| *** Swami has quit IRC | 00:01 | |
| *** mlavalle has quit IRC | 00:07 | |
| *** thorst has quit IRC | 00:08 | |
| *** thorst has joined #openstack-nova | 00:09 | |
| *** rrecio_ has quit IRC | 00:15 | |
| *** vladikr has quit IRC | 00:17 | |
| *** Apoorva has quit IRC | 00:17 | |
| *** thorst has quit IRC | 00:17 | |
| *** mmedvede has quit IRC | 00:19 | |
| *** yassou has quit IRC | 00:22 | |
| *** mingyu_ has joined #openstack-nova | 00:24 | |
| *** mriedem has quit IRC | 00:24 | |
| *** mriedem has joined #openstack-nova | 00:24 | |
| *** ducttape_ has joined #openstack-nova | 00:25 | |
| *** sdague has quit IRC | 00:26 | |
| *** mingyu has quit IRC | 00:26 | |
| *** hoangcx has joined #openstack-nova | 00:28 | |
| *** jamesdenton has joined #openstack-nova | 00:29 | |
| *** Apoorva has joined #openstack-nova | 00:30 | |
| *** Apoorva has quit IRC | 00:35 | |
| *** mingyu_ has quit IRC | 00:35 | |
| *** mingyu has joined #openstack-nova | 00:36 | |
| *** diogogmt has joined #openstack-nova | 00:37 | |
| *** mingyu has quit IRC | 00:37 | |
| *** kaisers has quit IRC | 00:42 | |
| *** kaisers has joined #openstack-nova | 00:43 | |
| *** jamesdenton has quit IRC | 00:43 | |
| *** Jeffrey4l has joined #openstack-nova | 00:47 | |
| *** rrecio has joined #openstack-nova | 00:49 | |
| *** felix23ma has joined #openstack-nova | 00:50 | |
| *** gjayavelu has quit IRC | 00:52 | |
| *** mmedvede has joined #openstack-nova | 00:54 | |
| *** Sukhdev has joined #openstack-nova | 00:56 | |
| *** rrecio has quit IRC | 00:56 | |
| *** hieulq has quit IRC | 01:00 | |
| *** hieulq has joined #openstack-nova | 01:00 | |
| *** annp has joined #openstack-nova | 01:01 | |
| *** Sukhdev has quit IRC | 01:01 | |
| *** annp has quit IRC | 01:01 | |
| *** pwnall1 has quit IRC | 01:03 | |
| *** ssurana has quit IRC | 01:03 | |
| *** rama_y has joined #openstack-nova | 01:08 | |
| *** gongysh has joined #openstack-nova | 01:12 | |
| *** phuongnh has joined #openstack-nova | 01:12 | |
| openstackgerrit | Eli Qiao proposed openstack/nova: Libvirt: Correct PERF_EVENTS_CPU_FLAG_MAPPING https://review.openstack.org/365842 | 01:13 |
|---|---|---|
| *** thorst has joined #openstack-nova | 01:14 | |
| *** thorst has quit IRC | 01:16 | |
| *** thorst has joined #openstack-nova | 01:16 | |
| *** wangqun has joined #openstack-nova | 01:19 | |
| *** kevinz has joined #openstack-nova | 01:22 | |
| *** annp has joined #openstack-nova | 01:22 | |
| *** felix23ma has quit IRC | 01:22 | |
| *** browne has quit IRC | 01:22 | |
| *** nic has quit IRC | 01:27 | |
| *** jamesdenton has joined #openstack-nova | 01:29 | |
| *** amotoki has joined #openstack-nova | 01:29 | |
| *** lixiaoy1 has joined #openstack-nova | 01:30 | |
| *** thorst has quit IRC | 01:32 | |
| *** amotoki has quit IRC | 01:33 | |
| *** raunak has joined #openstack-nova | 01:35 | |
| *** raunak has quit IRC | 01:40 | |
| *** felix23ma has joined #openstack-nova | 01:40 | |
| *** ljxiash has joined #openstack-nova | 01:41 | |
| *** baoli has joined #openstack-nova | 01:43 | |
| *** liuyulong|away is now known as liuyulong | 01:43 | |
| *** yamahata has quit IRC | 01:44 | |
| *** ljxiash has quit IRC | 01:45 | |
| *** amotoki has joined #openstack-nova | 01:46 | |
| openstackgerrit | Jianghua Wang proposed openstack/nova: XenAPI: resolve the fetch_bandwidth failure https://review.openstack.org/365838 | 01:50 |
| *** liangy has quit IRC | 01:54 | |
| *** EinstCrazy has joined #openstack-nova | 01:55 | |
| *** ducttape_ has quit IRC | 01:57 | |
| *** ijw has joined #openstack-nova | 01:58 | |
| *** EinstCrazy has quit IRC | 01:58 | |
| *** ijw has quit IRC | 01:58 | |
| *** ljxiash has joined #openstack-nova | 01:59 | |
| *** ijw has joined #openstack-nova | 01:59 | |
| *** ljxiash has quit IRC | 01:59 | |
| *** ducttape_ has joined #openstack-nova | 01:59 | |
| *** ljxiash has joined #openstack-nova | 01:59 | |
| *** EinstCrazy has joined #openstack-nova | 02:03 | |
| *** amotoki has quit IRC | 02:05 | |
| *** ijw has quit IRC | 02:09 | |
| *** unicell has quit IRC | 02:14 | |
| *** mriedem has quit IRC | 02:17 | |
| *** amotoki has joined #openstack-nova | 02:20 | |
| *** baoli has quit IRC | 02:21 | |
| *** priteau has joined #openstack-nova | 02:22 | |
| *** akshai has joined #openstack-nova | 02:23 | |
| *** thorst has joined #openstack-nova | 02:24 | |
| *** baoli has joined #openstack-nova | 02:26 | |
| *** felix23ma has quit IRC | 02:26 | |
| *** priteau has quit IRC | 02:27 | |
| *** felix23ma has joined #openstack-nova | 02:28 | |
| *** amotoki has quit IRC | 02:28 | |
| *** hoangcx has quit IRC | 02:29 | |
| *** thorst has quit IRC | 02:29 | |
| *** salv-orlando has joined #openstack-nova | 02:30 | |
| *** akshai has quit IRC | 02:30 | |
| *** felix23ma has quit IRC | 02:31 | |
| *** salv-orl_ has quit IRC | 02:33 | |
| *** felix23ma has joined #openstack-nova | 02:34 | |
| *** baoli has quit IRC | 02:35 | |
| *** amotoki has joined #openstack-nova | 02:41 | |
| *** akshai has joined #openstack-nova | 02:42 | |
| *** thorst has joined #openstack-nova | 02:48 | |
| *** thorst has quit IRC | 02:48 | |
| *** tjones has left #openstack-nova | 02:48 | |
| *** amotoki has quit IRC | 02:48 | |
| *** thorst has joined #openstack-nova | 02:48 | |
| *** dimtruck is now known as zz_dimtruck | 02:50 | |
| *** shangxdy has joined #openstack-nova | 02:53 | |
| *** hoangcx has joined #openstack-nova | 02:53 | |
| *** mingyu has joined #openstack-nova | 02:53 | |
| *** vishwanathj has quit IRC | 02:54 | |
| *** jamesdenton has quit IRC | 02:58 | |
| *** ljxiash has quit IRC | 03:02 | |
| *** mingyu has quit IRC | 03:02 | |
| *** jogo has joined #openstack-nova | 03:03 | |
| *** tongli has quit IRC | 03:03 | |
| *** ljxiash has joined #openstack-nova | 03:04 | |
| *** mingyu has joined #openstack-nova | 03:05 | |
| *** ducttape_ has quit IRC | 03:05 | |
| *** felix23ma has quit IRC | 03:06 | |
| *** felix23ma has joined #openstack-nova | 03:08 | |
| openstackgerrit | javeme proposed openstack/nova: server_group_members quota exceeded silently https://review.openstack.org/337251 | 03:10 |
| *** Guest42361 has joined #openstack-nova | 03:17 | |
| Guest42361 | hi | 03:17 |
| *** Guest42361 has quit IRC | 03:19 | |
| *** Guest42361 has joined #openstack-nova | 03:19 | |
| *** tbachman has quit IRC | 03:22 | |
| *** felix23ma has quit IRC | 03:25 | |
| *** ducttape_ has joined #openstack-nova | 03:26 | |
| *** ljxiash_ has joined #openstack-nova | 03:26 | |
| *** ljxiash has quit IRC | 03:27 | |
| *** wuhg has joined #openstack-nova | 03:31 | |
| *** Apoorva has joined #openstack-nova | 03:32 | |
| *** sdake_ has joined #openstack-nova | 03:32 | |
| *** sdake has quit IRC | 03:35 | |
| *** gouthamr has quit IRC | 03:37 | |
| *** Apoorva has quit IRC | 03:37 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs: Add support for tag instances when boot https://review.openstack.org/366469 | 03:38 |
| *** ducttape_ has quit IRC | 03:39 | |
| *** fragatin_ has joined #openstack-nova | 03:44 | |
| *** gouthamr has joined #openstack-nova | 03:45 | |
| *** links has joined #openstack-nova | 03:45 | |
| *** fragatin_ has quit IRC | 03:46 | |
| *** fragatina has quit IRC | 03:46 | |
| *** ljxiash_ has quit IRC | 03:48 | |
| *** ljxiash has joined #openstack-nova | 03:49 | |
| *** coolsvap_ has joined #openstack-nova | 03:50 | |
| *** fragatina has joined #openstack-nova | 03:53 | |
| *** ljxiash has quit IRC | 03:53 | |
| *** markvoelker has quit IRC | 03:54 | |
| *** felix23ma has joined #openstack-nova | 03:56 | |
| *** fragatina has quit IRC | 03:57 | |
| *** shangxdy has quit IRC | 04:00 | |
| *** rdo_ has quit IRC | 04:07 | |
| *** mingyu has quit IRC | 04:09 | |
| *** ljxiash has joined #openstack-nova | 04:11 | |
| *** unicell has joined #openstack-nova | 04:11 | |
| *** rama_y has quit IRC | 04:12 | |
| *** ManishD has joined #openstack-nova | 04:18 | |
| *** unicell1 has joined #openstack-nova | 04:20 | |
| *** unicell has quit IRC | 04:20 | |
| *** amotoki has joined #openstack-nova | 04:20 | |
| *** gouthamr has quit IRC | 04:21 | |
| *** ljxiash has quit IRC | 04:23 | |
| *** ljxiash has joined #openstack-nova | 04:23 | |
| *** ManishD has quit IRC | 04:24 | |
| *** akshai has quit IRC | 04:27 | |
| *** ljxiash has quit IRC | 04:27 | |
| *** stevemar has quit IRC | 04:28 | |
| *** lixiaoy1 has quit IRC | 04:30 | |
| *** 17WAAKU0S has joined #openstack-nova | 04:37 | |
| *** 17WAAKU0S has quit IRC | 04:38 | |
| *** mingyu has joined #openstack-nova | 04:40 | |
| *** thorst has quit IRC | 04:43 | |
| *** thorst has joined #openstack-nova | 04:45 | |
| *** akshai has joined #openstack-nova | 04:48 | |
| *** thorst has quit IRC | 04:50 | |
| *** bkopilov_ has joined #openstack-nova | 04:50 | |
| *** prateek has joined #openstack-nova | 04:55 | |
| *** markvoelker has joined #openstack-nova | 04:55 | |
| *** thorst has joined #openstack-nova | 04:56 | |
| *** markvoelker has quit IRC | 04:59 | |
| *** thorst has quit IRC | 05:00 | |
| *** Jeffrey4l_ has joined #openstack-nova | 05:00 | |
| *** mingyu has quit IRC | 05:02 | |
| *** claudiub has joined #openstack-nova | 05:03 | |
| *** Jeffrey4l has quit IRC | 05:04 | |
| *** claudiub|2 has joined #openstack-nova | 05:05 | |
| *** mamandle has joined #openstack-nova | 05:05 | |
| *** claudiub has quit IRC | 05:09 | |
| *** moshele has joined #openstack-nova | 05:09 | |
| *** tonytan_brb has quit IRC | 05:09 | |
| *** ljxiash has joined #openstack-nova | 05:10 | |
| *** yamahata has joined #openstack-nova | 05:11 | |
| *** jaosorior has joined #openstack-nova | 05:12 | |
| openstackgerrit | huan proposed openstack/nova: xenapi: support the hotplug of a neutron port https://review.openstack.org/365505 | 05:14 |
| openstackgerrit | Prateek Arora proposed openstack/nova: Handle ConnectFailure from keystone when neutron is down https://review.openstack.org/327706 | 05:17 |
| *** harlowja_at_home has joined #openstack-nova | 05:18 | |
| *** openstackgerrit has quit IRC | 05:18 | |
| *** mingyu has joined #openstack-nova | 05:18 | |
| *** openstackgerrit has joined #openstack-nova | 05:19 | |
| *** hoangcx has quit IRC | 05:20 | |
| *** akshai has quit IRC | 05:24 | |
| *** hoangcx has joined #openstack-nova | 05:25 | |
| *** thorst has joined #openstack-nova | 05:28 | |
| *** akshai has joined #openstack-nova | 05:29 | |
| *** Sukhdev has joined #openstack-nova | 05:29 | |
| *** thorst has quit IRC | 05:29 | |
| *** EinstCrazy has quit IRC | 05:30 | |
| *** EinstCrazy has joined #openstack-nova | 05:31 | |
| *** thorst has joined #openstack-nova | 05:32 | |
| *** ratailor has joined #openstack-nova | 05:34 | |
| *** mingyu has quit IRC | 05:36 | |
| *** unicell has joined #openstack-nova | 05:38 | |
| *** unicell1 has quit IRC | 05:38 | |
| *** akshai has quit IRC | 05:39 | |
| *** ducttape_ has joined #openstack-nova | 05:39 | |
| *** wuhg has quit IRC | 05:41 | |
| *** ducttape_ has quit IRC | 05:43 | |
| *** bhagyashri has joined #openstack-nova | 05:46 | |
| *** sdake_ is now known as sdake | 05:46 | |
| *** ekuris has joined #openstack-nova | 05:47 | |
| *** lixiaoy1 has joined #openstack-nova | 05:47 | |
| *** raunak has joined #openstack-nova | 05:49 | |
| *** Sukhdev has quit IRC | 05:51 | |
| *** Apoorva has joined #openstack-nova | 05:52 | |
| *** harlowja_at_home has quit IRC | 05:54 | |
| *** mingyu has joined #openstack-nova | 05:56 | |
| *** Apoorva has quit IRC | 05:56 | |
| *** mingyu has quit IRC | 05:58 | |
| *** mjura has joined #openstack-nova | 06:01 | |
| *** nkrinner_afk is now known as nkrinner | 06:03 | |
| *** ManishD has joined #openstack-nova | 06:03 | |
| *** paul-carlton2 has joined #openstack-nova | 06:04 | |
| *** stevemar has joined #openstack-nova | 06:06 | |
| *** EinstCrazy has quit IRC | 06:06 | |
| *** EinstCrazy has joined #openstack-nova | 06:10 | |
| *** tonytan4ever has joined #openstack-nova | 06:10 | |
| *** stevemar has quit IRC | 06:11 | |
| *** ttrumm has joined #openstack-nova | 06:12 | |
| *** tonytan4ever has quit IRC | 06:15 | |
| *** sahid has joined #openstack-nova | 06:15 | |
| *** kevinz has quit IRC | 06:16 | |
| *** sdake has quit IRC | 06:18 | |
| *** oanson has joined #openstack-nova | 06:18 | |
| *** shangxdy has joined #openstack-nova | 06:22 | |
| *** kevinz has joined #openstack-nova | 06:25 | |
| *** thorst has quit IRC | 06:26 | |
| slaweq__ | amotoki: hello | 06:27 |
| slaweq__ | thx for Your fixes in https://review.openstack.org/#/c/358344/6/api-ref/source/v2/quotas.inc@44 | 06:27 |
| slaweq__ | should I change this order of parameters there also? | 06:27 |
| slaweq__ | or it will be done in this "another patch" about what You wrote there? | 06:27 |
| *** edand has joined #openstack-nova | 06:28 | |
| *** thorst has joined #openstack-nova | 06:28 | |
| *** thorst has quit IRC | 06:28 | |
| *** ttrumm_ has joined #openstack-nova | 06:30 | |
| *** praneshp has quit IRC | 06:30 | |
| amotoki | slaweq__: wrong channel? this is #-nova | 06:31 |
| prateek | sahid, can you please have a look at https://review.openstack.org/#/c/327706/, i have made the changes as per yesterday's comments | 06:31 |
| *** ttrumm has quit IRC | 06:32 | |
| slaweq__ | amotoki: sorry | 06:32 |
| *** zenoway has joined #openstack-nova | 06:33 | |
| slaweq__ | my fault :) | 06:33 |
| amotoki | slaweq__: np. let's move to -neutron | 06:33 |
| slaweq__ | amotoki: ok | 06:33 |
| *** hussain_c has quit IRC | 06:33 | |
| *** raunak has quit IRC | 06:33 | |
| *** lixiaoy1 has quit IRC | 06:34 | |
| *** pcaruana has joined #openstack-nova | 06:34 | |
| *** edand has quit IRC | 06:35 | |
| *** lixiaoy1 has joined #openstack-nova | 06:39 | |
| *** ducttape_ has joined #openstack-nova | 06:40 | |
| *** sticker has quit IRC | 06:42 | |
| *** stevemar has joined #openstack-nova | 06:44 | |
| *** ducttape_ has quit IRC | 06:45 | |
| *** lpetrut has joined #openstack-nova | 06:48 | |
| *** andreas_s has joined #openstack-nova | 06:54 | |
| *** EinstCrazy has quit IRC | 06:54 | |
| *** mamandle_ has joined #openstack-nova | 06:55 | |
| *** markvoelker has joined #openstack-nova | 06:56 | |
| *** mamandle has quit IRC | 06:56 | |
| *** mamandle_ is now known as mamandle | 06:57 | |
| *** devananda|travel has quit IRC | 07:00 | |
| *** thorst has joined #openstack-nova | 07:00 | |
| *** markvoelker has quit IRC | 07:01 | |
| openstackgerrit | Eli Qiao proposed openstack/nova: Libvirt: Correct PERF_EVENTS_CPU_FLAG_MAPPING https://review.openstack.org/365842 | 07:01 |
| *** tesseract- has joined #openstack-nova | 07:02 | |
| *** EinstCrazy has joined #openstack-nova | 07:03 | |
| *** sdake has joined #openstack-nova | 07:03 | |
| *** thorst has quit IRC | 07:06 | |
| *** garyk1 has joined #openstack-nova | 07:07 | |
| openstackgerrit | Hussain Chachuliya proposed openstack/nova: Log warning for negative config option value https://review.openstack.org/366535 | 07:08 |
| *** markus_z has joined #openstack-nova | 07:08 | |
| garyk1 | with latest code when i spin up a instance with kvm i get message on the console 'atkbd serio0: Use 'setkeycodes'…' Any idea what may be causing this? | 07:12 |
| garyk1 | mdbooth: ^ | 07:14 |
| *** matrohon has joined #openstack-nova | 07:16 | |
| *** avolkov has quit IRC | 07:18 | |
| *** mamandle_ has joined #openstack-nova | 07:21 | |
| *** hshiina has joined #openstack-nova | 07:22 | |
| *** trisq has joined #openstack-nova | 07:23 | |
| *** mamandle has quit IRC | 07:24 | |
| *** mamandle_ is now known as mamandle | 07:24 | |
| *** EinstCrazy has quit IRC | 07:27 | |
| *** slaweq__ has quit IRC | 07:31 | |
| *** armax has quit IRC | 07:31 | |
| *** paul-carlton2 has quit IRC | 07:34 | |
| *** abalutoiu has joined #openstack-nova | 07:35 | |
| *** thorst has joined #openstack-nova | 07:36 | |
| *** thorst has quit IRC | 07:36 | |
| *** EinstCrazy has joined #openstack-nova | 07:36 | |
| *** lpetrut has quit IRC | 07:38 | |
| *** woodster_ has quit IRC | 07:39 | |
| *** avolkov has joined #openstack-nova | 07:40 | |
| *** ducttape_ has joined #openstack-nova | 07:41 | |
| *** ducttape_ has quit IRC | 07:45 | |
| *** tonytan4ever has joined #openstack-nova | 07:47 | |
| *** mamandle has quit IRC | 07:49 | |
| *** tonytan4ever has quit IRC | 07:52 | |
| *** lpetrut has joined #openstack-nova | 07:52 | |
| *** ihrachys has joined #openstack-nova | 07:52 | |
| *** roeyc has joined #openstack-nova | 07:53 | |
| *** garyk1 has quit IRC | 07:58 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-nova | 08:00 | |
| *** moshele has quit IRC | 08:01 | |
| *** moshele has joined #openstack-nova | 08:01 | |
| *** thorst has joined #openstack-nova | 08:01 | |
| *** EinstCrazy has quit IRC | 08:03 | |
| *** lixiaoy1 has quit IRC | 08:04 | |
| *** lixiaoy1 has joined #openstack-nova | 08:04 | |
| *** mvk has quit IRC | 08:07 | |
| *** stevemar_ has joined #openstack-nova | 08:07 | |
| *** mvk has joined #openstack-nova | 08:08 | |
| *** yassine has joined #openstack-nova | 08:08 | |
| *** yassine is now known as Guest10379 | 08:09 | |
| *** mamandle has joined #openstack-nova | 08:09 | |
| *** thorst_ has joined #openstack-nova | 08:10 | |
| *** thorst__ has joined #openstack-nova | 08:10 | |
| *** stevemar_ has quit IRC | 08:12 | |
| *** EinstCrazy has joined #openstack-nova | 08:12 | |
| *** thorst has quit IRC | 08:12 | |
| *** thorst_ has quit IRC | 08:14 | |
| *** lpetrut has quit IRC | 08:16 | |
| *** cdent has joined #openstack-nova | 08:16 | |
| *** hieulq has quit IRC | 08:19 | |
| *** annp has quit IRC | 08:19 | |
| *** annp has joined #openstack-nova | 08:19 | |
| *** hieulq has joined #openstack-nova | 08:19 | |
| *** lucas-dinner is now known as lucasagomes | 08:22 | |
| *** ttrumm_ has quit IRC | 08:22 | |
| *** rdo_ has joined #openstack-nova | 08:25 | |
| *** priteau has joined #openstack-nova | 08:27 | |
| *** liverpooler has joined #openstack-nova | 08:29 | |
| *** salv-orl_ has joined #openstack-nova | 08:29 | |
| *** thorst__ has quit IRC | 08:32 | |
| *** danpb has joined #openstack-nova | 08:32 | |
| *** salv-orlando has quit IRC | 08:32 | |
| *** claudiub has joined #openstack-nova | 08:32 | |
| *** Guest42361 has quit IRC | 08:33 | |
| *** openstackgerrit has quit IRC | 08:34 | |
| *** EinstCra_ has joined #openstack-nova | 08:34 | |
| *** openstackgerrit has joined #openstack-nova | 08:34 | |
| *** thorst has joined #openstack-nova | 08:34 | |
| *** thorst has quit IRC | 08:35 | |
| *** claudiub|2 has quit IRC | 08:35 | |
| *** ihrachys has quit IRC | 08:36 | |
| *** EinstCrazy has quit IRC | 08:37 | |
| *** derekh has joined #openstack-nova | 08:40 | |
| *** ttrumm has joined #openstack-nova | 08:44 | |
| *** lpetrut has joined #openstack-nova | 08:44 | |
| *** ftersin has quit IRC | 08:44 | |
| *** sudipto has joined #openstack-nova | 08:45 | |
| *** garyk1 has joined #openstack-nova | 08:45 | |
| *** sudipto_ has joined #openstack-nova | 08:45 | |
| *** thorst has joined #openstack-nova | 08:46 | |
| *** hshiina has quit IRC | 08:47 | |
| *** thorst has quit IRC | 08:49 | |
| *** paul-carlton2 has joined #openstack-nova | 08:51 | |
| *** diga has joined #openstack-nova | 08:52 | |
| *** karthiks has quit IRC | 08:53 | |
| *** ratailor has quit IRC | 08:53 | |
| *** links has quit IRC | 08:53 | |
| *** ratailor has joined #openstack-nova | 08:54 | |
| *** prateek has quit IRC | 08:54 | |
| *** priteau has quit IRC | 08:55 | |
| *** EinstCra_ has quit IRC | 08:55 | |
| *** links has joined #openstack-nova | 08:55 | |
| *** karthiks has joined #openstack-nova | 08:55 | |
| *** prateek has joined #openstack-nova | 08:55 | |
| *** EinstCrazy has joined #openstack-nova | 08:56 | |
| *** markvoelker has joined #openstack-nova | 08:57 | |
| *** lixiaoy1 has quit IRC | 08:58 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 08:58 |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 08:58 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 08:58 |
| prateek | cdent, can you please take a look at https://review.openstack.org/#/c/327706/ when you have some free time ? | 09:01 |
| *** markvoelker has quit IRC | 09:01 | |
| cdent | yup, will open up a tab with it now, just finishing up some placement api stuff first | 09:02 |
| *** code-R has joined #openstack-nova | 09:02 | |
| *** mamandle has quit IRC | 09:02 | |
| prateek | cdent, thanks | 09:02 |
| *** mamandle has joined #openstack-nova | 09:03 | |
| *** code-R_ has joined #openstack-nova | 09:05 | |
| mdbooth | cdent: Hey, I was thinking about that bug from last night. I think that query is complex enough to have its own unit tests, which is probably the best way to get useful, deterministic tests. | 09:05 |
