review.openstack.org/209018 | 00:38 |
| *** Jeffrey4l has quit IRC | 00:38 | |
| *** heyongli has quit IRC | 00:39 | |
| *** heyongli has joined #openstack-nova | 00:40 | |
| *** su_zhang has joined #openstack-nova | 00:41 | |
| *** eglynn has quit IRC | 00:45 | |
| *** jamielennox is now known as jamielennox|away | 00:46 | |
| *** su_zhang has quit IRC | 00:46 | |
| *** xyang1 has quit IRC | 00:47 | |
| *** jamielennox|away is now known as jamielennox | 00:47 | |
| *** daemontool has quit IRC | 00:48 | |
| *** daemontool has joined #openstack-nova | 00:48 | |
| *** heyongli has quit IRC | 00:50 | |
| *** heyongli has joined #openstack-nova | 00:50 | |
| *** Piet has joined #openstack-nova | 00:51 | |
| *** ijw has quit IRC | 00:52 | |
| *** dims has joined #openstack-nova | 00:52 | |
| *** hemna is now known as hemnafk | 00:53 | |
| *** mc_nair has quit IRC | 00:57 | |
| *** jamielennox is now known as jamielennox|away | 00:58 | |
| *** heyongli has quit IRC | 01:00 | |
| *** haomaiwang has quit IRC | 01:01 | |
| *** zhangjn has joined #openstack-nova | 01:01 | |
| *** heyongli has joined #openstack-nova | 01:01 | |
| *** haomaiwang has joined #openstack-nova | 01:01 | |
| *** EinstCrazy has joined #openstack-nova | 01:02 | |
| *** stackdump has quit IRC | 01:02 | |
| *** arunkant_ has joined #openstack-nova | 01:03 | |
| arunkant_ | Hi...looking for information on running nova unit tests ...keep getting this error with pydev "error: cannot switch to a different thread" . Seems related to eventlet threading. | 01:05 |
| *** jamielennox|away is now known as jamielennox | 01:06 | |
| *** su_zhang has joined #openstack-nova | 01:10 | |
| *** heyongli has quit IRC | 01:10 | |
| *** heyongli has joined #openstack-nova | 01:11 | |
| *** ssurana has quit IRC | 01:16 | |
| *** ssurana has joined #openstack-nova | 01:16 | |
| *** dims has quit IRC | 01:17 | |
| *** Daisy has joined #openstack-nova | 01:17 | |
| *** Daisy has quit IRC | 01:19 | |
| *** Daisy has joined #openstack-nova | 01:20 | |
| *** heyongli has quit IRC | 01:20 | |
| *** otter768 has joined #openstack-nova | 01:20 | |
| eliqiao | hi all, any changes made to nova v2.1 , I am adding a new keypair to nova, but failed and api told me Keypair data is invalid: failed to generate fingerprint | 01:21 |
| *** ssurana has quit IRC | 01:21 | |
| *** heyongli has joined #openstack-nova | 01:21 | |
| *** gjayavel_ has quit IRC | 01:22 | |
| *** arunkant_ has quit IRC | 01:23 | |
| *** otter768 has quit IRC | 01:25 | |
| *** Jeffrey4l has joined #openstack-nova | 01:26 | |
| *** electrocucaracha has quit IRC | 01:27 | |
| *** electrocucaracha has joined #openstack-nova | 01:27 | |
| *** Daisy has quit IRC | 01:28 | |
| *** pratikmallya has quit IRC | 01:30 | |
| *** heyongli has quit IRC | 01:31 | |
| *** heyongli has joined #openstack-nova | 01:31 | |
| *** Nathan has joined #openstack-nova | 01:34 | |
| *** Daisy has joined #openstack-nova | 01:34 | |
| *** rfolco has joined #openstack-nova | 01:35 | |
| *** Daisy has quit IRC | 01:39 | |
| *** heyongli has quit IRC | 01:41 | |
| *** salv-orlando has joined #openstack-nova | 01:41 | |
| *** heyongli has joined #openstack-nova | 01:42 | |
| *** Nathan has quit IRC | 01:43 | |
| *** Kevin_Zheng has joined #openstack-nova | 01:43 | |
| *** wverdugo500 has quit IRC | 01:44 | |
| *** Nathan has joined #openstack-nova | 01:45 | |
| *** mdrabe has joined #openstack-nova | 01:45 | |
| junjie | hi all, i am new here | 01:46 |
| *** zhenguo has joined #openstack-nova | 01:50 | |
| *** su_zhang has quit IRC | 01:50 | |
| *** heyongli has quit IRC | 01:51 | |
| *** pratikmallya has joined #openstack-nova | 01:52 | |
| *** heyongli has joined #openstack-nova | 01:52 | |
| eliqiao | junjie: hi welcome to nova. | 01:53 |
| *** tobe has joined #openstack-nova | 01:55 | |
| *** otter768 has joined #openstack-nova | 01:57 | |
| *** salv-orlando has quit IRC | 01:59 | |
| alex_xu_ | eliqiao: you need upgrade your python-novclient I guess | 02:00 |
| alex_xu_ | junjie: welcome~ | 02:00 |
| junjie | thanks | 02:00 |
| *** pixelbeat has quit IRC | 02:00 | |
| eliqiao | alex_xu_: it's so strange, I just use pdb do some debuging , it passed. | 02:00 |
| *** takedakn has quit IRC | 02:01 | |
| *** haomaiwang has quit IRC | 02:01 | |
| *** haomaiwang has joined #openstack-nova | 02:01 | |
| *** heyongli has quit IRC | 02:01 | |
| *** Daisy has joined #openstack-nova | 02:02 | |
| *** heyongli has joined #openstack-nova | 02:02 | |
| *** Daisy has quit IRC | 02:09 | |
| *** Daisy has joined #openstack-nova | 02:09 | |
| *** breitz2 has quit IRC | 02:11 | |
| *** breitz has joined #openstack-nova | 02:12 | |
| *** heyongli has quit IRC | 02:12 | |
| *** heyongli has joined #openstack-nova | 02:12 | |
| *** seanw1 has quit IRC | 02:12 | |
| *** takedakn has joined #openstack-nova | 02:13 | |
| *** seanw1 has joined #openstack-nova | 02:13 | |
| *** mrkz has quit IRC | 02:13 | |
| *** Daisy has quit IRC | 02:14 | |
| *** unicell has quit IRC | 02:17 | |
| *** heyongli has quit IRC | 02:22 | |
| *** heyongli has joined #openstack-nova | 02:23 | |
| *** browne has quit IRC | 02:23 | |
| *** ndipanov has joined #openstack-nova | 02:23 | |
| *** klkumar has joined #openstack-nova | 02:25 | |
| *** salv-orlando has joined #openstack-nova | 02:27 | |
| *** Daisy has joined #openstack-nova | 02:28 | |
| *** gyee has quit IRC | 02:31 | |
| *** heyongli has quit IRC | 02:32 | |
| *** heyongli has joined #openstack-nova | 02:33 | |
| *** Jeffrey4l has quit IRC | 02:33 | |
| *** Jeffrey4l has joined #openstack-nova | 02:34 | |
| *** mtanino has quit IRC | 02:34 | |
| *** gongysh has quit IRC | 02:35 | |
| *** gongysh has joined #openstack-nova | 02:35 | |
| *** yonglihe has joined #openstack-nova | 02:40 | |
| *** heyongli has quit IRC | 02:42 | |
| *** Daisy has quit IRC | 02:42 | |
| *** heyongli has joined #openstack-nova | 02:44 | |
| *** rfolco has quit IRC | 02:44 | |
| *** otter768 has quit IRC | 02:44 | |
| *** inteq has joined #openstack-nova | 02:45 | |
| *** salv-orlando has quit IRC | 02:49 | |
| *** Marga_ has quit IRC | 02:50 | |
| *** Marga_ has joined #openstack-nova | 02:51 | |
| *** heyongli has quit IRC | 02:53 | |
| *** heyongli has joined #openstack-nova | 02:54 | |
| *** pratikmallya has quit IRC | 02:54 | |
| *** Marga_ has quit IRC | 02:55 | |
| *** venkat_p has joined #openstack-nova | 03:00 | |
| *** heyongli has quit IRC | 03:00 | |
| *** haomaiwang has quit IRC | 03:01 | |
| *** ccarmack has joined #openstack-nova | 03:01 | |
| *** haomaiwang has joined #openstack-nova | 03:02 | |
| *** mdrabe has quit IRC | 03:03 | |
| *** Daisy has joined #openstack-nova | 03:04 | |
| *** Daisy has quit IRC | 03:08 | |
| *** jwcroppe has quit IRC | 03:16 | |
| *** ndipanov has quit IRC | 03:16 | |
| *** salv-orlando has joined #openstack-nova | 03:19 | |
| *** baoli has joined #openstack-nova | 03:19 | |
| *** otter768 has joined #openstack-nova | 03:22 | |
| openstackgerrit | venkatesh proposed openstack/nova: make the driver.Scheduler as abstract class https://review.openstack.org/212016 | 03:23 |
| *** browne has joined #openstack-nova | 03:24 | |
| *** boris-42 has joined #openstack-nova | 03:24 | |
| *** Jeffrey4l has quit IRC | 03:25 | |
| *** hparekh has quit IRC | 03:26 | |
| *** Daisy has joined #openstack-nova | 03:32 | |
| *** Jeffrey4l has joined #openstack-nova | 03:32 | |
| *** gjayavelu has joined #openstack-nova | 03:34 | |
| *** baoli has quit IRC | 03:36 | |
| *** Daisy has quit IRC | 03:37 | |
| *** gjayavelu has quit IRC | 03:38 | |
| *** hparekh has joined #openstack-nova | 03:39 | |
| *** salv-orlando has quit IRC | 03:40 | |
| *** haomaiwang has quit IRC | 03:51 | |
| *** Daisy has joined #openstack-nova | 03:54 | |
| *** jyuso1 has joined #openstack-nova | 03:55 | |
| *** baoli has joined #openstack-nova | 03:57 | |
| *** gcb has quit IRC | 03:57 | |
| *** Daisy has quit IRC | 03:57 | |
| *** Sree has joined #openstack-nova | 03:58 | |
| *** links has joined #openstack-nova | 03:59 | |
| *** stackdump has joined #openstack-nova | 04:00 | |
| *** yonglihe has quit IRC | 04:00 | |
| *** stackdump has quit IRC | 04:04 | |
| *** klkumar has quit IRC | 04:05 | |
| *** Nathan__ has joined #openstack-nova | 04:06 | |
| *** klkumar has joined #openstack-nova | 04:08 | |
| *** Nathan has quit IRC | 04:09 | |
| *** Nathan__ is now known as Nathan | 04:09 | |
| *** emagana has quit IRC | 04:10 | |
| *** jwcroppe has joined #openstack-nova | 04:17 | |
| *** subscope has joined #openstack-nova | 04:18 | |
| *** fawadkhaliq has joined #openstack-nova | 04:20 | |
| *** gjayavelu has joined #openstack-nova | 04:24 | |
| *** otter768 has quit IRC | 04:29 | |
| *** electrocucaracha has quit IRC | 04:30 | |
| *** electrocucaracha has joined #openstack-nova | 04:30 | |
| *** emagana has joined #openstack-nova | 04:33 | |
| openstackgerrit | Tang Chen proposed openstack/nova-specs: Support triggering crash dump in an instance https://review.openstack.org/229255 | 04:38 |
| *** salv-orlando has joined #openstack-nova | 04:40 | |
| *** emagana has quit IRC | 04:41 | |
| *** wuhg has joined #openstack-nova | 04:43 | |
| *** jwcroppe has quit IRC | 04:46 | |
| *** snikitin1 has joined #openstack-nova | 04:48 | |
| *** salv-orlando has quit IRC | 04:49 | |
| *** amotoki has joined #openstack-nova | 04:51 | |
| *** baoli has quit IRC | 04:51 | |
| *** samos123 has joined #openstack-nova | 04:52 | |
| *** jnclyz12483 has joined #openstack-nova | 04:54 | |
| *** vilobhmm11 has joined #openstack-nova | 04:59 | |
| *** samos123 has quit IRC | 04:59 | |
| *** jwcroppe has joined #openstack-nova | 05:01 | |
| *** subscope has quit IRC | 05:02 | |
| *** electrocucaracha has quit IRC | 05:02 | |
| *** vilobhmm11 has quit IRC | 05:02 | |
| *** gcb has joined #openstack-nova | 05:05 | |
| *** subscope has joined #openstack-nova | 05:05 | |
| *** samos123 has joined #openstack-nova | 05:07 | |
| *** samos123 has quit IRC | 05:07 | |
| *** zhangjn has quit IRC | 05:09 | |
| *** Marga_ has joined #openstack-nova | 05:14 | |
| *** Marga_ has quit IRC | 05:14 | |
| *** Marga_ has joined #openstack-nova | 05:15 | |
| *** zhangjn has joined #openstack-nova | 05:16 | |
| openstackgerrit | junxu proposed openstack/nova-specs: Allow force stop server in nova rest api https://review.openstack.org/195418 | 05:16 |
| *** jerrygb_ has quit IRC | 05:18 | |
| *** vilobhmm11 has joined #openstack-nova | 05:20 | |
| *** Samos123 has joined #openstack-nova | 05:20 | |
| *** pratikmallya has joined #openstack-nova | 05:26 | |
| *** zhangjn has quit IRC | 05:29 | |
| *** venkat_p has left #openstack-nova | 05:30 | |
| *** pratikmallya has quit IRC | 05:31 | |
| *** zzzeek has quit IRC | 05:34 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: doc: update threading.rst https://review.openstack.org/199013 | 05:35 |
| *** Samos123 has quit IRC | 05:36 | |
| *** heyongli has joined #openstack-nova | 05:40 | |
| *** zzzeek has joined #openstack-nova | 05:42 | |
| *** gjayavelu has left #openstack-nova | 05:43 | |
| *** gjayavelu has quit IRC | 05:43 | |
| *** heyongli has quit IRC | 05:44 | |
| *** heyongli has joined #openstack-nova | 05:44 | |
| *** Samos123 has joined #openstack-nova | 05:46 | |
| *** zzzeek has quit IRC | 05:47 | |
| *** nikhil_k has joined #openstack-nova | 05:48 | |
| *** nikhil has quit IRC | 05:48 | |
| *** zzzeek has joined #openstack-nova | 05:50 | |
| *** heyongli has quit IRC | 05:54 | |
| *** heyongli has joined #openstack-nova | 05:55 | |
| *** zzzeek has quit IRC | 05:58 | |
| *** klkumar has quit IRC | 06:00 | |
| *** oomichi_ has quit IRC | 06:03 | |
| *** heyongli has quit IRC | 06:05 | |
| *** jinxing has joined #openstack-nova | 06:05 | |
| *** heyongli has joined #openstack-nova | 06:05 | |
| *** zhangjn has joined #openstack-nova | 06:08 | |
| *** Nathan has quit IRC | 06:09 | |
| openstackgerrit | Andrew Bogott proposed openstack/nova: Add disk_minimum_physical_free_space_ratio option to disk filter. https://review.openstack.org/242251 | 06:12 |
| *** chinmaya has joined #openstack-nova | 06:12 | |
| *** cfriesen__ has quit IRC | 06:12 | |
| *** heyongli has quit IRC | 06:15 | |
| *** heyongli has joined #openstack-nova | 06:15 | |
| *** vilobhmm111 has joined #openstack-nova | 06:18 | |
| *** edtubill has joined #openstack-nova | 06:19 | |
| *** vilobhmm11 has quit IRC | 06:19 | |
| *** haomaiwang has joined #openstack-nova | 06:20 | |
| *** heyongli has quit IRC | 06:21 | |
| *** heyongli has joined #openstack-nova | 06:21 | |
| *** Jeffrey4l has quit IRC | 06:24 | |
| *** unicell has joined #openstack-nova | 06:24 | |
| *** heyongli has quit IRC | 06:25 | |
| *** heyongli has joined #openstack-nova | 06:25 | |
| *** subscope has quit IRC | 06:26 | |
| *** TakashiN has joined #openstack-nova | 06:27 | |
| *** heyongli has quit IRC | 06:27 | |
| *** yonglihe has joined #openstack-nova | 06:27 | |
| *** Nathan has joined #openstack-nova | 06:28 | |
| *** salv-orlando has joined #openstack-nova | 06:29 | |
| *** otter768 has joined #openstack-nova | 06:30 | |
| *** TakashiN has left #openstack-nova | 06:30 | |
| *** TakashiN has joined #openstack-nova | 06:31 | |
| *** openstackgerrit has quit IRC | 06:31 | |
| *** TakashiN has quit IRC | 06:31 | |
| *** openstackgerrit has joined #openstack-nova | 06:32 | |
| *** chinmaya has quit IRC | 06:34 | |
| *** otter768 has quit IRC | 06:34 | |
| *** yonglihe has quit IRC | 06:35 | |
| *** Samos123 has quit IRC | 06:35 | |
| *** yonglihe has joined #openstack-nova | 06:36 | |
| *** Jeffrey4l has joined #openstack-nova | 06:36 | |
| *** david-lyle has joined #openstack-nova | 06:40 | |
| *** salv-orlando has quit IRC | 06:41 | |
| *** tonytan4ever has quit IRC | 06:44 | |
| *** edtubill has quit IRC | 06:44 | |
| *** david-lyle has quit IRC | 06:45 | |
| *** yonglihe has quit IRC | 06:46 | |
| *** yonglihe has joined #openstack-nova | 06:46 | |
| *** praneshp has joined #openstack-nova | 06:47 | |
| praneshp | jaypipes: ping. I was looking for some help on https://github.com/openstack/cookbook-openstack-common, what’s the right irc for that? Saw your name on the author list | 06:48 |
| *** vilobhmm111 has quit IRC | 06:54 | |
| *** yonglihe has quit IRC | 06:56 | |
| *** yonglihe has joined #openstack-nova | 06:56 | |
| *** jamielennox is now known as jamielennox|away | 06:59 | |
| *** sudipto has joined #openstack-nova | 06:59 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** Marga_ has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-nova | 07:01 | |
| *** lsmola has joined #openstack-nova | 07:02 | |
| *** baoli has joined #openstack-nova | 07:03 | |
| *** unicell1 has joined #openstack-nova | 07:03 | |
| *** unicell has quit IRC | 07:04 | |
| *** josecastroleon has joined #openstack-nova | 07:05 | |
| *** yonglihe has quit IRC | 07:06 | |
| *** yonglihe has joined #openstack-nova | 07:06 | |
| *** baoli has quit IRC | 07:08 | |
| *** Jeffrey4l has quit IRC | 07:09 | |
| *** e0ne has joined #openstack-nova | 07:10 | |
| *** yonglihe has quit IRC | 07:16 | |
| *** yonglihe has joined #openstack-nova | 07:17 | |
| *** salv-orlando has joined #openstack-nova | 07:17 | |
| *** nkrinner has joined #openstack-nova | 07:18 | |
| *** fawadkhaliq has quit IRC | 07:19 | |
| *** salv-orlando has quit IRC | 07:19 | |
| *** salv-orlando has joined #openstack-nova | 07:20 | |
| *** e0ne has quit IRC | 07:21 | |
| *** Jeffrey4l has joined #openstack-nova | 07:21 | |
| openstackgerrit | lalit dagre proposed openstack/nova: Flavor object which support both DB with active migration https://review.openstack.org/213041 | 07:25 |
| openstackgerrit | lalit dagre proposed openstack/nova: Creates flavor* tables in API database https://review.openstack.org/201606 | 07:25 |
| *** Jeffrey4l has quit IRC | 07:26 | |
| *** yonglihe has quit IRC | 07:27 | |
| *** yonglihe has joined #openstack-nova | 07:27 | |
| *** takedakn has quit IRC | 07:28 | |
| *** snikitin1 has quit IRC | 07:33 | |
| *** fawadkhaliq has joined #openstack-nova | 07:34 | |
| *** Daisy has joined #openstack-nova | 07:36 | |
| *** romainh has joined #openstack-nova | 07:36 | |
| *** yonglihe has quit IRC | 07:37 | |
| *** yonglihe has joined #openstack-nova | 07:37 | |
| *** Jeffrey4l has joined #openstack-nova | 07:40 | |
| *** takedakn has joined #openstack-nova | 07:42 | |
| *** alex_klimov has joined #openstack-nova | 07:43 | |
| *** sahid has joined #openstack-nova | 07:44 | |
| *** rk4n has joined #openstack-nova | 07:44 | |
| *** yonglihe has quit IRC | 07:47 | |
| *** yonglihe has joined #openstack-nova | 07:47 | |
| *** jichen has joined #openstack-nova | 07:48 | |
| *** Jeffrey4l has quit IRC | 07:49 | |
| bauzas | good morning Nova | 07:50 |
| bauzas | oops, my manners | 07:50 |
| *** bauzas is now known as bauwser | 07:51 | |
| *** fawadkhaliq has quit IRC | 07:52 | |
| *** Guest65445 is now known as tdurakov | 07:56 | |
| *** yonglihe has quit IRC | 07:57 | |
| *** yonglihe has joined #openstack-nova | 07:58 | |
| *** fawadkhaliq has joined #openstack-nova | 07:58 | |
| *** ihrachys has joined #openstack-nova | 07:58 | |
| *** pratikmallya has joined #openstack-nova | 08:00 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-nova | 08:01 | |
| *** rdopiera has joined #openstack-nova | 08:05 | |
| *** praneshp_ has joined #openstack-nova | 08:05 | |
| *** RuiChen has joined #openstack-nova | 08:05 | |
| *** praneshp has quit IRC | 08:06 | |
| *** praneshp_ is now known as praneshp | 08:06 | |
| *** yonglihe has quit IRC | 08:08 | |
| *** markmc has joined #openstack-nova | 08:08 | |
| *** yonglihe has joined #openstack-nova | 08:08 | |
| openstackgerrit | Raga lahari proposed openstack/nova: Support for new --status parameter to "nova-manage vm list" command https://review.openstack.org/242370 | 08:09 |
| openstackgerrit | lvdongbing proposed openstack/nova: Add oslo_config PortOpt support https://review.openstack.org/239569 | 08:09 |
| *** matrohon has joined #openstack-nova | 08:11 | |
| *** haomaiwang has quit IRC | 08:11 | |
| *** josecastroleon has quit IRC | 08:12 | |
