| *** gouthamr_ has joined #openstack-nova | 00:04 | |
| *** gouthamr has quit IRC | 00:08 | |
| *** shaohe_feng has quit IRC | 00:09 | |
| *** woodster_ has quit IRC | 00:09 | |
| openstackgerrit | Mathieu Gagné proposed openstack/nova-specs: Update metadata-service-network-info according to implementation https://review.openstack.org/337819 | 00:10 |
|---|---|---|
| *** shaohe_feng has joined #openstack-nova | 00:10 | |
| *** hoangcx has joined #openstack-nova | 00:13 | |
| *** sdake has joined #openstack-nova | 00:14 | |
| *** mtanino has quit IRC | 00:16 | |
| *** lixiaoy1 has joined #openstack-nova | 00:18 | |
| *** trisq has joined #openstack-nova | 00:18 | |
| *** shaohe_feng has quit IRC | 00:19 | |
| *** Swami has quit IRC | 00:20 | |
| *** shaohe_feng has joined #openstack-nova | 00:20 | |
| *** thorst has quit IRC | 00:23 | |
| *** thorst has joined #openstack-nova | 00:24 | |
| *** fragatin_ has quit IRC | 00:27 | |
| *** Apoorva has quit IRC | 00:27 | |
| *** fragatina has joined #openstack-nova | 00:28 | |
| mgagne | markstur: afaik, capacity is enforced at the scheduler level. nova hypervisor-show won't show you the effective "available" vcpus as hypervisor could be part of an aggregate with different cpu_allocation_ratio. cpu_allocation_ratio config is read from machine running nova-scheduler, it's not read from compute node, even if it defaults to 1:16. | 00:29 |
| *** shaohe_feng has quit IRC | 00:29 | |
| *** browne has quit IRC | 00:30 | |
| *** shaohe_feng has joined #openstack-nova | 00:31 | |
| mgagne | markstur: so yea, it causes a lot of headaches if you wish to monitor your effective cloud capacity as you would have to reimplement the same logic nova-scheduler uses. This means find a way to detect default cpu_allocation_ratio (16), check if scheduler itself is configured with a different value in nova.conf or if hypervisor is part of an aggregate with a different cpu_allocation_ratio. Even more fun if it's part of 2 aggregates with 2 | 00:32 |
| mgagne | different ratios which makes little sense but is "technically possible". | 00:32 |
| *** thorst has quit IRC | 00:32 | |
| markstur | but the math would not be very smart about it, right? | 00:33 |
| markstur | I think the scheduler doesn't know that some existing vCPUs should have more weight and will allow ligher vCPUs to pile on. | 00:34 |
| markstur | mgagne, ^^ | 00:34 |
| markstur | Or perhaps "makes little sense" is enough said. :) | 00:35 |
| mgagne | markstur: that's why while it's technically possible to have multiple ratio, I wouldn't recommend it as there is no known way (to me) to make SOME physical cores 1:1 and others 1:16 on the same machine. | 00:35 |
| mgagne | markstur: that's why we (my employer) do not mix them but create different aggregates with specific ratios. | 00:36 |
| markstur | mgagne, Sure at run time they are not relatively throttled. But I was just looking at the ability to oversubcribe beyond the agreed | 00:36 |
| mgagne | markstur: do you want to oversubscribe at schedule time or by some other ways? | 00:37 |
| markstur | mgagne, We just want to have some VMs that use a 1:1 ratio and others to have a 4:1 ratio. And I was trying to explain how it is not a good idea to let those fellas share one hypervisor. | 00:37 |
| *** shaohe_feng has quit IRC | 00:40 | |
| markstur | Time for me to run. Thanks for the reply! | 00:40 |
| *** shaohe_feng has joined #openstack-nova | 00:40 | |
| mgagne | markstur: cya ;) | 00:42 |
| *** ijw has joined #openstack-nova | 00:46 | |
| *** ijw has quit IRC | 00:46 | |
| *** ijw has joined #openstack-nova | 00:46 | |
| *** su_zhang has quit IRC | 00:48 | |
| *** su_zhang has joined #openstack-nova | 00:48 | |
| *** su_zhang has quit IRC | 00:49 | |
| *** shaohe_feng has quit IRC | 00:50 | |
| *** shaohe_feng has joined #openstack-nova | 00:53 | |
| *** su_zhang has joined #openstack-nova | 00:53 | |
| *** cheneydc has joined #openstack-nova | 00:55 | |
| *** su_zhang has quit IRC | 00:57 | |
| *** trisq has quit IRC | 00:58 | |
| *** shaohe_feng has quit IRC | 01:00 | |
| mriedem_afk | mikal: this looks good - https://review.openstack.org/#/c/353793/ - just a few small things inline | 01:01 |
| *** vishwanathj has quit IRC | 01:02 | |
| *** shaohe_feng has joined #openstack-nova | 01:02 | |
| *** vishwanathj has joined #openstack-nova | 01:02 | |
| mriedem_afk | mikal: and what's up with this? https://review.openstack.org/#/c/348066/ | 01:03 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add more vd2 unit tests https://review.openstack.org/349298 | 01:09 |
| *** shaohe_feng has quit IRC | 01:10 | |
| *** shaohe_feng has joined #openstack-nova | 01:11 | |
| *** nic has quit IRC | 01:12 | |
| *** maurosr has quit IRC | 01:12 | |
| *** raunak has quit IRC | 01:13 | |
| *** kaisers_ has quit IRC | 01:14 | |
| *** sdake has quit IRC | 01:14 | |
| *** thorst has joined #openstack-nova | 01:17 | |
| *** thorst has quit IRC | 01:17 | |
| *** thorst has joined #openstack-nova | 01:17 | |
| *** shaohe_feng has quit IRC | 01:21 | |
| *** sdake has joined #openstack-nova | 01:21 | |
| *** shaohe_feng has joined #openstack-nova | 01:22 | |
| *** gongysh has joined #openstack-nova | 01:23 | |
| *** trisq has joined #openstack-nova | 01:24 | |
| *** thorst has quit IRC | 01:26 | |
| *** mriedem_afk has quit IRC | 01:29 | |
| *** fragatina has quit IRC | 01:30 | |
| *** shaohe_feng has quit IRC | 01:31 | |
| *** hoangcx has quit IRC | 01:31 | |
| *** shaohe_feng has joined #openstack-nova | 01:31 | |
| *** wangqun has joined #openstack-nova | 01:32 | |
| *** ociuhandu has joined #openstack-nova | 01:32 | |
| *** salv-orlando has joined #openstack-nova | 01:32 | |
| *** cheneydc has quit IRC | 01:32 | |
| *** cheneydc has joined #openstack-nova | 01:33 | |
| *** ljxiash has joined #openstack-nova | 01:34 | |
| *** ociuhandu has quit IRC | 01:34 | |
| *** EinstCrazy has joined #openstack-nova | 01:38 | |
| *** ljxiash has quit IRC | 01:39 | |
| *** code-R has joined #openstack-nova | 01:40 | |
| *** shaohe_feng has quit IRC | 01:41 | |