| cdent | mdbooth: I think we're all still wondering about that | 09:06 |
| mdbooth | Incidentally, I haven't work on db stuff in nova for a bit. It's great to see we're finally freeing it from the db.sqlalchemy.api cage. Much easier to read in context. | 09:06 |
| cdent | nobody was able to replicate failure locally | 09:06 |
| cdent | yeah, a decision was made that new objects in the API db would not be tied to that old way, and the placement stuff is following that pattern | 09:07 |
| *** code-R has quit IRC | 09:08 | |
| cdent | I agree that a unit test would be useful, but I think people are working at two not necessarily collaborating goals: get something working and released; understanding what's going on and having proper tests | 09:08 |
| *** stevemar_ has joined #openstack-nova | 09:08 | |
| cdent | if you've followed some of the changes that happened since (our) yesterday evening, there were some tests added but I'm not sure they actually test anything other than confirming that local runs of tests don't show the problem (on the old code) | 09:09 |
| cdent | mdbooth: so, to sum: If you've got some ideas they are _very_ welcome | 09:09 |
| mdbooth | Like I was saying last night, I think a problem that you're going to hit in writing a test of the whole method is that the failure is going to be non-deterministic | 09:10 |
| mdbooth | Because it depends on the returned row order | 09:10 |
| mdbooth | But if you test the query separately, you can isolate the failure | 09:11 |
| mdbooth | Deterministically isolate it, that is | 09:11 |
| mdbooth | Because you can specifically check that the result is as expected for the input, including testing for duplicate rows | 09:11 |
| *** mvk has quit IRC | 09:12 | |
| *** stevemar_ has quit IRC | 09:13 | |
| cdent | yeah, agree, but what I'm saying is that locally we have been unable to generate the multiple rows: we've not figured out the correct inputs to do so. which means we don't actually understand the problem | 09:13 |
| *** ihrachys has joined #openstack-nova | 09:19 | |
| *** yamahata has quit IRC | 09:21 | |
| *** moshele has quit IRC | 09:21 | |
| *** moshele has joined #openstack-nova | 09:21 | |
| cdent | to some extent the fact that we've identified "oh, it didn't work because of an oversight in the creation of the sql" and we've fixed that is perhaps enough. It's just weird that we had had tests that got the right results with the wrong sql. | 09:21 |
| cdent | but because we've been rushing around there hasn't been clear heads to really investigate it properly | 09:21 |
| *** moshele has quit IRC | 09:23 | |
| *** moshele has joined #openstack-nova | 09:24 | |
| sfinucan | moshele: Say, are VFs managed by nova or neutron? I forget where the line is | 09:24 |
| openstackgerrit | Matthew Booth proposed openstack/nova: imagecache: Hold shared storage registry lock while running https://review.openstack.org/366239 | 09:26 |
| *** ihrachys has quit IRC | 09:26 | |
| moshele | sfinucan: it depends nova (libvirt) put the mac and vlan https://github.com/openstack/nova/blob/master/nova/virt/libvirt/vif.py#L304-L321 | 09:26 |
| sfinucan | moshele: So they both mess with it? | 09:27 |
| moshele | sfinucan: neutron side control the admin_state port_secuirty and qos | 09:27 |
| moshele | sfinucan: yes | 09:27 |
| *** ManishD has quit IRC | 09:30 | |
| openstackgerrit | Dao Cong Tien proposed openstack/nova: Stop using mox in unit/virt/xenapi/test_vmops.py (1) https://review.openstack.org/344000 | 09:33 |
| sfinucan | I ask because I'd like to know if this is possible https://review.openstack.org/#/c/342001/4/specs/newton/approved/nova-api-sriov-pcipt-stats.rst | 09:33 |
| sfinucan | namely, is it possible for nova to have an inventory of available vs. used PCI devices as things stand? | 09:33 |
| sfinucan | moshele: ^ | 09:33 |
| *** wznoinsk has quit IRC | 09:35 | |
| prateek | cdent, regarding your latest comment on the patch https://review.openstack.org/#/c/327706/ | 09:35 |
| cdent | prateek: yes? | 09:35 |
| prateek | i didn't plan to write the test initially but asked in the community and someone told me that it would need an api test to be written like this | 09:36 |
| *** tuanla_ has joined #openstack-nova | 09:36 | |
| prateek | cdent, i can look around a bit to find out what exactly the test does | 09:36 |
| *** EinstCrazy has quit IRC | 09:37 | |
| cdent | prateek: Do you remember my original comment about it not actually testing anything? | 09:37 |
| prateek | cdent, yes :) | 09:37 |
| prateek | and i still think that the test and 4 test above that written exactly same way can be done away with | 09:38 |
| cdent | Maybe another review can come along and clarify the situation there. Maybe alex_xu or sdague will have an opinion. from my standpoint the test is useless. | 09:38 |
| cdent | yeah | 09:38 |
| cdent | weird :) | 09:38 |
| prateek | def test_resource_not_authorized(self): | 09:38 |
| prateek | class Controller(object): | 09:38 |
| prateek | def index(self, req): | 09:38 |
| prateek | raise exception.Unauthorized() | 09:38 |
| prateek | req = webob.Request.blank('/tests') | 09:38 |
| prateek | app = fakes.TestRouter(Controller()) | 09:38 |
| prateek | self.assertRaises( | 09:38 |
| prateek | exception.Unauthorized, req.get_response, app) | 09:38 |
| prateek | def test_connection_failed(self): | 09:39 |
| prateek | class Controller(object): | 09:39 |
| prateek | def index(self, req): | 09:39 |
| prateek | raise exception.NeutronConnectionFailed() | 09:39 |
| prateek | req = webob.Request.blank('/tests') | 09:39 |
| prateek | app = fakes.TestRouter(Controller()) | 09:39 |
| prateek | self.assertRaises( | 09:39 |
| prateek | exception.NeutronConnectionFailed, req.get_response, app) | 09:39 |
| prateek | one i have written and one is written by sahid | 09:39 |
| *** EinstCrazy has joined #openstack-nova | 09:39 | |
| prateek | sahid, can you please help me and cdent understand how this work ? | 09:39 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs: Add support for tag instances when boot https://review.openstack.org/366469 | 09:39 |
| *** roeyc has quit IRC | 09:41 | |
| *** mvk has joined #openstack-nova | 09:41 | |
| *** danpawlik1 has left #openstack-nova | 09:41 | |
| openstackgerrit | Dao Cong Tien proposed openstack/nova: Stop using mox in compute/test_hypervisors.py https://review.openstack.org/326814 | 09:42 |
| sahid | prateek: I think what we wnat here is to ensure that the excpetion raised by controller is not going to be translated in error 500 | 09:43 |
| sahid | let me try to add comment in the review | 09:44 |
| *** abalutoiu_ has joined #openstack-nova | 09:45 | |
| cdent | prateek: for future ref: when you want to send that many lines you should use http://paste.openstack.org/ | 09:45 |
| prateek | sahid, thanks for adding the comment, can you help me out a bit in understanding that, in the test_wsgi method, we pass app to req.get_response which has a controller which raises NeutronConnectionFailed , so how did we hit the API ? | 09:47 |
| openstackgerrit | Dao Cong Tien proposed openstack/nova: Stop using mox in virt/xenapi/image/test_glance.py https://review.openstack.org/327457 | 09:47 |
| *** sdake has quit IRC | 09:47 | |
| prateek | cdent, ok, would take care from next time | 09:47 |
| *** abalutoiu has quit IRC | 09:48 | |
| *** wuhg has joined #openstack-nova | 09:49 | |
| prateek | sahid, so i guess what i understand from your comment is we intentionally raise that exception and pass that to the API method we have given whether we are able to get back that exception from API | 09:50 |
| prateek | sahid, is that correct ? | 09:50 |
| *** gszasz has joined #openstack-nova | 09:51 | |
| sahid | yeah i'm not totally sure of your question, pehraps i can try to give to you some pointer into the code | 09:51 |
| alex_xu | prateek: sahid , the exception translation is for v2 API, which we should clear up. For v2.1 API, we have decorator expected_error, it will turn all the non HTTP* exception to 500. so I agree with cdent, that test is useless | 09:51 |
| sahid | but what we want is juste to create a route which are going to be execute and so pass throught the api which normally translate any unreconzable exception to a 500 | 09:52 |
| openstackgerrit | Dao Cong Tien proposed openstack/nova: Stop using mox in virt/libvirt/test_driver.py (1) https://review.openstack.org/328070 | 09:52 |
| sahid | because you have added a new catch for NeutronConnectionFailed, it shoud to pass there and return the 504 | 09:52 |
| sahid | s/shoud to/should | 09:52 |
| prateek | sahid, i get it now , thanks | 09:53 |
| openstackgerrit | Dao Cong Tien proposed openstack/nova: Stop using mox in api/openstack/fakes.py https://review.openstack.org/325759 | 09:54 |
| *** gszasz has quit IRC | 09:56 | |
| alex_xu | We need clear up the code for v2 stuff. And we should have test ensure it will return 500. That will be another patch. And after that, I think we won't test all the exception, ensure they will be return as 500 | 09:56 |
| *** gszasz has joined #openstack-nova | 09:57 | |
| *** markvoelker has joined #openstack-nova | 09:57 | |
| * alex_xu leave an hour | 09:57 | |
| *** hoonetorg has quit IRC | 09:58 | |
| prateek | alex_xu, so shall we keep the test in the current commit and do the cleanup for review later or cleaning up should start now itself ? | 09:58 |
| *** gszasz has quit IRC | 09:59 | |
| *** thorst has joined #openstack-nova | 10:00 | |
| *** markvoelker has quit IRC | 10:02 | |
| *** thorst has quit IRC | 10:04 | |
| *** ihrachys has joined #openstack-nova | 10:05 | |
| *** ttrumm has quit IRC | 10:06 | |
| *** mamandle has quit IRC | 10:07 | |
| *** annp has quit IRC | 10:10 | |
| *** mamandle has joined #openstack-nova | 10:11 | |
| *** liuyulong is now known as liuyulong|away | 10:11 | |
| *** phuongnh has quit IRC | 10:11 | |
| *** hoangcx has quit IRC | 10:13 | |
| *** gszasz has joined #openstack-nova | 10:13 | |
| *** suxiang has joined #openstack-nova | 10:15 | |
| *** tbachman has joined #openstack-nova | 10:17 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] don't write existing equivalent allocations https://review.openstack.org/365069 | 10:17 |
| *** ftersin has joined #openstack-nova | 10:20 | |
| *** ttrumm has joined #openstack-nova | 10:22 | |
| *** zenoway has quit IRC | 10:23 | |
| openstackgerrit | Matthew Booth proposed openstack/nova: libvirt: Only check migration_flags for post-copy https://review.openstack.org/366659 | 10:24 |
| bauzas | ouch, 12 newton-rc-potential bugs, that's probably too much | 10:25 |
| *** akijak has joined #openstack-nova | 10:25 | |
| openstackgerrit | Alberto Planas proposed openstack/nova: Add migrate_uri for invoking the migration https://review.openstack.org/356558 | 10:27 |
| *** roeyc has joined #openstack-nova | 10:27 | |
| akijak | hi, I'd like to allow a normal user to do 'nova show' for any tenant. Is is possible for recent releases? | 10:28 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] don't write existing equivalent allocations https://review.openstack.org/365069 | 10:28 |
| openstackgerrit | Chris Dent proposed openstack/nova: More conservative inventory updates https://review.openstack.org/365073 | 10:29 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Correct serialization of inventory collections https://review.openstack.org/365633 | 10:29 |
| *** ttrumm has quit IRC | 10:30 | |
| sfinucan | akijak: see the channel message. That's a question for #openstack | 10:30 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Correct serialization of inventory collections https://review.openstack.org/365633 | 10:31 |
| akijak | sfinucan: I'm sorry. Thanks | 10:31 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Clarify commentary around logging trap https://review.openstack.org/365719 | 10:32 |
| *** akijak has left #openstack-nova | 10:32 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Add additional allocation tests https://review.openstack.org/366144 | 10:32 |
| *** ljxiash has quit IRC | 10:34 | |
| *** ttrumm has joined #openstack-nova | 10:34 | |
| *** tbachman has quit IRC | 10:35 | |
| *** sdague has joined #openstack-nova | 10:37 | |
| *** bsv has joined #openstack-nova | 10:37 | |
| garyk1 | mdbooth: around? | 10:37 |
| mdbooth | garyk1: I am indeed! How's it going? | 10:38 |
| *** akijak has joined #openstack-nova | 10:38 | |
| garyk1 | mdbooth: alls good thanks. maybe you can help here. i am running libvirt latest and greatest driver. when i spin up an instance each keystroke via vnc has the following error: 'atkbd serio0: Use 'setkeycodes'…' Any idea what may be causing this? | 10:38 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Add additional allocation tests https://review.openstack.org/366144 | 10:38 |
| garyk1 | that is, each keystroke this happens. its with all of the cirros images | 10:39 |
| mdbooth | garyk1: Qemu/vnc is a complete mystery to me, I'm afraid. danpb ^^^ ? | 10:39 |
| garyk1 | mdbooth: ok, thanks. danpb any idea? | 10:39 |
| *** ljxiash has joined #openstack-nova | 10:40 | |
| *** akijak has left #openstack-nova | 10:40 | |
| *** sdague has quit IRC | 10:41 | |
| *** trisq has quit IRC | 10:44 | |
| *** ljxiash has quit IRC | 10:44 | |
| *** dgonzalez has quit IRC | 10:45 | |
| danpb | mdbooth: garyk1 that's not going to be anything todo with qemu/vnc really -that simply looks like a guest OS setup issue | 10:46 |
| *** sdague has joined #openstack-nova | 10:46 | |
| mdbooth | https://bugzilla.redhat.com/show_bug.cgi?id=868237 | 10:46 |
| openstack | bugzilla.redhat.com bug 868237 in spice-gtk "spice-gtk sends 00 scancodes to inputs_scancode capable vms" [Unspecified,Closed: errata] - Assigned to marcandre.lureau | 10:46 |
| mdbooth | ^^^ Has similar error messages | 10:47 |
| mdbooth | Looks like it was a client issue in that case. | 10:47 |
| mdbooth | Spice client in that case. | 10:47 |
| mdbooth | Is it possible the vnc client is sending unexpected things? | 10:47 |
| *** thorst has joined #openstack-nova | 10:48 | |
| *** thorst has quit IRC | 10:48 | |
| danpb | depends what the full error message is - garyk1 truncated it in the example above | 10:48 |
| *** shangxdy has quit IRC | 10:48 | |
| mdbooth | garyk1: Are you seeing 'atkbd serio0: Unknown key pressed...'? | 10:49 |
| mdbooth | Would be just before the message above | 10:49 |
| danpb | that spice bug is specifically about sending scancodes with value of 0, not unknown scancodes in general | 10:50 |
| garyk1 | danpb: mdbooth: the message is Use 'setkeycodes 00 <keycode>' to make it known | 10:50 |
| mdbooth | garyk1: What comes before it? | 10:51 |
| garyk1 | this is afetr Unkbown key pressed (translated set 2, code 0x0…) | 10:51 |
| garyk1 | its for every key that is pressed | 10:51 |
| *** PaulMurray has joined #openstack-nova | 10:51 | |
| mdbooth | danpb: So that does look like the client is sending a 00 keycode, right? | 10:51 |
| garyk1 | danpb: mdbooth: code 0x0 on isa0060/serio0 | 10:51 |
| openstackgerrit | Stephen Finucane proposed openstack/nova-specs: Proposal for standardizing hardware NIC capabilities https://review.openstack.org/341341 | 10:51 |
| openstackgerrit | Stephen Finucane proposed openstack/nova-specs: Standardize capabilities using Enums https://review.openstack.org/309762 | 10:51 |
| mdbooth | garyk1: I'm going to go out on a limb and guess it might be an issue in then vnc client. | 10:52 |
| garyk1 | ok. thanks | 10:53 |
| *** ttrumm_ has joined #openstack-nova | 10:54 | |
| *** dikonoor has joined #openstack-nova | 10:55 | |
| *** dgonzalez has joined #openstack-nova | 10:57 | |
| *** ttrumm has quit IRC | 10:57 | |
| *** rtheis has joined #openstack-nova | 10:57 | |
| *** ljxiash has joined #openstack-nova | 11:03 | |
| *** GB21 has joined #openstack-nova | 11:05 | |
| *** dikonoor has quit IRC | 11:05 | |
| *** slagle has joined #openstack-nova | 11:06 | |
| *** amotoki has quit IRC | 11:06 | |
| *** ljxiash has quit IRC | 11:07 | |
| *** wangqun has quit IRC | 11:08 | |
| *** stevemar_ has joined #openstack-nova | 11:09 | |
| *** zerda2 has quit IRC | 11:09 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Correct serialization of inventory collections https://review.openstack.org/365633 | 11:10 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Add additional allocation tests https://review.openstack.org/366144 | 11:10 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Allow inventory to violate allocations https://review.openstack.org/365068 | 11:10 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] don't write existing equivalent allocations https://review.openstack.org/365069 | 11:10 |
| openstackgerrit | Chris Dent proposed openstack/nova: More conservative inventory updates https://review.openstack.org/365073 | 11:10 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Clarify commentary around logging trap https://review.openstack.org/365719 | 11:10 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Example & Parameter verification of os-security-group-default-rules.inc https://review.openstack.org/366681 | 11:10 |
| *** tuanla_ has quit IRC | 11:11 | |
| *** gmann has quit IRC | 11:13 | |
| *** stevemar_ has quit IRC | 11:13 | |
| *** paul-carlton has joined #openstack-nova | 11:13 | |
| *** dikonoor has joined #openstack-nova | 11:15 | |
| *** sahid has quit IRC | 11:16 | |
| *** baoli has joined #openstack-nova | 11:17 | |
| *** lucasagomes is now known as lucas-hungry | 11:20 | |
| mdbooth | danpb: In the live migration meeting yesterday, you mentioned that the libvirt driver is already handling events generated by libvirt. I see the loop in guest.get_job_info() which polls, but from your description I was expecting something asynchronous. | 11:20 |
| mdbooth | Have I missed something? | 11:20 |
| mdbooth | Sorry, the loop is in _live_migration_monitor() | 11:21 |
| mdbooth | It polls guest.get_job_info() | 11:21 |
| *** ttrumm_ has quit IRC | 11:23 | |
| mdbooth | dansmith: Are you awake, yet? | 11:24 |
| *** mfedosin has joined #openstack-nova | 11:26 | |
| *** gongysh has quit IRC | 11:26 | |
| *** rtheis_ has joined #openstack-nova | 11:29 | |
| openstackgerrit | Rikimaru Honjo proposed openstack/nova: Add checks for preventing unexpected detaching volume https://review.openstack.org/354617 | 11:30 |
| *** rtheis has quit IRC | 11:30 | |
| *** ttrumm has joined #openstack-nova | 11:33 | |
| kashyap | mdbooth: It's 04:30 AM in Portland :-) | 11:36 |
| mdbooth | kashyap: Slackers. | 11:36 |
| *** amotoki has joined #openstack-nova | 11:37 | |
| *** cdelatte has joined #openstack-nova | 11:38 | |
| *** ttrumm has quit IRC | 11:38 | |
| *** sudipto has quit IRC | 11:39 | |
| *** sudipto_ has quit IRC | 11:39 | |
| *** ttrumm has joined #openstack-nova | 11:40 | |
| *** cdent has quit IRC | 11:41 | |
| *** tonytan4ever has joined #openstack-nova | 11:49 | |
| *** amotoki has quit IRC | 11:50 | |
| *** salv-orl_ has quit IRC | 11:52 | |
| *** salv-orlando has joined #openstack-nova | 11:52 | |
| *** tonytan4ever has quit IRC | 11:53 | |
| *** baoli_ has joined #openstack-nova | 11:54 | |
| *** PaulMurray has quit IRC | 11:55 | |
| *** eandersson has joined #openstack-nova | 11:57 | |
| *** jaypipes has joined #openstack-nova | 11:57 | |
| *** baoli has quit IRC | 11:57 | |
| *** amotoki has joined #openstack-nova | 11:58 | |
| *** markvoelker has joined #openstack-nova | 11:58 | |
| *** coolsvap_ is now known as coolsvap | 12:01 | |
| *** markvoelker has quit IRC | 12:03 | |
| mdbooth | kashyap: Can you remind me how, when I'm logged in to a host, to jump onto a guest network? | 12:04 |
| *** sean-k-mooney has quit IRC | 12:04 | |
| *** links has quit IRC | 12:04 | |
| mdbooth | Some invocation to change the network namespace of the shell, right? | 12:04 |
| *** sahid has joined #openstack-nova | 12:05 | |
| * kashyap is parsing mdbooth's first sentence | 12:05 | |
| kashyap | What do you mean "jump onto a guest network"? | 12:05 |
| mdbooth | I want to ping an instance which is using neutron | 12:05 |
| *** acabot has quit IRC | 12:05 | |
| kashyap | Ah, Neutron is involved here | 12:05 |
| kashyap | mdbooth: You're asking `ip netns exec` commands I guess | 12:06 |
| mdbooth | Ah! | 12:06 |
| *** devananda has joined #openstack-nova | 12:06 | |
| kashyap | There should be two of them: | 12:06 |
| mdbooth | That should be enough :) | 12:06 |
| *** praveen has joined #openstack-nova | 12:06 | |