| *** yonglihe has quit IRC | 08:18 | |
| *** yonglihe has joined #openstack-nova | 08:18 | |
| *** abhishekk has joined #openstack-nova | 08:20 | |
| alex_xu_ | bauwser: morning | 08:22 |
| tangchen_ | Hi there, :) | 08:23 |
| tangchen_ | Do you guys have time to discuss about how to deprecate host maintenance mode in nova ? | 08:24 |
| *** Nathan__ has joined #openstack-nova | 08:25 | |
| tangchen_ | 1. There is an API. 2. Xen is using it. So I don't know how to deprecate it. | 08:25 |
| *** kiran-r has joined #openstack-nova | 08:27 | |
| *** Nathan has quit IRC | 08:27 | |
| *** Nathan__ is now known as Nathan | 08:27 | |
| *** yonglihe has quit IRC | 08:28 | |
| *** yonglihe has joined #openstack-nova | 08:28 | |
| *** otter768 has joined #openstack-nova | 08:31 | |
| *** rk4n has quit IRC | 08:32 | |
| *** rk4n has joined #openstack-nova | 08:34 | |
| *** otter768 has quit IRC | 08:36 | |
| *** jinxing has quit IRC | 08:36 | |
| *** agireud has quit IRC | 08:37 | |
| *** zigo has quit IRC | 08:37 | |
| *** agireud has joined #openstack-nova | 08:37 | |
| *** praneshp has quit IRC | 08:37 | |
| *** Daisy has quit IRC | 08:38 | |
| *** jinxing has joined #openstack-nova | 08:38 | |
| *** ZZelle has quit IRC | 08:38 | |
| *** yonglihe has quit IRC | 08:38 | |
| *** ZZelle has joined #openstack-nova | 08:38 | |
| *** Daisy has joined #openstack-nova | 08:38 | |
| *** yonglihe has joined #openstack-nova | 08:39 | |
| *** zigo has joined #openstack-nova | 08:42 | |
| *** openstackstatus has quit IRC | 08:42 | |
| *** Daisy has quit IRC | 08:43 | |
| *** lpetrut has joined #openstack-nova | 08:43 | |
| *** jlanoux has joined #openstack-nova | 08:44 | |
| *** gszasz has joined #openstack-nova | 08:46 | |
| *** takedakn has quit IRC | 08:48 | |
| *** yonglihe has quit IRC | 08:48 | |
| *** yonglihe has joined #openstack-nova | 08:49 | |
| *** Daisy has joined #openstack-nova | 08:52 | |
| *** jwcroppe has quit IRC | 08:53 | |
| *** jwcroppe has joined #openstack-nova | 08:54 | |
| *** Daisy has quit IRC | 08:54 | |
| *** Nathan has quit IRC | 08:54 | |
| *** browne has quit IRC | 08:55 | |
| *** subscope has joined #openstack-nova | 08:57 | |
| *** subscope has quit IRC | 08:58 | |
| *** yonglihe has quit IRC | 08:58 | |
| *** jwcroppe has quit IRC | 08:58 | |
| *** yassine has joined #openstack-nova | 08:59 | |
| *** yonglihe has joined #openstack-nova | 08:59 | |
| *** dgonzalez has quit IRC | 09:03 | |
| *** dgonzalez has joined #openstack-nova | 09:03 | |
| *** jichen has quit IRC | 09:07 | |
| *** yonglihe has quit IRC | 09:07 | |
| *** jinxing_ has joined #openstack-nova | 09:08 | |
| *** josecastroleon has joined #openstack-nova | 09:08 | |
| *** jinxing has quit IRC | 09:08 | |
| *** deepakcs has joined #openstack-nova | 09:09 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: glance:add helper method to get client version https://review.openstack.org/222150 | 09:10 |
| *** links has quit IRC | 09:11 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: Use EngineFacade from oslo_db.enginefacade https://review.openstack.org/221310 | 09:11 |
| *** derekh has joined #openstack-nova | 09:11 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'instance_info' and 'instance_extra' https://review.openstack.org/225065 | 09:15 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'cell_mapping' https://review.openstack.org/226788 | 09:15 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'instance_mapping' https://review.openstack.org/228979 | 09:15 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'network' https://review.openstack.org/240253 | 09:15 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'key_pair' and 'cell' https://review.openstack.org/239915 | 09:15 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'certificate' and 'pci_device' https://review.openstack.org/240275 | 09:16 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: WIP: enginefacade: 'floating_ip' https://review.openstack.org/240513 | 09:16 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: WIP: enginefacade: 'fixed_ip' https://review.openstack.org/240811 | 09:16 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'service' https://review.openstack.org/241178 | 09:16 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'host_mapping' https://review.openstack.org/242035 | 09:16 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'dnsdomain' and 'ec2' https://review.openstack.org/242019 | 09:16 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'compute_node' https://review.openstack.org/241189 | 09:16 |
| kashyap | tangchen_: I think most folks interested in it are in US east coast timezone | 09:18 |
| *** subscope has joined #openstack-nova | 09:18 | |
| *** jistr has joined #openstack-nova | 09:19 | |
| *** pixelbeat has joined #openstack-nova | 09:20 | |
| *** achanda has quit IRC | 09:21 | |
| *** e0ne has joined #openstack-nova | 09:27 | |
| *** jnclyz12483 has quit IRC | 09:31 | |
| johnthetubaguy | tangchen_: so I can cover some of this, I think we just add the deprecation log messages | 09:32 |
| johnthetubaguy | tangchen_: the API side of things is a little tricker | 09:32 |
| johnthetubaguy | tangchen_: its probably worth a very small spec to agree the details | 09:32 |
| *** Samos123 has joined #openstack-nova | 09:36 | |
| *** Sree has quit IRC | 09:36 | |
| *** Sree has joined #openstack-nova | 09:37 | |
| *** cdent has joined #openstack-nova | 09:38 | |
| *** haomaiwa_ has joined #openstack-nova | 09:40 | |
| *** jianghuaw has joined #openstack-nova | 09:41 | |
| *** rk4n has quit IRC | 09:41 | |
| *** haomaiwa_ has quit IRC | 09:45 | |
| openstackgerrit | Sam Yaple proposed openstack/nova: Fix attibute error when cloning raw images in Ceph https://review.openstack.org/237799 | 09:49 |
| *** njirap has joined #openstack-nova | 09:53 | |
| *** josecastroleon has quit IRC | 09:55 | |
| *** jwcroppe has joined #openstack-nova | 09:55 | |
| *** tobe has quit IRC | 09:55 | |
| *** josecastroleon has joined #openstack-nova | 09:56 | |
| *** zhenguo has quit IRC | 09:58 | |
| tangchen_ | kashyap, johnthetubaguy: OK, I know. I agree we cannot simply remove all the source code. I'll try to post a BP and a spec to list up all the things I can think of. | 09:58 |
| johnthetubaguy | tangchen_: sounds great, thank you, spec is a good way forward, I totally support the idea, just need to think through some of the details | 09:59 |
| *** zhangjn has quit IRC | 10:01 | |
| *** zhangjn has joined #openstack-nova | 10:03 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: test: use enforce_type=True parameter in method CONF.set_override https://review.openstack.org/242416 | 10:03 |
| *** zhangjn has quit IRC | 10:04 | |
| *** zhangjn has joined #openstack-nova | 10:08 | |
| *** kiran-r has quit IRC | 10:08 | |
| *** fawadkhaliq has quit IRC | 10:10 | |
| *** sudipto has quit IRC | 10:10 | |
| *** Samos123 has quit IRC | 10:11 | |
| *** mitz has quit IRC | 10:12 | |
| *** pedroalvarez has quit IRC | 10:12 | |
| *** pedroalvarez has joined #openstack-nova | 10:12 | |
| *** mitz has joined #openstack-nova | 10:14 | |
| *** abhishekk_ has joined #openstack-nova | 10:14 | |
| *** rajesht_ has joined #openstack-nova | 10:14 | |
| *** jyuso1 has quit IRC | 10:15 | |
| *** jyuso1 has joined #openstack-nova | 10:16 | |
| *** abhishekk has quit IRC | 10:16 | |
| *** rajesht has quit IRC | 10:16 | |
| *** huanxie has joined #openstack-nova | 10:19 | |
| *** tdurakov has quit IRC | 10:20 | |
| *** Guest13749 has joined #openstack-nova | 10:20 | |
| *** jwcroppe has quit IRC | 10:23 | |
| *** BadCub has quit IRC | 10:27 | |
| *** BadCub has joined #openstack-nova | 10:27 | |
| *** tpeoples has quit IRC | 10:28 | |
| *** tpeoples has joined #openstack-nova | 10:29 | |
| *** otter768 has joined #openstack-nova | 10:32 | |
| *** otter768 has quit IRC | 10:36 | |
| *** Daisy has joined #openstack-nova | 10:37 | |
| *** ihrachys has quit IRC | 10:37 | |
| *** ihrachys has joined #openstack-nova | 10:38 | |
| *** zhangjn has quit IRC | 10:40 | |
| *** EinstCrazy has quit IRC | 10:41 | |
| *** Daisy has quit IRC | 10:41 | |
| *** ferest has joined #openstack-nova | 10:43 | |
| *** danpb has joined #openstack-nova | 10:45 | |
| *** ferest has quit IRC | 10:47 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'dnsdomain' and 'ec2' https://review.openstack.org/242019 | 10:48 |
| *** jinxing_ has quit IRC | 10:48 | |
| *** jamielennox|away is now known as jamielennox | 10:49 | |
| *** jianghuaw has quit IRC | 10:50 | |
| *** Sree has quit IRC | 10:51 | |
| *** njirap has quit IRC | 10:57 | |
| *** mhickey has joined #openstack-nova | 10:58 | |
| openstackgerrit | Alexander Schmidt proposed openstack/nova-specs: Libvirt: AIO mode for disk devices https://review.openstack.org/232514 | 11:02 |
| *** aysyd has joined #openstack-nova | 11:04 | |
| PaulMurray | anyone know how to trigger recheck for check-citrix-xenserver ? | 11:05 |
| *** aix has quit IRC | 11:07 | |
| *** aysyd has quit IRC | 11:08 | |
| *** aysyd has joined #openstack-nova | 11:09 | |
| *** mmdurrant has quit IRC | 11:09 | |
| kashyap | 'recheck xen'? | 11:10 |
| *** romainh has quit IRC | 11:11 | |
| *** jamielennox is now known as jamielennox|away | 11:11 | |
| PaulMurray | kashyap, I think that does dsvm-tempest-xen | 11:13 |
| kashyap | Oops, sorry for the misinfo then. | 11:13 |
| PaulMurray | oh, I found it in the comment from the job (duh) it says xenserver: recheck | 11:14 |
| PaulMurray | thanks anyway | 11:14 |
| *** chinmaya has joined #openstack-nova | 11:21 | |
| *** achanda has joined #openstack-nova | 11:22 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: cells: Create BDM only when device_name is not null https://review.openstack.org/242447 | 11:26 |
| *** achanda has quit IRC | 11:26 | |
| *** mwagner has quit IRC | 11:27 | |
| *** dims has joined #openstack-nova | 11:27 | |
| Anticimex | what's the best way to utilize multiple rbd pools on one nova-compute? | 11:28 |
| Anticimex | i found a post from seb from september: http://www.sebastien-han.fr/blog/2015/09/15/openstack-nova-configure-multiple-ceph-backend-on-one-hypervisor/ | 11:28 |
| Anticimex | a bit interesting approach, is there no native way? | 11:28 |
| *** smatzek has joined #openstack-nova | 11:33 | |
| *** pratikmallya has quit IRC | 11:34 | |
| *** subscope has quit IRC | 11:41 | |
| *** ferest has joined #openstack-nova | 11:41 | |
| *** josecastroleon has quit IRC | 11:42 | |
| andrearosa | is there a way to change the assigne and drafter for a blueprint? the original author is not working on it and I have inherited his job. I'd like to make some changes to the blueprint. can some admin here change those fields or do I need to put a new blueprint? ref: https://blueprints.launchpad.net/nova/+spec/enhance-instance-operations-during-live-migration | 11:43 |
| *** subscope has joined #openstack-nova | 11:43 | |
| *** ferest has quit IRC | 11:46 | |
| bauwser | andrearosa: ask one of the LP nova-drivers, cc johnthetubaguy | 11:46 |
| johnthetubaguy | andrearosa: yeah, I can update that for you | 11:47 |
| bauwser | I'm getting crazy with the cells job | 11:47 |
| andrearosa | ok thanks | 11:47 |
| johnthetubaguy | andrearosa: not sure I got the right launchpad id, did that work? | 11:48 |
| *** Marga_ has joined #openstack-nova | 11:49 | |
| andrearosa | johnthetubaguy: nope it is assigned to caporosso and I am andrea-rosa-m | 11:49 |
| johnthetubaguy | andrearosa: that makes a lot more sense, launchpad search was thawting me | 11:49 |
| bauwser | I actually use the Foundation stuff | 11:50 |
| andrearosa | johnthetubaguy: worked now, thanks | 11:50 |
| johnthetubaguy | andrearosa: np | 11:50 |
| johnthetubaguy | andrearosa: are you getting a spec up for this one? | 11:50 |
| andrearosa | yes the spec is already up, I need to ref the BP in the commit | 11:51 |
| andrearosa | https://review.openstack.org/226199 | 11:51 |
| *** zhangjn has joined #openstack-nova | 11:53 | |
| *** EinstCrazy has joined #openstack-nova | 11:53 | |
| *** vladikr has quit IRC | 11:53 | |
| *** jerrygb has joined #openstack-nova | 11:54 | |
| PaulMurray | johnthetubaguy, can we talk about https://blueprints.launchpad.net/nova/+spec/volume-ops-when-shelved for a moment | 11:54 |
| johnthetubaguy | andrearosa: cools | 11:54 |
| johnthetubaguy | PaulMurray: sure | 11:54 |
| PaulMurray | I put a question on the mail list - seems to me the only way to synchronize the operations is | 11:55 |
| PaulMurray | using task states | 11:55 |
| PaulMurray | does that seem reasonable? | 11:55 |
| PaulMurray | At least for shevled_offloaded | 11:56 |
| *** ferest has joined #openstack-nova | 11:57 | |
| PaulMurray | johnthetubaguy, if so my plan would be to make that part of the spec for this and then make the detach boot volume one depend on it | 11:58 |
| *** sudipto has joined #openstack-nova | 11:58 | |
| *** sudipto has quit IRC | 11:59 | |
| johnthetubaguy | PaulMurray: so maybe, not sure I get all the implications | 12:02 |
| johnthetubaguy | PaulMurray: we do a lot of synchronisation on hidden task states, but yeah, need to think through that a bit more | 12:03 |
| *** RuiChen has quit IRC | 12:03 | |
| PaulMurray | johnthetubaguy, at the moment in compute manager the volume attach/detach ops are synchronised on instance id, so there are no task states associated | 12:04 |
| *** RuiChen has joined #openstack-nova | 12:04 | |
| PaulMurray | and operations will queue and take turn | 12:04 |
| PaulMurray | they are asynchronous | 12:04 |
| PaulMurray | if done in the API they would be done synchronously with the API calls | 12:05 |
| *** baoli has joined #openstack-nova | 12:05 | |
| PaulMurray | If using task states (like detaching, attaching) to make sure only one happens at a time | 12:05 |
| PaulMurray | a second one would fail and return error instead of queing | 12:06 |
| *** josecastroleon has joined #openstack-nova | 12:06 | |
| PaulMurray | so the API would de different in that respect | 12:06 |
| PaulMurray | But only for shelved_offloaded | 12:06 |
| *** rk4n has joined #openstack-nova | 12:07 | |
| PaulMurray | (this is all a bit more involved than I had hoped for - but there you go). | 12:07 |
| johnthetubaguy | PaulMurray: OK, I am missing a bit, why do we need that when we are shelved_offloaded? | 12:09 |
| *** baoli has quit IRC | 12:09 | |
| *** rfolco has joined #openstack-nova | 12:09 | |
| PaulMurray | The proposal for detaching boot volumes is to only do it when shelved | 12:10 |
| johnthetubaguy | PaulMurray: yeah, I am cool with that bit | 12:10 |
| johnthetubaguy | PaulMurray: but why do we need the task states to synchronise the attach/detach again? | 12:10 |
| PaulMurray | I need some synchronization around it | 12:11 |
| PaulMurray | otherwise its possible for two attach boot volume operations to collide | 12:11 |
| *** ferest has quit IRC | 12:11 | |
| PaulMurray | (maybe) | 12:11 |
| PaulMurray | It would also be good to allow any volume to be attached and detach if we can do it with boot volumes | 12:12 |
| PaulMurray | (again - maybe) | 12:12 |
| PaulMurray | I may be able to synchronize the boot volume on its own bdm state (the bdm will exist) | 12:13 |
| PaulMurray | But that only works for a boot volume | 12:13 |
| *** Daisy has joined #openstack-nova | 12:13 | |
| PaulMurray | BTW detaching just works - its only attaching that is a problem for other volumes | 12:14 |
| *** Daisy has quit IRC | 12:14 | |
| *** romainh has joined #openstack-nova | 12:16 | |
| *** jwcroppe has joined #openstack-nova | 12:19 | |
| *** jichen has joined #openstack-nova | 12:19 | |
| *** IanGovett has joined #openstack-nova | 12:20 | |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Remove flavors.get_all_flavors() only used in tests https://review.openstack.org/242465 | 12:22 |
| *** Sree has joined #openstack-nova | 12:22 | |
| johnthetubaguy | PaulMurray: oh, now I see, because it doesn't go to the compute, we can't do the simple locking on instance-uuid, I get your point | 12:23 |
| johnthetubaguy | PaulMurray: yeah, some hidden task state makes sense, like attaching-boot-volume | 12:23 |
| *** aysyd has quit IRC | 12:23 | |
| *** jwcroppe has quit IRC | 12:23 | |
| *** aysyd has joined #openstack-nova | 12:24 | |
| *** achanda has joined #openstack-nova | 12:24 | |
| PaulMurray | johnthetubaguy, I'm going to write the spec and see what people think - if I can do the general case I think the boot from volume will be an extension of it | 12:26 |
| *** rajesht_ has quit IRC | 12:26 | |
| PaulMurray | thanks for being a sounding board | 12:26 |
| PaulMurray | johnthetubaguy, do you happen to know when ndipanov will be back online? | 12:28 |
| *** achanda has quit IRC | 12:30 | |
| PaulMurray | bauwser, do you know ? ^^^ | 12:30 |
| openstackgerrit | Zoltan Arnold Nagy proposed openstack/nova-specs: Encryption support for rbd-backed volumes https://review.openstack.org/239798 | 12:31 |
| *** wuhg has quit IRC | 12:31 | |
| *** otter768 has joined #openstack-nova | 12:32 | |
| johnthetubaguy | PaulMurray: I think he is away all this week on holiday, I was assuming Monday, but I don't know so | 12:32 |
| *** lpetrut has quit IRC | 12:34 | |
| johnthetubaguy | PaulMurray: I would rather keep existing attach/detach using the nova-compute local lock, that seems nice and simple. Its really just about the shelved case where there is no owner of the VM, I guess, although that sounds messy now I typed that out | 12:34 |
| *** rotbeard has joined #openstack-nova | 12:34 | |
| *** zigo has quit IRC | 12:36 | |
| *** yonglihe has joined #openstack-nova | 12:36 | |
| *** zigo has joined #openstack-nova | 12:37 | |
| *** otter768 has quit IRC | 12:37 | |
| *** weihan has joined #openstack-nova | 12:38 | |
| *** RuiChen has quit IRC | 12:40 | |
| *** RuiChen has joined #openstack-nova | 12:43 | |
| *** Sree has quit IRC | 12:43 | |
| *** mgoddard_ has joined #openstack-nova | 12:43 | |
| *** Sree has joined #openstack-nova | 12:43 | |
| *** mgoddard has quit IRC | 12:44 | |