| *** shaohe_feng has joined #openstack-nova | 01:41 | |
| *** amotoki has joined #openstack-nova | 01:42 | |
| *** salv-orlando has quit IRC | 01:43 | |
| *** stewie925 has quit IRC | 01:43 | |
| *** parora has joined #openstack-nova | 01:43 | |
| *** prateek_ has quit IRC | 01:45 | |
| *** amotoki has quit IRC | 01:46 | |
| *** rfolco has quit IRC | 01:47 | |
| *** Sukhdev has joined #openstack-nova | 01:48 | |
| auggy | omfg... finally got a mox test to pass FO REALZIES!! | 01:50 |
| * auggy wonders how many she broke in the process | 01:51 | |
| *** shaohe_feng has quit IRC | 01:51 | |
| *** ljxiash has joined #openstack-nova | 01:52 | |
| *** ljxiash has quit IRC | 01:54 | |
| *** ljxiash has joined #openstack-nova | 01:54 | |
| *** unicell has quit IRC | 01:55 | |
| *** shaohe_feng has joined #openstack-nova | 01:56 | |
| *** code-R has quit IRC | 01:57 | |
| openstackgerrit | Tuan Luong-Anh proposed openstack/nova: Using assertIsNone() instead of assertEqual(None) https://review.openstack.org/361847 | 01:58 |
| *** shaohe_feng has quit IRC | 02:02 | |
| *** hoangcx has joined #openstack-nova | 02:02 | |
| *** shaohe_feng has joined #openstack-nova | 02:02 | |
| *** vishwanathj has quit IRC | 02:03 | |
| *** EinstCra_ has joined #openstack-nova | 02:03 | |
| *** vishwanathj has joined #openstack-nova | 02:03 | |
| *** EinstCrazy has quit IRC | 02:06 | |
| openstackgerrit | melanie witt proposed openstack/nova: Move cell message queue switching and add caching https://review.openstack.org/326906 | 02:07 |
| darrenc | Could I get a few nova cores to review https://review.openstack.org/#/c/352290/1? Thanks! | 02:09 |
| *** gongysh has quit IRC | 02:11 | |
| *** shaohe_feng has quit IRC | 02:12 | |
| *** shaohe_feng has joined #openstack-nova | 02:13 | |
| *** ijw has quit IRC | 02:14 | |
| *** ijw has joined #openstack-nova | 02:14 | |
| *** ijw has quit IRC | 02:16 | |
| *** ducttape_ has joined #openstack-nova | 02:18 | |
| *** shaohe_feng has quit IRC | 02:22 | |
| *** thorst has joined #openstack-nova | 02:24 | |
| *** shaohe_feng has joined #openstack-nova | 02:25 | |
| *** kevinz has joined #openstack-nova | 02:27 | |
| *** code-R has joined #openstack-nova | 02:27 | |
| *** EinstCrazy has joined #openstack-nova | 02:28 | |
| *** code-R_ has joined #openstack-nova | 02:30 | |
| *** gouthamr_ has quit IRC | 02:30 | |
| *** thorst has quit IRC | 02:31 | |
| *** takashin has joined #openstack-nova | 02:31 | |
| *** amotoki has joined #openstack-nova | 02:31 | |
| *** EinstCra_ has quit IRC | 02:32 | |
| *** magic has joined #openstack-nova | 02:32 | |
| *** Sukhdev has quit IRC | 02:32 | |
| *** magic is now known as Guest33431 | 02:32 | |
| *** shaohe_feng has quit IRC | 02:32 | |
| *** ljxiash has quit IRC | 02:32 | |
| *** shaohe_feng has joined #openstack-nova | 02:33 | |
| *** code-R has quit IRC | 02:33 | |
| *** hoangcx has quit IRC | 02:34 | |
| *** xiaoyang has quit IRC | 02:35 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: api-ref: Fix wrong parameters in os-volumes.inc https://review.openstack.org/347112 | 02:35 |
| *** ljxiash has joined #openstack-nova | 02:35 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Add a hacking rule for string interpolation at logging https://review.openstack.org/334844 | 02:35 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: api-ref: Fix 'id' (attachment_id) parameters https://review.openstack.org/349863 | 02:35 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Fix an error in archiving 'migrations' table https://review.openstack.org/323684 | 02:35 |
| *** ljxiash has quit IRC | 02:36 | |
| *** ljxiash has joined #openstack-nova | 02:36 | |
| *** zerda2 has quit IRC | 02:36 | |
| *** Sukhdev has joined #openstack-nova | 02:37 | |
| *** bdeschenes has quit IRC | 02:39 | |
| *** esmiurium has quit IRC | 02:41 | |
| *** Sukhdev has quit IRC | 02:42 | |
| *** shaohe_feng has quit IRC | 02:43 | |
| *** mlavalle has joined #openstack-nova | 02:43 | |
| *** shaohe_feng has joined #openstack-nova | 02:43 | |
| *** hoangcx has joined #openstack-nova | 02:44 | |
| *** esmiurium has joined #openstack-nova | 02:45 | |
| *** code-R_ has quit IRC | 02:46 | |
| *** armax has quit IRC | 02:46 | |
| *** code-R has joined #openstack-nova | 02:46 | |
| *** shaohe_feng has quit IRC | 02:53 | |
| *** ociuhandu has joined #openstack-nova | 02:53 | |
| *** shaohe_feng has joined #openstack-nova | 02:54 | |
| *** vishwanathj has quit IRC | 02:55 | |
| *** vishwanathj has joined #openstack-nova | 02:55 | |
| *** amotoki has quit IRC | 02:58 | |
| *** amotoki has joined #openstack-nova | 03:00 | |
| *** ducttape_ has quit IRC | 03:02 | |
| *** kaisers_ has joined #openstack-nova | 03:02 | |
| *** Kevin_Zheng has quit IRC | 03:03 | |
| *** shaohe_feng has quit IRC | 03:03 | |
| *** vishwanathj has quit IRC | 03:04 | |
| *** shaohe_feng has joined #openstack-nova | 03:04 | |
| *** vishwanathj has joined #openstack-nova | 03:04 | |
| *** marst has quit IRC | 03:06 | |
| *** ociuhandu has quit IRC | 03:07 | |
| *** daemontool has quit IRC | 03:07 | |
| *** kaisers_ has quit IRC | 03:08 | |
| *** vishwanathj has quit IRC | 03:08 | |
| *** vishwanathj has joined #openstack-nova | 03:08 | |
| *** amotoki has quit IRC | 03:09 | |
| *** daemontool has joined #openstack-nova | 03:11 | |
| *** ijw has joined #openstack-nova | 03:11 | |
| *** yamahata has quit IRC | 03:13 | |
| *** shaohe_feng has quit IRC | 03:13 | |
| *** shaohe_feng has joined #openstack-nova | 03:14 | |
| *** ijw has quit IRC | 03:15 | |
| *** sudipto has joined #openstack-nova | 03:18 | |
| *** sudipto_ has joined #openstack-nova | 03:18 | |
| *** baoli_ has quit IRC | 03:19 | |
| *** vishwanathj is now known as vishwanathj_zzz | 03:19 | |
| *** baoli has joined #openstack-nova | 03:19 | |
| *** marst has joined #openstack-nova | 03:20 | |
| *** tonytan_brb has quit IRC | 03:21 | |
| *** shaohe_feng has quit IRC | 03:24 | |
| *** klkumar has joined #openstack-nova | 03:24 | |
| *** shaohe_feng has joined #openstack-nova | 03:25 | |
| *** vishwanathj_zzz has quit IRC | 03:29 | |