| kashyap | $ ip netns | 12:06 |
| kashyap | Should show | 12:06 |
| mdbooth | I know which man pages to read now | 12:06 |
| *** devananda is now known as devananda|travel | 12:06 | |
| * mdbooth was looking at 'net' | 12:06 | |
| *** amotoki has quit IRC | 12:07 | |
| *** acabot has joined #openstack-nova | 12:07 | |
| kashyap | ip netns exec qrouter-$UUID ping [...] | 12:07 |
| mdbooth | kashyap: Thanks! My first attempt is failing, but I'll debug for a bit and maybe get back. | 12:09 |
| kashyap | Sure | 12:09 |
| *** stevemar_ has joined #openstack-nova | 12:10 | |
| *** cdent has joined #openstack-nova | 12:10 | |
| praveen | Hi can someone point me the timings for spec discussion? I wanted to propose a change into nova.. The change is very small not sure if that requires a spec/blueprint or Can directly fix as a defect.. | 12:12 |
| kashyap | mdbooth: I'm sure you must've already checked, but if you're using DevStack, ensure you have the security group rules. I run this silly script as part of my DevStack setup -- https://kashyapc.fedorapeople.org/virt/openstack/Neutron-setup-post-DevStack.bash | 12:14 |
| mdbooth | kashyap: I have, of course, not done that | 12:14 |
| kashyap | :-) | 12:14 |
| * mdbooth gets on it | 12:14 | |
| *** stevemar_ has quit IRC | 12:14 | |
| *** ducttape_ has joined #openstack-nova | 12:16 | |
| *** links has joined #openstack-nova | 12:16 | |
| openstackgerrit | Andrey Volkov proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | 12:17 |
| *** ducttape_ has quit IRC | 12:18 | |
| *** pradk has joined #openstack-nova | 12:18 | |
| *** ducttape_ has joined #openstack-nova | 12:18 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Optional separate database for placement API https://review.openstack.org/362766 | 12:20 |
| *** EinstCrazy has quit IRC | 12:21 | |
| *** tongli has joined #openstack-nova | 12:22 | |
| *** Hazelesque_ is now known as Hazelesque | 12:22 | |
| *** PaulMurray has joined #openstack-nova | 12:22 | |
| *** hoonetorg has joined #openstack-nova | 12:23 | |
| *** lucas-hungry is now known as lucasagomes | 12:23 | |
| markus_z | danpb: If time permits, would you take a look at this regression fix please? https://review.openstack.org/#/c/335132/ | 12:23 |
| *** gongysh has joined #openstack-nova | 12:24 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: Reorder scheduler options https://review.openstack.org/350085 | 12:27 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Group scheduler options https://review.openstack.org/349959 | 12:27 |
| mdbooth | Is there any reason I would not be able to connect to the vnc port of a local qemu process? | 12:27 |
| mdbooth | domain xml has port='5900' autoport='yes' listen='127.0.0.1' | 12:27 |
| mdbooth | telnet to localhost 5900 is refused | 12:27 |
| *** liverpooler has quit IRC | 12:28 | |
| mdbooth | Ah, nm. Was still in network namespace :) | 12:28 |
| *** karthiks has quit IRC | 12:31 | |
| *** liangy has joined #openstack-nova | 12:31 | |
| *** mingyu has joined #openstack-nova | 12:31 | |
| * cdent misses the days of simple networking | 12:31 | |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: block_device: fix propagation of instance AZ to Cinder https://review.openstack.org/366724 | 12:31 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Example & Parameter verification of os-security-group-default-rules.inc https://review.openstack.org/366681 | 12:31 |
| *** edmondsw has joined #openstack-nova | 12:31 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Remove deprecated configuration option network_device_mtu https://review.openstack.org/319543 | 12:31 |
| *** markvoelker has joined #openstack-nova | 12:32 | |
| *** ducttape_ has quit IRC | 12:33 | |
| mdbooth | kashyap: Imagine for a second that I know little of neutron. It'll be a stretch, but try anyway. | 12:33 |
| kashyap | Let's try | 12:33 |
| mdbooth | My guest is failing to get an ip via dhcp. Where should I look first? Environment is devstack. | 12:33 |
| *** amotoki has joined #openstack-nova | 12:33 | |
| *** trisq has joined #openstack-nova | 12:34 | |
| kashyap | mdbooth: What do you in the guest serial console log? | 12:35 |
| mdbooth | When I originally logged in there was no ip address on eth0 | 12:35 |
| mdbooth | so I manually did 'ifup eth0' | 12:35 |
| mdbooth | It does 'Sending discover...' for a bit | 12:35 |
| mdbooth | Then dies with 'No lease, failing' | 12:36 |
| *** mmedvede has quit IRC | 12:36 | |
| mdbooth | Ah... could this be because it's on the 'public' network? | 12:37 |
| kashyap | Ah | 12:37 |
| kashyap | $ ip netns exec qdhcp-$UUID ip a | 12:37 |
| kashyap | ^ What does that say? | 12:37 |
| mdbooth | 10.1.0.2/24 | 12:38 |
| mdbooth | And the guest is on 172.x.x.x | 12:38 |
| kashyap | mdbooth: Can you run this https://raw.githubusercontent.com/larsks/neutron-diag/master/gather-network-info | 12:38 |
| kashyap | It just collects Neutron networking info | 12:38 |
| mdbooth | I trust you, man | 12:39 |
| mdbooth | I'm just gonna run it :) | 12:39 |
| *** diga has quit IRC | 12:40 | |
| *** praveen has left #openstack-nova | 12:42 | |
| *** woodster_ has joined #openstack-nova | 12:42 | |
| kashyap | There are actually two diagnostics scripts: | 12:43 |
| kashyap | https://github.com/larsks/neutron-diag | 12:43 |
| mdbooth | kashyap: Sorted. The instance wasn't on the subnet with dhcp. | 12:43 |
| kashyap | Ah-ha | 12:43 |
| mdbooth | I now have networking :) | 12:43 |
| mdbooth | kashyap: Thank you, sir! | 12:43 |
| kashyap | I didn't help much :-) | 12:43 |
| mdbooth | Well I didn't have any idea what to poke until you told me :) | 12:44 |
| mdbooth | woot, I have 2 instances which can ping each other :) Revolutionary. | 12:45 |
| kashyap | Excellent | 12:45 |
| mdbooth | I literally don't think I've ever done that before. | 12:45 |
| *** karthiks has joined #openstack-nova | 12:45 | |
| kashyap | If I'm in a hurry w/o floating IP, I just SSH via the namespace $ sudo ip netns exec qdhcp-$UUID ssh cirros@10.1.0.8 | 12:46 |
| * mdbooth is just connecting directly to the vnc port | 12:46 | |
| *** vladikr has joined #openstack-nova | 12:47 | |
| *** pradk has quit IRC | 12:47 | |
| paul-carlton2 | William18.,. | 12:47 |
| cdent | jaypipes: responded to your -2 on https://review.openstack.org/#/c/365069/ | 12:47 |
| *** MVenesio has joined #openstack-nova | 12:47 | |
| *** MVenesio has quit IRC | 12:47 | |
| *** mgoddard_ has joined #openstack-nova | 12:48 | |
| *** slaweq_ has joined #openstack-nova | 12:49 | |
| slaweq_ | hello, can I ask someone from Nova core reviewers to check https://review.openstack.org/#/c/363926/ ? | 12:49 |
| slaweq_ | thx in advance | 12:49 |
| *** jaosorior has quit IRC | 12:49 | |
| sdague | cdent / jaypipes with the placement job using the placement api optionally, I wonder if there is some way to figure out that it's working in the tempest test run besides looking at logs | 12:50 |
| *** jaosorior has joined #openstack-nova | 12:50 | |
| alex_xu | nova api meeting is in 10 mins at #openstack-meeting-4 | 12:51 |
| cdent | sdague: does it make any sense to either: a) add a simple tempest test that gets a list of resource providers or something b) add a post test hook that does a curl doing the same thing? | 12:51 |
| *** tblakes has joined #openstack-nova | 12:51 | |
| *** wuhg has quit IRC | 12:51 | |
| *** mgoddard has quit IRC | 12:52 | |
| *** GB21 has quit IRC | 12:56 | |
| rpodolyaka | hi all! would appreciate a second core review on https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bug/1606496, if someone has a minute | 12:57 |
| *** thorst has joined #openstack-nova | 12:57 | |
| edleafe | cdent: sdague: jaypipes: Sorry I wasn't available much yesterday; lots of non-OpenStack stuff to handle. Good work on identifying the SQL bug. | 12:58 |
| edleafe | Where can I dive in today? | 12:58 |
| *** gmann has joined #openstack-nova | 12:58 | |
| *** mdrabe has joined #openstack-nova | 12:59 | |
| *** syjulian_ has joined #openstack-nova | 13:00 | |
| cdent | edleafe: I posted a bit of a status update to os-dev this morning which points to the main chunk of code under review. And then if you search for stuff owned by me there's a fair few minor fixups for buglets. | 13:00 |
| cdent | I suspec the most important thing to be doing now is making it go and seeing what happens | 13:00 |
| *** roeyc has quit IRC | 13:00 | |
| *** claudiub|2 has joined #openstack-nova | 13:00 | |
| alex_xu | nova api meeting is running | 13:01 |
| edleafe | cdent: ok, thanks. So is it multiple RPs or multiple consumers that are needed to see the problem? | 13:01 |
| openstackgerrit | Sean Dague proposed openstack/nova: Additional logging for placement API https://review.openstack.org/366186 | 13:01 |
| *** jwcroppe has joined #openstack-nova | 13:01 | |
| *** mmedvede has joined #openstack-nova | 13:02 | |
| cdent | edleafe: "problem" should be fixed, but we've never been too clear on the combination of inputs to make it go because we couldn't make it happen in local tests, only in "real" runs. but basically: multiple resource providers, each with multiple (three) classes of resources, with > 1 allocation | 13:03 |
| *** claudiub has quit IRC | 13:03 | |
| edleafe | cdent: I see. That will be fun to set up locally. | 13:03 |
| bauzas | cdent: tbh, the most important thing for now is to remember we're in a RC period | 13:04 |
| bauzas | cdent: which means most of us should look at bugfixes and bug reports | 13:04 |
| cdent | bauzas: all the placement stuff is bugs and bugfixes | 13:05 |
| bauzas | cdent: but they are not regressions | 13:05 |
| *** garyk1 has quit IRC | 13:06 | |
| bauzas | if I eat all my time to look at the placement changes, then I couldn't help others wanting to fix regressions | 13:06 |
| sdague | cdent: for the reproduce it was 1 rp, > 1 resource classes, and > 1 consumer (so there can be allocations that won't be deleted when a new allocation comes in) | 13:06 |
| cdent | bauzas: it doesn't work well enough yet and we committed to making writing inventory and allocations work and be scrutable enough to make work. That's going to mean adding more code, now. | 13:06 |
| bauzas | and we also have 43 new bugs, and 14 rc possible changes | 13:06 |
| *** pilgrimstack1 has joined #openstack-nova | 13:06 | |
| sdague | which also explains why it was only happening some times in tests | 13:07 |
| *** gongysh has quit IRC | 13:07 | |
| *** pilgrimstack has quit IRC | 13:07 | |
| sdague | because you needed 2 vms up at once to expose it | 13:07 |
| bauzas | cdent: okay, it's your opinion, I prefer to look at Launchpad | 13:07 |
| *** PaulMurray has quit IRC | 13:07 | |
| cdent | please do bauzas | 13:08 |
| *** roeyc has joined #openstack-nova | 13:08 | |
| edleafe | sdague: for tests can we add some dummy consumer records? | 13:08 |
| sdague | edleafe: there is a test now | 13:08 |
| edleafe | sdague: ah, good - still catching up | 13:08 |
| cdent | yeah, edleafe I think the problem is sorted mostly well enough now | 13:08 |
| sdague | https://review.openstack.org/#/c/366245/8/nova/tests/functional/db/test_resource_provider.py | 13:09 |
| edleafe | sdague: awesome | 13:09 |
| bhagyashri | dansmith: Hi, could you please review https://review.openstack.org/#/c/345191/ ? | 13:10 |
| *** stevemar_ has joined #openstack-nova | 13:10 | |
| *** andrey-mp has joined #openstack-nova | 13:11 | |
| *** vishwanathj has joined #openstack-nova | 13:11 | |
| *** links has quit IRC | 13:12 | |
| openstackgerrit | Sam Betts proposed openstack/nova: [WIP] Support Ironic interface attach/detach in nova virt https://review.openstack.org/364413 | 13:13 |
| *** xyang1 has joined #openstack-nova | 13:13 | |
| *** stevemar_ has quit IRC | 13:15 | |
| *** PaulMurray has joined #openstack-nova | 13:17 | |
| andrey-mp | bauzas: what do you think - when you can finish this: https://review.openstack.org/#/c/365740/ development is stuck in EC2-API due to this. is it possible to fix it or maybe its simplier to revert something first? | 13:17 |
| bauzas | andrey-mp: on it | 13:17 |
| *** mingyu has quit IRC | 13:17 | |
| bauzas | andrey-mp: I did a bit of triage this morning, but I'm now full speed on writing a new PS for the above | 13:18 |
| *** ducttape_ has joined #openstack-nova | 13:18 | |
| *** mingyu has joined #openstack-nova | 13:18 | |
| *** hoangcx has joined #openstack-nova | 13:19 | |
| andrey-mp | bauzas: ok, thank you! | 13:19 |
| *** ratailor has quit IRC | 13:19 | |
| *** gmann_ has joined #openstack-nova | 13:19 | |
| *** gmann_ has quit IRC | 13:19 | |
| *** ducttape_ has quit IRC | 13:20 | |
| *** su_zhang has joined #openstack-nova | 13:21 | |
| *** oanson has quit IRC | 13:21 | |
| *** hemanthm|away is now known as hemanthm | 13:21 | |
| *** mingyu has quit IRC | 13:22 | |
| syjulian_ | gibi: hey gibi do you have a moment? | 13:25 |
| *** SS-SP has joined #openstack-nova | 13:26 | |
| *** tonytan4ever has joined #openstack-nova | 13:26 | |
| *** SS-SP has quit IRC | 13:26 | |
| *** SS-SP has joined #openstack-nova | 13:28 | |
| *** jamesdenton has joined #openstack-nova | 13:28 | |
| SS-SP | Hi All, | 13:28 |
| SS-SP | we have an issue | 13:28 |
| *** karthiks has quit IRC | 13:28 | |
| SS-SP | When we try to boot a VM, multiple Ip's are assigning to it and the VM is going to error state | 13:29 |
| SS-SP | can anyone help us | 13:29 |
| *** efried has quit IRC | 13:29 | |
| SS-SP | with the error message: | 13:29 |
| SS-SP | | {"message": "Exceeded maximum number of retries. Exceeded max scheduling attempts 3 for instance a8b953a3-85c4-46b2-a0bd-dd312c1d36fa. Last exception: 500 Internal Server Error: The server has either erred or is incapable of performing the requested operation. (HTTP 5", "code": 500, "details": " File \"/usr/lib/python2.7/dist-packages/nova/conductor/manager.py\", line 393, in build_instances | | | 13:30 |
| *** mgoddard_ has quit IRC | 13:30 | |
| *** mgoddard has joined #openstack-nova | 13:30 | |
| gibi | syjulian_: on a meeting right now, I will ping you in about an hour | 13:30 |
| *** kfarr has joined #openstack-nova | 13:31 | |
| *** wangqun has joined #openstack-nova | 13:31 | |
| *** adisky has quit IRC | 13:31 | |
| *** tbachman has joined #openstack-nova | 13:33 | |
| *** syjulian_ has quit IRC | 13:34 | |
| *** bsv has quit IRC | 13:34 | |
| *** hamzy has quit IRC | 13:34 | |
| openstackgerrit | Prateek Arora proposed openstack/nova: Handle ConnectFailure from keystone when neutron is down https://review.openstack.org/327706 | 13:35 |
| *** sudipto_ has joined #openstack-nova | 13:35 | |
| *** sudipto has joined #openstack-nova | 13:35 | |
| *** ttrumm has quit IRC | 13:35 | |
| andrey-mp | SS-SP: do you have a bigger stack? | 13:35 |
| *** syjulian_ has joined #openstack-nova | 13:36 | |
| *** ttrumm has joined #openstack-nova | 13:36 | |
| prateek | alex_xu, I have made the changes as per your comments on https://review.openstack.org/#/c/327706/, can you please take a look ? | 13:38 |
| syjulian_ | gibi: ok cool | 13:38 |
| *** tbachman has quit IRC | 13:39 | |
| *** tbachman has joined #openstack-nova | 13:39 | |
| SS-SP | <andrey-mp> we used flavor with memory:512 disk:1 vcpu:1 | 13:40 |
| *** mriedem has joined #openstack-nova | 13:40 | |
| prateek | cdent, https://review.openstack.org/#/c/327706/ the test problem has been fixed , if you can just have a look now ? | 13:41 |
| *** ttrumm has quit IRC | 13:41 | |
| prateek | mriedem, sahid has reviewed and approved https://review.openstack.org/#/c/327706/, can you please take a look now ? | 13:41 |
| andrey-mp | SS-SP: i mean stack trace of an error | 13:41 |
| *** annegentle has joined #openstack-nova | 13:42 | |
| *** hoangcx has quit IRC | 13:42 | |
| *** gmann has quit IRC | 13:42 | |
| *** krtaylor has joined #openstack-nova | 13:42 | |
| mdbooth | danpb: Can you recommend a simple way to put memory stress on a guest running a cirros image? | 13:43 |
| *** takedakn has joined #openstack-nova | 13:44 | |
| mdbooth | My plan was to try to get 'stress' working on one | 13:44 |
| mdbooth | But if there's a simple hack I can use to dirty lots of memory continuously that would be awesome | 13:44 |
| SS-SP | <andrey-mp>We didnt get any error log in nova-api.log | 13:44 |
| danpb | mdbooth: imho just biuld your own initrd containing a custom init.c that does yuour stress | 13:44 |
| dansmith | cdent: are you working on addressing this? https://review.openstack.org/#/c/365068/7 | 13:44 |
| SS-SP | but in nova show: | 13:45 |
| mdbooth | danpb: Yes, I was thinking less effort than that :) | 13:45 |
| bhagyashri | mikal: Hi, could you please review https://review.openstack.org/#/c/346026/ ? | 13:45 |
| *** hoangcx has joined #openstack-nova | 13:45 | |
| SS-SP | {"message": "Exceeded maximum number of retries. Exceeded max scheduling attempts 3 for instance 91f31b5d-4e40-4dd4-853a-7bfc62e5789e. Last exception: 500 Internal Server Error: The server has either erred or is incapable of performing the requested operation. (HTTP 5", "code": 500, "details": " File \"/usr/lib/python2.7/dist-packages/nova/conductor/manager.py\", line 393, in build_instances | | | 13:45 |
| danpb | mdbooth: you can steal my stress code if you want http://git.qemu.org/?p=qemu.git;a=blob;f=tests/migration/stress.c;hb=HEAD | 13:45 |
| *** karthiks has joined #openstack-nova | 13:45 | |
| mdbooth | danpb: Awesome, thanks | 13:45 |
| cdent | dansmith: no I'm in the midst of writing GET /resource_providers/{uuid}/allocations because jaypipes says we need that | 13:45 |
| *** ducttape_ has joined #openstack-nova | 13:45 | |
| dansmith | cdent: sounds familiar :) | 13:45 |
| danpb | mdbooth: with that stress program migration will never complete unless using post-copy, or throttling by >=95% | 13:46 |
| cdent | dansmith: ikr | 13:46 |
| mdbooth | danpb: Does it work out how much ram the guest has? | 13:46 |
| * cdent shrugs | 13:46 | |
| *** eharney has joined #openstack-nova | 13:46 | |
| dansmith | cdent: okay, the log warning is probably too small to justify having to rebase your thing so can you just throw a warning into the bottom patch when you submit? | 13:46 |
| mdbooth | danpb: tbh, I just need to force post-copy to be used. My idle cirros instance doesn't touch it. | 13:46 |
| dansmith | just needs to be a single line change I think | 13:46 |
| danpb | mdbooth: it takes ramsize=GB on kernel command line | 13:47 |
| sdague | dansmith: well, we need to bring back a chunk of the code right? | 13:47 |
| frickler | SS-SP: getting multiple IPs is a known issue: https://bugs.launchpad.net/nova/+bug/1597596 but that doesn't explain why your instance gets rescheduled in the first place | 13:47 |
| openstack | Launchpad bug 1597596 in OpenStack Compute (nova) "network not always cleaned up when spawning VMs" [Medium,In progress] - Assigned to Aihua Edward Li (aihuaedwardli) | 13:47 |
| cdent | yeah, dansmith sdague is right | 13:47 |
| sdague | https://review.openstack.org/#/c/365068/7/nova/objects/resource_provider.py | 13:47 |
| cdent | we removed the detector | 13:47 |
| mdbooth | danpb: I'll be compiling my own anyway, so I can hack that :) | 13:47 |
| mdbooth | Thanks | 13:47 |
| sdague | which would also be needed for jay's wanted follow on notification | 13:47 |
| sdague | cdent: if you are fine with that, I can respin that | 13:47 |
| cdent | dansmith: if you can do that, I can make your your GET | 13:47 |
| dansmith | cdent: sdague: ah, I'm squashing two things.. the notification and this, yah.. was thinking that part got done | 13:47 |
| cdent | or sdague | 13:47 |
| dansmith | sdague: I'm also happy to do it, but whatever you want | 13:48 |
| *** prateek has quit IRC | 13:48 | |
| mdbooth | dansmith: Got a minute to talk about the post-copy networking thing and rpc? | 13:48 |
| sdague | dansmith: if you want to run with it, it's all yours | 13:48 |
| dansmith | sdague: okay | 13:48 |
| dansmith | mdbooth: not really, but you can talk and I can smile-n-nod if you want :D | 13:48 |
| andrey-mp | SS-SP: ok, maybe it's another bug... I meant this - https://bugs.launchpad.net/nova/+bug/1620248 | 13:49 |
| openstack | Launchpad bug 1620248 in OpenStack Compute (nova) "Can't rename instance right after creation (regression)" [High,In progress] - Assigned to Sylvain Bauza (sylvain-bauza) | 13:49 |
| mdbooth | dansmith: Don't tell me that's what you normally do | 13:49 |
| * mdbooth tries anyway | 13:49 | |
| dansmith | heh | 13:49 |
| *** ljxiash has joined #openstack-nova | 13:49 | |