| *** mgoddard__ has joined #openstack-nova | 12:44 | |
| *** thorst has joined #openstack-nova | 12:45 | |
| mnestratov|2 | johnthetubaguy: Hi! Could we discuss https://blueprints.launchpad.net/nova/+spec/libvirt-vzstorage-volume-support a bit ? | 12:46 |
| *** mnestratov|2 is now known as mnestratov | 12:46 | |
| johnthetubaguy | mnestratov: so I am stepping away for lunch | 12:47 |
| johnthetubaguy | mnestratov: is it a quick question? | 12:47 |
| mnestratov | in https://etherpad.openstack.org/p/mitaka-nova-spec-review-tracking it is marked as don not appove | 12:47 |
| mnestratov | as do not approve | 12:48 |
| *** mgoddard_ has quit IRC | 12:48 | |
| mnestratov | it is not clear why? | 12:48 |
| johnthetubaguy | mnestratov: yes, do not approve as a specless blueprint, because there is a spec in review, only approve after the spec has merged | 12:48 |
| johnthetubaguy | mnestratov: thats the list of specless blueprint approvals, basically | 12:48 |
| mnestratov | year but there is a spec | 12:48 |
| johnthetubaguy | mnestratov: so I am clearly not communiting this well | 12:49 |
| johnthetubaguy | mnestratov: my comment is, there is a spec, lets approved the blueprint once we get the spec merged | 12:50 |
| mnestratov | ah, ok. makes sence | 12:50 |
| johnthetubaguy | mnestratov: sorry for the confusion | 12:50 |
| mnestratov | no problem | 12:50 |
| johnthetubaguy | mnestratov: appreciate you reaching out quickly to resolve that :) | 12:51 |
| * johnthetubaguy bravely runs away to lunch | 12:51 | |
| *** weihan has quit IRC | 12:51 | |
| mnestratov | i thought it was some sort of miscommunication | 12:51 |
| *** irenab has joined #openstack-nova | 12:53 | |
| *** deepakcs has quit IRC | 12:55 | |
| *** jwcroppe has joined #openstack-nova | 12:56 | |
| gibi | bauwser: hi! do you have a minute or two for questions around request_spec, filter_properties and instance_groups? | 12:56 |
| *** boris-42 has quit IRC | 12:58 | |
| *** huanxie has quit IRC | 12:59 | |
| *** markmc has quit IRC | 13:01 | |
| *** markmc has joined #openstack-nova | 13:02 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Enable options for oslo.reports https://review.openstack.org/216917 | 13:04 |
| *** mwagner has joined #openstack-nova | 13:06 | |
| bauwser | PaulMurray: sorry was out to lunch | 13:07 |
| *** thorst has quit IRC | 13:07 | |
| *** lbelivea has quit IRC | 13:07 | |
| *** ndipanov has joined #openstack-nova | 13:07 | |
| *** baoli has joined #openstack-nova | 13:07 | |
| PaulMurray | bauwser, thats ok - do you know when ndipanov is back? | 13:08 |
| bauwser | PaulMurray: so ndipanov will be back on Tues 10/11 | 13:08 |
| PaulMurray | thanks | 13:08 |
| bauwser | PaulMurray: but he'll be on the Japanese TZ | 13:08 |
| bauwser | IIUC | 13:08 |
| bauwser | for the week | 13:08 |
| *** jyuso1 has quit IRC | 13:08 | |
| bauwser | gibi: hi sure | 13:08 |
| gibi | bauwser: great. :) Let me explain what I'd like to achieve | 13:09 |
| gibi | bauwser: I'm trying to understand how the filter properties converted to request spec and how it is converted back to filter properties during select_destination. | 13:09 |
| gibi | bauwser: basically I want to rebase https://review.openstack.org/#/c/148275/ as part of rebasing the soft-affinty server-group policy patches. | 13:09 |
| *** cfriesen__ has joined #openstack-nova | 13:10 | |
| gibi | bauwser: what I need is a new field num_instance_per_host in the filter properties as the soft affinity po | 13:10 |
| gibi | licy is implemented with a weigher that sort the hosts by this number. | 13:10 |
| bauwser | gibi: did you see my last patches about that? | 13:10 |
| gibi | bauwser: I don't think so | 13:10 |
| bauwser | gibi: okay, sec | 13:10 |
| bauwser | gibi: okay, sechttps://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/request-spec-object-mitaka,n,z | 13:10 |
| *** mmdurrant has joined #openstack-nova | 13:11 | |
| bauwser | oops | 13:11 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Avoid intermediate Token auth plugin https://review.openstack.org/229368 | 13:11 |
| bauwser | s/okay, sec/ | 13:11 |
| *** abhishekk_ has quit IRC | 13:11 | |
| gibi | bauwser: thanks, I have to check these as I guess these are touching the same places | 13:11 |
| bauwser | gibi: so, my last changes are about moving where I'm returning a primitive dict | 13:12 |
| PaulMurray | johnthetubaguy, agreed - could make existing cases use the task states even though using local lock - but wont propose that. Keep existing as is. | 13:13 |
| bauwser | gibi: since I didn't want to modify all the filters by the same change, I'm providing a compat decorator in the next patch | 13:13 |
| bauwser | and then moving the back tranformation into the decorator instead | 13:13 |
| gibi | bauwser: ohh I see | 13:13 |
| PaulMurray | johnthetubaguy, incidentally, philday started to introduce attaching and detaching task states for existing case before - but abandoned. Seems it was | 13:13 |
| bauwser | gibi: and then, when modifying the filter to use the object, then also remove the deco | 13:14 |
| PaulMurray | possible to try and start some operations (like resize) in the middle of a detach/attach. But I think that has all changed long since. | 13:14 |
| bauwser | gibi: then, once all of that is done, I'm changing *where* the Spec object is hydrated by https://review.openstack.org/199205 | 13:15 |
| *** ndipanov has quit IRC | 13:15 | |
| gibi | bauwser: as far as I understand the RequestSpec obj has an InstanceGroup field. But that is not the original group given by the user in the scheduler hint but a new one created during some compatibility code here https://github.com/openstack/nova/blob/master/nova/objects/request_spec.py#L157 | 13:15 |
| *** ndipanov has joined #openstack-nova | 13:15 | |
| bauwser | right | 13:15 |
| bauwser | since I was having a dict payload by the scheduler, I was needing to hypdrate a new InstanceGroup object | 13:16 |
| bauwser | hence the from_primitives() helper | 13:16 |
| gibi | why don't we simply reuse the original instance group that was provided in the scheduler_hints? | 13:16 |
| bauwser | but by mid-term, we should remove that from_primitives helper in the conductor and directly create the Spec bject directly | 13:16 |
| bauwser | gibi: mmm? you say the instance_group object is provided in the hints ? | 13:17 |
| bauwser | where then ? | 13:17 |
| gibi | the id of the group is provided | 13:18 |
| bauwser | gibi: because AFAIK it's a filter_props key | 13:18 |
| bauwser | gibi: but it would need to call the DB to get it, right? | 13:18 |
| bauwser | I mean an InstanceGroup.get_by_id | 13:18 |
| bauwser | which calls the DB | 13:18 |
| bauwser | if so, I'm opposed to that | 13:19 |
| bauwser | because we want to avoid to call the DB for that | 13:19 |
| bauwser | if it's already provided by the filter_props dict | 13:19 |
| bauwser | gibi: ^ | 13:19 |
| gibi | bauwser: oh, OK. You are right we would need a db call | 13:19 |
| bauwser | gibi: like I said, it's a temporary work | 13:20 |
| gibi | bauwser: anyhow, besides hosts and policies I would need another information in the filter_properties for the soft affinty filter | 13:20 |
| bauwser | gibi: because by the end, the Spec object will be passed by the conductor | 13:20 |
| bauwser | so we'll be possibily direct return the InstanceGroup object directly by the conductor | 13:21 |
| bauwser | gibi: mmm, what kind of information you'd like to add ? | 13:21 |
| gibi | number of instance per host in a given group | 13:22 |
| bauwser | gibi: tbh, we decided in Kilo to carefully review all the changes which were adding more keys to the filt_props dict - because of the Spec object | 13:22 |
| bauwser | gibi: I see | 13:22 |
| bauwser | gibi: how could you make sure this field would be accurate? | 13:22 |
| gibi | it is the patch calculating the information https://review.openstack.org/#/c/147517/51/nova/objects/instance_group.py,cm | 13:23 |
| bauwser | I mean, the number of instances | 13:23 |
| bauwser | ie. the conductor passing the object to the scheduler in a distributed and concurrent world | 13:23 |
| gibi | bauwser: yes, it has the same concurrency problem as the instance group already has | 13:24 |
| bauwser | gibi: and you know we don't have a consensus yet on that ? | 13:24 |
| gibi | bauwser: a concurrent scheduling can make the hosts field of the filter_properties invalid | 13:24 |
| bauwser | for example, jaypipes was pretty concerned about that | 13:24 |
| bauwser | gibi: I'm reading your spec now, to make sure I haven't missed anything | 13:25 |
| gibi | bauwser: I'm awar of jaypipes concerns about the whole instance group concept | 13:25 |
| gibi | s/awar/aware/ | 13:25 |
| bauwser | I mean, it's a weigher, so probably it's less invasive than a filter, but still the decision could be a bit unaccurate | 13:26 |
| gibi | bauwser: agree, it has the same unaccuracy as the instance group has in case of parallel scheduling for the same group | 13:26 |
| bauwser | but don't get me wrong - I don't want a scheduler not making mistakes, I certainly prefer a scheduler able to make quick decisions with a reasonable number of retries | 13:27 |
| *** jinxing has joined #openstack-nova | 13:27 | |
| bauwser | so, from a technical standpoint, you're now stuck because some bad man provided a new NovaObject instead of leaving you just adding a new key in an unformated dict ? :) | 13:28 |
| *** tellesnobrega_ has joined #openstack-nova | 13:28 | |
| bauwser | the short answer would be 'you could add a new field to the object you want to amend' | 13:28 |
| gibi | bauwser: either I add a field to the unformatted dict, or I would need to query the original object that has the db | 13:28 |
| *** aix has joined #openstack-nova | 13:29 | |
| bauwser | the latter is performance-concernable | 13:29 |
| bauwser | I guess | 13:29 |
| bauwser | lots of guys already complain that we do ComputeNode.get_all in the scheduler codepath | 13:30 |
| gibi | bauwser: this is a calculated value from the hosts and the members of the InstanceGroup but the compatibility code creates a new InstanceGroup object that does not have members | 13:30 |
| gibi | bauwser: therefore I cannot recalculate | 13:30 |
| bauwser | gibi: oh I see, then why aren't you proposing to modify the helper to match your need? | 13:31 |
| bauwser | gibi: oh sec | 13:31 |
| *** edmondsw has joined #openstack-nova | 13:32 | |
| bauwser | gibi: I guess the problem is https://review.openstack.org/#/c/148275/51/nova/scheduler/weights/affinity.py,cm L69 right? | 13:32 |
| *** thorst has joined #openstack-nova | 13:32 | |
| gibi | bauwser: right, this is the calculated value I need for the weighers | 13:33 |
| *** jwcroppe has quit IRC | 13:33 | |
| bauwser | gibi: how do you set that value ? | 13:34 |
| bauwser | I can review the series for sure | 13:34 |
| gibi | bauwser: it is calculated here https://review.openstack.org/#/c/147517/51/nova/objects/instance_group.py,cm and added to the filter_properties here https://review.openstack.org/#/c/148277/52/nova/scheduler/utils.py,cm L337 | 13:36 |
| *** fnordahl has quit IRC | 13:37 | |
| *** xyang1 has joined #openstack-nova | 13:37 | |
| *** josecastroleon has quit IRC | 13:38 | |
| *** moshele has joined #openstack-nova | 13:38 | |
| *** josecastroleon has joined #openstack-nova | 13:38 | |
| *** fnordahl has joined #openstack-nova | 13:39 | |
| openstackgerrit | Sam Yaple proposed openstack/nova: Fix attibute error when cloning raw images in Ceph https://review.openstack.org/237799 | 13:40 |
| bauwser | gibi: I see | 13:40 |
| *** yonglihe has quit IRC | 13:40 | |
| bauwser | well | 13:40 |
| bauwser | I'm pretty concerned by that, but I'll try to explain why from my best | 13:41 |
| *** baoli_ has joined #openstack-nova | 13:41 | |
| gibi | bauwser: I'm totally open for alternatives | 13:41 |
| *** yonglihe_ has joined #openstack-nova | 13:41 | |
| *** ericksonsantos has quit IRC | 13:41 | |
| *** Nathan has joined #openstack-nova | 13:42 | |
| bauwser | gibi: so, the problem is that you calculate a discrete value (well, a keyed dict of hosts having discrete values) once in the lifetime and then passing it | 13:42 |
| bauwser | gibi: for that calculation, you call the world to get the full list of instances with a specific filter | 13:43 |
| bauwser | say that one cloud is running 1000 nodes with 100 instances each, I leave you imagine the full scan | 13:43 |
| gibi | bauwser: the calculation is limited to a single instance group | 13:44 |
| *** baoli has quit IRC | 13:44 | |
| gibi | the group already has the members stored so there is no full scan at all | 13:44 |
| bauwser | instances = objects.InstanceList.get_by_filters(self._context, filters=filters) isn't a full scan ? | 13:45 |
| gibi | no because the filter is filtering for uuids | 13:45 |
| gibi | or it depends what do you mean by full scan | 13:45 |
| gibi | it has the same performance as InstanceGroup.get_hosts() and that is already called in the scheduling code path | 13:46 |
| *** ericksonsantos has joined #openstack-nova | 13:47 | |
| *** tellesnobrega_ has quit IRC | 13:47 | |
| gibi | here https://review.openstack.org/#/c/148277/52/nova/scheduler/utils.py,cm L334 | 13:48 |
| bauwser | gibi: agreed | 13:48 |
| bauwser | gibi: but that doesn't mean it's a good thing :) | 13:49 |
| gibi | bauwser: ok :) then what to do? :) | 13:49 |
| *** diogogmt has quit IRC | 13:49 | |
| bauwser | anyway, we have indexes on uuid, so that's a bit okay | 13:49 |
| *** inteq has quit IRC | 13:49 | |
| *** pradk has joined #openstack-nova | 13:50 | |
| bauwser | gibi: okay, so you calculate once the values then passing them to the namedtuple, right? | 13:50 |
| gibi | right, the same way as group_hosts | 13:50 |
| *** chinmaya has quit IRC | 13:50 | |
| *** salv-orlando has quit IRC | 13:51 | |
| *** salv-orlando has joined #openstack-nova | 13:51 | |
| bauwser | gibi: and then, that namedtuple is casted as a dict for becoming filt_props['group'] right? | 13:52 |
| gibi | ohh I just found that you added hosts as a field to the InstanceGroup during the request spec work | 13:52 |
| *** moshele has quit IRC | 13:52 | |
| gibi | bauwser: right | 13:52 |
| bauwser | gibi: yup because members was the list of uuids | 13:52 |
| bauwser | gibi: so I had to amend the object to carry the needed information | 13:53 |
| *** subscope has quit IRC | 13:53 | |
| bauwser | gibi: see my point? | 13:53 |
| *** diogogmt has joined #openstack-nova | 13:53 | |
| gibi | bauwser: so basically you had the same problem as I had and you added a field for it. Then maybe I can add a field too | 13:53 |
| bauwser | performance-unwise, if you want to pass something to the scheduler, you now have to add a new field to the instance_group object | 13:53 |
| gibi | bauwser: I see | 13:53 |
| *** paul-carlton has joined #openstack-nova | 13:54 | |
| *** vladikr has joined #openstack-nova | 13:54 | |
| gibi | bauwser: so you suggest to adding a new field to the InstsanceGroup to store this calculated value a same way as you added hosts | 13:54 |
| bauwser | but the main difference between the hosts field and the dict you want to add is that one is pretty small and quite unchanging, while the latter is a bit questionable about its accuracy and its performance | 13:54 |
| gibi | bauwser: yes, you are right | 13:55 |
| *** e0ne has quit IRC | 13:55 | |
| gibi | bauwser: so it will be more problematic to add it to the object | 13:55 |
| bauwser | gibi: that's my concerns yeds | 13:55 |
| bauwser | so, I'm wondering how to actually get that correctly | 13:55 |
| gibi | but this is a deadlock as either I calculate the value but that has performance impact or I store the value but that will look ugly | 13:56 |
| bauwser | I'm thinking of an alternative approach | 13:56 |
| bauwser | we do have the list of instances per host in the scheduler memory | 13:56 |
| bauwser | in the HostManager | 13:56 |
| bauwser | and the list of instances per host is thus given in host_state passed in filters | 13:56 |
| *** e0ne has joined #openstack-nova | 13:56 | |
| *** dshaw has joined #openstack-nova | 13:57 | |
| bauwser | any change to an instance makes an RPC cast to the scheduler, so that's a bit less racy | 13:57 |
| bauwser | what you get in the HostState is a pretty accurate value of the list of instances per host | 13:57 |
| gibi | yes, but I have to filter that by the group members | 13:57 |
| bauwser | gibi: that's then just a live calculation that can be done in the weigher, no ? | 13:58 |
| bauwser | ie. filter one list by another | 13:58 |
| gibi | weigher does not have the list of memeber is a group right now | 13:58 |
| bauwser | gibi: no, but it gets the instance_group object from the spec, right? | 13:59 |
| bauwser | which contains the list of members and hosts | 13:59 |
| gibi | but that is a ~fake object which does not have memebers just hosts and policies | 13:59 |
| bauwser | gibi: on the other hand, the weigher knows the host_state | 13:59 |
| *** mdrabe has joined #openstack-nova | 13:59 | |
| bauwser | gibi: what if you pass the members to the fake object? | 14:00 |
| gibi | bauwser: that might be doable :) | 14:00 |
| bauwser | when hydrating the Spec object? | 14:00 |
| bauwser | lemme show you how you can get the list of instances per host | 14:00 |
| *** gszasz has quit IRC | 14:01 | |
| bauwser | https://github.com/openstack/nova/blob/master/nova/scheduler/host_manager.py#L582-L601 | 14:01 |
| bauwser | gibi: ^ | 14:01 |
| *** edleafe is now known as figleaf | 14:02 | |
| bauwser | so I'd see 2 things : | 14:02 |
| *** dustins has joined #openstack-nova | 14:03 | |
| gibi | bauwser: I see | 14:03 |
| bauwser | #1 add the list of members in the filt_props dict, so that the helper hydrate the list of members in the nested InstanceGroup object of the ReqSpec object (with the backwards method for returning a dict keyed by members) | 14:03 |
| bauwser | #2 in the weigher, get the list of members from the spec object, and match that with the list of hosts | 14:04 |
| bauwser | oops | 14:04 |