| *** mingyu has joined #openstack-nova | 03:29 | |
| *** vishwanathj_zzz has joined #openstack-nova | 03:29 | |
| *** thorst has joined #openstack-nova | 03:30 | |
| *** TravT has joined #openstack-nova | 03:30 | |
| *** ManishD has joined #openstack-nova | 03:33 | |
| *** shaohe_feng has quit IRC | 03:34 | |
| *** shaohe_feng has joined #openstack-nova | 03:35 | |
| *** thorst has quit IRC | 03:36 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: WIP: Use admin context to create neutron ports https://review.openstack.org/357540 | 03:41 |
| *** vishwanathj_zzz has quit IRC | 03:44 | |
| *** code-R has quit IRC | 03:44 | |
| *** code-R_ has joined #openstack-nova | 03:44 | |
| *** shaohe_feng has quit IRC | 03:44 | |
| *** devkulkarni has joined #openstack-nova | 03:44 | |
| *** vishwanathj_zzz has joined #openstack-nova | 03:44 | |
| *** shaohe_feng has joined #openstack-nova | 03:44 | |
| *** salv-orlando has joined #openstack-nova | 03:45 | |
| *** baoli has quit IRC | 03:48 | |
| *** ljxiash has quit IRC | 03:48 | |
| *** links has joined #openstack-nova | 03:51 | |
| *** vishwanathj_zzz has quit IRC | 03:53 | |
| *** zz_dimtruck is now known as dimtruck | 03:53 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements https://review.openstack.org/362595 | 03:53 |
| *** vishwanathj_zzz has joined #openstack-nova | 03:53 | |
| *** salv-orlando has quit IRC | 03:53 | |
| *** shaohe_feng has quit IRC | 03:54 | |
| *** vishwanathj_zzz has quit IRC | 03:55 | |
| *** armax has joined #openstack-nova | 03:55 | |
| *** vishwanathj_zzz has joined #openstack-nova | 03:55 | |
| *** shaohe_feng has joined #openstack-nova | 03:57 | |
| *** ljxiash has joined #openstack-nova | 03:57 | |
| *** prateek_ has joined #openstack-nova | 03:58 | |
| *** mingyu has quit IRC | 03:58 | |
| *** parora has quit IRC | 04:01 | |
| *** ljxiash has quit IRC | 04:02 | |
| *** fragatina has joined #openstack-nova | 04:02 | |
| *** ducttape_ has joined #openstack-nova | 04:02 | |
| *** shaohe_feng has quit IRC | 04:05 | |
| *** shaohe_feng has joined #openstack-nova | 04:05 | |
| *** garyk1 has joined #openstack-nova | 04:05 | |
| *** raunak has joined #openstack-nova | 04:07 | |
| *** fragatina has quit IRC | 04:07 | |
| *** stewie925 has joined #openstack-nova | 04:08 | |
| *** ducttape_ has quit IRC | 04:09 | |
| *** lixiaoy1 has quit IRC | 04:09 | |
| *** lixiaoy1 has joined #openstack-nova | 04:10 | |
| *** hoangcx has quit IRC | 04:10 | |
| *** code-R_ has quit IRC | 04:10 | |
| *** amotoki has joined #openstack-nova | 04:12 | |
| *** ducttape_ has joined #openstack-nova | 04:13 | |
| *** shaohe_feng has quit IRC | 04:15 | |
| *** ducttape_ has quit IRC | 04:15 | |
| *** shaohe_feng has joined #openstack-nova | 04:16 | |
| *** devkulkarni has quit IRC | 04:17 | |
| ManishD | Good Morning Nova | 04:22 |
| *** tonytan4ever has joined #openstack-nova | 04:22 | |
| *** mlavalle has quit IRC | 04:23 | |
| *** rama_y has joined #openstack-nova | 04:25 | |
| *** shaohe_feng has quit IRC | 04:25 | |
| *** hoangcx has joined #openstack-nova | 04:25 | |
| *** shaohe_feng has joined #openstack-nova | 04:26 | |
| *** tonytan4ever has quit IRC | 04:26 | |
| *** prateek_ has quit IRC | 04:26 | |
| *** code-R has joined #openstack-nova | 04:30 | |
| *** raunak has quit IRC | 04:33 | |
| *** prateek has joined #openstack-nova | 04:34 | |
| *** sudipto_ has quit IRC | 04:34 | |
| *** thorst has joined #openstack-nova | 04:35 | |
| *** sudipto has quit IRC | 04:35 | |
| *** shaohe_feng has quit IRC | 04:35 | |
| *** shaohe_feng has joined #openstack-nova | 04:37 | |
| *** vishwanathj_zzz has quit IRC | 04:38 | |
| *** code-R_ has joined #openstack-nova | 04:38 | |
| *** vishwanathj_zzz has joined #openstack-nova | 04:38 | |
| *** ratailor has joined #openstack-nova | 04:40 | |
| *** code-R has quit IRC | 04:41 | |
| *** thorst has quit IRC | 04:41 | |
| *** salv-orlando has joined #openstack-nova | 04:41 | |
| *** shaohe_feng has quit IRC | 04:43 | |
| *** shaohe_feng has joined #openstack-nova | 04:44 | |
| *** shaohe_feng has quit IRC | 04:46 | |
| *** vishwanathj_zzz has quit IRC | 04:47 | |
| *** shaohe_feng has joined #openstack-nova | 04:47 | |
| *** vishwanathj_zzz has joined #openstack-nova | 04:47 | |
| *** shaohe_feng has quit IRC | 04:49 | |
| *** shaohe_feng has joined #openstack-nova | 04:49 | |
| *** lixiaoy1 has quit IRC | 04:50 | |
| *** vishwanathj_zzz has quit IRC | 04:50 | |
| *** vishwanathj_zzz has joined #openstack-nova | 04:50 | |
| *** kaisers_ has joined #openstack-nova | 04:52 | |
| *** dimtruck is now known as zz_dimtruck | 04:53 | |
| *** shaohe_feng has quit IRC | 04:56 | |
| *** kaisers_ has quit IRC | 04:57 | |
| *** shaohe_feng has joined #openstack-nova | 04:57 | |
| openstackgerrit | Prateek Arora proposed openstack/nova: Delete traces of in-progress snapshot on VM being deleted https://review.openstack.org/294513 | 04:57 |
| *** mingyu has joined #openstack-nova | 04:58 | |
| *** ayoung has quit IRC | 04:58 | |
| *** code-R_ has quit IRC | 05:03 | |
| *** mingyu has quit IRC | 05:03 | |
| *** jdurgin has quit IRC | 05:03 | |
| *** code-R has joined #openstack-nova | 05:03 | |
| *** jdurgin has joined #openstack-nova | 05:04 | |
| *** lixiaoy1 has joined #openstack-nova | 05:04 | |
| *** vishwanathj_zzz has quit IRC | 05:06 | |
| *** Sukhdev has joined #openstack-nova | 05:06 | |
| *** vishwanathj_zzz has joined #openstack-nova | 05:07 | |
| *** sudipto_ has joined #openstack-nova | 05:07 | |
| *** sudipto has joined #openstack-nova | 05:07 | |
| *** jaosorior has joined #openstack-nova | 05:10 | |
| *** vishwanathj_zzz has quit IRC | 05:11 | |
| *** vishwanathj_zzz has joined #openstack-nova | 05:11 | |
| *** Sukhdev has quit IRC | 05:12 | |
| *** ljxiash has joined #openstack-nova | 05:15 | |
| *** vishwanathj_zzz has quit IRC | 05:15 | |
| *** vishwanathj_zzz has joined #openstack-nova | 05:15 | |
| *** lifeless_ is now known as lifeless | 05:15 | |
| *** ducttape_ has joined #openstack-nova | 05:16 | |
| *** vishwanathj_zzz has quit IRC | 05:17 | |
| *** vishwanathj_zzz has joined #openstack-nova | 05:17 | |