| mdbooth | dansmith: So, looking at the stuff which needs to happen at the point the migration source decides to kick off post-copy, it's all remote calls already | 13:50 |
| mdbooth | all done via network_api | 13:50 |
| mdbooth | So, there's code which currently runs on the source, and code which currently runs on the dest | 13:50 |
| mdbooth | IIUC, we can just move all of that to the source, because it'll call out to the dest anyway, so we're good | 13:50 |
| mdbooth | Anyway, there's still an issue | 13:51 |
| mdbooth | computemanager.post_live_migration_at_destination() does a bunch of networky stuff, called via rpc | 13:51 |
| SS-SP | <andrey-mp>ok..But Why its trying to assign multiple ip's(from same ip Address range) | 13:51 |
| mdbooth | We need to tell it to not do that if we already did | 13:51 |
| dansmith | mdbooth: yeah | 13:52 |
| mdbooth | The existing arguments to it don't obviously allow that | 13:52 |
| mdbooth | And as it's remote, we really need to know in advance if it can even be told not to do it | 13:52 |
| mdbooth | Otherwise we basically don't support post-copy to that destination | 13:52 |
| andrey-mp | SS-SP: sorry, I don't know ) i thought that this is an issue with updating server right after creation | 13:53 |
| mdbooth | dansmith: Not sure how to approach that. | 13:53 |
| *** annegent_ has joined #openstack-nova | 13:53 | |
| *** pradk has joined #openstack-nova | 13:53 | |
| dansmith | mdbooth: don't we have a pre-migration call that gets made to both the source and destination? | 13:54 |
| danpb | mdbooth: moving code from method on the source, to method on the dest is going to cause you pain with upgrades | 13:54 |
| danpb | mdbooth: because you'll have source running old code, and dest running new code, and so both will contain the same logic | 13:55 |
| dansmith | so the destination needs to say definitively if it supports it or not, and then we treat the absence of that as a no | 13:55 |
| *** efried has joined #openstack-nova | 13:55 | |
| mdbooth | danpb: Yeah, that's what I'm looking for guidance on. | 13:55 |
| danpb | mdbooth: i'd not touch that existing logic at all | 13:55 |
| danpb | mdbooth: basically we just need hooks in the virt driver to trigger callback in the compute manager to run | 13:56 |
| mdbooth | dansmith: Will look into that, thanks. | 13:56 |
| danpb | self.network_api.setup_networks_on_host(context, instance, prev_host, teardown=True) | 13:56 |
| danpb | self.network_api.setup_networks_on_host(context, instance, self.host) | 13:56 |
| danpb | on the source + dest, when post-copy switches over | 13:56 |
| mdbooth | danpb: Yeah, and migration_instance_Start() and migration_instance_finish() | 13:56 |
| mdbooth | That's all fine | 13:57 |
| mdbooth | The problem is that compute manager will still subsequently call post_live_migration_at_destination() | 13:57 |
| danpb | so you shouldn't need to change anything about RPC layer afaict | 13:57 |
| *** rcernin has joined #openstack-nova | 13:57 | |
| *** annegentle has quit IRC | 13:57 | |
| mdbooth | And that method does all the networky things that we just did | 13:57 |
| mdbooth | So we need to be able to tell it that we already did it | 13:57 |
| *** zenoway has joined #openstack-nova | 13:57 | |
| mdbooth | We can't not call it, because it also does other stuff | 13:57 |
| danpb | is it not possible to make setup_networks_on_host() into a no-op if it detects the work is already done | 13:58 |
| mdbooth | It's already a no-op in neutron, no? | 13:59 |
| mdbooth | Maybe not | 13:59 |
| mdbooth | Anyway, I didn't want to go there, tbh | 13:59 |
| mdbooth | I wanted to keep the same network calls with the same semantics | 13:59 |
| mdbooth | Just call them at the right times | 13:59 |
| mdbooth | dansmith: So, I don't think we can do that (info back from pre_live_migration) | 14:00 |
| *** anush has joined #openstack-nova | 14:00 | |
| dansmith | mdbooth: that's the point of that call, right? | 14:00 |
| mdbooth | Well, maybe we can, but we'd still require an rpc change to post_live_migration_at_destination() | 14:00 |
| mdbooth | because it doesn't take migrate_data as an argument | 14:01 |
| dansmith | sure | 14:01 |
| mdbooth | It really don't have an obviously fudgable argument | 14:01 |
| *** wangqun has quit IRC | 14:01 | |
| mdbooth | doesn't | 14:01 |
| *** ducttape_ has quit IRC | 14:02 | |
| *** mlavalle has joined #openstack-nova | 14:02 | |
| mdbooth | dansmith: So, making an rpc change to post_live_migration_at_destination() is fine (albeit a pita) | 14:03 |
| dansmith | mdbooth: of course | 14:03 |
| mdbooth | However, I'd need to know way before I called it what version it supports | 14:03 |
| mdbooth | Is that achievable? | 14:03 |
| dansmith | mdbooth: rpcapi knows, and can either fail the call or adjust the call/return based on the version yes | 14:03 |
| *** ayoung has quit IRC | 14:03 | |
| *** mamandle has quit IRC | 14:04 | |
| mdbooth | dansmith: Right, but I'd need to know that *before* the call | 14:04 |
| dansmith | mdbooth: so you could make it fail the call with some known exception and you retry it with something different | 14:04 |
| dansmith | mdbooth: how far before/ | 14:04 |
| danpb | the RPC version cap should be sufficient i'd think | 14:04 |
| mdbooth | dansmith: Right at the start of the live migration | 14:04 |
| *** liverpooler has joined #openstack-nova | 14:04 | |
| dansmith | mdbooth: you could also use SERVICE_VERSION to do this detection, and just refuse to do these kinds of migrations unless everything is upgraded | 14:04 |
| mdbooth | Perhaps | 14:04 |
| mdbooth | ... | 14:05 |
| danpb | eg, check if the version cap >= the desired version | 14:05 |
| mdbooth | danpb: Can you do that against an arbitrary target nova compute? | 14:05 |
| dansmith | mdbooth: you could use the pre calls to signal to the caller very early, right? | 14:05 |
| *** zul_ has joined #openstack-nova | 14:05 | |
| mdbooth | dansmith: Ok, so make 2 rpc changes? Assume that if pre_live_migration returns the data I need, then post also supports it | 14:06 |
| bauzas | alaski: not sure I got your comment on https://review.openstack.org/#/c/365740/3/nova/compute/api.py@2421 | 14:06 |
| danpb | mdbooth: the RPC version cap is cloud-wide | 14:06 |
| dansmith | mdbooth: right | 14:06 |
| bauzas | alaski: that instance update is done within a contextmanager, I can't really move that out from the conditional | 14:06 |
| *** rodrigods has quit IRC | 14:06 | |
| *** rodrigods has joined #openstack-nova | 14:06 | |
| mdbooth | danpb: That might be simpler. Is that the SERVICE_VERSION thing dansmith mentioned? | 14:06 |
| bauzas | alaski: unless you meant I should use a nested private helper I could call that would just do that | 14:07 |
| dansmith | mdbooth: no | 14:07 |
| dansmith | mdbooth: the rpc cap and the service version are somewhat related but different things | 14:08 |
| *** marst has quit IRC | 14:08 | |
| dansmith | mdbooth: but seriously, I think the right thing to do is have the pre calls return the sentinel about whether it's supported or not | 14:08 |
| mdbooth | Ok, thanks. | 14:08 |
| * mdbooth still needs to cause post-copy migrate to occur in his test environment | 14:09 | |
| *** prateek has joined #openstack-nova | 14:09 | |
| * mdbooth compiles danpb's stress bomb | 14:09 | |
| dansmith | sdague: so at one point, logging inside of db_api was a no-no because it can cause us to switch threads while we hold DB locks.. I'm not sure where that landed with the facade stuff, but it seems like that should still be off limits.. do you know? | 14:10 |
| *** READ10 has joined #openstack-nova | 14:11 | |
| sdague | dansmith: I don't | 14:11 |
| dansmith | zzzeek: jaypipes ^ do you know? I thought any time we're in a writer and have touched stuff that we're holding a transaction open, right? logging that could hit syslog via socket would be bad, yes? | 14:12 |
| dansmith | I'm thinking back to neutron having a lot of deadlocks over this kind of thing | 14:12 |
| sdague | dansmith: was that before we switched mysql drivers? | 14:12 |
| *** ajo_ has joined #openstack-nova | 14:12 | |
| jaypipes | dansmith: that was something to do with real threads IIRC, right? | 14:13 |
| dansmith | sdague: yes, but not sure that matters. | 14:13 |
| dansmith | jaypipes: no, not the logging thing | 14:13 |
| jaypipes | hmmm, /me thinks back... | 14:13 |
| dansmith | jaypipes: if I do a thing that grabs a lock on some table, then I switch out to do a log, which causes me to run another thread that goes to update the same table, it waits | 14:13 |
| *** GB21 has joined #openstack-nova | 14:14 | |
| dansmith | although maybe with the new backend and not real threads I would switch again when I go to wait and not deadlock? I dunno | 14:14 |
| dansmith | we have LOG statements in our db api, but they're few and in places that used to be outside transactions before the facade work, AFAIK | 14:14 |
| cdent | dansmith: Isn't that concern just in eventlet situations? and there shouldn't be any eventlet in the placement api (although some may have leaked in) | 14:15 |
| dansmith | sdague: I was planning to return info about exceeded RPs up the stack until we're back out of all the writers and then log it there, but was just making sure I wasn't overcomplicating things | 14:15 |
| jaypipes | dansmith: those log statements are outside the transaction boundary. if you are referring to the log statements I had originally included in the _inventory_changed() method? | 14:15 |
| dansmith | cdent: that's a fair point, but this in the object code which can technically be called from anywhere | 14:15 |
| dansmith | jaypipes: no I'm talking about what sdague wants, which is a log deep in the inner methods, which are @db_writers | 14:16 |
| *** BjoernT has joined #openstack-nova | 14:16 | |
| cdent | dansmith: yeah, just putting it out there in case we want to change that lack-of-restriction | 14:16 |
| dansmith | jaypipes: same goes for your firing of a notification though | 14:16 |
| jaypipes | gotcha | 14:16 |
| dansmith | cdent: yeah, well, you can do that when you're separate and can't be called from eventlet code :) | 14:16 |
| cdent | :) | 14:16 |
| dansmith | regardless, | 14:16 |
| dansmith | we probably don't want to fire a notification from inside a @db_writer anyway I think, | 14:17 |
| cdent | oh yeah, notifications will blow all that out | 14:17 |
| dansmith | so might as well do this the return method I think | 14:17 |
| *** diogogmt has quit IRC | 14:17 | |
| *** hoangcx has quit IRC | 14:17 | |
| *** tri2sing has joined #openstack-nova | 14:18 | |
| sdague | dansmith: so how about do that check, set a var, return the exception at the end of the block if we exceeded it | 14:19 |
| sdague | then just handle the logging in the exception handler | 14:19 |
| alaski | bauzas: instance.update is essentially a dict update, it's not doing anything that requires the context manager. | 14:19 |
| dansmith | sdague: we need to handle the case where multiple RPs are over cap, so I'm just making a list of info and returning that | 14:19 |
| bauzas | alaski: oh I see, fair point | 14:19 |
| dansmith | sdague: we don't want to raise anywhere because we don't want to abort the transaction | 14:19 |
| sdague | dansmith: oh, right any raise will roll us back | 14:20 |
| dansmith | yup | 14:20 |
| *** hoangcx has joined #openstack-nova | 14:20 | |
| *** zul_ has quit IRC | 14:20 | |
| *** zenoway has quit IRC | 14:20 | |
| *** pradk has quit IRC | 14:20 | |
| *** eandersson has quit IRC | 14:20 | |
| *** salv-orlando has quit IRC | 14:20 | |
| *** gszasz has quit IRC | 14:20 | |
| *** andrey-mp has quit IRC | 14:20 | |
| *** danpb has quit IRC | 14:20 | |
| *** bkopilov has quit IRC | 14:20 | |
| *** mirrorbo1 has quit IRC | 14:20 | |
| *** jgrassler has quit IRC | 14:20 | |
| *** hehx has quit IRC | 14:20 | |
| *** dulek has quit IRC | 14:20 | |
| *** lennyb has quit IRC | 14:20 | |
| *** cfriesen has quit IRC | 14:20 | |
| *** kswiatek has quit IRC | 14:20 | |
| *** darrenc has quit IRC | 14:20 | |
| *** tpatzig_ has quit IRC | 14:20 | |
| *** gstepanov__ has quit IRC | 14:20 | |
| *** raj_singh has quit IRC | 14:20 | |
| *** anteaya has quit IRC | 14:20 | |
| *** niedbalski has quit IRC | 14:20 | |
| *** eliqiao has quit IRC | 14:20 | |
| *** ejat has quit IRC | 14:20 | |
| *** TravT has joined #openstack-nova | 14:20 | |
| markus_z | PaulMurray: FYI: I checked the live-migration with serial console patches again locally, and they passed: https://review.openstack.org/#/c/275801/ | 14:20 |
| *** TravT has quit IRC | 14:20 | |
| markus_z | PaulMurray: Would be cool to get the merged for RC | 14:20 |
| *** su_zhang has quit IRC | 14:20 | |
| *** su_zhang has joined #openstack-nova | 14:21 | |
| *** TravT has joined #openstack-nova | 14:22 | |
| *** TravT has quit IRC | 14:22 | |
| *** TravT has joined #openstack-nova | 14:23 | |
| gibi | syjulian_: now I'm available | 14:23 |
| *** marst has joined #openstack-nova | 14:23 | |
| *** syjulian_ has quit IRC | 14:24 | |
| *** tblakes has quit IRC | 14:24 | |
| pumaranikar | mriedem, can you please review this https://review.openstack.org/#/c/365181/ | 14:25 |
| *** pradk has joined #openstack-nova | 14:25 | |
| *** su_zhang has quit IRC | 14:26 | |
| *** yamahata has joined #openstack-nova | 14:27 | |
| *** felix23ma has quit IRC | 14:27 | |
| mriedem | pumaranikar: i don't think we want to make that change in newton at this point, with 1.5 weeks to RC1 | 14:28 |
| openstackgerrit | MJWurtz proposed openstack/nova: Refresh info_cache after deleting floating IP https://review.openstack.org/357494 | 14:29 |
| *** lucasagomes is now known as lucas-afk | 14:29 | |
| *** salv-orlando has joined #openstack-nova | 14:30 | |
| *** EinstCrazy has joined #openstack-nova | 14:30 | |
| *** mdrabe has quit IRC | 14:30 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add GET /resource_providers/{uuid}/allocations https://review.openstack.org/366789 | 14:31 |
| cdent | jaypipes, dansmith: that's ^ a first stab at getting allocations, needs some commentary on response format and any desired tweaks (presumably in some contexts the results could get very long?) | 14:31 |
| *** syjulian_ has joined #openstack-nova | 14:32 | |
| cmartin | Could I get somebody to review this possibly? https://review.openstack.org/#/c/364337/ | 14:32 |
| *** sudipto_ has quit IRC | 14:32 | |
| *** sudipto has quit IRC | 14:32 | |
| dansmith | cdent: so, um, _set_inventory() doesn't seem to actually replace inventory for things we already have.. certainly I'm wrong about that right? | 14:33 |
| cdent | ugh | 14:33 |
| dansmith | cdent: it seems to only look for new classes added or removed, but | 14:33 |
| *** annegent_ has quit IRC | 14:33 | |
| dansmith | er, wait, maybe the & of the two sets is "things that are in both old and new, to be updated" ? | 14:34 |
| cdent | so what it is supposed to do is figure out add, delete, update | 14:34 |
| cdent | and then do each separately | 14:34 |
| dansmith | ah, nm, I had a typo | 14:34 |
| * cdent schwews | 14:34 | |
| *** dikonoor has quit IRC | 14:35 | |
| *** markus_z has quit IRC | 14:35 | |
| cdent | dansmith: I need to take an air break, but if we can sync up when I get back so I can know where to put some energy that would be awesome. | 14:35 |
| pumaranikar | mriedem, agree. One query on your comment. Do I need to include schema change in the same patch ? I thought we do schema change in the release after model changes. | 14:35 |
| mriedem | pumaranikar: they would go together | 14:35 |
| cdent | having us all in and out of the same stack has my brain going ow | 14:35 |
| mriedem | there is no point in the model change w/o the schema change, that would just be confusing | 14:35 |
| pumaranikar | mriedem, ok i can add that to same patch the. thanks | 14:36 |
| pumaranikar | *then | 14:36 |
| *** cdent has quit IRC | 14:37 | |
| openstackgerrit | Dan Smith proposed openstack/nova: [placement] Allow inventory to violate allocations https://review.openstack.org/365068 | 14:38 |
| dansmith | see if that looks good ^ ... brbr | 14:38 |
| *** GB21 has quit IRC | 14:39 | |
| *** eandersson has joined #openstack-nova | 14:39 | |
| *** ducttape_ has joined #openstack-nova | 14:39 | |
| *** kfarr has quit IRC | 14:40 | |
| *** rfolco has joined #openstack-nova | 14:40 | |
| *** jdurgin1 has joined #openstack-nova | 14:40 | |
| *** zenoway has joined #openstack-nova | 14:40 | |
| *** salv-orlando has quit IRC | 14:40 | |
| *** tblakes has joined #openstack-nova | 14:41 | |
| *** zul_ has joined #openstack-nova | 14:42 | |
| *** andrey-mp has joined #openstack-nova | 14:42 | |
| *** danpb has joined #openstack-nova | 14:42 | |
| *** bkopilov has joined #openstack-nova | 14:42 | |
| *** mirrorbo1 has joined #openstack-nova | 14:42 | |
| *** hehx has joined #openstack-nova | 14:42 | |
| *** dulek has joined #openstack-nova | 14:42 | |
| *** lennyb has joined #openstack-nova | 14:42 | |
| *** cfriesen has joined #openstack-nova | 14:42 | |
| *** kswiatek has joined #openstack-nova | 14:42 | |
| *** darrenc has joined #openstack-nova | 14:42 | |
| *** tpatzig_ has joined #openstack-nova | 14:42 | |
| *** gstepanov__ has joined #openstack-nova | 14:42 | |
| *** raj_singh has joined #openstack-nova | 14:42 | |
| *** anteaya has joined #openstack-nova | 14:42 | |
| *** niedbalski has joined #openstack-nova | 14:42 | |
| *** eliqiao has joined #openstack-nova | 14:42 | |
| *** ejat has joined #openstack-nova | 14:42 | |
| edleafe | cdent: suggestion: use set1.intersection(set2) instead of set1 & set2 - it's much less confusing to readers | 14:42 |
| *** zz_dimtruck is now known as dimtruck | 14:42 | |
| *** hoonetorg has quit IRC | 14:43 | |
| *** jgrassler has joined #openstack-nova | 14:44 | |
| *** zenoway has quit IRC | 14:45 | |
| *** Swami has joined #openstack-nova | 14:45 | |
| *** mdrabe has joined #openstack-nova | 14:46 | |
| *** hamzy has joined #openstack-nova | 14:46 | |
| *** Swami_ has joined #openstack-nova | 14:46 | |
| *** ekuris has quit IRC | 14:46 | |
| *** takedakn has quit IRC | 14:47 | |
| *** ayoung has joined #openstack-nova | 14:48 | |
| *** takedakn has joined #openstack-nova | 14:48 | |
| *** hoangcx has quit IRC | 14:48 | |
| *** gszasz has joined #openstack-nova | 14:49 | |
| jaypipes | dansmith: just finishing up cdent review and will re-review your after. | 14:49 |
| jaypipes | (oh, and another meeting...) | 14:49 |
| *** moshele has quit IRC | 14:49 | |
| *** takedakn has quit IRC | 14:50 | |
| *** Swami has quit IRC | 14:50 | |
| *** kevinz has quit IRC | 14:50 | |
| openstackgerrit | Swaminathan Vasudevan proposed openstack/nova: Implement setup_networks_on_host for Neutron networks https://review.openstack.org/275073 | 14:50 |
| dansmith | jaypipes: I can only assume you're banking up meetings in case of some apocalyptic scenario where suddenly meetings are in demand and you have a large supply | 14:52 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Group scheduler options https://review.openstack.org/349959 | 14:53 |
| *** karthiks has quit IRC | 14:54 | |
| *** amotoki has quit IRC | 14:56 | |
| *** danpb has left #openstack-nova | 14:56 | |
| *** hoonetorg has joined #openstack-nova | 14:56 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Add tests for safe_connect decorator https://review.openstack.org/366800 | 14:57 |
| *** hoangcx has joined #openstack-nova | 14:57 | |
| sdague | dansmith: lol | 14:57 |
| *** rrecio has joined #openstack-nova | 14:57 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: live-mig: Add claims and proper resource tracking https://review.openstack.org/244489 | 14:59 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: libvirt: pass libvirt reference methods to migration https://review.openstack.org/328997 | 14:59 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: libvirt: live-migrate updates NUMA and cpus in the XML https://review.openstack.org/286744 | 14:59 |
| *** mjura has quit IRC | 14:59 | |
| *** gouthamr has joined #openstack-nova | 14:59 | |
| *** _ducttape_ has joined #openstack-nova | 15:00 | |
| *** gouthamr_ has joined #openstack-nova | 15:01 | |
| jaypipes | dansmith: it's industry standard. | 15:01 |
| alaski | johnthetubaguy: can you take a look at https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bug/1606496 when you get some time | 15:01 |
| *** EinstCrazy has quit IRC | 15:01 | |
| dansmith | heh | 15:01 |
| *** hoangcx has quit IRC | 15:01 | |