| bauwser | match that with the list of instances the host belongs to | 14:04 |
| openstackgerrit | Andrea Rosa proposed openstack/nova-specs: Allow more instance operations during live migration https://review.openstack.org/226199 | 14:04 |
| *** rlrossit has joined #openstack-nova | 14:04 | |
| bauwser | gibi: does that work for you ? | 14:04 |
| gibi | bauwser: I think it is doable. | 14:04 |
| bauwser | coolness | 14:05 |
| dshaw | when installing and configuring a compute node, is /etc/nova/nova.conf or /etc/nova/nova-compute.conf the proper file to edit? | 14:05 |
| gibi | I will try put this together :) | 14:05 |
| gibi | bauwser: thanks a lot! | 14:05 |
| bauwser | sorry to have thrown your series under the bus, but I feel it's important for us to have stable and versioned APIs :) | 14:05 |
| gibi | bauwser: I totally agree that your work is more important than my series so no hard feelings | 14:05 |
| bauwser | gibi: the main thing is that we found an approach I feel that can be accurate and perfomance-wise | 14:06 |
| *** subscope has joined #openstack-nova | 14:07 | |
| gibi | bauwser: agree. For me the main thing is that I can move forward with my rebase meaningfully. :) | 14:07 |
| bauwser | what I like is that the calculation is made in the weigher, which can be optional for operators depending on their conf | 14:07 |
| gibi | make sense | 14:07 |
| *** klkumar has joined #openstack-nova | 14:08 | |
| gibi | anyhow, thanks again for your time! | 14:09 |
| *** ccarmack has quit IRC | 14:09 | |
| *** markus_z is now known as markymark | 14:09 | |
| *** pabardina has joined #openstack-nova | 14:10 | |
| *** nkrinner has quit IRC | 14:10 | |
| gibi | I owe you a beer or some nutella in Austin :) | 14:11 |
| bauwser | gibi: np | 14:11 |
| bauwser | the fun stuff is that I *hate* nutella but shhhhttt | 14:11 |
| gibi | lol | 14:11 |
| *** ccarmack has joined #openstack-nova | 14:12 | |
| *** Piet has quit IRC | 14:14 | |
| *** ctrath has joined #openstack-nova | 14:14 | |
| *** ndipanov has quit IRC | 14:14 | |
| *** Nathan has quit IRC | 14:16 | |
| *** mriedem has joined #openstack-nova | 14:18 | |
| bauwser | mriedem: morning | 14:19 |
| bauwser | mriedem: that cells job failure is getting me crazy | 14:19 |
| *** pabardina has left #openstack-nova | 14:19 | |
| mriedem | heh | 14:19 |
| openstackgerrit | garyk proposed openstack/nova: VMware: expand support for Opaque networks https://review.openstack.org/165750 | 14:19 |
| mriedem | did you put that try/except change? | 14:19 |
| mriedem | *push | 14:19 |
| openstackgerrit | jichenjc proposed openstack/nova: Make lock policy default to admin or owner https://review.openstack.org/242213 | 14:19 |
| bauwser | mriedem: nope, I tried an alternative | 14:20 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/nova: Nested Quota Driver: Create Column Allocated https://review.openstack.org/242500 | 14:20 |
| bauwser | mriedem: because the exception is just eaten by the cells rpcapi | 14:20 |
| bauwser | but that doesn't work too | 14:21 |
| bauwser | so, your change seems for some unknown reasons making the gate happy | 14:21 |
| *** Janice_Lee has joined #openstack-nova | 14:21 | |
| *** diogogmt has quit IRC | 14:21 | |
| *** paul-carlton has quit IRC | 14:21 | |
| bauwser | mriedem: because like tonyb stated, it shouldn't work because of the local DB refresh we do before calling the parent | 14:21 |
| *** salv-orlando has quit IRC | 14:22 | |
| bauwser | FWIW, I'm going to abandon my change (or spin off the try/except stuff if I can) | 14:23 |
| openstackgerrit | garyk proposed openstack/nova: VMware: fix trivial indentations https://review.openstack.org/242505 | 14:23 |
| *** ijuwang has joined #openstack-nova | 14:23 | |
| mriedem | bauwser: ok i see your change hit the cells failure too huh | 14:24 |
| bauwser | yup :( | 14:24 |
| mriedem | bauwser: so where does the exception get eaten in the cells code? | 14:24 |
| mriedem | btw, i meant doing the try/except in bdm_update_or_create_at_top in cells.messaing, not bdm.save() | 14:25 |
| bauwser | https://github.com/openstack/nova/blob/master/nova/cells/rpcapi.py#L421-L421 | 14:26 |
| *** salv-orlando has joined #openstack-nova | 14:26 | |
| openstackgerrit | dane-fichter proposed openstack/nova-specs: Nova Support of Glance Image Signing https://review.openstack.org/188874 | 14:26 |
| mriedem | yeah so that's when the bdm.save() calls it right? | 14:26 |
| *** diogogmt has joined #openstack-nova | 14:26 | |
| mriedem | i was thinking we do it in here https://github.com/openstack/nova/blob/master/nova/cells/messaging.py#L1167 | 14:26 |
| mriedem | change self.db.block_device_mapping_update_or_create to be a try/except with update/create | 14:26 |
| bauwser | yup, that was my other idea | 14:27 |
| bauwser | lemme just provide another thing | 14:27 |
| bauwser | mriedem: we also pop the BDM id, any idea why ? | 14:28 |
| mriedem | alaski: said that was to avoid collisions | 14:28 |
| mriedem | which makes me wonder how the update actually works | 14:28 |
| mriedem | since the id is the primary key | 14:29 |
| mriedem | if we pop that how does the db api know what it's updating | 14:29 |
| bauwser | mriedem: because it checks the device_name | 14:29 |
| mriedem | yeah, i guess it's making it's own UC | 14:29 |
| mriedem | but since the device_name changes, that doesn't really work | 14:29 |
| bauwser | oh I wonder something, since we nullify device_name in the second save(), it means that https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/api.py#L3945 is always False right? | 14:30 |
| bauwser | mriedem: ^ | 14:30 |
| *** burt has joined #openstack-nova | 14:31 | |
| mriedem | yeah | 14:31 |
| mriedem | iff the bdm.save() is getting there with bdm.device_name == None that is | 14:31 |
| bauwser | adding a bit of logging + removing that | 14:31 |
| mriedem | i'm not entirely sure if it is saved into the db that way | 14:31 |
| mriedem | yeah i was working o na logging patch too | 14:32 |
| bauwser | that's a bit strange | 14:33 |
| *** Sree has quit IRC | 14:33 | |
| bauwser | how can we nullify the device_name by this code ? | 14:33 |
| *** otter768 has joined #openstack-nova | 14:33 | |
| bauwser | also the model is weird, we say that device_name is like a PK but we drop it | 14:34 |
| bauwser | mriedem: https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/api.py#L3943-L3944 that comment is particulary weird | 14:35 |
| *** subscope has quit IRC | 14:35 | |
| *** pratikmallya has joined #openstack-nova | 14:36 | |
| mriedem | i think it'd be useful to have some logging in there that gets the bdms by instance uuid and then checks to see if any don't have a device_name set | 14:36 |
| *** smatzek has quit IRC | 14:38 | |
| bauwser | mriedem: so IIUC, we pop the id to allow api db and cell db to have same rows with different ids | 14:38 |
| *** otter768 has quit IRC | 14:38 | |
| bauwser | that means we semehow need to do a right lookup | 14:38 |
| bauwser | by using device_name for the lookup, we should use only the original one, not the one we want to set to | 14:39 |
| mriedem | yeah, so the other day i was talking about in cells.messaging.bdm_update_or_create_at_top if create is None we get the bdms from the db by instance uuid (like it does later in that method) and then try to correlate on device_name | 14:39 |
| mriedem | but as alaski pointed out's what the db api code was already doing | 14:39 |
| mriedem | but say the bdm device_name was set and we created a bdm, then later we set bdm.device_name=None and save it, that would trigger an update_or_create in the parent cell and since device_name isn't set, it'd create a 2nd bdm, right? | 14:40 |
| bauwser | that's my understanding but I could be wrong | 14:40 |
| mriedem | this doesn't fail in the non-cells case b/c the id isn't popped from the bdm object when we do the save, like it does in the api cell code | 14:41 |
| bauwser | okay, I'm gonna try something in the cells.messaging code | 14:41 |
| bauwser | mriedem: right, the problem is that the lookup is terribly bad in the cells case | 14:41 |
| bauwser | mriedem: so I'm going to do that scan for getting the right BDM | 14:42 |
| bauwser | in the cells.messaging code | 14:42 |
| bauwser | we do that anyway later in the code | 14:42 |
| bauwser | the only problem I see is that we still search for the next device_name | 14:43 |
| bauwser | speaking of https://github.com/openstack/nova/blob/master/nova/cells/messaging.py#L1183 | 14:43 |
| bauwser | mriedem: ^ | 14:43 |
| *** jwcroppe has joined #openstack-nova | 14:43 | |
| mriedem | yeah that's if create is False | 14:43 |
| bauwser | rigt | 14:44 |
| mriedem | if we get there and bdm.device_name was None, we'd fail to update | 14:44 |
| *** edtubill has joined #openstack-nova | 14:44 | |
| mriedem | also because volume_id isn't set in this test, snapshot_id is | 14:44 |
| mriedem | which is why i added https://review.openstack.org/#/c/241390/ | 14:44 |
| mriedem | but ^ is also failing in the cells job | 14:44 |
| *** dshaw has quit IRC | 14:44 | |
| bauwser | mriedem: but it fails because it's with the create=False case | 14:45 |
| bauwser | so I guess we should do this lookup in the create=None case | 14:45 |
| mriedem | yeah | 14:45 |
| bauwser | happy with writing it or me ? | 14:46 |
| *** achanda has joined #openstack-nova | 14:46 | |
| bauwser | like, exploding the db.update_or_create case by trying to lookup the BDM and then either update or create | 14:46 |
| bauwser | I feel we're pretty close | 14:46 |
| openstackgerrit | Erickson Filipe Guedes dos Santos proposed openstack/nova: Nested Quota: Set default values to subproject https://review.openstack.org/242514 | 14:47 |
| *** achanda has quit IRC | 14:51 | |
| *** mc_nair has joined #openstack-nova | 14:52 | |
| alaski | mriedem: bauwser do we need a similar check to https://github.com/openstack/nova/blob/master/nova/objects/block_device.py#L145 in save()? | 14:52 |
| johnthetubaguy | jaypipes: got libosinfo approved, a quick heads up, this patch has a -2 on it from you: https://review.openstack.org/#/c/149625/21 | 14:52 |
| *** alaski is now known as lascii | 14:52 | |
| bauwser | lascii: well, I tried this approach with no luck https://review.openstack.org/#/c/242447/1/nova/objects/block_device.py,cm | 14:53 |
| *** Sree has joined #openstack-nova | 14:53 | |
| snikitin | mriedem: Hi Matt! Yesterday you +2ed enginafacade patch. but we forgot to configure() facade for tests. Could you please look at it again? https://review.openstack.org/#/c/221310/ | 14:53 |
| lascii | bauwser: I'm thinking that rather than toggling 'create' maybe just don't send a message when device_name is not set | 14:55 |
| lascii | because the bdm can't be looked up properly in the parent cell | 14:55 |
| bauwser | lascii: that's what we discussed above ^ | 14:55 |
| lascii | mriedem has been mentioning that device_name gets nulled and reset, which I didn't realize was a thing that happens | 14:55 |
| bauwser | :) | 14:55 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: DNM: more debug logging for cells BDM race failure https://review.openstack.org/242517 | 14:55 |
| mriedem | bauwser: lascii: dansmith: ^ | 14:55 |
| bauwser | lascii: I'm just writing a change | 14:56 |
| lascii | bauwser: okay, cool. I wasn't quite following it all | 14:56 |
| *** burgerk has joined #openstack-nova | 14:57 | |
| *** Sree has quit IRC | 14:58 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Add service records for nova-api services https://review.openstack.org/242206 | 14:58 |
| *** pratikma_ has joined #openstack-nova | 14:59 | |
| *** pumaranikar has joined #openstack-nova | 14:59 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: cells: bdm_update_or_create_at_top needs to check snapshot/image id https://review.openstack.org/241390 | 14:59 |
| bauwser | lascii: mriedem: ^ | 15:00 |
| bauwser | shamely reusing mriedem's change | 15:00 |
| *** Piet has joined #openstack-nova | 15:01 | |
| mriedem | bauwser: cool, that's basically what i just added in the db api https://review.openstack.org/#/c/242517/1/nova/db/sqlalchemy/api.py | 15:01 |
| *** pratikmallya has quit IRC | 15:01 | |
| bauwser | lascii: mriedem: of course I need to modify some unittests, but I first want to check if that passes | 15:01 |
| bauwser | heh :) | 15:02 |
| mriedem | i would actually move that to the db api if it works | 15:02 |
| mriedem | but yeah let's see | 15:02 |
| bauwser | mriedem: agreed | 15:02 |
| mriedem | hey, high five for synergy! | 15:02 |
| bauwser | I don't care about the all-star credits :) | 15:02 |
| bauwser | if that works with you, I'm fine :) | 15:02 |
| bauwser | I just want to stop recheck bug | 15:02 |
| *** alejandrito has joined #openstack-nova | 15:04 | |
| *** alex_klimov has quit IRC | 15:06 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: DNM: more debug logging for cells BDM race failure https://review.openstack.org/242517 | 15:07 |
| *** smatzek has joined #openstack-nova | 15:08 | |
| *** Marga_ has quit IRC | 15:10 | |
| *** Marga_ has joined #openstack-nova | 15:10 | |
| *** jichen has quit IRC | 15:11 | |
| *** lbelivea has joined #openstack-nova | 15:12 | |
| bauwser | mriedem: FWIW I usually prefer to call a traceback.print_stack() but that sounds a little pervasive :p | 15:13 |
| mriedem | i just copied what was in nova.cmd.network | 15:13 |
| mriedem | the traceback is probably not going to be helpful since it'll just traceback to an rpc call i think | 15:14 |
| mriedem | the logging in the bdm object is probably more helpful | 15:14 |
| mriedem | since there was a question about where the bdm.create was initially called from | 15:14 |
| bauwser | right | 15:14 |
| *** dshaw has joined #openstack-nova | 15:14 | |
| bauwser | it's just friday and I'm just a little pissed off by the cells job failure :) | 15:15 |
| *** Marga_ has quit IRC | 15:15 | |
| *** whenry has joined #openstack-nova | 15:16 | |
| openstackgerrit | Derek Higgins proposed openstack/nova: Handle a NeutronClientException 404 Error for floating ips https://review.openstack.org/242158 | 15:19 |
| *** amotoki has quit IRC | 15:19 | |
| *** wearemy369 has joined #openstack-nova | 15:20 | |
| *** tongli has joined #openstack-nova | 15:20 | |
| *** mtanino has joined #openstack-nova | 15:21 | |
| mriedem | it's ok to be angry | 15:22 |
| mriedem | embrace it | 15:22 |
| *** edtubill has quit IRC | 15:23 | |
| johnthetubaguy | mriedem: thanks for running the meeting yesterday, all seemed to go OK | 15:24 |
| mriedem | johnthetubaguy: yeah, i stumbled through a lot of the release process change stuff | 15:25 |
| mriedem | i hadn't read all of the emails in the [release] threads yet | 15:25 |
| *** klkumar has quit IRC | 15:25 | |
| bauwser | well, it just seems the release cadence has to be per-project now, amirite? | 15:26 |
| bauwser | incl. the stable branches | 15:27 |
| mriedem | that was one of the many things | 15:27 |
| *** tjones has joined #openstack-nova | 15:27 | |
| mriedem | basically yeah for stable we own doing our releases as we see fit | 15:27 |
| mriedem | so, i guess i/we will need to be keeping track of when we get some important fixes backported to stable so we can cut a release | 15:28 |
| mriedem | speaking of, garyk had a critical fix on stable/liberty i was supposed to look at | 15:28 |
| *** sfinucan has joined #openstack-nova | 15:31 | |
| *** diogogmt has quit IRC | 15:32 | |
| *** sileht has quit IRC | 15:32 | |
| mriedem | where is superdan? | 15:32 |
| *** dansmith is now known as superdan | 15:33 | |
| superdan | here :) | 15:33 |
| mriedem | whew | 15:33 |
| *** Janice_Lee has quit IRC | 15:33 | |
| johnthetubaguy | mriedem: it was more a highlight them to folks thing, only just read through them myself yesterday | 15:34 |
| *** READ10 has joined #openstack-nova | 15:34 | |
| *** paul-carlton has joined #openstack-nova | 15:34 | |
| mriedem | i did have a question, the nova mitaka release schedule said spec proposal freeze is 12/3 but i think the meeting agenda said 12/1 | 15:35 |
| mriedem | so i was saying 12/1 during the meeting | 15:35 |
| *** tonytan4ever has joined #openstack-nova | 15:37 | |
| *** edtubill has joined #openstack-nova | 15:37 | |
| *** juzuluag has quit IRC | 15:38 | |
| *** klkumar has joined #openstack-nova | 15:38 | |
| *** diogogmt has joined #openstack-nova | 15:38 | |
| *** dave-mccowan has quit IRC | 15:39 | |
| johnthetubaguy | mriedem: oops | 15:40 |
| johnthetubaguy | mriedem: I think 12/3 is the deadline | 15:40 |
| johnthetubaguy | mriedem: thats the thursday at least | 15:40 |
| mriedem | ok | 15:41 |
| mriedem | as long as the wiki is correct | 15:42 |
| *** electrocucaracha has joined #openstack-nova | 15:43 | |
| pkholkin | mriedem: Hello Matt! Thanks a lot for you help with enginefacade! We have updated your patch once again, we forgot to configure enginefacade for tests | 15:44 |
| pkholkin | please look https://review.openstack.org/#/c/221310/30 | 15:44 |
| pkholkin | now everything seems ok | 15:44 |
| mriedem | pkholkin: why didn't the unit tests fail on patch set 29 then? | 15:45 |
| pkholkin | modified only nova/tests/fixtures.py as it was before | 15:45 |
| pkholkin | mriedem: I think it took default settings | 15:45 |
| pkholkin | but it should be configured too | 15:45 |
| mriedem | is there a way to detect when that's not set and blow up? | 15:46 |
| *** e0ne has quit IRC | 15:46 | |
| pkholkin | no, we can check some private field 'started' https://github.com/openstack/oslo.db/blob/master/oslo_db/sqlalchemy/enginefacade.py#L220 , but it is not needed | 15:47 |
| mriedem | superdan: you were on the original change for this backport, it looks clean, i think we should get it in before juno-eol https://review.openstack.org/#/c/221876/ | 15:47 |
| pkholkin | mriedem: now it is configured everywhere I think | 15:47 |
| mriedem | pkholkin: i commented in the review | 15:48 |
| superdan | mriedem: done | 15:48 |
| *** dane-fichter has joined #openstack-nova | 15:49 | |