| *** unicell has joined #openstack-nova | 05:18 | |
| *** unicell has quit IRC | 05:19 | |
| *** ducttape_ has quit IRC | 05:21 | |
| *** armax has quit IRC | 05:22 | |
| *** unicell has joined #openstack-nova | 05:24 | |
| *** vishwanathj_zzz has quit IRC | 05:26 | |
| *** vishwanathj_zzz has joined #openstack-nova | 05:26 | |
| *** bkopilov has joined #openstack-nova | 05:27 | |
| *** mamandle has joined #openstack-nova | 05:30 | |
| *** sudswas__ has joined #openstack-nova | 05:33 | |
| *** code-R_ has joined #openstack-nova | 05:34 | |
| *** sudipto has quit IRC | 05:35 | |
| *** sudipto has joined #openstack-nova | 05:35 | |
| *** sudipto_ has quit IRC | 05:35 | |
| *** code-R has quit IRC | 05:38 | |
| *** vishwanathj_zzz has quit IRC | 05:38 | |
| *** vishwanathj_zzz has joined #openstack-nova | 05:38 | |
| *** thorst has joined #openstack-nova | 05:39 | |
| *** vishwanathj_zzz has quit IRC | 05:40 | |
| *** sridharg has joined #openstack-nova | 05:40 | |
| *** vishwanathj_zzz has joined #openstack-nova | 05:40 | |
| *** code-R_ has quit IRC | 05:40 | |
| *** kevinz has quit IRC | 05:41 | |
| *** code-R has joined #openstack-nova | 05:41 | |
| *** kevinz has joined #openstack-nova | 05:41 | |
| *** swatson has left #openstack-nova | 05:42 | |
| *** armax has joined #openstack-nova | 05:46 | |
| *** thorst has quit IRC | 05:46 | |
| *** sdake has quit IRC | 05:47 | |
| *** nkrinner_afk is now known as nkrinner | 05:48 | |
| *** sdake has joined #openstack-nova | 05:50 | |
| *** kaisers_ has joined #openstack-nova | 05:56 | |
| *** su_zhang has joined #openstack-nova | 05:56 | |
| *** paul-carlton2 has joined #openstack-nova | 05:58 | |
| *** pbandark has joined #openstack-nova | 05:59 | |
| *** mingyu has joined #openstack-nova | 06:00 | |
| *** mingyu has quit IRC | 06:04 | |
| *** Sukhdev has joined #openstack-nova | 06:06 | |
| *** andreas_s has joined #openstack-nova | 06:07 | |
| *** salv-orl_ has joined #openstack-nova | 06:08 | |
| *** ekuris has joined #openstack-nova | 06:08 | |
| *** kevinz has quit IRC | 06:09 | |
| *** kevinz has joined #openstack-nova | 06:09 | |
| *** salv-orlando has quit IRC | 06:11 | |
| *** salv-orlando has joined #openstack-nova | 06:12 | |
| *** su_zhang has quit IRC | 06:13 | |
| *** su_zhang has joined #openstack-nova | 06:14 | |
| *** salv-orl_ has quit IRC | 06:17 | |
| *** salv-orlando has quit IRC | 06:17 | |
| *** mjura_ has joined #openstack-nova | 06:17 | |
| *** mjura has joined #openstack-nova | 06:17 | |
| *** su_zhang has quit IRC | 06:18 | |
| *** moshele has joined #openstack-nova | 06:19 | |
| *** rcernin has joined #openstack-nova | 06:19 | |
| openstackgerrit | gengchc2 proposed openstack/nova: Fix a typo in the driver.py file https://review.openstack.org/361861 | 06:19 |
| *** ijw has joined #openstack-nova | 06:20 | |
| *** mjura has quit IRC | 06:21 | |
| *** amotoki has quit IRC | 06:22 | |
| *** ijw has quit IRC | 06:25 | |
| openstackgerrit | melanie witt proposed openstack/nova: Move cell message queue switching and add caching https://review.openstack.org/326906 | 06:26 |
| *** pcaruana has joined #openstack-nova | 06:26 | |
| *** paul-carlton2 has quit IRC | 06:28 | |
| *** armax has quit IRC | 06:28 | |
| *** paul-carlton2 has joined #openstack-nova | 06:28 | |
| *** praneshp has quit IRC | 06:32 | |
| ManishD | HI paul-carlton2 | 06:33 |
| *** artom has quit IRC | 06:34 | |
| *** lpetrut has joined #openstack-nova | 06:36 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: Should not be able to sort instances based on joined tables https://review.openstack.org/359134 | 06:37 |
| openstackgerrit | Bhagyashri Shewale proposed openstack/nova: Return 400 error for non-existing snapshot_id https://review.openstack.org/360941 | 06:38 |
| *** ManishD has quit IRC | 06:39 | |
| *** Manish has joined #openstack-nova | 06:39 | |
| *** artom has joined #openstack-nova | 06:44 | |
| *** sdake has quit IRC | 06:44 | |
| *** thorst has joined #openstack-nova | 06:44 | |
| *** amotoki has joined #openstack-nova | 06:46 | |
| *** sudipto has quit IRC | 06:48 | |
| *** sudswas__ has quit IRC | 06:48 | |
| *** Sukhdev has quit IRC | 06:51 | |
| *** liverpooler has joined #openstack-nova | 06:51 | |
| *** thorst has quit IRC | 06:51 | |
| *** markus_z has joined #openstack-nova | 06:52 | |
| *** mingyu has joined #openstack-nova | 06:53 | |
| *** amotoki has quit IRC | 06:56 | |
| *** rama_y has quit IRC | 06:57 | |
| *** sudipto has joined #openstack-nova | 07:00 | |
| *** sudswas__ has joined #openstack-nova | 07:00 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Don't pass argument sqlite_db in method set_defaults https://review.openstack.org/350947 | 07:01 |
| *** tesseract- has joined #openstack-nova | 07:07 | |
| *** sheel has joined #openstack-nova | 07:08 | |
| *** avolkov` has joined #openstack-nova | 07:09 | |
| *** avolkov has quit IRC | 07:09 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova-specs: Abort Cold Migration https://review.openstack.org/334732 | 07:10 |
| *** jlanoux has joined #openstack-nova | 07:13 | |
| *** paul-carlton2 has quit IRC | 07:15 | |
| *** code-R has quit IRC | 07:18 | |
| *** jlanoux has quit IRC | 07:18 | |
| *** ducttape_ has joined #openstack-nova | 07:18 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Return instance of Guest from method write_instance_config https://review.openstack.org/265146 | 07:19 |
| *** jlanoux has joined #openstack-nova | 07:19 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Don't use 'updated_at' to check service's status https://review.openstack.org/297085 | 07:19 |
| *** rmart04 has joined #openstack-nova | 07:20 | |
| *** lpetrut has quit IRC | 07:21 | |
| *** rkrum has quit IRC | 07:21 | |
| *** ducttape_ has quit IRC | 07:23 | |
| *** Dinesh_Bhor has quit IRC | 07:23 | |
| *** Dinesh_Bhor has joined #openstack-nova | 07:24 | |
| *** sameo has joined #openstack-nova | 07:28 | |
| *** paul-carlton2 has joined #openstack-nova | 07:28 | |
| *** matrohon has joined #openstack-nova | 07:33 | |