| *** gszasz has quit IRC | 15:02 | |
| *** matrohon has quit IRC | 15:03 | |
| *** sdake has joined #openstack-nova | 15:03 | |
| zzzeek | dansmith: I have the vaguest recollection of talking to you about that at some point | 15:03 |
| *** ducttape_ has quit IRC | 15:03 | |
| zzzeek | dansmith: and i think it hinged on, "does eventlet context switch for file IO?" which i know the answer is "no" | 15:04 |
| *** gongysh has joined #openstack-nova | 15:04 | |
| *** diogogmt has joined #openstack-nova | 15:04 | |
| dansmith | zzzeek: right, but log to syslog will because socket right? | 15:04 |
| dansmith | I remember the file io thing | 15:04 |
| *** hoangcx has joined #openstack-nova | 15:04 | |
| zzzeek | dansmith: if it is using the posix calls for the syslog then I might say no, because the "socket" part doesnt happen in python at all | 15:05 |
| *** gouthamr has quit IRC | 15:05 | |
| dansmith | ah | 15:05 |
| zzzeek | dansmith: but, this is in the category of neither file IO or socket IO then, it's "system call magic" | 15:06 |
| zzzeek | dansmith: but im just conjecturing. | 15:06 |
| zzzeek | dansmith: how do we do syslog logging, w/ normal PYthon logging module ? | 15:06 |
| dansmith | zzzeek: well, the follow-on here will make a call to rabbit to send a notification, so I need to bring that out anyway, but I'd like to know for future | 15:06 |
| dansmith | zzzeek: oslo.log, but it wraps logging I think | 15:06 |
| clarkb | python syslog is via socket and udp/tcp iirc | 15:06 |
| zzzeek | dansmith: ah OK syslog handler might do the TCP in python | 15:07 |
| clarkb | oslo logging at one point proposed a change to use syslog syscall which would block | 15:07 |
| zzzeek | dansmith: OK python sysloghandler is using socket() | 15:07 |
| *** sdake has quit IRC | 15:08 | |
| zzzeek | dansmith: so yes, logging w/ that handler will incur eventlet context switch. But i still don't see what that has to do with the DB transaction being interrupted. it's just two io streams | 15:08 |
| dansmith | clarkb: there's not actually a syscall for syslog() messages, right? you mean the libc syscall I assume? | 15:08 |
| clarkb | dansmith: sorry yes | 15:08 |
| *** andrey-mp has left #openstack-nova | 15:08 | |
| dansmith | zzzeek: if the other thread goes to do something that locks the same table.. that was the justification given for previous issues.. | 15:09 |
| zzzeek | dansmith: hmmmm | 15:09 |
| zzzeek | dansmith: but...another process could do that anyway. | 15:09 |
| *** READ10 has quit IRC | 15:10 | |
| dansmith | zzzeek: right, but the thing holding the lock isn't blocked on a thing that is blocked for the same lock | 15:10 |
| dansmith | zzzeek: I wasn't involved in the diagnosis or fixing, but the commandment came out to "not do that because of this" | 15:10 |
| zzzeek | dansmith: OK but we're on the pymysql driver which won't block on IO | 15:10 |
| zzzeek | dansmith: that was it, becasue ancient mysqldb driver would block | 15:10 |
| zzzeek | which breaks eventlet | 15:10 |
| *** annegentle has joined #openstack-nova | 15:10 | |
| dansmith | zzzeek: okay, I'll buy that.. but we can still be run with mysqldb if so configured right? what about postgres? | 15:11 |
| zzzeek | dansmith: in eventlet world, anything that blocks on IO needs to yield for things to do the right thing. | 15:11 |
| dansmith | yep | 15:11 |
| zzzeek | dansmith: the postgresql driver should get patched into the async version of it....der....need to see what eventlet does... | 15:12 |
| dansmith | I wonder if we should just log a warning if you use a backend like mysqldb and say "we know dragons lie here" | 15:12 |
| dansmith | or just outright refuse it | 15:12 |
| dansmith | clarkb: er, I meant "the libc call" above, heh :) | 15:13 |
| bauzas | alaski: I wonder if it shouldn't be better to raise an InstanceNotFound if we can't find the BuildRequest in https://review.openstack.org/#/c/365740/4/nova/compute/api.py | 15:13 |
| clarkb | I parsed :) | 15:13 |
| dansmith | clarkb: :) | 15:13 |
| clarkb | ehatever os.syslog ot whatever the method is maps to | 15:13 |
| bauzas | alaski: that way, in case users would see a 404, they would ask again | 15:13 |
| zzzeek | dansmith: i stil need to remember this. if using mysqldb, greenlet A starts mysqldb transaction that locks an object. then it yields due to syslog IO. then greenlet B tries to lock object. this blocks waiting for lock. w/ pymysql, lock wait would yield back to A so work continues, OK. right this is greenlet BS :) | 15:13 |
| bauzas | alaski: that would mean we wouldn't need to work on a race condition | 15:14 |
| *** mdrabe has quit IRC | 15:14 | |
| zzzeek | dansmith: but! | 15:14 |
| zzzeek | dansmith: but but! i think eventlet even throws mysqldb operations into a thread pool ! doesn't it ? | 15:14 |
| *** mdrabe has joined #openstack-nova | 15:14 | |
| bauzas | alaski: and it would help the code to be easier | 15:14 |
| *** hieulq_ has joined #openstack-nova | 15:14 | |
| *** ljxiash has quit IRC | 15:14 | |
| dansmith | zzzeek: so, apparently thread pooling with mysqldb breaks eventlet too in some other way, so we always recommended it be disabled.. it was way faster, but people like rax could run for a total of 7 minutes before it broke or something | 15:15 |
| dansmith | alaski: you remember that discussion with comstud? | 15:15 |
| zzzeek | OHHHh | 15:15 |
| zzzeek | well | 15:15 |
| *** gszasz has joined #openstack-nova | 15:15 | |
| zzzeek | yeah you're flipping switches until this breaks, pretty much :) | 15:15 |
| *** diogogmt has quit IRC | 15:15 | |
| dansmith | heh | 15:15 |
| zzzeek | dansmith: well yes, the mysqldb threadpool thing performs terribly if you look how it works | 15:16 |
| zzzeek | dansmith: because every dinky little io call is like, put it in the queue, etc. | 15:16 |
| dansmith | zzzeek: well, if you're sure pymysql is good here, that's fine with me, but we might want to think about some warnings if people configure for the other ones, especially if we're opening the gates on things we know will break them | 15:16 |
| cfriesen | sfinucan: got a sec? I'm seeing an interesting HT packing issue. | 15:16 |
| sfinucan | cfriesen: hit me | 15:16 |
| zzzeek | dansmith: if you're using syslog logging handler in python w/ eventlet, you definitely need eventlet-compliant DB workings, which means either pymysql or psycopg2-green or if non-green driver then it's in a thread pool | 15:17 |
| *** artom has quit IRC | 15:17 | |
| dansmith | zzzeek: yeah | 15:17 |
| zzzeek | dansmith: shoutout to "hey let's stop using eventlet at this level :) " | 15:17 |
| * dansmith sprays the protesters with the firehose | 15:17 | |
| dansmith | mriedem: you following this? might be good to either log a warning or at least put a reno in place to warn aboutthis | 15:18 |
| cfriesen | one of our testers tried a bunch of 2-numa-node instances with one cpu per node, with the 'prefer' policy. It grabbed one sibling from a new core each time, leaving the remaining sibling. this of course makes it harder to schedule 'require' instances since there are no empty cores left. | 15:18 |
| zzzeek | dansmith: I actually *like* eventlet / gevent, but the thing im doing w/ them, I dont do monkeypatching. im using the APIs explicitly,and things with DBs are in threads | 15:18 |
| cfriesen | it seems like "prefer" should try to use siblings on cores that are already partially-allocated, preferring to leave unused cores for 'require' instances. | 15:19 |
| *** diogogmt has joined #openstack-nova | 15:19 | |
| sfinucan | cfriesen: hmm, that depends :) | 15:20 |
| sfinucan | by definition, it will prefer free cores | 15:20 |
| *** sahid has quit IRC | 15:20 | |
| sfinucan | but I see the advantages of modifying it to work as you say | 15:20 |
| sfinucan | *fully free | 15:20 |
| *** ociuhandu has joined #openstack-nova | 15:21 | |
| mriedem | dansmith: nope, been debugging some internal stuff all morning | 15:21 |
| cfriesen | sfinucan: according to http://docs.openstack.org/admin-guide/compute-flavors.html "prefer" simply means that thread siblings are preferred. it makes no statement about interactions between separate instances. | 15:22 |
| *** _ducttape_ has quit IRC | 15:22 | |
| *** roeyc has quit IRC | 15:22 | |
| *** ducttape_ has joined #openstack-nova | 15:22 | |
| mriedem | dansmith: warn if you're not using pymysql or psycopg2-green? | 15:22 |
| mriedem | dansmith: is that up to nova, or oslo.db? | 15:23 |
| cfriesen | definitely, with multiple guests vcpus on a single guest numa node we would prefer to fill up host cores first...but any vcpus left over after filling up host cores could be preferentially placed on partially-allocated cores I think. | 15:23 |
| *** praneshp has joined #openstack-nova | 15:23 | |
| *** sahid has joined #openstack-nova | 15:23 | |
| sfinucan | cfriesen: aye, and if you're booting a four core instance and there are two free cores, it will take that | 15:23 |
| sfinucan | *those | 15:23 |
| *** kevinz has joined #openstack-nova | 15:23 | |
| sfinucan | thus preferring the siblings | 15:23 |
| *** beagles is now known as beagles_brb | 15:24 | |
| dansmith | mriedem: we could warn in oslo.db, but the offenses that break those would be in nova, so I'd warn in nova | 15:24 |
| *** lucas-afk is now known as lucasagomes | 15:24 | |
| *** sdake has joined #openstack-nova | 15:24 | |
| *** syjulian_ has quit IRC | 15:25 | |
| *** andreas_s has quit IRC | 15:25 | |
| mriedem | ok, i haven't been up to date on what's going on, so don't know what's new that will break those | 15:25 |
| sfinucan | cfriesen: but, ignoring semantics, packing would definitely make sense for both other cases though | 15:26 |
| cfriesen | sfinucan: okay....we'll open a bug so it doesn't get forgotten. might even propose a fix depending on how it gets prioritized. | 15:26 |
| dansmith | mriedem: nothing particularly new, but we now wrap db_api calls that do logging with the facade stuff, which could end up getting us into trouble.. nothing really new in newton necessarily | 15:26 |
| sfinucan | cfriesen: Sure. I could fix that now, but I don't know if it's a bug or a feature request (i.e. specless BP needed) | 15:26 |
| dansmith | mriedem: I was just about to add a log call deep in the db code, and started thinking about this and re-opened the conversation | 15:26 |
| dansmith | mriedem: we're not testing with mysqldb anyway AFAIK, so the warning is also just "good luck..." | 15:27 |
| *** EinstCrazy has joined #openstack-nova | 15:27 | |
| *** burgerk has joined #openstack-nova | 15:27 | |
| cfriesen | sfinucan: you're thinking maybe feature request due to cross-instance interference? | 15:27 |
| *** armax has joined #openstack-nova | 15:27 | |
| cfriesen | sfinucan: cause there's no other end-user-visible change other than you can better mix prefer with isolate/require | 15:28 |
| sfinucan | cfriesen: exactly. As things stand, booting two four core instances on an eight-core host will result in optimal performance for both guests | 15:28 |
| sfinucan | actually, booting two three core instances might be a better example | 15:28 |
| cfriesen | sfinucan: until you try to boot a "require" instance and it fails, even though you've got spare cpus | 15:29 |
| mriedem | dansmith: right we're not using mysqldb since quite awhile, because oslo.db needed to get to py3 compliance i think | 15:29 |
| mriedem | which was pymysql | 15:29 |
| *** sahid has quit IRC | 15:29 | |
| sfinucan | cfriesen: aye, but it really does depend on your use case | 15:29 |
| dansmith | mriedem: but we can still configure it AFAIK, right? | 15:29 |
| *** nkrinner is now known as nkrinner_afk | 15:29 | |
| cfriesen | sfinucan: agreed. pick the tradeoff. :) | 15:30 |
| sfinucan | all I'm saying is this is going to affect people, and it's not "broken" per se | 15:30 |
| *** kevinz has quit IRC | 15:30 | |
| cfriesen | fair enough | 15:30 |
| sfinucan | ...so it's probably a feature request like "make the require and isolate thread policies easier to schedule successfully" :) | 15:31 |
| *** kevinz has joined #openstack-nova | 15:31 | |
| mriedem | dansmith: yeah, i think so, rax uses mysqldb | 15:31 |
| dansmith | right | 15:31 |
| cfriesen | are you thinking a tuning knob? (tweak for performance vs packing) Or a "blessing" from the core team as to which way to go? | 15:31 |
| *** sahid has joined #openstack-nova | 15:31 | |
| *** stevemar_ has joined #openstack-nova | 15:33 | |
| sfinucan | cfriesen: the latter | 15:34 |
| *** stevemar_ has quit IRC | 15:34 | |
| *** baoli_ has quit IRC | 15:34 | |
| sfinucan | we've already got too many tuning knobs :) | 15:34 |
| *** rama_y has joined #openstack-nova | 15:35 | |
| cfriesen | sfinucan: I suppose an argument could be made that if they really care about performance they should use "isolate" since "prefer" can already result in sharing cores...this would just make it more likely. | 15:35 |
| stevemar | mriedem: o/ | 15:35 |
| *** krtaylor has quit IRC | 15:35 | |
| stevemar | mriedem: is there a reason none of the contrib managers are defined in client.py https://github.com/openstack/python-novaclient/blob/master/novaclient/v2/client.py ? | 15:36 |
| stevemar | mriedem: i wanted to implement action-list in OSC, but theres no way for me to call the manager https://github.com/openstack/python-novaclient/blob/master/novaclient/v2/contrib/instance_action.py | 15:37 |
| *** unicell1 has joined #openstack-nova | 15:38 | |
| *** david-lyle_ has joined #openstack-nova | 15:38 | |
| *** david-lyle_ has quit IRC | 15:38 | |
| *** hoonetorg has quit IRC | 15:38 | |
| *** Sukhdev has joined #openstack-nova | 15:38 | |
| *** su_zhang has joined #openstack-nova | 15:39 | |
| sfinucan | cfriesen: yup, but it still makes for (potentially) worse performance for these people | 15:39 |
| sfinucan | who did nothing wrong and don't deserve this :P | 15:39 |
| *** unicell has quit IRC | 15:40 | |
| sfinucan | but that's bikeshedding. Definitely file the bug and assign me. I'll take a look at it this evening | 15:40 |
| sfinucan | cfriesen: It would probably involve a change beyond https://review.openstack.org/#/c/342709 (which still needs core review :'() | 15:40 |
| *** paul-carlton has quit IRC | 15:41 | |
| *** salv-orlando has joined #openstack-nova | 15:41 | |
| *** moshele has joined #openstack-nova | 15:41 | |
| *** mingyu has joined #openstack-nova | 15:41 | |
| * sfinucan better bike home now before it gets too dark | 15:42 | |
| *** lpetrut has quit IRC | 15:42 | |
| *** hoonetorg has joined #openstack-nova | 15:42 | |
| *** syjulian_ has joined #openstack-nova | 15:43 | |
| *** praneshp has quit IRC | 15:43 | |
| mriedem | stevemar: not sure, probably models the old v2 api behavior where extensions were put into a contrib dir | 15:44 |
| openstackgerrit | Jianghua Wang proposed openstack/nova: Failed to create image from volume backed instance https://review.openstack.org/366825 | 15:45 |
| *** gongysh has quit IRC | 15:45 | |
| stevemar | mriedem: if i toss up a patch to expose them, would it get -2'ed? :) | 15:45 |
| mriedem | not necessarily | 15:45 |
| openstackgerrit | Dan Smith proposed openstack/nova: More conservative inventory updates https://review.openstack.org/365073 | 15:47 |
| *** salv-orlando has quit IRC | 15:48 | |
| *** paul-carlton2 has quit IRC | 15:50 | |
| *** ociuhandu has quit IRC | 15:52 | |
| PaulMurray | sfinucan, cfriesen just reading back over your discussion about core packing..... it looks exactly the same as spreading vs stacking in host resource usage. There are operators that prefer to do one or the other | 15:52 |
| PaulMurray | so both options are interesting | 15:52 |
| *** artom has joined #openstack-nova | 15:54 | |
| *** artom has quit IRC | 15:55 | |
| *** annegentle has quit IRC | 15:55 | |
| *** artom has joined #openstack-nova | 15:55 | |
| *** moshele has quit IRC | 15:55 | |
| sdague | mriedem / dansmith / jaypipes / bauzas - the missing unit tests for the safe_connect decorator - https://review.openstack.org/#/c/366800/ | 15:56 |
| sdague | should be quick | 15:56 |
| *** salv-orlando has joined #openstack-nova | 15:57 | |
| mriedem | stevemar: so instance actions seems like it should be in the main client, | 15:58 |
| mriedem | stevemar: there are other things in contrib that shouldn't, like the host-evacuate things, those aren't resource managers, they are helper scripts | 15:59 |
| jaypipes | sdague: on it. | 15:59 |
| mriedem | stevemar: we could move instance_actions up a level and provide an alias to the v2/contrib | 15:59 |
| *** annegentle has joined #openstack-nova | 15:59 | |
| mriedem | stevemar: why can't you just get the instance actions manager from the contrib directory though? | 15:59 |
| mriedem | because you only have the Client object? | 16:00 |
| stevemar | mriedem: right, we only have access to the client | 16:00 |
| mriedem | honestly i'm not even sure how those contrib entries are loaded | 16:00 |
| stevemar | mriedem: https://github.com/openstack/python-openstackclient/blob/master/openstackclient/compute/client.py#L69-L79 | 16:00 |
| stevemar | mriedem: i'm trying to figure that out now | 16:01 |
| mriedem | the discover_extensions method it looks like | 16:01 |
| *** pcaruana has quit IRC | 16:01 | |
| sfinucan | PaulMurray: but how do we do it without YAK (yet-another-knob) | 16:01 |
| *** su_zhang has quit IRC | 16:02 | |
| cfriesen | PaulMurray: agreed, both are interesting. the downside of spreading is that it leaves the host in a state where only other 'prefer' instances can be scheduled on it. 'isolate' and 'require' instances are out of luck. | 16:02 |
| *** cdent has joined #openstack-nova | 16:02 | |
| stevemar | mriedem: yeah "_discover_via_contrib_path: | 16:02 |
| *** su_zhang has joined #openstack-nova | 16:02 | |
| stevemar | mriedem: well that is icky | 16:02 |
| cdent | dansmith, jaypipes: Where we at? Something I should do? | 16:03 |
| cdent | sdague: thanks for writing that description of allocations thing to mailing list: writingstuffdown++ | 16:04 |
| *** hieulq_ has quit IRC | 16:04 | |
| jaypipes | cdent: just wrapped up some meetings. still reviewing... | 16:04 |
| cdent | jaypipes: ✔ | 16:04 |
| *** EinstCrazy has quit IRC | 16:05 | |
| *** beagles_brb is now known as beagles | 16:05 | |
| *** mgoddard_ has joined #openstack-nova | 16:05 | |
| *** abalutoiu_ has quit IRC | 16:05 | |
| mriedem | stevemar: looks like the shell discovers the extensions after creating the client, but you'd have to do something similar i guess | 16:06 |
| mriedem | or we just move the instance actions out of contrib and alias it | 16:06 |
| mriedem | seems like that should be in the mainline path anyway, i'm not sure why it's considered an 'extension' | 16:06 |
| *** su_zhang_ has joined #openstack-nova | 16:06 | |
| stevemar | mriedem: yeah, probably just do that | 16:06 |
| mriedem | again, v2 wonkiness | 16:06 |
| cdent | edleafe: on [t 1V65], yeah, I think you're right. That's either jay-code or me cargoing jay code | 16:06 |
| purplerbot | <edleafe> cdent: suggestion: use set1.intersection(set2) instead of set1 & set2 - it's much less confusing to readers [2016-09-07 14:42:15] [n 1V65] | 16:06 |
| stevemar | i know all about version api wonkiness | 16:06 |
| mriedem | sdague wants to kill extensions and at some point we need to do that in novaclient too probably | 16:06 |
| stevemar | mriedem: yeah, we did that in keystone 2 cycles ago | 16:06 |
| *** su_zhang has quit IRC | 16:07 | |
| sdague | stevemar: there is just a lot of cruft that needs cleaning out | 16:07 |
| *** rama_y has quit IRC | 16:07 | |
| sfinucan | cfriesen: looks like a definite spec, IMO | 16:07 |
| sfinucan | the best way to get both sides of the argument | 16:07 |
| stevemar | sdague: worms in a garden and all that | 16:07 |
| sdague | stevemar: so, I would honestly suggest we move stuff you care about out of contrib | 16:07 |
| cdent | edleafe: I can look into adjusting some of that stuff when I know if dansmith is holding any code out | 16:07 |
| *** rama_y has joined #openstack-nova | 16:07 | |
| *** EinstCrazy has joined #openstack-nova | 16:07 | |
| sdague | instead of trying to figure out loading all the stuff there | 16:07 |
| *** su_zhang_ has quit IRC | 16:07 | |
| dansmith | cdent: I have nothing unsubmitted right now | 16:08 |
| *** su_zhang has joined #openstack-nova | 16:08 | |
| sfinucan | cfriesen: a la https://review.openstack.org/#/c/364468/ | 16:08 |
| *** su_zhang has quit IRC | 16:08 | |