| dane-fichter | danpb: please check my responses to your comments on https://review.openstack.org/#/c/188874/ I hope they will clarify our approach | 15:50 |
| pkholkin | mriedem: for working in production everything is good because of parse_args, as for tests I think that it would be better to configure facade too, also as far as I understand default setting are the same as for main database | 15:50 |
| pkholkin | maybe the problems can happen with api database, but tests don;t fail | 15:51 |
| pkholkin | by default it takes this https://github.com/openstack/oslo.db/blob/master/oslo_db/sqlalchemy/enginefacade.py#L132 | 15:52 |
| mriedem | maybe we need a comment in the database fixture when configuring the session to justify it | 15:52 |
| *** mdrabe has quit IRC | 15:52 | |
| mriedem | like, we do this when configuring any other service to use the db, so also do it in the db fixture so we're testing the same code path | 15:53 |
| *** dave-mccowan has joined #openstack-nova | 15:53 | |
| mriedem | *with the same db options | 15:53 |
| mriedem | i just don't want to forget why it's there in a year | 15:54 |
| pkholkin | mriedem: ok, understand) we will add a comment right now | 15:54 |
| superdan | markymark: are you going to rev the config spec? | 15:54 |
| *** Sree has joined #openstack-nova | 15:55 | |
| markymark | superdan: yeah, I'll push another patch set for that very important sentence. Takes a few minutes. | 15:55 |
| superdan | okay | 15:55 |
| superdan | I'll wait to review then | 15:56 |
| markymark | superdan: thanks | 15:56 |
| *** vilobhmm11 has joined #openstack-nova | 15:57 | |
| *** READ10 has quit IRC | 15:58 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova-specs: Centralize Config Options https://review.openstack.org/227948 | 15:59 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: Use EngineFacade from oslo_db.enginefacade https://review.openstack.org/221310 | 16:00 |
| pkholkin | mriedem: added note message, please look https://review.openstack.org/#/c/221310/ | 16:00 |
| *** rdopiera has quit IRC | 16:00 | |
| mriedem | pkholkin: +2 | 16:01 |
| mriedem | jaypipes: ^ | 16:01 |
| markymark | superdan: done: https://review.openstack.org/#/c/227948/ Thanks for reviewing. | 16:01 |
| *** Sree has quit IRC | 16:01 | |
| pkholkin | mriedem: great!!! thank you) | 16:01 |
| superdan | markymark: yeah, looking now | 16:01 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'instance_info' and 'instance_extra' https://review.openstack.org/225065 | 16:01 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'cell_mapping' https://review.openstack.org/226788 | 16:01 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'instance_mapping' https://review.openstack.org/228979 | 16:01 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'key_pair' and 'cell' https://review.openstack.org/239915 | 16:01 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'network' https://review.openstack.org/240253 | 16:01 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'certificate' and 'pci_device' https://review.openstack.org/240275 | 16:01 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: WIP: enginefacade: 'floating_ip' https://review.openstack.org/240513 | 16:01 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: WIP: enginefacade: 'fixed_ip' https://review.openstack.org/240811 | 16:02 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'service' https://review.openstack.org/241178 | 16:02 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'dnsdomain' and 'ec2' https://review.openstack.org/242019 | 16:02 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'host_mapping' https://review.openstack.org/242035 | 16:02 |
| *** sudipto has joined #openstack-nova | 16:02 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova: enginefacade: 'compute_node' https://review.openstack.org/241189 | 16:02 |
| markymark | PaulMurray: You're leading the live-migrate team to glory, right? | 16:03 |
| markymark | PaulMurray: I'm asking because we have two different tags for bugs in that area "live-migration" and "live-migrate". Would be cool if we use just one. Preferences? | 16:05 |
| markymark | PaulMurray: Would make my life easier when I'm doing bug-skimming. | 16:06 |
| *** pratikma_ has quit IRC | 16:07 | |
| markymark | pkholkin: I've seen you change that often to "live-migrate". What's the background? | 16:07 |
| *** zzzeek has joined #openstack-nova | 16:08 | |
| *** Janice_Lee has joined #openstack-nova | 16:09 | |
| *** mdrabe has joined #openstack-nova | 16:09 | |
| *** jaypipes is now known as leakypipes | 16:09 | |
| *** zzzeek has quit IRC | 16:09 | |
| *** zzzeek has joined #openstack-nova | 16:10 | |
| openstackgerrit | Sudipta Biswas proposed openstack/nova: glance:add helper method to get client version https://review.openstack.org/222150 | 16:10 |
| leakypipes | mriedem, pkholkin: +W'd | 16:11 |
| *** hemnafk is now known as hemna | 16:12 | |
| diana_clarke | I only have a few more days of work before I go on maternity leave. For the most part, my patches in the review queue are pretty boring, but one is a prerequisite for someone else's patch. | 16:13 |
| diana_clarke | Should I abandon it before I leave, so they aren't blocked by me not answering feedback on it? And encourage the other person to submit it instead? Or just leave it as-is? Thanks! | 16:13 |
| *** pumaranikar has quit IRC | 16:13 | |
| *** klkumar has quit IRC | 16:13 | |
| pkholkin | leakypipes: wow, thanks!) | 16:13 |
| *** pumaranikar has joined #openstack-nova | 16:14 | |
| snikitin | leakypipes: thanks! | 16:14 |
| superdan | mriedem: bauwser: so what's the current thinking on the cells stuff? | 16:14 |
| superdan | I missed the convo this morning, but it seemed like maybe you were shifting plans | 16:14 |
| *** emagana has joined #openstack-nova | 16:15 | |
| bauwser | superdan: so we tried to remove the update_or_create call by something better like | 16:15 |
| superdan | diana_clarke: don't abandon, anyone else can revise your patch if need be | 16:15 |
| bauwser | find the related BDM, and if so, update it, else create it | 16:15 |
| *** romainh has quit IRC | 16:15 | |
| superdan | bauwser: the cells messaging call or inside the object? | 16:15 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Replaced deprecated timeutils methods https://review.openstack.org/241179 | 16:15 |
| diana_clarke | superdan: okie dokie. thanks | 16:15 |
| mriedem | superdan: in cells messaging https://review.openstack.org/#/c/241390/2/nova/cells/messaging.py | 16:16 |
| bauwser | superdan: so we made 2 changes, one for me in the cells.messaging stuff, one from mriedem in the db api | 16:16 |
| mriedem | superdan: i have a separate debug patch here https://review.openstack.org/#/c/242517/ | 16:16 |
| bauwser | for the moment, we need to see if the cells job is fine before trying to modify more than that | 16:16 |
| mriedem | that as bauwser says does the same check in the db api | 16:16 |
| superdan | hmm | 16:16 |
| *** electrocucaracha has quit IRC | 16:17 | |
| superdan | so, that comment says you don't know the bdm id, but that wouldn't help anyway, since they're going to be different in the parent vs. child database anyway | 16:17 |
| mriedem | lascii: we don't test resize in the cells job b/c of flavors right? tempest has it's own flavors created in devstack that aren't synced to cells | 16:17 |
| mriedem | right? | 16:17 |
| bauwser | superdan: yup because since there are 2 DBs, the ID is not the same | 16:17 |
| superdan | I'm not sure if using one of those three is always enough to identify it, but if so, then that might be cool | 16:17 |
| lascii | mriedem: right | 16:18 |
| mriedem | ok | 16:18 |
| *** rk4n has quit IRC | 16:18 | |
| bauwser | superdan: so, the main problem is that for finding the BDM, the db api method was checking the device_name as a key | 16:18 |
| *** jinxing has quit IRC | 16:18 | |
| *** klkumar has joined #openstack-nova | 16:18 | |
| superdan | bauwser: yeah, I'm just worried about changing that assumption everywhere | 16:19 |
| bauwser | superdan: but since we *want* to modify the device_name to none, then there are no ways to get the right BDM to update which means it's always a create | 16:19 |
| superdan | lascii: you okay with changing that to be one of the three uuids? | 16:19 |
| BobBall | I've got 2 kilo installs; one of them has the os-hypervisor extension enabled but the other does not. I'm getting confused with all of the config options that exist orr are deprecated etc. Any clues for a config option that I should look at? The symptom is just that "nova hypervisor-list" returns a 404 from /v2/.../os-hypervisors | 16:19 |
| bauwser | superdan: for the moment, this is just a DNM | 16:19 |
| superdan | bauwser: yeah, I know | 16:19 |
| *** rk4n has joined #openstack-nova | 16:19 | |
| bauwser | superdan: but I agree with you that it's questionable | 16:19 |
| superdan | bauwser: just wondering if it's a legit path to pursue if it "seems" to work | 16:19 |
| *** venkat_p has joined #openstack-nova | 16:19 | |
| mriedem | jroll: so you need to get flavors into the cells db that are non-default right? | 16:19 |
| bauwser | superdan: maybe the long-term plan is to find a clear path for finding a BDM if not by the fields | 16:20 |
| *** achanda has joined #openstack-nova | 16:20 | |
| superdan | bauwser: sure, long term | 16:20 |
| mriedem | jroll: we kind of need the same just for the cells job b/c the flavors created in devstack aren't in the cells db, so we don't test any resize operations | 16:20 |
| leakypipes | DNM? | 16:20 |
| lascii | in a meeting atm, not following well enough to answer smartly | 16:20 |
| mriedem | leakypipes: do not merge | 16:20 |
| bauwser | Do Not Merge | 16:20 |
| jroll | mriedem: doeeeeet | 16:20 |
| *** venkat_p has quit IRC | 16:21 | |
| jroll | mriedem: I really just need it for occassional testing | 16:21 |
| *** Janice_Lee has quit IRC | 16:23 | |
| bauwser | mriedem: meh 241390 is currently off the weeds for cells | 16:23 |
| bauwser | sooooo | 16:24 |
| mriedem | so, i have to leave for an appt | 16:25 |
| mriedem | back in probably 1.5 hours | 16:25 |
| mriedem | need my annual high powered puffs of air shot into my eyeballs | 16:25 |
| mriedem | will check back on the test results when i'm done | 16:25 |
| *** derekh has quit IRC | 16:26 | |
| bauwser | mriedem: actively chasing down your change for the DB API | 16:26 |
| bauwser | mriedem: but it seems superdan was right, the lookup is not right | 16:26 |
| leakypipes | mriedem: lol, that's funny. I have an opthamologist appt this afternoon at 3pm... | 16:27 |
| mriedem | high five | 16:27 |
| *** mriedem is now known as mriedem_away | 16:27 | |
| bauwser | FWIW I had a dentist appointment this morning, can I high-five too ? | 16:27 |
| leakypipes | cores: https://review.openstack.org/#/c/222150/ should be an easy win. Been through a number of iterations and I'm happy with it now. | 16:28 |
| *** ihrachys has quit IRC | 16:29 | |
| *** smatzek_ has joined #openstack-nova | 16:30 | |
| *** pratikmallya has joined #openstack-nova | 16:31 | |
| *** mhickey has quit IRC | 16:32 | |
| *** mhickey_ has joined #openstack-nova | 16:32 | |
| *** electrocucaracha has joined #openstack-nova | 16:32 | |
| *** otter768 has joined #openstack-nova | 16:34 | |
| *** smatzek has quit IRC | 16:35 | |
| *** sudipto_ has joined #openstack-nova | 16:35 | |
| *** vilobhmm11 has quit IRC | 16:35 | |
| *** vilobhmm11 has joined #openstack-nova | 16:36 | |
| *** emagana has quit IRC | 16:37 | |
| *** HenryG has quit IRC | 16:37 | |
| *** sudipto has quit IRC | 16:38 | |
| openstackgerrit | Ludovic Beliveau proposed openstack/nova: Support configuration of multiple pci_alias within an array https://review.openstack.org/242555 | 16:38 |
| *** HenryG has joined #openstack-nova | 16:39 | |
| *** otter768 has quit IRC | 16:39 | |
| *** e0ne has joined #openstack-nova | 16:39 | |
| PaulMurray | markymark, sorry - I was in a call | 16:39 |
| *** pratikma_ has joined #openstack-nova | 16:39 | |
| *** pratikmallya has quit IRC | 16:39 | |
| *** sudipto__ has joined #openstack-nova | 16:39 | |
| PaulMurray | markymark, are those tags both used for the same thing? | 16:40 |
| markymark | PaulMurray: yes, the "live-migration" is the "official" one which gets autocompletion the other one is often used by pkholkin | 16:40 |
| *** matrohon has quit IRC | 16:41 | |
| *** sileht has joined #openstack-nova | 16:41 | |
| *** sudipto_ has quit IRC | 16:42 | |
| PaulMurray | markymark, same thing has happened with the priority title - i called it live migration - johnthetubaguy called it live migrate | 16:43 |
| *** pratikma_ is now known as pratikmallya | 16:43 | |
| PaulMurray | I really don't mind - if there is already an "official" one I would go with that | 16:43 |
| PaulMurray | I expect its something that will continue to be confused | 16:43 |
| markymark | PaulMurray: I have no strong preference either, just focusing on one would be cool. | 16:45 |
| *** achanda has quit IRC | 16:45 | |
| *** emagana has joined #openstack-nova | 16:45 | |
| *** su_zhang has joined #openstack-nova | 16:45 | |
| *** mgoddard_ has joined #openstack-nova | 16:45 | |
| markymark | PaulMurray: OK, let's use "live-migration" then and switch those bugs to the proper tag: https://bugs.launchpad.net/nova/+bugs?field.searchtext=&orderby=-importance&field.status%3Alist=NEW&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&assignee_option=an | 16:45 |
| markymark | y&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.tag=live-migrate&field.tags_combinator=ANY&field.has_cve.used=&field.omit_dupes.used=&field.omit_dupes=on&field.affects_me.used=&field.has_patch.used=&field.has_branches.used=&field.has_branches=on&field.has_no_branches.used=&field.has_no_branches=on&field.has_blueprints.used=&field.has_blueprints=on&field.ha | 16:45 |
| markymark | s_no_blueprints.used=&field.has_no_blueprints=on&search=Search | 16:45 |
| markymark | whoops, that was bigger than I thought :-| | 16:45 |
| markymark | PaulMurray: http://bit.ly/1XUHNkg <= Would you change those to the correct tag then? | 16:47 |
| *** su_zhang has quit IRC | 16:47 | |
| markymark | PaulMurray: I'm using then "live-migration" in bug skimming in the future. | 16:47 |
| *** mgoddard__ has quit IRC | 16:48 | |
| PaulMurray | ok | 16:48 |
| *** sudipto__ has quit IRC | 16:49 | |
| *** vilobhmm11 has quit IRC | 16:49 | |
| *** Marga_ has joined #openstack-nova | 16:49 | |
| *** vilobhmm11 has joined #openstack-nova | 16:49 | |
| *** dshaw has quit IRC | 16:51 | |
| markymark | PaulMurray: Cool cool, thanks! I've added also an entry to https://wiki.openstack.org/wiki/Nova/BugTriage#Step_2:_Triage_Tagged_Bugs | 16:51 |
| *** pratikma_ has joined #openstack-nova | 16:52 | |
| *** vilobhmm11 has quit IRC | 16:54 | |
| *** ccesario has quit IRC | 16:54 | |
| *** tjones has quit IRC | 16:54 | |
| superdan | lascii: johnthetubaguy: https://review.openstack.org/#/c/242206/ | 16:55 |
| superdan | is that similar to what you had in mind? | 16:55 |
| *** pratikmallya has quit IRC | 16:55 | |
| *** achanda has joined #openstack-nova | 16:56 | |
| *** tjones has joined #openstack-nova | 16:56 | |
| *** Sree has joined #openstack-nova | 16:57 | |
| superdan | johnthetubaguy: lascii that will make api also fail to start if you try to start one older than every other api in the deployment | 16:57 |
| *** Sree has quit IRC | 17:01 | |
| *** mgoddard_ has quit IRC | 17:01 | |
| *** mgoddard has joined #openstack-nova | 17:01 | |
| openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/nova: Added new scheduler filter: AggregateTypeExtraSpecsAffinityFilter https://review.openstack.org/189279 | 17:04 |
| PaulMurray | markymark, did you sneak in a commitment to triage all the bugs behind my back then? | 17:04 |
| *** unicell1 has quit IRC | 17:04 | |
| *** jlanoux has quit IRC | 17:04 | |
| *** unicell has joined #openstack-nova | 17:05 | |
| *** ctrath has quit IRC | 17:05 | |
| *** ccarmack has quit IRC | 17:05 | |
| *** jdurgin has quit IRC | 17:06 | |
| *** rk4n has quit IRC | 17:06 | |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/nova: Nested Quota Driver: Get Hierarchy info https://review.openstack.org/242568 | 17:06 |
| lascii | lascii: which three uuids are you referencing above? | 17:07 |
| lascii | dangit | 17:08 |
| superdan | I see that baby exhaustion is in full effect | 17:08 |
| lascii | superdan: which three uuids are you referencing above? | 17:08 |
| lascii | it has a bit of a hold on me atm | 17:08 |
| superdan | lascii: the volume, snapshot, or .. one other | 17:09 |
| superdan | lemme get a link | 17:09 |
| superdan | lascii: https://review.openstack.org/#/c/241390/2/nova/cells/messaging.py,cm | 17:09 |
| *** ccarmack has joined #openstack-nova | 17:11 | |
| *** ctrath has joined #openstack-nova | 17:12 | |
| *** browne has joined #openstack-nova | 17:15 | |
| *** ssurana has joined #openstack-nova | 17:16 | |
| lascii | I'm not sure that snapshot_id or image_id are safe to treat as unique here | 17:17 |
| superdan | yeah, that's my concern | 17:17 |
| *** romainh has joined #openstack-nova | 17:18 | |
| *** sahid has quit IRC | 17:19 | |
| lascii | I would like to use device_name/volume_id, but that fails if device name is ever nulled. I was thinking that we could keep it from ever being null in the parent cell db | 17:19 |
| lascii | we perform a check for that in the object create method, but not in save | 17:20 |
| lascii | in the spirit of these other PoC patches let me push something up with what I'm thinking | 17:21 |
| *** e0ne has quit IRC | 17:21 | |
| superdan | lascii: so, | 17:21 |
| johnthetubaguy | superdan: loving the API service record stuff, we should maybe add a specless blueprint for this, process wise | 17:22 |
| superdan | lascii: I wonder if we could get away with stashing something in connection_info, potentially hidden by the object from the upper layers and use that until we get a uuid? | 17:22 |
| *** rlrossit has left #openstack-nova | 17:22 | |