| openstackgerrit | Prateek Arora proposed openstack/nova: Delete traces of in-progress snapshot on VM being deleted https://review.openstack.org/294513 | 07:37 |
| *** bhagyashri has joined #openstack-nova | 07:46 | |
| *** bhagyashri has quit IRC | 07:47 | |
| *** bhagyashri has joined #openstack-nova | 07:47 | |
| *** thorst has joined #openstack-nova | 07:49 | |
| *** Oku_OS-away is now known as Oku_OS | 07:49 | |
| *** kaisers1 has quit IRC | 07:50 | |
| *** pilgrimstack has joined #openstack-nova | 07:51 | |
| *** kaisers has quit IRC | 07:52 | |
| *** tanee_away is now known as tanee | 07:52 | |
| *** kaisers has joined #openstack-nova | 07:53 | |
| *** kaisers1 has joined #openstack-nova | 07:53 | |
| *** maurosr has joined #openstack-nova | 07:53 | |
| *** cdent has joined #openstack-nova | 07:55 | |
| *** code-R has joined #openstack-nova | 07:55 | |
| openstackgerrit | melanie witt proposed openstack/nova: Move cell message queue switching and add caching https://review.openstack.org/326906 | 07:55 |
| *** paul-carlton2 has quit IRC | 07:56 | |
| *** thorst has quit IRC | 07:57 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** lhinds_ has joined #openstack-nova | 08:02 | |
| *** zzzeek has joined #openstack-nova | 08:02 | |
| *** lhinds_ has left #openstack-nova | 08:04 | |
| *** yassine has joined #openstack-nova | 08:05 | |
| *** yassine is now known as Guest17609 | 08:06 | |
| *** Guest17609 is now known as yassou | 08:06 | |
| *** derekh has joined #openstack-nova | 08:07 | |
| *** kaisers_ has quit IRC | 08:07 | |
| *** kaisers_ has joined #openstack-nova | 08:08 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] Testing latest u-c https://review.openstack.org/303345 | 08:10 |
| *** magic has joined #openstack-nova | 08:18 | |
| *** ducttape_ has joined #openstack-nova | 08:19 | |
| *** magic is now known as Guest99973 | 08:19 | |
| *** willise has joined #openstack-nova | 08:19 | |
| *** phuongnh has joined #openstack-nova | 08:19 | |
| *** Guest33431 has quit IRC | 08:22 | |
| *** tonytan4ever has joined #openstack-nova | 08:23 | |
| *** ducttape_ has quit IRC | 08:23 | |
| *** paul-carlton2 has joined #openstack-nova | 08:26 | |
| *** tonytan4ever has quit IRC | 08:28 | |
| *** roeyc has joined #openstack-nova | 08:28 | |
| openstackgerrit | Andrey Volkov proposed openstack/nova: Add test fixture for cell0 https://review.openstack.org/301864 | 08:32 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs: Support volume_type as a BDM parameter when boot instances https://review.openstack.org/362698 | 08:34 |
| *** zenoway has joined #openstack-nova | 08:34 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs: Support volume_type as a BDM parameter when boot instances https://review.openstack.org/362698 | 08:35 |
| *** RuiChen has joined #openstack-nova | 08:38 | |
| *** tojuvone has left #openstack-nova | 08:40 | |
| *** markd_ has joined #openstack-nova | 08:40 | |
| *** tojuvone_ has joined #openstack-nova | 08:40 | |
| *** moshele has quit IRC | 08:41 | |
| *** moshele has joined #openstack-nova | 08:42 | |
| *** lennyb has quit IRC | 08:42 | |
| bauzas | cdent: hello | 08:43 |
| *** tojuvone_ is now known as tojuvone | 08:43 | |
| johnthetubaguy | bauzas: have you looked at cells and the new context managers, I have a question here I am not sure about: https://review.openstack.org/#/c/328491/7 | 08:44 |
| bauzas | johnthetubaguy: just saw yesterday alaski's patch | 08:45 |
| bauzas | johnthetubaguy: but sure, how can I help ? | 08:45 |
| johnthetubaguy | bauzas: I am curious if my worries are valid or not, basically | 08:45 |
| bauzas | (note: 2 PTO weeks + 1 off-site conference week means that I'm a bit out :p ) | 08:46 |
| johnthetubaguy | Its more a documentation concern really | 08:46 |
| bauzas | johnthetubaguy: okay, lemme review it | 08:46 |
| *** daemontool has quit IRC | 08:46 | |
| johnthetubaguy | bauzas: thanks | 08:46 |
| bauzas | cdent: in case you also want me to help you for the patch, lemme know | 08:46 |
| bauzas | johnthetubaguy: I actually wonder if we could find some other way to automatically add hosts within a cell | 08:49 |
| *** daemontool has joined #openstack-nova | 08:49 | |
| *** antonyfm has joined #openstack-nova | 08:49 | |
| johnthetubaguy | bauzas: long term, maybe, but as an operator I actually quite like being explicit about adding hosts, so we need both options I feel | 08:50 |
| openstackgerrit | jichenjc proposed openstack/nova: Add usage of the allow_resize_same_host https://review.openstack.org/362708 | 08:50 |
| johnthetubaguy | bauzas: its optional to do the conversion at this point, I think? | 08:50 |
| bauzas | johnthetubaguy: well, the problem is that if we don't have a HostMapping, no instances can be booted for some host | 08:53 |
| bauzas | johnthetubaguy: so, say I'm adding host A to cell1 | 08:53 |
| bauzas | johnthetubaguy: then the service will automatically be created | 08:53 |
| bauzas | johnthetubaguy: but there will be no HostMapping item being created, so no way to boot an instance for it | 08:54 |
| johnthetubaguy | bauzas: agreed, thats the problem | 08:54 |
| bauzas | so, here the proposal is to run a discover command, and I think we could merge it for Newton | 08:55 |
| bauzas | but I'd rather prefer to discuss on how to automatically add computes to cells | 08:55 |
| johnthetubaguy | yes, scheduler picks it, possibly, but we fail to know where the host lives, so the boot fails | 08:55 |
| bauzas | that means the compute would need to find out which cell it belongs to | 08:56 |
| bauzas | johnthetubaguy: zactly | 08:56 |
| *** thorst has joined #openstack-nova | 08:56 | |
| johnthetubaguy | so I am OK with the cmd as a workaround for newton | 08:56 |
| bauzas | yup, given we only support one cell atm | 08:57 |
| johnthetubaguy | my problem with the patch, is we need to describe where the user needs to run that cmd, I belive | 08:57 |
| johnthetubaguy | I think it will break if you run it on the child cell | 08:57 |
| johnthetubaguy | but not 100% sure | 08:57 |
| bauzas | sec, verifying it | 08:58 |