| dansmith | cdent: I just pushed an update to the conservative inventory updates patch | 16:08 |
| dansmith | cdent: and was going to start on doing the same for allocations on top of your GET patch when it looks like that is mostly set | 16:08 |
| edleafe | cdent: yeah, it's not the end of the world or anything | 16:08 |
| *** mgoddard has quit IRC | 16:08 | |
| *** su_zhang has joined #openstack-nova | 16:08 | |
| cdent | dansmith: have you seen/investigated https://review.openstack.org/#/c/365633/5 | 16:09 |
| cdent | that's the fix of the inventory serialization | 16:09 |
| dansmith | cdent: I also put the current list of things I think we need to land for newton in the etherpad | 16:09 |
| *** paul-carlton2 has joined #openstack-nova | 16:09 | |
| *** su_zhang has quit IRC | 16:09 | |
| dansmith | cdent: I think sdague wants release-blocking bugs for each | 16:09 |
| *** tesseract- has quit IRC | 16:09 | |
| sdague | dansmith: or at least all patches we think we need to land associated to a bug | 16:09 |
| *** su_zhang has joined #openstack-nova | 16:09 | |
| * cdent looks at etherpad | 16:09 | |
| dansmith | sdague: yeah, two of the patches in that list are actually for one bug | 16:09 |
| sdague | mostly, I'm not sure where our FFE stops | 16:10 |
| dansmith | cdent: I will let the api peeps decide on that serialization thing | 16:10 |
| sdague | and it helps in priortizing | 16:10 |
| dansmith | sdague: these are all bugs in code that is landed now | 16:10 |
| dansmith | sdague: so not FFE related | 16:10 |
| *** akshai has joined #openstack-nova | 16:10 | |
| *** su_zhang has quit IRC | 16:11 | |
| cdent | dansmith: the thing I just linked (https://review.openstack.org/#/c/365633/ ) is the change that makes it so the inventory writing code can be less complex (by including the resource provider in an accessible) and includes the de-complexifying of the resource tracker | 16:11 |
| *** su_zhang has joined #openstack-nova | 16:11 | |
| cdent | the allocation serialization question is in https://review.openstack.org/#/c/366789/ | 16:12 |
| dansmith | cdent: yeah I know | 16:12 |
| cdent | (the de-complexifying may be wrong now, as a result of the changes you've recently made) | 16:12 |
| cdent | dansmith: ah, okay, it wasn't clear from your response if we were talking of the same stuff | 16:12 |
| cdent | do you think we need that change or only want it? | 16:13 |
| mriedem | sdague: +2 | 16:13 |
| mriedem | thanks | 16:13 |
| dansmith | cdent: well, you're changing the API, so I think if we're going to do it, we better do it soon, yeah | 16:14 |
| cdent | wow, that was magic | 16:14 |
| cfriesen | sfinucan: okay, can go that route | 16:14 |
| cdent | I went to etherpad tab and it started typing what I was going to type. amazeballs | 16:15 |
| alaski | dansmith: yeah, I remember the thread pooling/eventlet thing from way back. comstud had a fix proposed to eventlet which I don't think they ever merged. rax runs with a patched eventlet to avoid it | 16:15 |
| *** wznoinsk has joined #openstack-nova | 16:15 | |
| *** ajo_ has quit IRC | 16:15 | |
| *** su_zhang has quit IRC | 16:16 | |
| dansmith | alaski: right | 16:16 |
| *** ajo_ has joined #openstack-nova | 16:16 | |
| *** EinstCrazy has quit IRC | 16:16 | |
| dansmith | cdent: we should get anything else we have on the etherpad so we can discuss what is critical and not | 16:17 |
| dansmith | what is there is pretty much what is in my head at the moment | 16:17 |
| cdent | dansmith: I'll add what I know | 16:18 |
| *** raunak has joined #openstack-nova | 16:18 | |
| *** gszasz has quit IRC | 16:18 | |
| *** raunak has quit IRC | 16:19 | |
| *** code-R_ has quit IRC | 16:19 | |
| *** syjulian_ has quit IRC | 16:20 | |
| alaski | bauzas: I think you're misunderstanding the race | 16:20 |
| *** raunak has joined #openstack-nova | 16:20 | |
| alaski | bauzas: and we really shouldn't return a 404 if at all possible | 16:20 |
| alaski | if the buildrequest is deleted things are fine, we know there's an instance in a cell at that point that can be updated | 16:20 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: WIP: Update BuildRequest if instance currently being scheduled https://review.openstack.org/365740 | 16:20 |
| *** mgoddard has joined #openstack-nova | 16:21 | |
| alaski | the issue is that the buildrequest update may not be picked up by the process working on the instance boot. unless it looks up the buildrequest after the update it will not be aware that something changed | 16:21 |
| *** akshai_ has joined #openstack-nova | 16:21 | |
| *** mgoddard_ has quit IRC | 16:21 | |
| *** akshai has quit IRC | 16:21 | |
| bauzas | alaski: right, I see | 16:21 |
| bauzas | alaski: so the instance boots on proc A, then goes to conductor which destroys the buildrequest | 16:22 |
| *** syjulian_ has joined #openstack-nova | 16:22 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Additional logging for placement API https://review.openstack.org/366186 | 16:22 |
| bauzas | alaski: in the meantime, another API worker could lookup the buildrequest, that's what you want to say? | 16:23 |
| openstackgerrit | Tom Patzig proposed openstack/nova: refresh instances_path in shared stor testfile chk https://review.openstack.org/366857 | 16:23 |
| *** thorst has quit IRC | 16:24 | |
| alaski | bauzas: yeah, and remember that the order in conductor will be create instance in cell, map instance to cell, remove buildrequest. so not finding the buildrequest is good, that means the instance is either in a cell or deleted | 16:24 |
| *** kevinz has quit IRC | 16:24 | |
| *** thorst has joined #openstack-nova | 16:24 | |
| bauzas | alaski: right, I wrote that in my last PS I just uploaded | 16:24 |
| *** raunak has quit IRC | 16:25 | |
| bauzas | alaski: the real problem I saw when trying to verify that (whether the instance is still there or not) means that I need to lookup the instance with all the needed attributes | 16:25 |
| *** mingyu has quit IRC | 16:25 | |
| bauzas | alaski: so, again, trying to look at the instance mapping, and if not, failback to the simple DB call | 16:25 |
| alaski | bauzas: my concern is this: compute/api creates a buildrequest for a future instance, it sends that to conductor which picks a cell and creates an instance based on the buildrequest, but the buildrequest was never refreshed so any updates will be missed | 16:25 |
| *** krtaylor has joined #openstack-nova | 16:26 | |
| alaski | bauzas: you mean the expected_attrs? | 16:27 |
| alaski | let me looks at the latest PS. I think we're talking about two different things | 16:27 |
| *** raunak has joined #openstack-nova | 16:28 | |
| bauzas | alaski: sorry, it seems I'm very tired by that FF :/ | 16:28 |
| *** baoli has joined #openstack-nova | 16:29 | |
| bauzas | alaski: so, trying to get your concern, I don't see a problem if we always verify first if the instance has a cell mapping | 16:29 |
| bauzas | because it's litterally the action made right after creating the instance record in the cell, and before deleting the buildrequest | 16:30 |
| alaski | yeah, there's no problem doing that | 16:30 |
| alaski | everything above line 2424 looks fine | 16:31 |
| bauzas | the race I can see is when _get_server() returned a nested buildrequest.instance because at that time, the instance record in the cell wasn't done | 16:31 |
| bauzas | so, we could say "eh look, this instance doesn't have an id, so it's not yet created in the child cell" | 16:31 |
| bauzas | so, L2424 and below are for that case | 16:32 |
| cdent | dansmith: is this still relevant? https://review.openstack.org/#/c/364547/ | 16:32 |
| bauzas | where we verify whether there is an instance mapping *now* | 16:32 |
| *** thorst has quit IRC | 16:32 | |
| alaski | bauzas: down to line 2437 things look good to me. it's below that where we're not on the same page | 16:32 |
| dansmith | cdent: yeah | 16:33 |
| *** nic has joined #openstack-nova | 16:33 | |
| dansmith | cdent: iddn | 16:33 |
| bauzas | alaski: I see, you disagree with my comment | 16:33 |
| dansmith | cdent: didn't realize that didn't merge | 16:33 |
| *** jaosorior has quit IRC | 16:33 | |
| *** gus has quit IRC | 16:33 | |
| bauzas | alaski: okay, then I'll update that, and do the instance mapping (and instance record) lookup | 16:33 |
| alaski | bauzas: right. I think doing the lookup is better. if instance is found then update it, if not then it was deleted so raise that | 16:34 |
| *** ayoung has quit IRC | 16:35 | |
| alaski | bauzas: but there's still another race, which I mentioned above. This update won't be picked up by the boot process if buildrequest is updated | 16:35 |
| bauzas | alaski: but I need to keep the else statement like in L2418 because we agreed to not have the cells migration done pre-Newton | 16:35 |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: block_device: fix propagation of instance AZ to Cinder https://review.openstack.org/366724 | 16:35 |
| *** yamahata has quit IRC | 16:35 | |
| alaski | bauzas: yep | 16:35 |
| *** ayoung has joined #openstack-nova | 16:36 | |
| bauzas | alaski: oh, because we pass down the BuildReq to the conductor without looking it up again ? | 16:36 |
| alaski | bauzas: yep | 16:36 |
| bauzas | if so, I think I understand your concern, lemme look at the code | 16:37 |
| bauzas | I thought we were fetching the buildrequest record, not passing it thru RPC | 16:37 |
| alaski | cool. I think we should have a second patch for fixing that, just wanted to raise the issue. | 16:38 |
| alaski | well, https://review.openstack.org/#/c/319379/ was planning to pass via RPC | 16:38 |
| alaski | which I think is the right thing to do | 16:38 |
| dansmith | cdent: your get patch is RP-based, not consumer-based | 16:39 |
| dansmith | cdent: which means I have to pull the entire list of allocations for every consumer of my resources in order to check to see if my consumer is good | 16:39 |
| dansmith | cdent: which also means I can't generically compare allocations for all resource providers that have allocations for a given consumer | 16:40 |
| dansmith | cdent: I was expecting a GET /allocations/$consumer_uuid | 16:40 |
| bauzas | alaski: well, I now fully understand the race you mention | 16:40 |
| bauzas | alaski: I think I agree with you about passing it down thru RPC, I just wonder how we could prevent that race but having a synchronized section | 16:41 |
| *** burt has joined #openstack-nova | 16:41 | |
| bauzas | (and where...) | 16:41 |
| *** gus has joined #openstack-nova | 16:42 | |
| cdent | dansmith: I asked jaypipes what he wanted, that's what he said he wanted. I think you/we will need to consult with him on what he has in mind | 16:42 |
| cdent | dansmith: but a) we can change it pretty easily b) we can have both pretty easily | 16:42 |
| mriedem | alaski: check it out https://bugs.launchpad.net/nova/+bug/1621138 | 16:43 |
| openstack | Launchpad bug 1621138 in OpenStack Compute (nova) "block_device_mappings column is not large enough in table build_requests" [Undecided,New] - Assigned to Kenneth Burger (burgerk) | 16:43 |
| dansmith | jaypipes: you're not okay with GET /allocations/$uuid ? | 16:43 |
| mriedem | Test is apparently not big enough | 16:43 |
| mriedem | burgerk: how many volumes are you booting the instance with? | 16:43 |
| *** akshai_ has quit IRC | 16:43 | |
| edleafe | cdent: maybe GET /allocations/$rp_id/$consumer_id as a refinement? | 16:43 |
| bauzas | mriedem: alaski: FWIW, I also saw that one https://bugs.launchpad.net/nova/+bug/1620684 but I struggled finding time to look further | 16:44 |
| openstack | Launchpad bug 1620684 in OpenStack Compute (nova) "nova list --status soft_deleted is not showing soft deleted Instances" [Undecided,In progress] - Assigned to Anusha Unnam (anusha-unnam) | 16:44 |
| cdent | edleafe: that would violate the current grammar | 16:44 |
| burgerk | mriedem: broke at ~ 80 volumes, want to go up to 128 | 16:44 |
| cdent | so we'd either need to change it a lot, or not do that | 16:44 |
| alaski | bauzas: a distributed lock would solve this :) But I think what we'll need to do is lookup the buildrequest after we create the instance and then move any updates to the instance | 16:44 |
| bauzas | mriedem: half-related, we now have 14 rc-potential bugs | 16:44 |
| alaski | bauzas: sdague nacked that one | 16:44 |
| mriedem | burgerk: jesus, 128 volumes for an instance? | 16:44 |
| sdague | alaski: was it needed for something? | 16:45 |
| sdague | it seemed like a random API change | 16:45 |
| bauzas | alaski: mmm, I thought the reporter was mentioning some regression on the API qide | 16:45 |
| alaski | sdague: not really, just mentioning it because bauzas brought up the bug | 16:45 |
| *** lucasagomes is now known as lucas-dinner | 16:45 | |
| bauzas | sdague: context is https://bugs.launchpad.net/nova/+bug/1620684 | 16:46 |
| openstack | Launchpad bug 1620684 in OpenStack Compute (nova) "nova list --status soft_deleted is not showing soft deleted Instances" [Undecided,In progress] - Assigned to Anusha Unnam (anusha-unnam) | 16:46 |
| bauzas | I just made a brief pass on our open bugs, and saw that one | 16:46 |
| bauzas | so I actually wonder if that's a regression or not | 16:46 |
| *** diogogmt has quit IRC | 16:46 | |
| bauzas | sdague: but if you say that's a new feature, then... :) | 16:46 |
| alaski | bauzas: I don't believe that's a regression | 16:46 |
| mriedem | burgerk: will MEDIUMTEXT work for you? | 16:47 |
| alaski | bauzas: or if it is it's a regression from adding a soft-delete state | 16:47 |
| alaski | mriedem: looking | 16:47 |
| bauzas | alaski: well, the bug description left me think that | 16:47 |
| burgerk | mreidem: yes, that was going to be my proposal | 16:47 |
| edleafe | cdent: how would you represent a filter on both in the current grammar? Or is that not possible. | 16:47 |
| bauzas | alaski: the reporter was saying he was getting an empty list when asking for soft-deleted | 16:47 |
| bauzas | while the instance was still on soft-delete (of cours)e | 16:48 |
| bauzas | but I should reproduce | 16:48 |
| bauzas | that's #2 task I have :) | 16:48 |
| sdague | bauzas: there is no indication that it's a regression in the bug | 16:48 |
| bauzas | sdague: I agree | 16:48 |
| mriedem | heh, unrelated, why is hypervisor_type mediumtext? | 16:48 |
| mriedem | hypervisor_type = Column(MediumText(), nullable=False) | 16:48 |
| alaski | bauzas: buildrequest lookups won't affect soft-delete because only scheduling instances are buildrequests. I don't think it's related to cells changes | 16:48 |
| *** rama_y has quit IRC | 16:48 | |
| bauzas | hence me needing to test whether we had that too in Mitaka | 16:49 |
| cdent | edleafe: the grammar issue is that, as currently written, good grammar is /allocations/{consumer_id} and /resource_providers/{rp_uuid} so we shouldn't mix those. filters, one could do lots of things but it depends on the use case. I'm not really sure what you're asking? | 16:49 |
| bauzas | alaski: okay, tbh, it wasn't an hard call, I just put the rc potential tag to make sure I wasn't forgetting it | 16:49 |
| *** Apoorva has joined #openstack-nova | 16:49 | |
| bauzas | either way, like I said, I need to test on Mitaka and see whether it was a regression or not | 16:49 |
| bauzas | if not, then I'll remove the rc tag | 16:50 |
| bauzas | and if that's an API change, I'll even close the bug | 16:50 |
| mriedem | burgerk: ok ping me when you have a change up | 16:51 |
| *** unicell1 has quit IRC | 16:51 | |
| edleafe | cdent: it's more efficient to only return the data that is relevant. If you want the allocations for a particular RP and consumer, shouldn't that be possible> | 16:51 |
| edleafe | ? | 16:51 |
| dansmith | mriedem: a change for what? altering the column definition in BDMs? | 16:51 |
| melwitt | that soft-deleted bug isn't a regression and there's been a patch up for a long time that would enable --status soft_deleted to work https://review.openstack.org/#/c/258472/ but afaik it's never worked | 16:52 |
| mriedem | dansmith: making build_requsts.block_device_mappings MEDIUMTEXT rather than TEXT | 16:52 |
| bauzas | melwitt: oh, thanks for helping ! | 16:52 |
| mriedem | since it's a serialized blob of all bdms in the build request | 16:52 |
| bauzas | melwitt: so, that's just a looong story bug | 16:52 |
| melwitt | bauzas: yeah, unfortunately | 16:52 |
| dansmith | mriedem: okay, technically that'd be a migration we'd want to not do, but it's a small table so maybe it's not so bad | 16:52 |
| bauzas | melwitt: looking for duplicates then | 16:53 |
| dansmith | mriedem: i.e. a table rewrite | 16:53 |
| mriedem | dansmith: if we're going to do it we should probably do it in newton | 16:53 |
| mriedem | although build_requests was around since mitaka right? | 16:53 |
| dansmith | mriedem: yeah, but it's also just things that are waiting for scheduling anyway right? | 16:53 |
| mriedem | and those are supposed to be short-lived resources | 16:53 |
| mriedem | and hard-deleted | 16:54 |
| dansmith | yep | 16:54 |
| alaski | okay, just caught up | 16:54 |
| *** syjulian_ has quit IRC | 16:54 | |
| alaski | yeah, the table should have almost nothing in it at any given time | 16:54 |
| *** GB21 has joined #openstack-nova | 16:54 | |
| mriedem | just looking at where we use mediumtext on other things, it's clob-y type things like this too | 16:54 |
| mriedem | i think i asked about this when that column was added actually, if it'd be big enough | 16:54 |
| * dansmith looks up clob-y | 16:54 | |
| *** lpetrut has joined #openstack-nova | 16:54 | |
| mriedem | clobish | 16:55 |
| bauzas | melwitt: well https://bugs.launchpad.net/nova/+bug/1526715 is technically not a duplicate of https://bugs.launchpad.net/nova/+bug/1620684 but that's enough for me to close the bug and ask to create a spec because it will need a microversion as sdague stated correctly | 16:55 |
| openstack | Launchpad bug 1526715 in OpenStack Compute (nova) "disallow non-admin search for soft-delete instance" [Low,In progress] - Assigned to jichenjc (jichenjc) | 16:55 |
| openstack | Launchpad bug 1620684 in OpenStack Compute (nova) "nova list --status soft_deleted is not showing soft deleted Instances" [Undecided,In progress] - Assigned to Anusha Unnam (anusha-unnam) | 16:55 |
| * dansmith look sup clobish | 16:55 | |
| bauzas | melwitt: thanks: | 16:55 |
| alaski | mriedem: 80 volumes should be enough for anyone :) | 16:55 |
| dansmith | alaski: seriously | 16:55 |
| sdague | bauzas: works for me | 16:55 |
| *** Swami_ has quit IRC | 16:55 | |
| melwitt | bauzas: hm, I guess I didn't think it would need a microversion | 16:55 |
| mriedem | btw, it's hilarious that compute_nodes.hypervisor_type is a mediumtext column | 16:56 |
| cdent | edleafe: if what I wanted was all the allocations by consumer_id against just one resource provider I'd probably leave it as it: get the allocaitons against all the resource providers cuz it's small. Where efficiency might come into place is when getting all or lots of the allocations against a particular resource provider (GET /resource_providers/{uuid}/allocations) | 16:56 |
| cdent | edleafe: in case I'd think some kind of filtering would be useful, ad query params | 16:57 |
| melwitt | bauzas: the behavior is caused by inconsistent behavior between 'deleted' and 'soft-deleted' and the fix just makes it consistent. so I thought it would just be a bug fix | 16:57 |
| *** zenoway has joined #openstack-nova | 16:57 | |
| cdent | I do not, however, thing that the representation should change (such that only some fields are displayed in the output) | 16:57 |
| bauzas | melwitt: not really :/ | 16:57 |
| cdent | that would be an early optimization | 16:57 |
| bauzas | melwitt: because the experience would change | 16:57 |
| bauzas | consistency is certainly a good concern, but changing what the API returns based on some arguments needs a microversion I guess | 16:58 |
| cdent | edleafe: if we eventually discovered that the movement of a small number of bytes around was having real drag on the system, we could explore what to do, but since at the moment we've not even worried about tcp connection overhead, I don't think it is something worth worrying about | 16:58 |
| melwitt | bauzas: I can see how it would be debatable | 16:58 |
| bauzas | :) | 16:58 |
| sdague | regardless, I don't think it's a post FF bug to address | 16:58 |
| alaski | agreed. though I also agree with melwitt that it's more a bug fix than feature | 16:59 |