| superdan | johnthetubaguy: I *knew* you were going to say that :) | 17:23 |
| johnthetubaguy | PaulMurray: markymark: yeah, oops, I wish I could remember what the live-migrate API is called for longer than 5 mins, and just stick to that one | 17:23 |
| superdan | johnthetubaguy: we could do that, or we could attach it to this: https://review.openstack.org/#/c/229062/ | 17:23 |
| johnthetubaguy | superdan: :) | 17:23 |
| openstackgerrit | Ludovic Beliveau proposed openstack/nova: Update binding:profile for SR-IOV ports https://review.openstack.org/242573 | 17:23 |
| johnthetubaguy | superdan: sure, I buy that too | 17:23 |
| superdan | johnthetubaguy: okay | 17:24 |
| lascii | superdan: stashing what though? I'm not sure there's a clear identifier that will always work atm | 17:24 |
| superdan | lascii: if we generate a uuid on the child cell and put it into connection_info, then we could use that in the parent cell to disambiguate | 17:25 |
| superdan | lascii: we'd need some graceful handling of the upgrade situation, but presumably all BDMs are in both places at the point at which you roll the new code, so as long as we're graceful about it not being there... | 17:25 |
| openstackgerrit | Andrew Laski proposed openstack/nova: DNM Cells: don't sync up bdm with no device_name https://review.openstack.org/242577 | 17:25 |
| lascii | superdan: oh I see, I didn't realize you meant store in the db as well | 17:26 |
| superdan | how many parallel hail mary patches are we up to now? | 17:26 |
| superdan | lascii: yeah, store it | 17:26 |
| *** krtaylor has quit IRC | 17:26 | |
| superdan | lascii: if we did that, then we could add the column to the DB and start migrating into that when it's available | 17:27 |
| lascii | superdan: why connection_info, because it's a text field? I would think device_name if we went that route | 17:28 |
| lascii | oh, nvm | 17:28 |
| superdan | lascii: just because we can stash things in there without breaking anything else | 17:28 |
| superdan | presumably we don't want to hijack device_name | 17:28 |
| lascii | yeah, just realized that would be bad | 17:28 |
| superdan | this aside, that cells path is not using objects, right? | 17:28 |
| lascii | it if | 17:29 |
| lascii | bah, *is | 17:29 |
| superdan | okay | 17:29 |
| *** stackdump has joined #openstack-nova | 17:29 | |
| superdan | I think we could add uuid to the object, and strip that in/out during load/save without anything above noticing | 17:29 |
| superdan | or | 17:30 |
| superdan | we just add the uuid column, be graceful about it, and say the bug isn't fixed until you roll that | 17:30 |
| *** subscope has joined #openstack-nova | 17:30 | |
| lascii | we did something similar to avoid instance.save() loops with cells so there's precedent for hiding info. My concern is I don't want a temp fix to steal motivation for a longer fix | 17:32 |
| superdan | no, the temp fix would come with the long-term fix I think | 17:32 |
| superdan | but we can certainly do the long fix and just make it graceful so the bug doesn't go away until you expand your schema | 17:33 |
| *** klkumar has quit IRC | 17:34 | |
| lascii | I'm not sure I understand the temp fix coming with the long term fix. | 17:35 |
| lascii | But I'm good to whip something up so we can debate on a review | 17:35 |
| superdan | the temp fix would be something we could backport | 17:36 |
| *** klkumar has joined #openstack-nova | 17:36 | |
| superdan | but I'm good with not doing that | 17:38 |
| *** zhangjn has quit IRC | 17:38 | |
| superdan | I can put up the db and base object change for us to consider if you want | 17:38 |
| lascii | sure, if you're itching to do it | 17:38 |
| superdan | I'm not itching to do it, but I would like to see this stop failing soon :) | 17:38 |
| openstackgerrit | Maxim Nestratov proposed openstack/nova-specs: libvirt: runtime image type detection https://review.openstack.org/240163 | 17:39 |
| *** krtaylor has joined #openstack-nova | 17:40 | |
| *** mc_nair has quit IRC | 17:41 | |
| lascii | superdan: okay. if you're good to take it I'll leave it with you so I can get lunch and then write/revise some cells specs | 17:41 |
| *** unicell has quit IRC | 17:42 | |
| superdan | lascii: I'll do that part and we can circle back with mriedem to see what he thinks and maybe get him do to the higher level updates on top to actually use it | 17:42 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add service records for nova-api services https://review.openstack.org/242206 | 17:43 |
| *** lpetrut has joined #openstack-nova | 17:44 | |
| *** mc_nair has joined #openstack-nova | 17:48 | |
| *** mhickey_ has quit IRC | 17:50 | |
| *** rk4n has joined #openstack-nova | 17:51 | |
| *** pradk has quit IRC | 17:52 | |
| *** lpetrut has quit IRC | 17:52 | |
| *** vilobhmm11 has joined #openstack-nova | 17:54 | |
| *** IanGovett has quit IRC | 17:54 | |
| *** jistr has quit IRC | 17:54 | |
| *** cdent has quit IRC | 17:56 | |
| *** dane-fichter has quit IRC | 17:57 | |
| *** Sree has joined #openstack-nova | 17:57 | |
| *** tonytan4ever has quit IRC | 17:58 | |
| *** tonytan4ever has joined #openstack-nova | 18:00 | |
| *** Sree has quit IRC | 18:02 | |
| *** su_zhang has joined #openstack-nova | 18:02 | |
| *** yassine has quit IRC | 18:03 | |
| *** su_zhang has quit IRC | 18:03 | |
| *** su_zhang has joined #openstack-nova | 18:04 | |
| *** romainh has left #openstack-nova | 18:05 | |
| *** mriedem_away is now known as mriedem | 18:09 | |
| mriedem | roar i'm back! | 18:09 |
| mriedem | i don't need to read scrollback do i? | 18:09 |
| mriedem | my left eye is official f'ed, soon i'll be wearing a patch like an old sea captain | 18:10 |
| *** achanda has quit IRC | 18:12 | |
| *** unicell has joined #openstack-nova | 18:12 | |
| *** danpb has quit IRC | 18:13 | |
| *** ctrath has quit IRC | 18:13 | |
| lascii | time for one of these then http://www.amazon.com/Smiffys-Parrot-Inches-Elastic-Holder/dp/B003BR9F4G/ref=sr_1_4 | 18:14 |
| kashyap | mriedem: Ugh, is that a coincidence, I just came back from the doc. about a screwed up left eye! | 18:16 |
| mriedem | no | 18:16 |
| mriedem | it's contageous | 18:16 |
| *** tonytan_brb has joined #openstack-nova | 18:17 | |
| *** browne has quit IRC | 18:17 | |
| *** tonytan_brb has quit IRC | 18:17 | |
| mriedem | like cells DNM patches | 18:17 |
| kashyap | Crazy, I'm typing this with one eye closed - not fun | 18:17 |
| mriedem | bauwser: so https://review.openstack.org/#/c/241390/ actually hit the cells failure | 18:17 |
| mriedem | have you dug into the logs? | 18:17 |
| *** ccarmack has quit IRC | 18:18 | |
| *** klkumar has quit IRC | 18:19 | |
| *** PaulMurray is now known as ptm_away | 18:19 | |
| *** tonytan4ever has quit IRC | 18:19 | |
| *** ctrath has joined #openstack-nova | 18:20 | |
| *** jichen has joined #openstack-nova | 18:20 | |
| *** e0ne has joined #openstack-nova | 18:21 | |
| *** cdent has joined #openstack-nova | 18:21 | |
| *** tjones has quit IRC | 18:21 | |
| *** lsmola has quit IRC | 18:21 | |
| *** klkumar has joined #openstack-nova | 18:21 | |
| *** ccarmack has joined #openstack-nova | 18:24 | |
| *** cdent has quit IRC | 18:24 | |
| *** dsanders has joined #openstack-nova | 18:24 | |
| *** tjones has joined #openstack-nova | 18:25 | |
| *** salv-orlando has quit IRC | 18:25 | |
| *** salv-orlando has joined #openstack-nova | 18:26 | |
| *** stackdump has quit IRC | 18:28 | |
| *** stackdump has joined #openstack-nova | 18:29 | |
| *** rotbeard has quit IRC | 18:29 | |
| *** jdurgin has joined #openstack-nova | 18:31 | |
| *** pratikma_ has quit IRC | 18:32 | |
| *** achanda has joined #openstack-nova | 18:34 | |
| *** otter768 has joined #openstack-nova | 18:35 | |
| *** achanda has quit IRC | 18:35 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Remove skips for resize tests from tempest-dsvm-cells-rc https://review.openstack.org/242594 | 18:37 |
| mriedem | my debug logging change hit the cells failure http://logs.openstack.org/17/242517/2/check/gate-tempest-dsvm-cells/feadce1/console.html#_2015-11-06_16_48_44_720 so i'll start digging into the logs there | 18:38 |
| *** emagana has quit IRC | 18:38 | |
| *** otter768 has quit IRC | 18:40 | |
| superdan | mriedem: did you read the scrollback between lascii and I? | 18:43 |
| mriedem | yeah, stashing uuid in connection_info ftw | 18:43 |
| superdan | well, | 18:44 |
| superdan | lascii: nak'd that in favor of just doing the full uuid column approach | 18:44 |
| superdan | which isn't backportable, but I'm okay with that | 18:44 |
| bauwser | mriedem: I'm back (was afk for dinner), scrolling back | 18:45 |
| *** paul-carlton has quit IRC | 18:46 | |
| mriedem | cells doesn't support rolling upgrades anyway | 18:47 |
| bauwser | mriedem: so yeah I saw the change hitted the cells issue, it seems the problem is probably that it didn't found the good BDM | 18:48 |
| *** achanda has joined #openstack-nova | 18:48 | |
| lascii | well, I'm fine with connection_info in the short term, as long as we also move towards uuids | 18:49 |
| openstackgerrit | Tiago Mello proposed openstack/nova: libvirt: Handle unsupported Spice feature per arch https://review.openstack.org/203809 | 18:50 |
| *** mnestratov has quit IRC | 18:54 | |
| superdan | mriedem: no, not rolling, but we can't suddenly break everyone's rows in their DBs | 18:55 |
| superdan | mriedem: I meant backportable to libuerty | 18:56 |
| *** ildikov has joined #openstack-nova | 18:56 | |
| bauwser | mriedem: uh, just reading the result for your eye, hope it can go better | 18:56 |
| mriedem | superdan: so you know how yesterday we were looking at what was creating the bdm in the first place? | 18:58 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/compute/api.py#L1281 | 18:58 |
| mriedem | and how the cells api code overrides that and noops it? | 18:58 |
| *** Sree has joined #openstack-nova | 18:58 | |
| mriedem | i added this logging and hit the failure https://review.openstack.org/#/c/242517/2/nova/compute/api.py | 18:58 |
| mriedem | but that log message doesn't show up | 18:58 |
| bauwser | mriedem: superdan: lascii: what do you think of the idea to provide the original device_name when calling save() ? | 18:58 |
| mriedem | oh nvm http://logs.openstack.org/17/242517/2/check/gate-tempest-dsvm-cells/feadce1/logs/screen-n-cell-child.txt.gz#_2015-11-06_16_32_41_043 | 18:59 |
| mriedem | different log file | 18:59 |
| bauwser | so we could do a lookup on the original one if device_name is within the updates | 18:59 |
| *** browne has joined #openstack-nova | 18:59 | |
| mriedem | bauwser: i feel like a uuid would be just as easy | 19:00 |
| mriedem | unless we know the prevoius value in the object before saving it off | 19:00 |
| mriedem | which i'm not sure we do | 19:00 |
| bauwser | mriedem: it needs a migration, right? | 19:00 |
| mriedem | the uuid can be stashed in bdm.connection_info, which is a json string | 19:00 |
| mriedem | so the wild west | 19:00 |
| bauwser | mriedem: I mean, I'm fine with an uuid, I just feel a bit torn if it needs to add a migration | 19:00 |
| superdan | uuid is a much better way to go I think | 19:01 |
| superdan | bauwser: why? adding a uuid is easy | 19:01 |
| mriedem | uuid is pretty simple | 19:01 |
| superdan | much easier than a new UC | 19:01 |
| mriedem | column add should be quick | 19:01 |
| mriedem | and then we don't need to monkey with enforcing constraints and logic in the code | 19:01 |
| bauwser | superdan: because it's a bit invasive, but okay | 19:01 |
| superdan | it's not as straightforward in the code because BDMs are so eff'd up anyway, but I'm getting there | 19:01 |
| bauwser | k, fair enough then :) | 19:01 |
| superdan | bauwser: it's also the right approach, and not that invasive, IMHO | 19:01 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Replaced deprecated timeutils methods https://review.openstack.org/241179 | 19:02 |
| *** su_zhang has quit IRC | 19:02 | |
| superdan | it's just not backportable | 19:02 |
| bauwser | yeah that | 19:02 |
| superdan | well, that's why I was suggesting a connection_info hack | 19:02 |
| superdan | but the more I think about it, | 19:02 |
| superdan | it's just easier if we only have two things to consider, instead of three | 19:02 |
| bauwser | agreed on the right approach, we just left it out in our previous discussions because of the migration stuff | 19:02 |
| superdan | 1. original 2. hack 3. uuid | 19:02 |
| *** Sree has quit IRC | 19:02 | |
| lascii | superdan: +1 | 19:03 |
| bauwser | but if we consider adding an uuid is acceptable, then I'm +1 to it | 19:03 |
| superdan | I'm almost done with it | 19:03 |
| bauwser | it's even much better from a conceptual PoV | 19:03 |
| bauwser | superdan: cool | 19:04 |
| *** praneshp has joined #openstack-nova | 19:04 | |
| *** emagana has joined #openstack-nova | 19:05 | |
| superdan | man, 11am and I haven't had a shred to eat | 19:05 |
| * superdan hungers | 19:05 | |
| bauwser | heh | 19:05 |
| superdan | you might even say I'm: https://www.youtube.com/watch?v=oOg5VxrRTi0 | 19:05 |
| mriedem | c'mon | 19:06 |
| bauwser | man, I have a very precise idea in mind, but that's unfortunately a french movie, so I guess the joke is unbearable | 19:06 |
| mriedem | more like https://www.youtube.com/watch?v=2ssCL292DQA | 19:06 |
| mriedem | ooo la la | 19:07 |
| mriedem | nobody puts cells in the corner | 19:07 |
| *** e0ne has quit IRC | 19:08 | |
| mriedem | bauwser: btw, you can see here that the first time we create the bdm in the api code the device_name isn't set http://logs.openstack.org/17/242517/2/check/gate-tempest-dsvm-cells/feadce1/logs/screen-n-cell-child.txt.gz#_2015-11-06_16_45_08_042 | 19:10 |
| bauwser | wtf? | 19:11 |
| bauwser | you speak of the first save() we do ? | 19:12 |
| bauwser | in the manager? | 19:12 |
| bauwser | compute manager | 19:12 |
| bauwser | mriedem: ^ | 19:12 |
| mriedem | no | 19:12 |
| mriedem | bauwser: this https://github.com/openstack/nova/blob/master/nova/compute/api.py#L1281 | 19:12 |
| mriedem | it happens in the child cell | 19:12 |
| *** e0ne has joined #openstack-nova | 19:12 | |
| bauwser | sec | 19:13 |
| bauwser | that's a compute.api stuff | 19:13 |
| bauwser | just finding the call path | 19:13 |
| *** aix has quit IRC | 19:15 | |
| bauwser | oh right, because it's done by the cells scheduler | 19:15 |
| bauwser | when creating the instance | 19:15 |
| *** smurke has joined #openstack-nova | 19:16 | |
| mriedem | yup, volume_id is set, device_name isn't | 19:16 |
| bauwser | because it doesn't get the user request, right? | 19:17 |
| bauwser | I mean the options | 19:17 |
| *** tonytan4ever has joined #openstack-nova | 19:17 | |
| mriedem | let me see if the test passes it in | 19:17 |
| mriedem | yeah tempest passes in the vda hard-coded device name on the request to create the server | 19:20 |
| mriedem | 'device_name': 'vda', | 19:20 |
| mriedem | i'm not really sure why it's not in http://logs.openstack.org/17/242517/2/check/gate-tempest-dsvm-cells/feadce1/logs/screen-n-cell-child.txt.gz#_2015-11-06_16_45_08_042 then | 19:20 |
| mriedem | on the initial create | 19:20 |
| bauwser | not sure how the base options are populated down to the cells scheduler | 19:21 |
| mriedem | interesting, look for "Calling bdm_update_or_create_at_top from child cell" here http://logs.openstack.org/17/242517/2/check/gate-tempest-dsvm-cells/feadce1/logs/screen-n-cell-child.txt.gz | 19:23 |
| mriedem | only shows up twice, and for different instances | 19:23 |
| mriedem | neither is from the test that failed | 19:24 |
| *** achanda has quit IRC | 19:24 | |
| mriedem | nothing hit the warning i added here https://review.openstack.org/#/c/242517/2/nova/db/sqlalchemy/api.py | 19:24 |
| mriedem | well, f it, uuid column it is | 19:25 |
| bauwser | mriedem: shouldn't we provide yet another DNM for just getting the list of BDMs ? | 19:25 |
| mriedem | where? | 19:25 |
| mriedem | https://review.openstack.org/#/c/242517/ is dnm for logging, whatever we want | 19:26 |
| bauwser | in the cells.messaging lovely method | 19:26 |
| *** pratikmallya has joined #openstack-nova | 19:27 | |
| mriedem | idk, i think i'm just going to way for dan's patch at this point | 19:28 |
| mriedem | *wait | 19:28 |
| *** electrocucaracha has quit IRC | 19:28 | |
| sdague | so, the cells job is at 50% failure rate which means tons of people are now burning all the CI resources trying to get their nova patches to pass | 19:28 |
| sdague | which is slowing down fixing this | 19:28 |
| sdague | can we make this non voting? | 19:28 |
| mriedem | we can skip it and unskip in whatever change we think fixes it | 19:28 |
| bauwser | fair point | 19:28 |
| mriedem | we don't need to make the job non-voting | 19:28 |
| bauwser | we can add it in the RC file | 19:29 |
| mriedem | right | 19:29 |
| bauwser | so yeah, we could just remove it in any change we'd like to test | 19:29 |
| sdague | at 50% fail, for a full week, it's no longer responsible to have it voting | 19:29 |
| *** achanda has joined #openstack-nova | 19:29 | |
| *** pratikma_ has joined #openstack-nova | 19:29 | |
| bauwser | sdague: it's failing for a single test, hence the rc file | 19:29 |
| mriedem | bauwser: just restore your skip patch | 19:29 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add uuid column to BlockDeviceMapping https://review.openstack.org/242602 | 19:30 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make BlockDeviceMapping object support uuid https://review.openstack.org/242603 | 19:30 |
| superdan | sdague: is it really 50%? | 19:30 |
| bauwser | sec | 19:30 |
| *** dave-mccowan has quit IRC | 19:30 | |
| superdan | sdague: we sometimes have a hard time reproducing it | 19:30 |
| sdague | superdan: yes | 19:30 |
| superdan | mriedem: bauwser: lascii: ^ | 19:30 |
| superdan | sdague: okay | 19:30 |
| mriedem | yeah looking | 19:30 |
| sdague | http://tinyurl.com/nr3c9ng | 19:30 |
| mriedem | so we'll skip with bauwser's change and unskip with superdan's | 19:30 |