| *** zenoway has quit IRC | 08:58 | |
| bauzas | tbh, I don't remember the internals of the MQ switching | 08:58 |
| johnthetubaguy | HostMapping add is what I am thinking about | 08:58 |
| *** cdent has quit IRC | 08:58 | |
| johnthetubaguy | maybe that always defaults to the API db, which all nodes should have, so it works? I am not sure | 08:59 |
| *** zenoway has joined #openstack-nova | 08:59 | |
| bauzas | johnthetubaguy: so , HostMapping.get_by_host reads the API DB | 08:59 |
| *** lixiaoy1 has quit IRC | 09:00 | |
| *** code-R_ has joined #openstack-nova | 09:00 | |
| *** links has quit IRC | 09:00 | |
| *** thorst has quit IRC | 09:02 | |
| johnthetubaguy | and we need that on all nodes anyways? I guess we do | 09:02 |
| johnthetubaguy | bauzas: ^ | 09:02 |
| *** code-R has quit IRC | 09:03 | |
| bauzas | johnthetubaguy: HostMapping is listing all hosts indeed | 09:04 |
| bauzas | that's the routing table | 09:04 |
| *** jlanoux has quit IRC | 09:05 | |
| bauzas | johnthetubaguy: tbc, my wonders are, given we're within a cell, is it possible to call out the API DB session ? | 09:05 |
| *** tuanla_ has joined #openstack-nova | 09:05 | |
| bauzas | johnthetubaguy: that's the whole purpose of the MQ/DB context switch | 09:05 |
| johnthetubaguy | I think it is in theory possible, if configured correctly | 09:05 |
| johnthetubaguy | I think we are trying to avoid such things though | 09:05 |
| *** lpetrut has joined #openstack-nova | 09:06 | |
| *** jaosorior is now known as jaosorior_lunch | 09:08 | |
| prateek | johnthetubaguy, can you please review https://review.openstack.org/294513 , whenever you have some free time | 09:08 |
| johnthetubaguy | prateek: fixated on all the priority features right now, hopefully get more time other end of this week | 09:09 |
| prateek | johnthetubaguy, thanks! i would get in touch EOW | 09:10 |
| bauzas | johnthetubaguy: agreed, we really try to refrain our cells world to call northbound | 09:11 |
| *** hamzy has quit IRC | 09:11 | |
| bauzas | johnthetubaguy: honestly, I'm torn | 09:11 |
| *** hamzy has joined #openstack-nova | 09:11 | |
| johnthetubaguy | bauzas: anyways, can catch up with the others when they are awake, was just curious | 09:12 |
| bauzas | johnthetubaguy: you have a valid point, but we should for Newton just mention where to run, not really implement | 09:12 |
| openstackgerrit | Claudiu Belu proposed openstack/nova: WIP: api: discoverable policy https://review.openstack.org/280539 | 09:13 |
| bauzas | johnthetubaguy: well it's more or less us agreeing about if we should be defensive with our patches or just explaining to operators where they need to run commands | 09:13 |
| *** bdeschenes has joined #openstack-nova | 09:14 | |
| bauzas | AFAICR, we have a couple of cell mng commands that are not preventing operators to trample their deployments if they run it intra-cell | 09:14 |
| *** links has joined #openstack-nova | 09:15 | |
| johnthetubaguy | bauzas: I think this is safe, if it fails how I think it does, just a likely a strange error message | 09:17 |
| *** ducttape_ has joined #openstack-nova | 09:19 | |
| *** jlanoux has joined #openstack-nova | 09:20 | |
| bauzas | johnthetubaguy: I just commented out | 09:20 |
| bauzas | johnthetubaguy: for Newton, given the very experimental level, I'd say docs is fine | 09:20 |
| bauzas | we don't yet support multi-cells | 09:20 |
| bauzas | but in Ocata, once multi-cells will become a reality, then indeed we should somehow fail gently and warn | 09:21 |
| bauzas | johnthetubaguy: anyway, given we have a lot of Cellv2Commands that are based on that assumption, it's more or less finding some way to wrap something around that would say "meh, your fault, run it elsewhere" | 09:22 |
| *** tbachman_ has joined #openstack-nova | 09:22 | |
| *** tbachman has quit IRC | 09:22 | |
| *** tbachman_ is now known as tbachman | 09:22 | |
| *** ducttape_ has quit IRC | 09:24 | |
| *** cdent has joined #openstack-nova | 09:27 | |
| *** jlanoux_ has joined #openstack-nova | 09:28 | |
| *** nkrinner has quit IRC | 09:29 | |
| *** jlanoux has quit IRC | 09:31 | |
| *** lpetrut1 has joined #openstack-nova | 09:34 | |
| *** lpetrut has quit IRC | 09:34 | |
| *** lpetrut1 is now known as lpetrut | 09:34 | |
| *** nkrinner has joined #openstack-nova | 09:35 | |
| *** ljxiash has quit IRC | 09:36 | |
| *** ljxiash has joined #openstack-nova | 09:37 | |
| *** madhaviy has joined #openstack-nova | 09:38 | |
| madhaviy | hi there, I want to create a server using admin token by passing key_name( nova keypair) of a some other user (Imported by some other user) | 09:40 |
| *** sahid has joined #openstack-nova | 09:40 | |
| madhaviy | is this possible? | 09:40 |
| *** ljxiash has quit IRC | 09:41 | |
| *** jaosorior_lunch is now known as jaosorior | 09:42 | |
| *** lennyb has joined #openstack-nova | 09:43 | |
| *** rwmjones_hols is now known as rwmjones | 09:43 | |
| *** oanson has joined #openstack-nova | 09:46 | |
| *** kaisers_ has quit IRC | 09:48 | |
| *** RuiChen has quit IRC | 09:49 | |
| bhagyashri | rpodolyaka: Hi, could you please review https://review.openstack.org/#/c/345191/ ? | 09:51 |
| *** roeyc has quit IRC | 09:55 | |
| *** thorst has joined #openstack-nova | 09:58 | |
| *** ihrachys has joined #openstack-nova | 10:00 | |
| *** liuyulong is now known as yulong|away | 10:00 | |
| *** jlanoux has joined #openstack-nova | 10:00 | |
| *** jlanoux_ has quit IRC | 10:03 | |
| *** GB21 has joined #openstack-nova | 10:03 | |
| *** amotoki_ has joined #openstack-nova | 10:03 | |
| *** thorst has quit IRC | 10:05 | |
| *** mnestratov has joined #openstack-nova | 10:08 | |
| *** wangqun has quit IRC | 10:08 | |
| *** daemontool has quit IRC | 10:09 | |
| *** daemontool has joined #openstack-nova | 10:09 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] Testing latest u-c https://review.openstack.org/303345 | 10:10 |
| *** hoangcx has quit IRC | 10:19 | |
| *** ducttape_ has joined #openstack-nova | 10:20 | |
| *** sridharg has quit IRC | 10:22 | |
| *** amotoki has joined #openstack-nova | 10:23 | |