| melwitt | yeah, we can sort that out later whether it's spec or not by a quick nova meeting discussion or something | 16:59 |
| *** annegentle has quit IRC | 16:59 | |
| melwitt | er, microversion or not | 16:59 |
| *** anush has quit IRC | 17:00 | |
| *** Sukhdev has quit IRC | 17:01 | |
| bauzas | melwitt: sdague: I marked it as opinion to leave the door open | 17:01 |
| bauzas | either way, it's certainly not something we want to merge for Newton | 17:01 |
| *** zenoway has quit IRC | 17:02 | |
| melwitt | yeah, definitely not | 17:02 |
| mriedem | if it worked in mitaka, then it's a bug fix imo | 17:02 |
| * bauzas goes afk for dinner | 17:02 | |
| * mriedem sets up mitaka devstack to recreate | 17:03 | |
| sdague | mriedem: https://github.com/openstack/nova/commit/09f2c4729456443c4874a8cadc53299817d6371a | 17:03 |
| sdague | mriedem: honestly, put that back on the bug reporter. I think we need them to indicate that this is a regression. The bug is just "I want to do this" | 17:04 |
| sdague | mriedem: https://github.com/openstack/nova/commit/09f2c4729456443c4874a8cadc53299817d6371a is where hypervisor_type becomes Text() | 17:04 |
| sdague | MediumText() was part of model sanity later, as I think the type might have shifted slightly in the grizzly data compaction | 17:04 |
| sdague | mriedem: note, that predates gerrit | 17:05 |
| *** derekh is now known as derekh_ark | 17:05 | |
| *** derekh_ark is now known as derekh_afk | 17:05 | |
| *** annegentle has joined #openstack-nova | 17:05 | |
| mriedem | and of course no real good info on why it was changed | 17:06 |
| mriedem | nice, merged with print statements in the runtime code | 17:07 |
| *** tblakes has quit IRC | 17:07 | |
| sdague | it's pre gerrit | 17:08 |
| sdague | this was when it was all just on vishy's laptop :) | 17:08 |
| *** fragatina has joined #openstack-nova | 17:08 | |
| jaypipes | dansmith: hey, sorry, grabbed some lunch... | 17:09 |
| dansmith | jaypipes: dedicated people eat lunch at their desk and keep working | 17:09 |
| mriedem | i don't suppose making that column smaller would give any perf benefit? | 17:09 |
| openstackgerrit | Dan Smith proposed openstack/nova: More conservative allocation updates https://review.openstack.org/366872 | 17:09 |
| dansmith | jaypipes: cdent: ^ | 17:09 |
| jaypipes | dansmith: so, yeah, the idea is that we need to get all the allocations for the resource provider and have the reporting client only send updated allocations when they are not known right? | 17:09 |
| dansmith | jaypipes: right, but the way the code is currently structured in the RT, we do that per instance, which means we have to fetch the entire set every time we do one instance, or try to do some caching that invalidates infrequently enough to not hide when things do need changing | 17:10 |
| jaypipes | dansmith: yes, I know, I should be more dedicated, but if I didn't walk up to get some lunch, you might never have seen this: https://twitter.com/jaypipes/status/773568094358933504 | 17:11 |
| *** Swami has joined #openstack-nova | 17:11 | |
| dansmith | heh | 17:11 |
| *** Sukhdev has joined #openstack-nova | 17:14 | |
| *** ajo_ has quit IRC | 17:14 | |
| *** ajo_ has joined #openstack-nova | 17:14 | |
| *** links has joined #openstack-nova | 17:14 | |
| *** PaulMurray has quit IRC | 17:15 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Correct serialization of inventory collections https://review.openstack.org/365633 | 17:15 |
| cdent | dansmith: that ^ is rebased onto your updated conservative updates | 17:15 |
| *** krtaylor has quit IRC | 17:16 | |
| dansmith | cdent: cool | 17:16 |
| cdent | is a smaller change than I remember | 17:16 |
| *** jdurgin1 has quit IRC | 17:16 | |
| * cdent is blurring everything together | 17:16 | |
| *** krtaylor has joined #openstack-nova | 17:16 | |
| dansmith | cdent: gotta keep up! rest is for after RC1 :) | 17:17 |
| *** annegentle has quit IRC | 17:17 | |
| dansmith | (rest as in sleep, not the other ten meanings we have for it) | 17:17 |
| cdent | dansmith: I think my error was that I wasn't resting after the last RC1 | 17:17 |
| *** ducttape_ has quit IRC | 17:18 | |
| *** mvk has quit IRC | 17:21 | |
| *** unicell has joined #openstack-nova | 17:25 | |
| *** sahid has quit IRC | 17:26 | |
| *** sudipto_ has joined #openstack-nova | 17:27 | |
| *** sudipto has joined #openstack-nova | 17:27 | |
| *** ayoung has quit IRC | 17:27 | |
| sdague | mriedem: unlikely, it's not like we are selecting by it often | 17:28 |
| sdague | or joining on it | 17:28 |
| *** links has quit IRC | 17:29 | |
| *** gouthamr_ is now known as gouthamr | 17:29 | |
| *** thorst has joined #openstack-nova | 17:29 | |
| *** ihrachys has quit IRC | 17:30 | |
| mriedem | sdague: yeah i guess the only thing that does filter on it is the ironic hash ring stuff | 17:30 |
| * cdent cuts the grass | 17:30 | |
| *** ssurana has joined #openstack-nova | 17:31 | |
| *** zenoway has joined #openstack-nova | 17:32 | |
| *** kfarr has joined #openstack-nova | 17:33 | |
| *** trisq has quit IRC | 17:34 | |
| sdague | https://review.openstack.org/#/c/366186/ is the consolidated logging I found helpful in tracking down the other issue. Though the capacity logging may not actually be that interesting. | 17:36 |
| *** zenoway has quit IRC | 17:36 | |
| *** Sukhdev has quit IRC | 17:36 | |
| sdague | so... on the allocations front | 17:36 |
| sdague | we're never allocating disk in the gate | 17:37 |
| sdague | is the special handling of 0 for disk in the flavor accounted for? | 17:37 |
| *** thorst has quit IRC | 17:38 | |
| *** yamahata has joined #openstack-nova | 17:38 | |
| *** ducttape_ has joined #openstack-nova | 17:38 | |
| cdent | sdague: I added that logging review to https://etherpad.openstack.org/p/placement-next so it doesn't get lost | 17:38 |
| sdague | cdent: cool | 17:38 |
| cdent | dansmith: ^^ (0 for disk)? | 17:38 |
| * cdent cuts grass tomorrow instead, can't be assed | 17:39 | |
| sdague | cdent: 0 for disk in the flavor means "use the size of the disk from glance" for... hysterical raisins (honestly, no idea why that is there like that) | 17:39 |
| cdent | sdague: yeah, I recall, I was just pinging dan into the conversation in case he had slipped away and didn't see that | 17:40 |
| openstackgerrit | Swaminathan Vasudevan proposed openstack/nova: Implement setup_networks_on_host for Neutron networks https://review.openstack.org/275073 | 17:40 |
| * cdent wonders where he can buy some hysterical raisins for his next party | 17:40 | |
| mriedem | dansmith: probably don't need to recheck https://review.openstack.org/#/c/364547/ since it's not merging b/c of the dependency | 17:40 |
| Swami | mriedem: ^^ | 17:41 |
| dansmith | mriedem: oh sorry | 17:41 |
| *** Sukhdev has joined #openstack-nova | 17:41 | |
| mriedem | i'm not sure that the dependency is necessary? | 17:41 |
| dansmith | mriedem: it's not I just did that to test | 17:42 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] Clarify commentary around logging trap https://review.openstack.org/365719 | 17:45 |
| *** paul-carlton2 has quit IRC | 17:45 | |
| mriedem | dansmith: i'll remove the depends-on and +W | 17:45 |
| dansmith | jsut removed | 17:45 |
| dansmith | mriedem: thanks | 17:46 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Cleanup allocation todo items https://review.openstack.org/364547 | 17:46 |
| dansmith | sorry missed that and thought it was due to gerrit maintenance on friday | 17:46 |
| *** harlowja has quit IRC | 17:46 | |
| dansmith | oh ffs | 17:46 |
| dansmith | I did the wrong one hah | 17:46 |
| mriedem | ha | 17:48 |
| mriedem | maybe you need some rest :) | 17:48 |
| dansmith | heh | 17:48 |
| dansmith | on a call and distracted :D | 17:48 |
| dansmith | I definitely need fewer calls, no doubt | 17:49 |
| *** gszasz has joined #openstack-nova | 17:49 | |
| openstackgerrit | Jay Pipes proposed openstack/nova-specs: Amend placement API with correct GET inventory https://review.openstack.org/366888 | 17:49 |
| *** harlowja has joined #openstack-nova | 17:49 | |
| jaypipes | cdent: https://review.openstack.org/366888 | 17:49 |
| jaypipes | cdent: I'll push a code fix for the placement API shortly. | 17:49 |
| jaypipes | cdent: not your fault. was just a simple oversight. | 17:50 |
| cdent | jaypipes: argh, anonymous dicts in lists! | 17:51 |
| *** su_zhang has joined #openstack-nova | 17:51 | |
| *** tblakes has joined #openstack-nova | 17:51 | |
| jaypipes | cdent: what is wrong with that? a dict is a record, simple as that. | 17:52 |
| *** fragatina has quit IRC | 17:52 | |
| *** fragatina has joined #openstack-nova | 17:53 | |
| cdent | now pretty much any time I want to actually get a specific piece of information out of the response, I have to pivot the list of dicts to a dict of dicts based on some key of some element in the list | 17:53 |
| *** ajo_ has quit IRC | 17:53 | |
| dansmith | I definitely like the anonymous dict approach less, but as long as they're consistent I don't care as much | 17:53 |
| jaypipes | cdent: or you use InventoryList.find() which is precisely why I wrote that method. | 17:53 |
| dansmith | cdent: right, that's why it sucks | 17:53 |
| *** ajo_ has joined #openstack-nova | 17:53 | |
| *** gszasz has quit IRC | 17:53 | |
| dansmith | but consistency wins every time | 17:53 |
| cdent | if I have an api response, I don't want to have to make it into an InventoryList, especially not client side, to have to do any work with it | 17:54 |
| jaypipes | dansmith, cdent: so do you want me to change it to be consistently *not* anon dicts? | 17:54 |
| *** hoangcx has quit IRC | 17:54 | |
| dansmith | jaypipes: were you saying that bauzas had a problem with that on spec review? | 17:54 |
| *** code-R has joined #openstack-nova | 17:55 | |
| *** zenoway has joined #openstack-nova | 17:55 | |
| cdent | brb | 17:56 |
| *** code-R_ has joined #openstack-nova | 17:56 | |
| *** hoangcx has joined #openstack-nova | 17:57 | |
| *** Sukhdev has quit IRC | 17:59 | |
| jaypipes | dansmith: yes, if I remember, bauzas wanted the anon dict format. | 17:59 |
| *** diogogmt has joined #openstack-nova | 17:59 | |
| dansmith | jaypipes: I wonder why | 17:59 |
| dansmith | bauzas: still around? | 18:00 |
| *** zenoway has quit IRC | 18:00 | |
| *** code-R has quit IRC | 18:00 | |
| cdent | sorry jaypipes, dansmith : dinner's on the table, will check for an update on this a bit later | 18:00 |
| *** cdent has quit IRC | 18:00 | |
| *** code-R_ has quit IRC | 18:02 | |
| *** fragatina has quit IRC | 18:03 | |
| *** fragatina has joined #openstack-nova | 18:04 | |
| sdague | mriedem: also, good news, in talking with mordred and bknudson, by our use of keystoneauth session object we don't have to work about token expiration | 18:09 |
| mriedem | cool | 18:10 |
| * mordred helped | 18:10 | |
| bauzas | dansmith: jaypipes: yup, wazup ? | 18:10 |
| jaypipes | bauzas: remember our discussions on the generic-resource-pools spec around the format of the returned JSON for the inventories collection? | 18:10 |
| jaypipes | bauzas: http://specs.openstack.org/openstack/nova-specs/specs/newton/approved/generic-resource-pools.html#get-resource-providers-uuid-inventories | 18:11 |
| jaypipes | bauzas: you had said you didn't like the use of a dict, keyed by resource class ID, and wanted to see anonymous dicts in a list, to be consistent with other OpenStack APIs. | 18:11 |
| bauzas | jaypipes: yup, it was a long time ago | 18:11 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: improve logging for shared storage check https://review.openstack.org/365140 | 18:11 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: nfs fix for xenial images https://review.openstack.org/329466 | 18:11 |
| jaypipes | bauzas: :) well, 2-3 months, yes. | 18:11 |
| bauzas | jaypipes: yup, my main concern was about the consistency | 18:12 |
| * bauzas needs to look at the review | 18:12 | |
| jaypipes | bauzas: anyway, we changed the PUT and POST to use anon dicts, like here: http://specs.openstack.org/openstack/nova-specs/specs/newton/approved/generic-resource-pools.html#put-resource-providers-uuid-inventories | 18:12 |
| jaypipes | bauzas: but forgot to do the GET response. | 18:12 |
| jaypipes | bauzas: but now we're questioning which way to go. | 18:13 |
| jaypipes | bauzas: and I mentioned to Dan you preferred the list of anonymous dicts approach. | 18:13 |
| *** ociuhandu has joined #openstack-nova | 18:13 | |
| *** akshai has joined #openstack-nova | 18:13 | |
| bauzas | okay, so that's because I think it's preferrable to have a list of keyed values identical | 18:13 |
| bauzas | in the formet GET method, I wasn't liking having keys not being stable | 18:14 |
| *** code-R has joined #openstack-nova | 18:14 | |
| bauzas | even if those keys were immutable, of course | 18:14 |
| *** code-R has quit IRC | 18:14 | |
| bauzas | by "stable", I mean having dicts having a list of defined keys | 18:15 |
| *** code-R has joined #openstack-nova | 18:15 | |
| bauzas | jaypipes: FWIW, +1d https://review.openstack.org/#/c/366888/1 | 18:16 |
| mordred | jaypipes: cdent may have his happiness about anonymous dicts improved by http://jmespath.org/ perhaps? | 18:17 |
| mordred | >>> jmespath.search('[].name', [{'name': 'foo'}, {'name': 'bar'}]) | 18:17 |
| mordred | ['foo', 'bar'] | 18:17 |
| mordred | works great | 18:17 |
| bauzas | jaypipes: dansmith: btw. given I'll be mostly looking at bug reports for the next week (due to the RC period), please ping me any patch you want me to look at needed for Newton | 18:17 |
| bauzas | jaypipes: dansmith: I can see lots of efforts around the resource-providers which I like, so in case you want me to review them... :) | 18:18 |
| *** sudipto_ has quit IRC | 18:19 | |
| *** sudipto has quit IRC | 18:19 | |
| *** hoangcx has quit IRC | 18:24 | |
| *** zul has quit IRC | 18:27 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Add tests for safe_connect decorator https://review.openstack.org/366800 | 18:28 |
| sdague | mriedem: the only actual issue was my assert_called, right? | 18:29 |
| mriedem | sdague: i think so | 18:30 |
| mriedem | the assert_not_called is in mock 2.0 and py3.5 | 18:30 |
| mriedem | i went over that with laski last week | 18:30 |
| mriedem | btw, why do we have both py 3.4 and 3.5 jobs? i'm sure someone has stated this at some point, | 18:31 |
| mriedem | but are we dropping the py3.4 half-baked stuff at some point? | 18:31 |
| sdague | yes | 18:31 |
| sdague | it's transitional | 18:32 |
| *** code-R_ has joined #openstack-nova | 18:32 | |
| sdague | because xenial has 3.5, and so most of the nodes will end up over there | 18:32 |
| clarkb | we only python3.4 on olderbranches now | 18:32 |
| jaypipes | bauzas: well, we are right now trying to achieve alignment on the format of the GET /r-p/{uuid}/inventories call. you read the above back and forth from cdent and dansmith yeah? | 18:34 |
| mriedem | clarkb: we're still running py34 jobs for nova on master https://review.openstack.org/#/c/357494/ | 18:34 |
| mriedem | gate-nova-python34-db | 18:34 |
| *** code-R has quit IRC | 18:35 | |
| clarkb | hrm probably need more cleanup in zuul.then | 18:35 |
| clarkb | the intent was to switch out I think since py34never really got aton of attention | 18:35 |
| clarkb | basically roll forward ratherthan support all the python3s | 18:36 |
| *** thorst has joined #openstack-nova | 18:39 | |
| *** ayoung has joined #openstack-nova | 18:39 | |
| *** su_zhang has quit IRC | 18:42 | |
| jaypipes | mriedem: +2 from me on sdague's safe_connect latest revision | 18:43 |
| *** thorst has quit IRC | 18:43 | |
| jaypipes | mriedem: I wasn't aware of the policy change to encourage docstrings on test methods, though. that's news to me. have we publicized that anywhere? | 18:43 |
| *** cdent has joined #openstack-nova | 18:45 | |
| cdent | mordred: is jmespath the ne jsonpath or jsonpath the new jmespath. jsonpath already used in lots of openstacky places, and I'm a fan, but in it (or jmespath) dealing with anonymous dicts still more complex than dicts of dicts | 18:46 |
| cdent | dicts of dicts just make more sense when we have unique "records" anyway... | 18:46 |
| mriedem | jaypipes: it comes up in irc every few months | 18:46 |
| cdent | but I'm agnostic at this stage | 18:47 |
| mriedem | jaypipes: otherwise no | 18:47 |
| cdent | jaypipes, dansmith: did you guys reach any conclusion? | 18:47 |
| jaypipes | cdent: dansmith | 18:47 |
| jaypipes | 's out for a bit. | 18:47 |
| * mriedem would love to hook up with a big ol' bag of anonymous dicts sometime | 18:48 | |
| *** code-R_ has quit IRC | 18:48 | |
| * cdent calls a data structure matchmaker for mriedem | 18:50 | |
| mriedem | also, we need a release steward to coordinate with the release CPL about the outgoing PTL | 18:50 |
| *** Apoorva_ has joined #openstack-nova | 18:51 | |
| cdent | mriedem: i'm pretty sure there's a czar for that | 18:51 |
| mriedem | good point | 18:52 |
| bauzas | mriedem: so, given I liked to be the release liaison, maybe I could help ? | 18:54 |
| mriedem | bauzas: you must take the title of czar now | 18:55 |
| mriedem | it's required | 18:55 |
| *** Apoorva has quit IRC | 18:55 | |
| *** zul_ has quit IRC | 18:55 | |
| bauzas | mriedem: well, I'm a French :po | 18:55 |
| *** hoangcx has joined #openstack-nova | 18:55 | |
| mriedem | bauzas: i was just joking, and making light of the ML thread on release stewards | 18:56 |
| bauzas | yup, I read that | 18:56 |
| *** tonytan4ever has quit IRC | 18:56 | |
| *** akshai has quit IRC | 18:59 | |
| cdent | jaypipes however we resolve this dict list thing, if we are going to be changing stuff we may as well do a sincere audit of all the request and response bodies and make sure we have consistency across the board. Or if not now, soon. | 18:59 |
| jaypipes | cdent: you mean just the placement API, right? :) | 19:00 |
| cdent | jaypipes: well, yes | 19:01 |
| cdent | but you know maybe the api-wg can make a czar for doing it across the _whole_ board | 19:01 |
| jaypipes | cdent: k, just making sure :) | 19:01 |
| *** sdake has quit IRC | 19:02 | |
| *** salv-orlando has quit IRC | 19:02 | |
| mriedem | dict usage czar? | 19:03 |
| *** raunak has quit IRC | 19:03 | |
| *** prateek has quit IRC | 19:04 | |
| *** lpetrut has quit IRC | 19:04 | |
| *** abalutoiu has joined #openstack-nova | 19:05 | |
| *** abalutoiu has quit IRC | 19:05 | |
| mriedem | armax: kevinbenton: were you guys going to try and get a requirements freeze FFE for https://review.openstack.org/#/c/365165/ ? | 19:05 |
| *** abalutoiu has joined #openstack-nova | 19:05 | |
| armax | no | 19:05 |
| armax | mriedem: I mean | 19:06 |
| armax | I’d love to get this in | 19:06 |
| armax | mriedem: but I thought you were going to champion and lobby for it :) | 19:06 |
| armax | mriedem: you need to be backed up? | 19:06 |
| mriedem | oh dear | 19:06 |
| armax | technically speaking the bug is with os-vif | 19:07 |
| armax | we’re good on our side :) | 19:07 |
| armax | do we have a communication breakdown here? | 19:07 |
| mriedem | we need a nova/neutron CPL for the PTL | 19:08 |
| mriedem | armax: i'll request the FFE | 19:08 |
| armax | mriedem: sorry | 19:08 |
| armax | mriedem: I thought this was on your plate | 19:08 |
| *** annegentle has joined #openstack-nova | 19:09 | |
| *** TravT has quit IRC | 19:09 | |
| mriedem | done | 19:10 |
| mriedem | np | 19:10 |
| *** clenimar has quit IRC | 19:10 | |
| *** sdake has joined #openstack-nova | 19:10 | |
| *** zul has joined #openstack-nova | 19:12 | |
| *** prateek has joined #openstack-nova | 19:12 | |
| *** salv-orlando has joined #openstack-nova | 19:13 | |
| *** derekh_afk is now known as derekh | 19:13 | |
| *** su_zhang has joined #openstack-nova | 19:13 | |
| *** lucas-dinner has quit IRC | 19:13 | |
| *** lucasagomes has joined #openstack-nova | 19:15 | |
| *** Apoorva_ has quit IRC | 19:16 | |
| *** Apoorva has joined #openstack-nova | 19:16 | |
| *** Sukhdev_ has joined #openstack-nova | 19:17 | |
| *** fragatin_ has joined #openstack-nova | 19:17 | |
| *** su_zhang has quit IRC | 19:18 | |
| *** fragatina has quit IRC | 19:21 | |
| *** rcernin has quit IRC | 19:21 | |
| *** ducttape_ has quit IRC | 19:23 | |
| auggy | bauzas: i like "release gaucho" myself | 19:26 |
| auggy | bauzas: then you can wear a cool hat | 19:27 |
| *** ducttape_ has joined #openstack-nova | 19:27 | |
| *** TravT has joined #openstack-nova | 19:27 | |