| superdan | sdague: just to be clear, several of us have been working on it all week | 19:31 |
| superdan | some hard, others (like me) not so hard | 19:31 |
| bauwser | https://review.openstack.org/#/c/241258/ is reopened | 19:31 |
| sdague | superdan: sure, which is fine. That's where non-voting would let that continue to get the fail rate back under control | 19:31 |
| sdague | but not cause a ton of collateral damage | 19:31 |
| mriedem | bauwser: you have to address that comment | 19:31 |
| mriedem | sdague: we're on it | 19:32 |
| superdan | mriedem: someone needs to put something on top of mine to actually use the uuid during the high-level update-or-create call I think | 19:32 |
| mriedem | i can do that in a sec | 19:32 |
| bauwser | mriedem: agreed, just wonder if we shouldn't just exclude the one we know failing | 19:32 |
| mriedem | let's get bauwser's skip patch approved | 19:32 |
| sdague | mriedem: by doing it as a skip, you do loose the data set | 19:32 |
| mriedem | bauwser: you have to skip it in both v1 and v2 | 19:32 |
| *** pratikmallya has quit IRC | 19:32 | |
| superdan | if we can control this via our own repo, then we can just unskip in mriedem's patch | 19:32 |
| superdan | which is probably okay | 19:32 |
| bauwser | mriedem: okay, spinning a new change | 19:32 |
| superdan | it means we won't be able to merge a non-unskipping fix and get wide data on it | 19:33 |
| sdague | superdan: sure, if you don't need the test volume to figure out if it's working or not | 19:33 |
| mriedem | yeah i want to control this in our repo, not project-config | 19:33 |
| bauwser | I just need to find my local branch :D | 19:33 |
| *** Marga_ has quit IRC | 19:33 | |
| *** Piet has quit IRC | 19:33 | |
| superdan | sdague: yeah, I don't really know what I prefer, tbh | 19:33 |
| superdan | I don't want to expose all of cells just because of this test | 19:33 |
| sdague | personally I've always found the non-voting to be better | 19:33 |
| superdan | if we're reasonably confident that we've fixed it, it won't hurt to unskip and then reskip if we're wrong | 19:33 |
| mriedem | we're going to have to skip it in stable/liberty anyway, so we need the rc change to backport | 19:34 |
| sdague | because you can then ask reviewers to look at a thing for a while | 19:34 |
| mriedem | since we aren't going to backport the db migration | 19:34 |
| sdague | but you can still know you got it | 19:34 |
| mriedem | b/c branchless tempest | 19:34 |
| superdan | sdague: unless something else regresses | 19:34 |
| superdan | I think I'd prefer to skip/unskip once.. if we get it wrong, then we should make it non-voting and wait for the test data | 19:35 |
| sdague | superdan: yeh, sure. | 19:35 |
| mriedem | superdan: +1 to that plan | 19:35 |
| superdan | this change we're going to make is the "right" one, so less questionably going to fix it | 19:35 |
| superdan | more definitely going to fix it | 19:35 |
| superdan | whatever, you know what I mean | 19:35 |
| mriedem | jinx | 19:35 |
| mriedem | you just jinxed us | 19:35 |
| *** openstackstatus has joined #openstack-nova | 19:35 | |
| *** ChanServ sets mode: +v openstackstatus | 19:35 | |
| sdague | but we need to be a bit more sensitive to the carnage this is causing everyone else, after 48 hrs in a state like this, it really needs to be skipped or nv | 19:36 |
| sdague | we're now on day 5 | 19:36 |
| mriedem | day 3 | 19:36 |
| mriedem | started on 11/3 | 19:36 |
| superdan | well, we wouldn't do that for the core jobs | 19:36 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Exclude all BDM checks for cells https://review.openstack.org/241258 | 19:37 |
| mriedem | anyway, we're skipping ^ | 19:37 |
| mriedem | so let's stop talking about it | 19:37 |
| superdan | mriedem: you +2 it and I'll +W.. I prolly can't verify the correctness of that rehex | 19:37 |
| superdan | er, regex | 19:37 |
| bauwser | once we're confident the issue is fixed, we can revert my change | 19:38 |
| superdan | bauwser: we have to put the revert in or on top of the new patches so we can get runs, but that's doable | 19:38 |
| bauwser | did I f*ck'd something ? | 19:38 |
| mriedem | bauwser: no it's ok | 19:38 |
| mriedem | yeah i can handle the revert | 19:38 |
| mriedem | we're also going to be backporting to stable | 19:38 |
| lascii | bauwser: I think they're saying we need a revert up to hang patches onto | 19:39 |
| bauwser | oh yeah | 19:39 |
| bauwser | can do, sec | 19:39 |
| bauwser | or mriedem | 19:39 |
| bauwser | so you or me ? | 19:39 |
| bauwser | :) | 19:39 |
| superdan | bauwser: let mriedem do it | 19:39 |
| mriedem | just hold up for 2 seconds | 19:39 |
| superdan | less confusion in the stacking | 19:39 |
| bauwser | ok | 19:39 |
| bauwser | fair point | 19:40 |
| mriedem | 1. i'm reviewing dan's changes, 2. we put a change on top of https://review.openstack.org/#/c/242603/ that uses it in parent cell update_or_create, 3. revert skip | 19:40 |
| -openstackstatus- NOTICE: Gerrit will be offline at 20:00-20:15 UTC today (starting 20 minutes from now) for scheduled project rename maintenance | 19:40 | |
| superdan | oh jeez | 19:40 |
| mriedem | heh | 19:40 |
| superdan | this all ends in 20 minutes | 19:40 |
| superdan | better hope that skip merges | 19:40 |
| bauwser | mriedem: +1 to that plan | 19:42 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make BlockDeviceMapping object support uuid https://review.openstack.org/242603 | 19:43 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add uuid column to BlockDeviceMapping https://review.openstack.org/242602 | 19:43 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add service records for nova-api services https://review.openstack.org/242206 | 19:45 |
| *** dave-mccowan has joined #openstack-nova | 19:46 | |
| *** daemontool has quit IRC | 19:47 | |
| *** salv-orl_ has joined #openstack-nova | 19:48 | |
| *** stackdump has quit IRC | 19:48 | |
| mriedem | superdan: bunch of comments in ps1 of https://review.openstack.org/#/c/242602/ | 19:51 |
| *** salv-orlando has quit IRC | 19:51 | |
| lascii | superdan: were you going to change the order of create/update in https://review.openstack.org/#/c/242206/5/nova/service.py,cm ? | 19:53 |
| superdan | lascii: yeah, I didn't see that until I pushed | 19:53 |
| superdan | because you silly people don't -1 | 19:53 |
| *** pratikma_ has quit IRC | 19:53 | |
| lascii | there you go :) | 19:53 |
| lascii | lgtm otherwise | 19:54 |
| *** e0ne has quit IRC | 19:54 | |
| superdan | mriedem: replied | 19:56 |
| superdan | lascii: trying to get the fix up before gerrit closes the curtain | 19:56 |
| *** Marga_ has joined #openstack-nova | 19:56 | |
| superdan | not that it matters | 19:56 |
| *** tonytan4ever has quit IRC | 19:57 | |
| *** pixelbeat has quit IRC | 19:57 | |
| superdan | lascii: I figured we might expose api services later, I just didn't want to get into the api crap in that patch | 19:58 |
| superdan | lascii: we also need to think about a few things, since they're not updating their records periodically | 19:59 |
| lascii | yeah, I didn't expect you too, was just mentioning it | 19:59 |
| superdan | so it might not make sense, or we might need to null out their states or something | 19:59 |
| *** lpetrut has joined #openstack-nova | 19:59 | |
| *** Sree has joined #openstack-nova | 19:59 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Add service records for nova-api services https://review.openstack.org/242206 | 19:59 |
| *** mgoddard_ has joined #openstack-nova | 20:00 | |
| lascii | superdan: fair point. it may make sense to consider liveness of api servers at some point, but no need to discuss now | 20:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Properly log BlockDeviceMappingList in _create_block_device_mapping https://review.openstack.org/242617 | 20:00 |
| superdan | lascii: I guess, I would have thought that would be unwelcome extra load from a lot of workers | 20:01 |
| superdan | lascii: perhaps we only do it if we're on a high-falutin' thing like zk | 20:01 |
| *** rk4n has quit IRC | 20:01 | |
| *** mmdurrant has quit IRC | 20:01 | |
| lascii | yeah. or maybe have a different checkin interval, dunno. | 20:02 |
| -openstackstatus- NOTICE: Gerrit is offline until 20:15 UTC today for scheduled project rename maintenance | 20:03 | |
| lascii | but everyone loves zookeeper so we can use it and ride off into the sunset | 20:03 |
| bauwser | ah snap | 20:03 |
| *** ChanServ changes topic to "Gerrit is offline until 20:15 UTC today for scheduled project rename maintenance" | 20:03 | |
| *** mgoddard has quit IRC | 20:03 | |
| bauwser | bye my comments | 20:03 |
| *** pratikmallya has joined #openstack-nova | 20:03 | |
| mmmpork | was there a memo about sillifying irc nicks that i missed?? ;) | 20:04 |
| *** Sree has quit IRC | 20:04 | |
| superdan | mmmpork: no, yours is already silly | 20:04 |
| superdan | so you're not on the cc list | 20:04 |
| mriedem | fart! | 20:04 |
| mmmpork | a girl's gotta have standards you know | 20:04 |
| mriedem | bye bye gerrit | 20:04 |
| mriedem | superdan: so block_device_mapping_destroy_by_instance_and_uuid - i'm thinking we don't need that | 20:05 |
| mmmpork | markymark: can i add my name to the project config funky bunch? | 20:05 |
| lascii | welp, lost my comments as well | 20:05 |
| bauwser | so, Gerrit is down for 20 mins, what's next on the TV ? | 20:05 |
| superdan | mriedem: did you see my reply? | 20:05 |
| *** pratikma_ has joined #openstack-nova | 20:05 | |
| lascii | mriedem: superdan I agree that we don't need to lookup by instance uuid | 20:05 |
| mriedem | superdan: yeah, something about possibly duplicate uuids in the parent cell | 20:05 |
| superdan | okay | 20:05 |
| mmmpork | markymark: i can either patch it and submit it or you can add my name, i just don't want to blow away the reviews people have done :/ | 20:05 |
| mriedem | which baffles me | 20:05 |
| mriedem | ok | 20:05 |
| superdan | if lascii thinks it's okay | 20:05 |
| lascii | superdan: we would have issues with instances as well, but we currently assume uuids will actually be unique | 20:06 |
| mriedem | when i was talking to odden about this earlier, he was wondering why we couldn't tell in the parent cell method that the id is unique within a given cell namespace or something | 20:06 |
| mriedem | to avoid pop'ing the id | 20:06 |
| superdan | lascii: but the parent has a UC on instance.uuid now | 20:06 |
| lascii | superdan: hmm, has it always, and do we actually handle a failure there? | 20:07 |
| mriedem | UC on instance.uuid was added in kilo | 20:07 |
| *** jroll is now known as tjroll | 20:07 | |
| * tjroll participates in casual friday, finally | 20:07 | |
| superdan | tjroll: +1 | 20:07 |
| superdan | tjroll: jtroll would be better, no? | 20:07 |
| mriedem | with a script to run before upgrade to kilo to see if you have null instance uuids | 20:07 |
| mriedem | or djroll | 20:08 |
| mriedem | if you're into that | 20:08 |
| tjroll | superdan: already registerd :( | 20:08 |
| superdan | bummer | 20:08 |
| *** pratikmallya has quit IRC | 20:08 | |
| mriedem | superdan: so didn't see the replies in the migration script about making it nullable | 20:08 |
| *** mmmpork is now known as mmmborkborkbork | 20:08 | |
| superdan | lascii: realized I had no test for that startup logic, which pointed out I had a majorly dumb bug in my code :) | 20:09 |
| superdan | lascii: (re: the service records one) | 20:09 |
| mriedem | superdan: i'm assuming we don't make the column non-nullable b/c we'd have to default the uuids on existing rows | 20:09 |
| *** mwagner has quit IRC | 20:09 | |
| lascii | superdan: is there a reason not to put a UC on bdm.uuid? with this in place we could create the bdm in the parent cell first, which we avoid now only because there's no device_name guaranteed at that point | 20:09 |
| superdan | mriedem: yeah | 20:09 |
| superdan | mriedem: we can't do that | 20:09 |
| mriedem | and we don't want to create the column, update all rows with a uuid, and then make it non-nullable in the migration? | 20:09 |
| superdan | lascii: because we have no default | 20:10 |
| superdan | lascii: it's kinda crazy that we have this insane loop between the parent and child just because there is no uuid in tthe table :( | 20:10 |
| superdan | mriedem: I guess we could, but it would require people to do this migration before rolling code | 20:11 |
| mriedem | lascii: how many bdm's are in rax production cloud dbs? | 20:11 |
| lascii | I am very aware | 20:11 |
| superdan | mriedem: right now, they can do it early, or while rolling out code knowing that it should be fast | 20:11 |
| lascii | mriedem: not sure, I can get a rough estimate from a region though | 20:11 |
| mriedem | updating every bdm entry in the db to add a uuid would be rough i suppose | 20:11 |
| mriedem | zzzeek: are you around? | 20:11 |
| bauwser | couldn't we add an uuid on the fly when saving ? | 20:11 |
| superdan | that's what I'm doing | 20:12 |
| bauwser | if noone is existing yet of course | 20:12 |
| superdan | they're talking about making it unique | 20:12 |
| bauwser | awesomeness | 20:12 |
| superdan | I really don't think that doing that at migration time is a good idea | 20:12 |
| bauwser | oh snap, missed that piece | 20:12 |
| superdan | I'd rather fix it after mitaka, which we can definitely do | 20:12 |
| superdan | mriedem: require them to do the online migration thing in mitaka, and then require that it's unique before we move to N, where we add the UC | 20:12 |
| superdan | like we did for flavors | 20:13 |
| bauwser | since it's an UUID, I feel it's less risky indeed | 20:13 |
| mriedem | sure, that's fine | 20:13 |
| superdan | then after that, we can drop this in cells and fix it up | 20:13 |
| *** salv-orl_ has quit IRC | 20:13 | |
| * mriedem takes a quick break | 20:14 | |
| lascii | huh, did not expect that. around 3 million bdm entries in one region | 20:14 |
| *** achanda has quit IRC | 20:15 | |
| superdan | region, meaning in the parent cell? | 20:15 |
| lascii | yes | 20:15 |
| superdan | can I just take a tangent for a quick set? | 20:15 |
| superdan | sec | 20:15 |
| lascii | most of those being deleted | 20:15 |
| superdan | I effing love that lascii can just jump on the RAX db (probably with root privs) and do a random query | 20:15 |
| superdan | right | 20:15 |
| tjroll | lol | 20:16 |
| lascii | well, a slave, but yeah :) | 20:16 |
| *** Nathan__ has joined #openstack-nova | 20:16 | |
| superdan | whatever :) | 20:16 |
| superdan | the only one I can get on has like 10 BDMs :) | 20:17 |
| lascii | scale up, it's the cloud | 20:18 |
| superdan | heh | 20:19 |
| *** praneshp has left #openstack-nova | 20:21 | |
| *** Piet has joined #openstack-nova | 20:21 | |
| sdague | superdan: sounds like you need some more equipment in your home data center | 20:22 |
| superdan | I most certainly do not | 20:22 |
| * superdan looks at his $300/mo power bill | 20:22 | |
| superdan | I do have a laser-etched sign outside the door to that room now though | 20:22 |
| superdan | my wife got certified on the laser at work and made it for me for my birthday | 20:23 |
| *** nihilifer has quit IRC | 20:23 | |
| *** rotbeard has joined #openstack-nova | 20:23 | |
| *** rfolco has quit IRC | 20:23 | |
| sdague | nice, though you should get one of these - https://glowforge.com/ | 20:24 |
| *** rotbeard has quit IRC | 20:24 | |
| *** rotbeard has joined #openstack-nova | 20:25 | |
| superdan | sdague: does it etch plate steel? if not, then it won't be as cool as mine :) | 20:25 |
| mmmborkborkbork | all hail the glowforge | 20:25 |
| sdague | superdan: they list titanium and aluminum in their etchable materials | 20:26 |
| sdague | https://glowforge.com/tech-specs | 20:27 |
| superdan | wow, that's scary | 20:27 |
| mriedem | sdague: where is the thing that decides that we don't run dsvm/tempest jobs on docs only or unit test only changes? | 20:30 |
| tjroll | mriedem: zuul/layout | 20:30 |
| tjroll | (in project-config) | 20:30 |
| mriedem | k | 20:30 |
| mriedem | yeah | 20:30 |
| tjroll | https://github.com/openstack-infra/project-config/blob/master/zuul/layout.yaml#L703 | 20:30 |
| mriedem | we should have to run unit tests/pep8 on these devstack rc files in the nova tree | 20:31 |
| mriedem | *shouldn't | 20:31 |
| *** tjones has quit IRC | 20:31 | |
| *** ildikov has quit IRC | 20:31 | |
| sdague | mriedem: so, probably, but the complexity of all the skips vs. the time saved, is pretty minor | 20:31 |
| sdague | the big win was dumping dsvm runs for things they couldn't possibly effect | 20:31 |
| mriedem | fine, you've talked me out of caring | 20:32 |
| sdague | heh | 20:32 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add service records for nova-api services https://review.openstack.org/242206 | 20:32 |
| *** tongli has quit IRC | 20:33 | |
| superdan | mriedem: did you see my reply about the update_or_create thing not letting you eff up the database if you provide a uuid and there are no matches? | 20:33 |
| mriedem | don't think so | 20:33 |
| mriedem | oh wait yeah | 20:33 |
| mriedem | i was replying to that when gerrit died | 20:33 |
| mriedem | i.e. we want to start using that thing in ambiguous ways | 20:34 |
| mriedem | like lance and gary | 20:34 |
| mriedem | s/lance/ace/ | 20:34 |
| superdan | um | 20:35 |
| openstackgerrit | jichenjc proposed openstack/nova: Add DiskFormat as Enum in fields https://review.openstack.org/229854 | 20:35 |
| openstackgerrit | jichenjc proposed openstack/nova: Make ImageMeta:disk_format use DiskFormatField https://review.openstack.org/231460 | 20:35 |
| *** aysyd has quit IRC | 20:35 | |
| dims | mriedem "project name migration work" (reason for Gerrit outage) | 20:36 |
| mriedem | i know | 20:36 |
| *** otter768 has joined #openstack-nova | 20:36 | |
| *** ChanServ changes topic to "This channel is for Nova development. For support of Nova deployments, please use #openstack. Please see: https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule" | 20:36 | |
| *** dims is now known as dimsum__ | 20:37 | |