| *** gszasz has joined #openstack-nova | 10:24 | |
| *** tonytan4ever has joined #openstack-nova | 10:24 | |
| *** amotoki_ has quit IRC | 10:25 | |
| *** ducttape_ has quit IRC | 10:25 | |
| *** salv-orlando has joined #openstack-nova | 10:27 | |
| *** tonytan4ever has quit IRC | 10:28 | |
| *** EinstCrazy has quit IRC | 10:31 | |
| *** EinstCrazy has joined #openstack-nova | 10:31 | |
| openstackgerrit | Bhagyashri Shewale proposed openstack/python-novaclient: Fix 'UnicodeEncodeError' for unicode values https://review.openstack.org/346026 | 10:32 |
| openstackgerrit | Chris Dent proposed openstack/nova: Optional separate database for placement API https://review.openstack.org/362766 | 10:33 |
| cdent | johnthetubaguy: ^ | 10:33 |
| cdent | that's the first section of https://etherpad.openstack.org/p/placement-db | 10:33 |
| *** amotoki_ has joined #openstack-nova | 10:33 | |
| johnthetubaguy | cdent: ah, good stuff, I will try get to that straight after lunch | 10:34 |
| cdent | cool, thanks | 10:34 |
| *** roeyc has joined #openstack-nova | 10:35 | |
| *** amotoki has quit IRC | 10:36 | |
| *** EinstCrazy has quit IRC | 10:36 | |
| *** phuongnh has quit IRC | 10:39 | |
| *** bdeschenes has quit IRC | 10:39 | |
| prateek | cdent, if you have some free time, can you please let me know what you mean what you meant by the comment on "This test doesn't test anything other than itself." https://review.openstack.org/#/c/327706/8/nova/tests/unit/api/openstack/test_wsgi.py | 10:42 |
| *** lixiaoy1 has joined #openstack-nova | 10:42 | |
| cdent | prateek: yeah, let me look at it again to remind myself | 10:42 |
| prateek | cdent, ok | 10:43 |
| cdent | prateek: what I mean is that no code outside of that test is actually called by the test. All it is testing is that your Controller raises an exception, which, to me at least, is useless. Or am I missing something? | 10:44 |
| cdent | brb | 10:45 |
| prateek | cdent, no you didn't , the controller was supposed to raise an exception and then a 504 was meant to be returned | 10:46 |
| prateek | and i messed up royally ! | 10:46 |
| prateek | cdent, i am wondering how test_resource_not_authorized which is on top of test i wrote got passed, it does the exactly same mistake i have done | 10:47 |
| cdent | hmmm, yeah. :( | 10:48 |
| *** sridharg has joined #openstack-nova | 10:48 | |
| *** GB21 has quit IRC | 10:49 | |
| cdent | I think there are probably a lot of tests like that | 10:49 |
| *** GB21 has joined #openstack-nova | 10:50 | |
| prateek | cdent, yeah would need a debugger to find out what actually is happening behind the scenes, would get back to you with how it works, thanks for your help | 10:51 |
| *** salv-orl_ has joined #openstack-nova | 10:52 | |
| *** sdague has joined #openstack-nova | 10:53 | |
| bauzas | cdent: as I said earlier, in case you need help for the etherpad implementation, lemme know | 10:54 |
| *** salv-orl_ has quit IRC | 10:54 | |
| cdent | bauzas: the first section is ready for review, after I get another coffee I'm going to settle in to do the additional database table, so I think it is under control but review on the first section would be very welcome: https://review.openstack.org/362766 | 10:55 |
| *** salv-orlando has quit IRC | 10:56 | |
| *** sridharg has quit IRC | 10:56 | |
| *** sridharg has joined #openstack-nova | 10:57 | |
| johnthetubaguy | cdent: given bad code is the only reason for that test to fail, it feels OK-ish to me. We have seen things were the raise or log hides the real exception, and it does help avoid that. I am curious if there is a better way for that though, they do feel lame. | 11:00 |
| cdent | johnthetubaguy: you mean it is testing the exception itself? | 11:01 |
| johnthetubaguy | cdent: testing that the code correctly raises the correct exception | 11:02 |
| cdent | it's not doing that | 11:02 |
| johnthetubaguy | oh... | 11:02 |
| cdent | it's testing _no_ code outside the test, except for anything that is called within the exception itself | 11:02 |
| johnthetubaguy | I see what you mean, its mocking a think and testing it raises | 11:03 |
| cdent | yes | 11:03 |
| johnthetubaguy | hmm, that does see wrong | 11:03 |
| cdent | we know from that test that raising a NeutronConnectionFailed() is possible, but in a complicated and not very robust way. That's all. | 11:03 |
| *** tbachman_ has joined #openstack-nova | 11:03 | |
| *** thorst has joined #openstack-nova | 11:03 | |
| johnthetubaguy | cdent: yeah | 11:04 |
| johnthetubaguy | cdent: I was looking at the wrong code, my bad | 11:05 |
| * cdent coffees up | 11:05 | |
| johnthetubaguy | cdent: justifies your comments about the framework being so opaque really | 11:05 |
| *** tbachman has quit IRC | 11:05 | |
| *** tbachman_ is now known as tbachman | 11:05 | |
| *** wznoinsk has quit IRC | 11:10 | |
| *** adisky has joined #openstack-nova | 11:10 | |
| *** cdent has quit IRC | 11:10 | |
| *** thorst has quit IRC | 11:11 | |
| *** hoangcx has joined #openstack-nova | 11:11 | |
| *** rtheis has joined #openstack-nova | 11:12 | |
| *** trisq has quit IRC | 11:12 | |
| *** cdent has joined #openstack-nova | 11:13 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: WIP: Enable cold migration with target host https://review.openstack.org/357105 | 11:13 |
| *** hoangcx2 has joined #openstack-nova | 11:15 | |
| cdent | sdague: in case you haven't seen it yet, I've updated the placement devstack stuff: https://review.openstack.org/#/c/342362/ | 11:16 |
| sdague | yep, I just pulled it down and am building the test env here | 11:16 |
| cdent | \o/ | 11:17 |
| *** hoangcx has quit IRC | 11:17 | |
| sdague | so far so good. It's odd to look at a service log without all the oslo.service dump in it though :) | 11:18 |
| *** ducttape_ has joined #openstack-nova | 11:21 | |
| *** ducttape_ has quit IRC | 11:26 | |
| *** bvanhav has joined #openstack-nova | 11:27 | |
| ratailor | can anyone please review it, https://review.openstack.org/#/c/312887/ | 11:28 |
| *** wznoinsk has joined #openstack-nova | 11:28 | |