| *** TravT has quit IRC | 19:28 | |
| *** mvk has joined #openstack-nova | 19:29 | |
| *** jvgrant has quit IRC | 19:29 | |
| *** _ducttape_ has joined #openstack-nova | 19:30 | |
| *** zul_ has joined #openstack-nova | 19:31 | |
| mriedem | wznoinsk: should intel nfv ci be running on stable/liberty patches at all? | 19:33 |
| mriedem | wznoinsk: i noticed intel nfv ci failed pretty hard here https://review.openstack.org/#/c/358152/ | 19:33 |
| *** ducttape_ has quit IRC | 19:34 | |
| mriedem | BobBall: citrix xenserver ci is failing devstack setup here http://dd6b71949550285df7dc-dda4e480e005aaa13ec303551d2d8155.r49.cf1.rackcdn.com/52/358152/1/check/dsvm-tempest-nova-network/9c13c23/logs/devstacklog.txt.gz | 19:34 |
| mriedem | ^ is stable/liberty | 19:34 |
| *** moshele has joined #openstack-nova | 19:37 | |
| *** _ducttape_ has quit IRC | 19:39 | |
| *** thorst has joined #openstack-nova | 19:39 | |
| *** ducttape_ has joined #openstack-nova | 19:40 | |
| *** hoangcx has quit IRC | 19:43 | |
| *** rama_y has joined #openstack-nova | 19:43 | |
| sdague | dansmith: what's the general approach / policy on __repr__ for nova objects? | 19:43 |
| *** lpetrut has joined #openstack-nova | 19:45 | |
| cdent | sdague: I had that question very early in the process of making resource provider objects, never figured it out | 19:45 |
| *** thorst has quit IRC | 19:48 | |
| dansmith | sdague: there is a general handler that tries to do something sane, but otherwise do whatever you want for each case | 19:49 |
| dansmith | sdague: by default it tries to use self.uuid or self.id, which almost all of our objects have | 19:49 |
| dansmith | sdague: same thing I told cdent when he asked the first time :) | 19:50 |
| sdague | dansmith: ok, the object lists seem to be a little odd | 19:50 |
| *** lpetrut has quit IRC | 19:50 | |
| sdague | let me propose a thing | 19:50 |
| sdague | which makes some of the debug a bit simpler | 19:50 |
| cdent | I should clarify: I never understood clearly what should be done... | 19:51 |
| dansmith | sdague: I think object lists should be just a repr that does a list of reprs, but if not.. by all means | 19:51 |
| sdague | dansmith: it doesn't seem to call the repr nested | 19:51 |
| dansmith | sdague: okay I'm looking | 19:51 |
| *** annegentle has quit IRC | 19:51 | |
| dansmith | sdague: hmm, yeah not set on o.vo.. we might've lost that in the transition or something | 19:52 |
| dansmith | or maybe I'm full of it | 19:52 |
| sdague | dansmith: ok, so the behavior you suggest is the one I want | 19:52 |
| sdague | glad I'm not crazy in thinking that it should be there | 19:52 |
| *** tonytan4ever has joined #openstack-nova | 19:53 | |
| sdague | I'll do a thing in the morning, it's nearly end of day here, and I have a meeting soon | 19:53 |
| *** fragatin_ has quit IRC | 19:53 | |
| dansmith | sdague: not for this reason, no | 19:53 |
| *** fragatina has joined #openstack-nova | 19:54 | |
| cdent | Do we have an existing process for commenting on existing code and talking about it? Or is the general way to talk about alternative things, to commit something for review? | 19:54 |
| cdent | The latter is often too heavyweight | 19:55 |
| *** Apoorva_ has joined #openstack-nova | 19:58 | |
| *** _ducttape_ has joined #openstack-nova | 19:59 | |
| *** lpetrut has joined #openstack-nova | 20:00 | |
| *** raunak has joined #openstack-nova | 20:01 | |
| *** jaypipes has quit IRC | 20:01 | |
| *** Apoorva has quit IRC | 20:01 | |
| *** rama_y has quit IRC | 20:02 | |
| *** ducttape_ has quit IRC | 20:02 | |
| *** openstackgerrit has quit IRC | 20:04 | |
| *** BjoernT has quit IRC | 20:04 | |
| *** openstackgerrit has joined #openstack-nova | 20:04 | |
| *** abalutoiu_ has joined #openstack-nova | 20:09 | |
| *** cdent has quit IRC | 20:09 | |
| *** salv-orlando has quit IRC | 20:10 | |
| *** tri2sing has quit IRC | 20:12 | |
| *** abalutoiu has quit IRC | 20:12 | |
| *** rfolco has quit IRC | 20:15 | |
| mriedem | Swami: i'm not understanding something in the new test in https://review.openstack.org/#/c/275073/44 | 20:16 |
| mriedem | with the assertRaises | 20:16 |
| Swami | yes | 20:16 |
| mriedem | sdague: dansmith: we talked about repr on list objects a long time ago, workaround in nova was doing list(listobj) in the nova log calls, but we had talked about making that native in the object list base class in ovo | 20:17 |
| Swami | mriedem: what is your question | 20:17 |
| dansmith | mriedem: ack | 20:17 |
| sdague | mriedem: yeh, | 20:17 |
| mriedem | Swami: it looks like you're asserting that the mock raises the exception from the side_effect, which is redundant | 20:17 |
| sdague | that seems about like I would expect | 20:17 |
| mriedem | that's like saying mock.foo = 'bar'; self.assertEqual('bar', mock.foo) | 20:18 |
| *** syjulian_ has joined #openstack-nova | 20:18 | |
| *** salv-orlando has joined #openstack-nova | 20:18 | |
| mriedem | Swami: in other words, why isn't the assertRaises on _update_port_with_migration_profile in the test? | 20:18 |
| mriedem | since _update_port_with_migration_profile should reraise the exception | 20:18 |
| Swami | mriedem: let me take look at it | 20:18 |
| Swami | since update_port is the one that raises the exception I have captured it in that way. | 20:19 |
| mriedem | Swami: but your update_port mock raises the exception b/c you tell it to with the side_effect | 20:21 |
| mriedem | so asserting that calling your mock raises the exception isn't helpful for the test | 20:21 |
| mriedem | Swami: i've got some comments inline in the test | 20:21 |
| Swami | ok will check it out | 20:22 |
| mriedem | i will pull this down and tinker with it also | 20:23 |
| Swami | thanks | 20:23 |
| *** su_zhang has joined #openstack-nova | 20:23 | |
| *** moshele has quit IRC | 20:23 | |
| openstackgerrit | Kenneth Burger proposed openstack/nova: Increase BDM column in build_requests table https://review.openstack.org/366955 | 20:25 |
| *** anush has joined #openstack-nova | 20:25 | |
| *** derekh has quit IRC | 20:25 | |
| burgerk | mriedem: alaski: https://review.openstack.org/#/c/366955 | 20:26 |
| dansmith | sdague: so the most important thing we still have pending is this: https://review.openstack.org/#/c/365068/ | 20:27 |
| dansmith | sdague: and you were +2 before, I added the warning you wanted | 20:27 |
| dansmith | sdague: we have other more important stuff, but that one is at the bottom of a stack and arguably the most important right now | 20:27 |
| alaski | burgerk: you need to add a new migration rather than modifying an existing one. Some people have already deployed that migration and wouldn't get your change | 20:27 |
| dansmith | sdague: so now or tomorrow morning if you could hit that agin that'd be cool | 20:28 |
| burgerk | alaski: ok | 20:28 |
| sdague | dansmith: ok, can we get that warning in there? | 20:28 |
| dansmith | sdague: I was just saying I did that | 20:28 |
| mriedem | alaski: better get your -1 credit :) | 20:29 |
| dansmith | sdague: https://review.openstack.org/#/c/365068/8/nova/objects/resource_provider.py L335 | 20:29 |
| *** lpetrut has quit IRC | 20:29 | |
| sdague | dansmith: oh, sorry, I didn't scroll down enough there | 20:29 |
| alaski | mriedem: heh. gotta rack up them points | 20:29 |
| sdague | I was expecting it up further, my bad | 20:29 |
| *** salv-orl_ has joined #openstack-nova | 20:29 | |
| dansmith | sdague: yeah I had to plumb it down low so we can use the same thing for notifications and not be blocking transactions for rpc calls :) | 20:29 |
| mriedem | alaski: if you get enough, you can turn them in for googly eyes at the ticket counter | 20:30 |
| dansmith | or plumb it up high? I dunno.. plumbing was involved. | 20:30 |
| mriedem | or a switchblade comb | 20:30 |
| dansmith | mriedem: lol | 20:30 |
| mriedem | Swami: i've got the test cleanup, just verifying it locally and then will push up and +2 | 20:30 |
| sdague | dansmith: ok, yeh, I'll probably tweak the message later with the actually capacity / usage | 20:30 |
| Swami | mriedem: thanks, great | 20:30 |
| sdague | but I think this is fine | 20:31 |
| sdague | for first go | 20:31 |
| sdague | at least you get some indication that things are wrong | 20:31 |
| dansmith | sdague: okay, I guess we'll want that in the notification anyway | 20:31 |
| dansmith | sdague: so we can do that when we add the notification | 20:31 |
| mriedem | alaski: is there a cellsv2 meeting in 28 minutes? | 20:32 |
| dansmith | I sure hope not | 20:32 |
| alaski | mriedem: not today | 20:32 |
| *** rama_y has joined #openstack-nova | 20:32 | |
| dansmith | woot | 20:32 |
| dansmith | if I have to be in a confined space with other people right now, I may bite. | 20:32 |
| mriedem | alaski: ok every other week now? | 20:32 |
| mriedem | i forgot | 20:32 |
| alaski | mriedem: yep. until we officially open O there's not much to discuss | 20:33 |
| *** salv-orlando has quit IRC | 20:33 | |
| dansmith | sdague: I wrote tests for the log thing.. you okay with me +Wing that? or should I find someone else? | 20:33 |
| dansmith | well, and wrote the log thing itself, but... | 20:33 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Implement setup_networks_on_host for Neutron networks https://review.openstack.org/275073 | 20:35 |
| sdague | I'm fine with it | 20:35 |
| sdague | jay was +2 earlier | 20:35 |
| *** thorst has joined #openstack-nova | 20:35 | |
| mriedem | Swami: now you just have to bribe all of the other nova cores for the next week | 20:36 |
| *** BjoernT has joined #openstack-nova | 20:37 | |
| Swami | mriedem: thanks, will you be able to ping others to add another +2. | 20:37 |
| mriedem | um | 20:37 |
| mriedem | well, | 20:37 |
| mriedem | johnthetubaguy would probably be the next best to look at it given he's been in the neutronv2 api code a lot recently | 20:37 |
| mriedem | but i'm not going to probably actively lobby him | 20:38 |
| Swami | mriedem: ok, I will look for him | 20:38 |
| mriedem | he's sure to be out tubaing in the streets of england right about now | 20:38 |
| mriedem | driving the neighbors crazy | 20:38 |
| Swami | mriedem: Let me check with him when he comes on line. | 20:39 |
| *** roeyc has joined #openstack-nova | 20:40 | |
| *** _ducttape_ has quit IRC | 20:40 | |
| *** ducttape_ has joined #openstack-nova | 20:41 | |
| *** BjoernT has quit IRC | 20:41 | |
| *** thorst has quit IRC | 20:42 | |
| *** ociuhandu has quit IRC | 20:43 | |
| openstackgerrit | Doug Hellmann proposed openstack/python-novaclient: Update reno for stable/newton https://review.openstack.org/366959 | 20:43 |
| openstackgerrit | Doug Hellmann proposed openstack/python-novaclient: Update reno for stable/newton https://review.openstack.org/366959 | 20:45 |
| *** GB21 has quit IRC | 20:46 | |
| dansmith | sdague: okay | 20:48 |
| *** prateek_ has joined #openstack-nova | 20:49 | |
| *** Apoorva_ has quit IRC | 20:49 | |
| *** gongysh has joined #openstack-nova | 20:49 | |
| mriedem | alaski: what are you thinking for a test on burgerk's db migration? just that the column type is changed? or also trying to insert >16Kb of data into it? | 20:49 |
| *** Apoorva has joined #openstack-nova | 20:50 | |
| dansmith | the former | 20:50 |
| alaski | mriedem: yeah, just that the column type changed | 20:50 |
| dansmith | we don't really test schema stuff | 20:50 |
| *** rtheis has joined #openstack-nova | 20:50 | |
| *** rrecio has quit IRC | 20:50 | |
| dansmith | the migration walk tests were really for when we migrated data | 20:50 |
| dansmith | maybe insert something in there and make sure it's still there post-migrate, | 20:50 |
| dansmith | but otherwise that's all I think | 20:50 |
| *** vishwana_ has joined #openstack-nova | 20:51 | |
| alaski | mriedem: mostly it's that the tests will fail if a new migration is added and no check is added for it | 20:51 |
| dansmith | i.e. to make sure we didn't do a DROP;ADD | 20:51 |
| *** liangy has quit IRC | 20:51 | |
| *** ociuhandu has joined #openstack-nova | 20:51 | |
| mriedem | dansmith: yeah that would be a decent test | 20:52 |
| *** vishwanathj has quit IRC | 20:52 | |
| *** prateek has quit IRC | 20:52 | |
| *** rtheis_ has quit IRC | 20:52 | |
| *** parora has joined #openstack-nova | 20:52 | |
| openstackgerrit | Doug Hellmann proposed openstack/os-vif: Update reno for stable/newton https://review.openstack.org/363115 | 20:53 |
| *** prateek_ has quit IRC | 20:55 | |
| sdague | dansmith: you mean, we shouldn't test the schema stuff | 20:55 |
| sdague | the test infrastruction in oslo.db requires it | 20:55 |
| dansmith | sdague: was that a question? | 20:55 |
| *** _ducttape_ has joined #openstack-nova | 20:55 | |
| sdague | it's just silly to test that sqla works :) | 20:56 |
| dansmith | sdague: should I fix this nit? https://review.openstack.org/#/c/366186/3/nova/api/openstack/placement/handlers/allocation.py | 20:57 |
| sdague | dansmith: if you like, if you don't I was going to do the repr for the list instead | 20:57 |
| sdague | and put that into there | 20:57 |
| sdague | that's part of what triggered my interest in list repr | 20:57 |
| dansmith | sdague: oh okay was just trying to clear the list | 20:57 |
| dansmith | gotcha | 20:57 |
| sdague | dansmith: either way, I'll work on that patch when I wake up in the morning | 20:58 |
| dansmith | sdague: okay cool | 20:58 |
| *** ducttape_ has quit IRC | 20:59 | |
| *** ssurana1 has joined #openstack-nova | 21:03 | |
| *** liangy has joined #openstack-nova | 21:03 | |
| *** ssurana has quit IRC | 21:06 | |
| *** _ducttape_ has quit IRC | 21:06 | |
| *** ociuhandu has quit IRC | 21:08 | |
| *** krtaylor has quit IRC | 21:11 | |
| *** roeyc has quit IRC | 21:11 | |
| *** ducttape_ has joined #openstack-nova | 21:12 | |
| openstackgerrit | Doug Hellmann proposed openstack/python-novaclient: Update reno for stable/newton https://review.openstack.org/364954 | 21:12 |
| *** mdrabe has quit IRC | 21:14 | |
| *** mdrabe has joined #openstack-nova | 21:14 | |
| *** rtheis_ has joined #openstack-nova | 21:14 | |
| *** mmedvede has quit IRC | 21:15 | |
| *** rtheis has quit IRC | 21:16 | |
| openstackgerrit | Doug Hellmann proposed openstack/os-vif: Update reno for stable/newton https://review.openstack.org/363115 | 21:18 |
| *** BjoernT has joined #openstack-nova | 21:19 | |
| *** edmondsw has quit IRC | 21:19 | |
| openstackgerrit | Doug Hellmann proposed openstack/os-vif: Update reno for stable/newton https://review.openstack.org/363115 | 21:19 |
| *** jheroux has quit IRC | 21:27 | |
| *** lpetrut has joined #openstack-nova | 21:30 | |
| *** ducttape_ has quit IRC | 21:31 | |
| *** abalutoiu has joined #openstack-nova | 21:32 | |
| *** syjulian_ has quit IRC | 21:33 | |
| *** roeyc has joined #openstack-nova | 21:34 | |
| *** abalutoiu_ has quit IRC | 21:34 | |
| *** kfarr has quit IRC | 21:34 | |
| *** roeyc has quit IRC | 21:34 | |
| *** ducttape_ has joined #openstack-nova | 21:36 | |
| *** hemna is now known as hemnafk | 21:37 | |
| *** su_zhang has quit IRC | 21:37 | |
| *** rtheis_ has quit IRC | 21:37 | |
| *** syjulian_ has joined #openstack-nova | 21:37 | |
| *** su_zhang has joined #openstack-nova | 21:38 | |
| *** su_zhang has quit IRC | 21:38 | |
| *** su_zhang has joined #openstack-nova | 21:38 | |
| *** thorst has joined #openstack-nova | 21:40 | |
| *** mmedvede has joined #openstack-nova | 21:41 | |
| *** mriedem has quit IRC | 21:41 | |
| *** claudiub|2 has quit IRC | 21:42 | |
| *** ducttape_ has quit IRC | 21:43 | |
| *** thorst has quit IRC | 21:48 | |
| *** Swami has quit IRC | 21:48 | |
| *** krtaylor has joined #openstack-nova | 21:53 | |
| *** su_zhang has quit IRC | 22:01 | |
| *** BjoernT has quit IRC | 22:02 | |
| *** ducttape_ has joined #openstack-nova | 22:03 | |
| *** gongysh has quit IRC | 22:03 | |
| *** burt has quit IRC | 22:04 | |
| *** rama_y has quit IRC | 22:04 | |
| *** su_zhang has joined #openstack-nova | 22:05 | |
| *** mfedosin has quit IRC | 22:07 | |
| *** tonytan4ever has quit IRC | 22:10 | |
| *** rama_y has joined #openstack-nova | 22:10 | |
| *** ducttape_ has quit IRC | 22:12 | |
| *** jamesdenton has quit IRC | 22:13 | |
| *** nic has quit IRC | 22:14 | |
| *** nic has joined #openstack-nova | 22:15 | |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Using get() method to prevent KeyError https://review.openstack.org/359424 | 22:16 |
| *** baoli_ has joined #openstack-nova | 22:17 | |
| *** baoli has quit IRC | 22:18 | |
| *** prateek_ has joined #openstack-nova | 22:18 | |
| *** ducttape_ has joined #openstack-nova | 22:19 | |
| *** tblakes has quit IRC | 22:19 | |
| *** burgerk_ has joined #openstack-nova | 22:19 | |
| *** lpetrut has quit IRC | 22:19 | |
| *** kevinz has joined #openstack-nova | 22:21 | |
| *** parora has quit IRC | 22:21 | |
| *** rama_y has quit IRC | 22:22 | |
| *** burgerk has quit IRC | 22:22 | |
| *** parora has joined #openstack-nova | 22:23 | |
| *** su_zhang has quit IRC | 22:23 | |
| *** burgerk_ has quit IRC | 22:24 | |
| *** sdake has quit IRC | 22:24 | |
| *** prateek_ has quit IRC | 22:25 | |
| *** kevinz has quit IRC | 22:26 | |
| *** BjoernT has joined #openstack-nova | 22:26 | |
| *** BjoernT has quit IRC | 22:26 | |
| *** dimtruck is now known as zz_dimtruck | 22:27 | |
| *** gongysh has joined #openstack-nova | 22:27 | |
| *** akshai has joined #openstack-nova | 22:30 | |
| *** fragatin_ has joined #openstack-nova | 22:34 | |
| *** abalutoiu has quit IRC | 22:36 | |
| *** fragatina has quit IRC | 22:37 | |
| *** mdrabe has quit IRC | 22:39 | |
| *** Guest10379 has quit IRC | 22:39 | |
| *** prateek_ has joined #openstack-nova | 22:39 | |
| *** mdrabe has joined #openstack-nova | 22:43 | |
| *** parora has quit IRC | 22:43 | |
| *** mlavalle has quit IRC | 22:46 | |
| *** ducttape_ has quit IRC | 22:48 | |
| *** thorst has joined #openstack-nova | 22:48 | |
| *** mdrabe has quit IRC | 22:51 | |
| *** thorst has quit IRC | 22:53 | |
| *** baoli_ has quit IRC | 22:54 | |
| *** zz_dimtruck is now known as dimtruck | 22:54 | |
| *** xyang1 has quit IRC | 22:55 | |
| *** ducttape_ has joined #openstack-nova | 22:55 | |
| *** pradk has quit IRC | 22:56 | |
| *** fragatin_ has quit IRC | 22:57 | |
| *** marst has quit IRC | 22:57 | |
| *** armax has quit IRC | 22:58 | |
| *** itlinux has joined #openstack-nova | 22:58 | |
| *** gouthamr has quit IRC | 22:58 | |
| *** harlowja has quit IRC | 23:03 | |
| *** gongysh has quit IRC | 23:08 | |
| *** syjulian_ has quit IRC | 23:09 | |
| *** itlinux has quit IRC | 23:10 | |
| *** tonytan4ever has joined #openstack-nova | 23:10 | |
| *** yamahata has quit IRC | 23:11 | |
| *** nic has quit IRC | 23:12 | |
| *** marst has joined #openstack-nova | 23:12 | |
| *** sdague has quit IRC | 23:12 | |
| *** tonytan4ever has quit IRC | 23:15 | |
| *** salv-orl_ has quit IRC | 23:16 | |
| *** salv-orlando has joined #openstack-nova | 23:17 | |
| *** ducttape_ has quit IRC | 23:17 | |
| *** beagles has left #openstack-nova | 23:18 | |
| *** salv-orlando has quit IRC | 23:21 | |
| *** akshai has quit IRC | 23:23 | |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Add image_meta when unshelving volume backend vm https://review.openstack.org/347063 | 23:23 |
| *** akshai has joined #openstack-nova | 23:24 | |
| *** diogogmt has quit IRC | 23:24 | |
| *** akshai has quit IRC | 23:25 | |
| *** anush has quit IRC | 23:26 | |
| *** claudiub|2 has joined #openstack-nova | 23:28 | |
| *** gouthamr has joined #openstack-nova | 23:34 | |
| *** gouthamr_ has joined #openstack-nova | 23:34 | |
| *** yamahata has joined #openstack-nova | 23:38 | |
| *** gouthamr has quit IRC | 23:38 | |
| *** gmann has joined #openstack-nova | 23:38 | |
| *** Swami has joined #openstack-nova | 23:40 | |
| *** tonytan4ever has joined #openstack-nova | 23:45 | |
| *** edmondsw has joined #openstack-nova | 23:46 | |
| *** ducttape_ has joined #openstack-nova | 23:46 | |
| *** ducttape_ has quit IRC | 23:48 | |
| *** ducttape_ has joined #openstack-nova | 23:48 | |
| *** akshai has joined #openstack-nova | 23:49 | |
| *** edmondsw has quit IRC | 23:49 | |
| *** armax has joined #openstack-nova | 23:50 | |
| *** thorst has joined #openstack-nova | 23:50 | |
| *** rfolco has joined #openstack-nova | 23:57 | |
| *** prateek_ has quit IRC | 23:57 | |
| *** akshai has quit IRC | 23:57 | |
| *** rfolco has quit IRC | 23:58 | |
| *** thorst has quit IRC | 23:58 | |
| *** rfolco has joined #openstack-nova | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!