| *** vilobhmm11 has quit IRC | 20:38 | |
| * mriedem goes back to reviewing what the ibm execs notes were from teh summit | 20:38 | |
| *** codebaus1 is now known as codebauss | 20:38 | |
| *** vilobhmm11 has joined #openstack-nova | 20:39 | |
| *** stackdump has joined #openstack-nova | 20:39 | |
| *** otter768 has quit IRC | 20:40 | |
| bauwser | good weekend, folks | 20:42 |
| *** bauwser is now known as bauzas | 20:42 | |
| mriedem | bauzas: later | 20:42 |
| *** lucas-afk has quit IRC | 20:43 | |
| *** Guest88047 is now known as jdg | 20:43 | |
| *** jdg is now known as jgriffith | 20:44 | |
| *** stackdump has quit IRC | 20:44 | |
| *** tjones has joined #openstack-nova | 20:45 | |
| superdan | mriedem: yeah, I could make the object always require a uuid and generate it if it reads a null one from the db | 20:48 |
| superdan | mriedem: the only problem would be if we pulled the record, grabbed the uuid, then didn't save and pulled it again and got a different one | 20:48 |
| mriedem | superdan: re: making the field nullable=False in the bdm objecT? | 20:48 |
| superdan | assuming we don't want to force the object to save on query | 20:48 |
| superdan | yes | 20:48 |
| mriedem | didn't we do a similar data conversion for flavors? | 20:49 |
| *** stackdump has joined #openstack-nova | 20:49 | |
| superdan | only on save, but it was less problematic there, | 20:49 |
| superdan | because we weren't making up an identifier | 20:49 |
| superdan | if you read twice and didn't save, you'd get the same result with flavors | 20:49 |
| *** lucasagomes has joined #openstack-nova | 20:49 | |
| mriedem | i guess i was thinking it'd be best to make the uuid field in the bdm object as nullable=False to avoid issues since we have nullable=True in the db schema | 20:50 |
| mriedem | that's how the instance.uuid stuff used to be | 20:50 |
| superdan | yep | 20:51 |
| openstackgerrit | Andrew Laski proposed openstack/nova-specs: Proposal for cells v2 scheduling interaction https://review.openstack.org/239995 | 20:51 |
| *** klkumar has quit IRC | 20:51 | |
| superdan | I agree, I'm just calling out the potential problem | 20:51 |
| superdan | but we should shoot for the moon and fix it if we uncover issues | 20:51 |
| *** daemontool has joined #openstack-nova | 20:52 | |
| *** mwagner has joined #openstack-nova | 20:52 | |
| *** salv-orlando has joined #openstack-nova | 20:53 | |
| mriedem | k | 20:55 |
| *** ccarmack has quit IRC | 20:55 | |
| openstackgerrit | Erickson Filipe Guedes dos Santos proposed openstack/nova: Nested Quota Driver https://review.openstack.org/242626 | 20:56 |
| *** harlowja_ has joined #openstack-nova | 20:56 | |
| *** harlowja has quit IRC | 20:56 | |
| *** BobBall is now known as BobBall_AWOL | 20:57 | |
| *** burgerk has quit IRC | 20:58 | |
| *** ccarmack has joined #openstack-nova | 20:59 | |
| *** ZZelle_ has joined #openstack-nova | 21:00 | |
| *** Sree has joined #openstack-nova | 21:00 | |
| *** ctrath has quit IRC | 21:01 | |
| *** openstackgerrit has quit IRC | 21:01 | |
| *** smatzek_ has quit IRC | 21:02 | |
| *** su_zhang has joined #openstack-nova | 21:02 | |
| *** openstackgerrit has joined #openstack-nova | 21:02 | |
| *** stackdump has quit IRC | 21:02 | |
| mriedem | gerrit is back, so back to work | 21:03 |
| dimsum__ | yay! | 21:03 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make BlockDeviceMapping object support uuid https://review.openstack.org/242603 | 21:03 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add uuid column to BlockDeviceMapping https://review.openstack.org/242602 | 21:03 |
| *** mriedem has left #openstack-nova | 21:03 | |
| openstackgerrit | jichenjc proposed openstack/nova: Add more 'actions' for server concepts doc https://review.openstack.org/236947 | 21:03 |
| *** mriedem has joined #openstack-nova | 21:04 | |
| leakypipes | guh, dilated eyes... can't read shit. | 21:04 |
| superdan | mriedem: ^ | 21:04 |
| mriedem | leakypipes: ha, i told them i wasn't going to do that | 21:04 |
| *** stackdump has joined #openstack-nova | 21:04 | |
| *** raildo is now known as raildo-afk | 21:04 | |
| mriedem | leakypipes: "are you going to do the extra thing to avoid dilating your eyes which costs extra money, or dilate your eyes?" "umm, can i do neither?" "..." | 21:04 |
| *** Sree has quit IRC | 21:05 | |
| superdan | I do the non-dilation thing | 21:05 |
| *** READ10 has joined #openstack-nova | 21:06 | |
| mriedem | i do too, when i have time and cash | 21:06 |
| superdan | heh | 21:06 |
| *** su_zhang has quit IRC | 21:06 | |
| *** ctrath has joined #openstack-nova | 21:09 | |
| *** thorst has quit IRC | 21:11 | |
| *** thorst has joined #openstack-nova | 21:12 | |
| mriedem | superdan: ok, comments in https://review.openstack.org/#/c/242602/ | 21:13 |
| mriedem | nothing serious | 21:13 |
| *** smurke has quit IRC | 21:14 | |
| *** READ10 has quit IRC | 21:14 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Make the ResourceTracker class configurable https://review.openstack.org/242636 | 21:14 |
| *** electrocucaracha has joined #openstack-nova | 21:16 | |
| *** thorst has quit IRC | 21:16 | |
| mriedem | i think your object compat check here is wrong https://review.openstack.org/#/c/242603/ | 21:16 |
| *** vladikr has quit IRC | 21:16 | |
| mriedem | otherwise LGTM | 21:16 |
| mriedem | figleaf: will you be offended if i proceduraly -2 that? | 21:17 |
| mriedem | since the bp isn't approved | 21:17 |
| superdan | mriedem: oops, yep, meant to fix that, thanks | 21:20 |
| figleaf | mriedem: sure | 21:20 |
| figleaf | I just wanted to get the code out there | 21:20 |
| mriedem | yeah, np | 21:20 |
| figleaf | mriedem: and it's so nice of you to think of my feelings. <3 | 21:21 |
| *** rotbeard has quit IRC | 21:21 | |
| mriedem | aww | 21:21 |
| *** thorst has joined #openstack-nova | 21:22 | |
| *** rfolco has joined #openstack-nova | 21:22 | |
| *** signed8bit has joined #openstack-nova | 21:22 | |
| mriedem | is there an easy way to refresh a series with dependent changes once those are updated in gerrit? | 21:24 |
| mriedem | w/o redoing my local topic branch | 21:24 |
| *** su_zhang has joined #openstack-nova | 21:24 | |
| superdan | just pull the new one down and then rebase your branch on it | 21:24 |
| mriedem | aight | 21:24 |
| *** slDabbler has joined #openstack-nova | 21:24 | |
| superdan | coming right up | 21:25 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make BlockDeviceMapping object support uuid https://review.openstack.org/242603 | 21:25 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add uuid column to BlockDeviceMapping https://review.openstack.org/242602 | 21:25 |
| *** lpetrut has quit IRC | 21:30 | |
| mriedem | do we want to be specific on bdm create? https://review.openstack.org/#/c/242602/4/nova/db/sqlalchemy/api.py | 21:33 |
| superdan | mriedem: yeah maybe so | 21:33 |
| mriedem | i was trying to think of how we handle this in bdm._create | 21:34 |
| superdan | we probably want to be diligent there so we don't get screwed by the less restrictive cells interface to these things | 21:34 |
| mriedem | yeah | 21:35 |
| *** subscope has quit IRC | 21:39 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Make BlockDeviceMapping object support uuid https://review.openstack.org/242603 | 21:39 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add uuid column to BlockDeviceMapping https://review.openstack.org/242602 | 21:39 |
| *** tonytan4ever has joined #openstack-nova | 21:39 | |
| *** jwcroppe has quit IRC | 21:40 | |
| *** jichen has quit IRC | 21:40 | |
| *** jwcroppe has joined #openstack-nova | 21:40 | |
| *** jwcroppe has quit IRC | 21:45 | |
| *** ZZelle_ has quit IRC | 21:46 | |
| *** jerrygb has quit IRC | 21:46 | |
| *** jerrygb has joined #openstack-nova | 21:47 | |
| mriedem | god i hate the logic in bdm_update_or_create_at_top | 21:49 |
| mgagne | mriedem: :D | 21:51 |
| *** tjones has left #openstack-nova | 21:51 | |
| mriedem | mgagne: don't tell me you guys have already fixed the cells bdm create races out of tree | 21:51 |
| superdan | mriedem: I hate most of the bdm code | 21:51 |
| mriedem | me 3 | 21:51 |
| mgagne | mriedem: only bug fix I made related to that: https://review.openstack.org/#/c/152997/ | 21:51 |
| *** jerrygb has quit IRC | 21:51 | |
| mriedem | i think i'll just push up what i have, but it feels very inefficient | 21:52 |
| lascii | phew, I'm not in git blame for that method | 21:52 |
| mgagne | in master: https://review.openstack.org/#/c/152257/ | 21:52 |
| *** rfolco has quit IRC | 21:52 | |
| superdan | mriedem: let's get it fixed and then we can start cutting out the nonsense | 21:52 |
| mgagne | mriedem: but your bug might not be related to mine which could only be triggered by booing on a volume without providing device name. | 21:53 |
| mgagne | mriedem: still waiting for Jenkins results for cells+neutron job so I can debug further | 21:54 |
| mriedem | hmm, since bauzas' skip change hasn't landed yet, i can't revert it in this series | 21:56 |
| *** alejandrito has quit IRC | 21:56 | |
| superdan | mriedem: well, no need until it lands anyway | 21:56 |
| mriedem | yeah | 21:56 |
| *** dave-mccowan has quit IRC | 21:57 | |
| *** Sree has joined #openstack-nova | 22:00 | |
| *** Sree has quit IRC | 22:05 | |
| *** pixelbeat has joined #openstack-nova | 22:07 | |
| openstackgerrit | Andrew Laski proposed openstack/nova-specs: WIP Cells: CellZero https://review.openstack.org/238665 | 22:08 |
| *** achanda has joined #openstack-nova | 22:08 | |
| lascii | superdan: decided to call the special cell cell0 for now. let the bikeshedding begin | 22:09 |
| *** jerrygb has joined #openstack-nova | 22:09 | |
| *** slDabbler has quit IRC | 22:10 | |
| *** jerrygb_ has joined #openstack-nova | 22:10 | |
| openstackgerrit | Andrew Laski proposed openstack/nova-specs: Cells: CellZero https://review.openstack.org/238665 | 22:11 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: cells: check for existing BDM based on uuid in bdm_update_or_create_at_top https://review.openstack.org/242649 | 22:12 |
| *** pixelbeat has quit IRC | 22:13 | |
| *** lascii is now known as alaski | 22:13 | |
| mriedem | CellZero (tm) | 22:13 |
| mriedem | superdan: alaski: so this was my thing, it's ugly, the logic bugs me | 22:13 |
| mriedem | https://review.openstack.org/#/c/242649/1/nova/cells/messaging.py | 22:13 |
| mriedem | no unit tests yet | 22:13 |
| *** ctrath has quit IRC | 22:14 | |
| *** jerrygb has quit IRC | 22:14 | |
| *** figleaf is now known as edleafe | 22:15 | |
| mriedem | at some point we'll need to update bdm_destroy_at_top or add a new one that passes the uuid | 22:15 |
| alaski | yeah, there's a lot to clean up I think | 22:16 |
| alaski | with the uuid in place I would like to have the bdm be created first in the parent cell, and then only updated from the children | 22:16 |
| mriedem | updating that change quick, the reduncant update code is bugging me | 22:20 |
| mriedem | if we found the bdm by uuid, don't try to lookup by device_name/vol_id | 22:21 |
| *** boris-42 has joined #openstack-nova | 22:21 | |
| *** mmmborkborkbork is now known as mmmpork | 22:22 | |
| mgagne | mriedem: only 1 test failing in gate-tempest-dsvm-full now: tempest.scenario.test_network_basic_ops.TestNetworkBasicOps.test_network_basic_ops[compute,id-f323b3ba-82f8-4db7-8ea6-6a895869ec49,network,smoke] | 22:22 |
| mgagne | mriedem: somehow it still complains about FIP | 22:22 |
| *** pratikma_ has quit IRC | 22:22 | |
| mgagne | mriedem: could be that tempest doesn't honour the "fixed" ssh connectivity config | 22:23 |
| mriedem | yeah | 22:23 |
| mriedem | that's what i was thinking | 22:23 |
| mgagne | mriedem: so I would ignore that test for now or fix it if it's simple | 22:23 |
| sfinucan | What do I do if a gate test keeps failing? Tests for this change keep flip-flopping :( https://review.openstack.org/#/c/237762/ | 22:24 |
| mriedem | mgagne: 5 failed tests http://logs.openstack.org/85/235485/6/check/gate-tempest-dsvm-neutron-full/2289378//console.html#_2015-11-06_21_56_10_455 | 22:24 |
| mriedem | but still, that's pretty damn good for cells + neutron | 22:24 |
| mriedem | with minimal changes | 22:24 |
| mriedem | alaski: ^ | 22:24 |
| mgagne | yea, trying to see each test results | 22:24 |
| mgagne | mriedem: ironic test is taking a LOT of time, probably caused by lot of timeout | 22:24 |
| mriedem | sfinucan: the cells fails are a known thing, there is a skip in the gate for that now | 22:24 |
| mriedem | we shouldn't be running ironic tests in the cells / neutron job | 22:25 |
| mriedem | i can disable those | 22:25 |
| mgagne | mriedem: cool. we can see to enable them back later as in our case, cells are the only way for us to run both virtual and metal instances | 22:25 |
| sfinucan | mriedem: OK, back to 'recheck' hell it seem, heh | 22:26 |
| sfinucan | Cheers | 22:26 |
| mriedem | sfinucan: https://review.openstack.org/#/c/241258/ | 22:26 |
| mriedem | just wait for that ^ | 22:26 |
| mriedem | otherwise you're just making noise | 22:26 |
| mriedem | with rechecks | 22:26 |
| sfinucan | mriedem: Ah, ok - no problem | 22:27 |
| sfinucan | anything in progress for xenserver? | 22:27 |
| *** achanda has quit IRC | 22:27 | |
| mriedem | mgagne: one of the failures was test_create_ebs_image_and_check_boot which we're skipping right now | 22:27 |
| mriedem | ^ | 22:27 |
| mriedem | sfinucan: ignore that one | 22:27 |
| mgagne | mriedem: also saw test_update_server_from_auto_to_manual | 22:28 |
| mgagne | mriedem: is there an easy way using some black magic to get the names of failing jobs and tests? | 22:28 |
| mriedem | i just grep for ... FAILED | 22:28 |
| mgagne | mriedem: I don't feel like digging through MB of logs in my browser | 22:28 |
| mriedem | in the console log | 22:28 |
| mriedem | or | 22:28 |
| mriedem | http://logs.openstack.org/85/235485/6/check/gate-tempest-dsvm-neutron-full/2289378/logs/testr_results.html.gz | 22:29 |
| alaski | mriedem: nice, 5 failures is really decent | 22:29 |
| mriedem | yar | 22:29 |
| mriedem | actually 4 | 22:29 |
| * sfinucan goes to drop colleague to airport | 22:29 | |
| *** sfinucan has quit IRC | 22:29 | |
| mriedem | one is the ebs global fail | 22:29 |
| alaski | even better | 22:29 |
| mriedem | mgagne: test_server_connectivity_pause_unpause is trying to create a floating IP | 22:30 |
| mriedem | so yeah, defintily some tempest bugs here | 22:30 |
| mgagne | mriedem: I'm poor, I don't have much ram. We should add ANT_SIZED_RAM=1 config so sc68cal doesn't feel alone in his misery | 22:31 |
| mriedem | mgagne: can't you just rip off resources from your own production cloud? | 22:32 |
| *** dustins has quit IRC | 22:33 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: cells: check for existing BDM based on uuid in bdm_update_or_create_at_top https://review.openstack.org/242649 | 22:33 |
| *** signed8bit has quit IRC | 22:33 | |
| mriedem | with ^ i'm going to let it run and check results after i get home, gotta pick up the kid - plus solo parenting tonight | 22:36 |
| mriedem | so pb + j's all around and lots of TV time! | 22:36 |
| andrewbogott | in my jenkins tests I’m seeeing a lot of 'ERROR_ON_CLONE=True’ — am I correct in thinking that that’s a failure of the CI infra and not actually a problem with my patch? | 22:36 |
| andrewbogott | I don’t have much practice understanding all these test logs :( | 22:36 |
| *** otter768 has joined #openstack-nova | 22:37 | |
| *** mriedem has quit IRC | 22:37 | |
| *** otter768 has quit IRC | 22:42 | |
| *** rk4n has joined #openstack-nova | 22:45 | |
| *** rk4n has quit IRC | 22:48 | |
| *** thorst has quit IRC | 22:54 | |
| *** thorst has joined #openstack-nova | 22:54 | |
| *** IanGovett has joined #openstack-nova | 22:55 | |
| superdan | alaski: heh, cool :) | 22:57 |
| *** su_zhang has quit IRC | 22:57 | |
| *** IanGovett1 has joined #openstack-nova | 22:58 | |
| *** thorst has quit IRC | 22:58 | |
| *** mdrabe has quit IRC | 23:00 | |
| *** IanGovett has quit IRC | 23:00 | |
| *** pixelbeat has joined #openstack-nova | 23:01 | |
| *** burt has quit IRC | 23:01 | |
| *** Sree has joined #openstack-nova | 23:01 | |
| *** IanGovett1 has quit IRC | 23:02 | |
| *** pratikmallya has joined #openstack-nova | 23:04 | |
| *** Sree has quit IRC | 23:06 | |
| *** achanda has joined #openstack-nova | 23:08 | |
| *** achanda has quit IRC | 23:08 | |
| *** edtubill has quit IRC | 23:09 | |
| *** thorst has joined #openstack-nova | 23:13 | |
| *** pixelbeat has quit IRC | 23:16 | |
| *** thorst has quit IRC | 23:17 | |
| openstackgerrit | Dan Smith proposed openstack/nova: cells: check for existing BDM based on uuid in bdm_update_or_create_at_top https://review.openstack.org/242649 | 23:19 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make BlockDeviceMapping object support uuid https://review.openstack.org/242603 | 23:19 |
| *** lbelivea has quit IRC | 23:24 | |
| *** su_zhang has joined #openstack-nova | 23:25 | |
| *** xyang1 has quit IRC | 23:26 | |
| *** pumaranikar has quit IRC | 23:27 | |
| *** achanda has joined #openstack-nova | 23:29 | |
| *** mgoddard_ has quit IRC | 23:30 | |
| *** salv-orlando has quit IRC | 23:33 | |
| *** tonytan4ever has quit IRC | 23:33 | |
| *** pratikmallya has quit IRC | 23:34 | |
| *** stackdump has quit IRC | 23:35 | |
| *** mc_nair has quit IRC | 23:35 | |
| *** thorst has joined #openstack-nova | 23:35 | |
| *** IanGovett has joined #openstack-nova | 23:37 | |
| *** achanda has quit IRC | 23:37 | |
| *** stackdump has joined #openstack-nova | 23:38 | |
| *** IanGovett has quit IRC | 23:38 | |
| *** IanGovett has joined #openstack-nova | 23:38 | |
| *** thorst has quit IRC | 23:40 | |
| *** IanGovett has quit IRC | 23:43 | |
| *** otter768 has joined #openstack-nova | 23:49 | |
| *** READ10 has joined #openstack-nova | 23:52 | |
| *** dimsum__ has quit IRC | 23:52 | |
| *** jerrygb_ has quit IRC | 23:56 | |
| *** rfolco has joined #openstack-nova | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!