| *** sudipto has quit IRC | 11:32 | |
| *** sudswas__ has quit IRC | 11:32 | |
| sdague | cdent: ok, so your devstack patch creates the endpoints | 11:32 |
| * cdent nods | 11:32 | |
| sdague | but it doesn't create the placement accounts, or put that info in the nova.conf | 11:32 |
| sdague | unless I'm missing something | 11:33 |
| cdent | "the placement accounts"? | 11:33 |
| cdent | create_service_user "placement" is in there | 11:33 |
| sdague | where are you adding the credentials to the nova.conf ? | 11:33 |
| cdent | I'm not, I don't even know what that means | 11:33 |
| *** amotoki_ has quit IRC | 11:33 | |
| sdague | https://review.openstack.org/#/c/358797/5/nova/conf/placement.py@37 | 11:34 |
| sdague | if we're going to talk to another service over rest | 11:34 |
| sdague | we need service user credentials | 11:34 |
| cdent | yeah, I've questioned multiple times why that stuff is even necessary, but nobody has taken the time to explain it to me | 11:34 |
| cdent | I was under the impression that a compute node knows its region so why do we need to have it in config | 11:35 |
| cdent | why can't we just ask keystone, and be done with it? | 11:35 |
| openstackgerrit | John Garbutt proposed openstack/nova: Api-ref: Improve os-migrateLive input parameters https://review.openstack.org/345652 | 11:35 |
| cdent | there are _so_many_ assumptions made by this crowd | 11:35 |
| sdague | http://logs.openstack.org/61/360761/1/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/25c13b2/logs/etc/nova/nova.conf.txt.gz | 11:35 |
| sdague | a compute node doesn't know it's region | 11:36 |
| cdent | weird | 11:36 |
| cdent | okay | 11:36 |
| sdague | but, regardless of that, it needs auth | 11:36 |
| *** amotoki has joined #openstack-nova | 11:36 | |
| *** Manish has quit IRC | 11:36 | |
| sdague | see that neutron section in the conf file | 11:36 |
| cdent | okay, but either: explain it to me like I'm five, or perhaps just fix it and then I'll know? | 11:37 |
| *** kaisers_ has joined #openstack-nova | 11:37 | |
| cdent | what I think you're going to tell me or show me is just going to make me cry | 11:37 |
| sdague | ok... where do you expect nova to get credentials to talk to another service? | 11:38 |
| *** thorst has joined #openstack-nova | 11:38 | |
| sdague | to generate a valid keystone token | 11:38 |
| cdent | I have no expectations about that, that's what I'm trying to tell you. | 11:38 |
| cdent | It is not something I've had to deal with, yet. | 11:38 |
| *** thorst_ has joined #openstack-nova | 11:39 | |
| sdague | ok, well when services are talking to each other over REST, we have 2 options | 11:39 |
| cdent | (as you're well aware the volume of tasks are such that most stuff just has to get dropped on the floor until it is critical) | 11:40 |
| sdague | if the call chain was user initiated, we can use the user's keystone token that started the process. This is how we talk to cinder, or glance, for instance | 11:40 |
| sdague | if it is not | 11:40 |
| sdague | then we need auth credentials for a service user in our config | 11:40 |
| sdague | so that we can make calls, not on behalf of a user, at any time | 11:40 |
| sdague | which we need to do with neutron, and hence the auth block there | 11:41 |
| sdague | https://review.openstack.org/#/c/358797/5/nova/conf/placement.py@37 - is jay enabling that for communication to the placement api | 11:41 |
| *** kaisers_ has quit IRC | 11:41 | |
| sdague | but it means we need the values in the config file, otherwise we're making calls with no auth, and they all explode | 11:41 |
| *** thorst has quit IRC | 11:42 | |
| cdent | I guess the surprise/crying comes in because of the need for this to be explicit rather than there being reasonable generic defaults which each service could override if it wanted to, because for the most part that looks like boilerplate | 11:43 |
| cdent | (the neutron chunk is what I mean by boilerplate) | 11:43 |
| openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix notify instance usage on evacuate https://review.openstack.org/362794 | 11:43 |
| cdent | sdague: in the context of this stuff, what difference does it make that currently the placement API's only auth handling is that the user has the admin role: https://github.com/openstack/nova/blob/master/nova/api/openstack/placement/handler.py#L140 ? | 11:46 |
| sdague | cdent: most of those things need to be pretty explicit during a real deploy, so if we aren't explicit about them, you end up encoding devstack defaults in nova code | 11:46 |
| sdague | cdent: that's the server | 11:47 |
| sdague | it still needs a valid keystone token to even get there | 11:47 |
| sdague | were would that come from? | 11:47 |
| cdent | sdague: yes, I know that. I'm just saying: is a 'create[d]_server_user' going to have that role? | 11:47 |
| *** lucasagomes is now known as lucas-hungry | 11:47 | |
| cdent | s/server/service/ | 11:48 |
| sdague | cdent: https://github.com/openstack-dev/devstack/blob/ecfc8a77c203c7832b8b7542342cbb8e258a7765/lib/keystone#L455-L467 | 11:49 |
| *** takashin has left #openstack-nova | 11:49 | |
| cdent | yes, I know how to read the code. What I'm asking is "hey, is this thing about the way the placement handles auth (either in devstack or on the server side) going to be an issue, do we need to worry about that?" | 11:50 |
| sdague | cdent: for right now, I think you just need to change it to create_service_user "placement" "admin" | 11:50 |
| sdague | at some point we're going to get all the service cross calls to work on the "service" role instead of the "admin" role, but that work has not yet started | 11:51 |
| sdague | others, I think checking role the way you did it should be fine | 11:51 |
| *** tblakes has joined #openstack-nova | 11:51 | |
| cdent | okay, I've made a note. Can you point me at a good example for correctly setting the conf for auth (in the [placement] section, right?)? | 11:53 |
| jroll | is there a reason not to build in using the service role and policy.json from the start, or is it just time constraints? | 11:53 |
| cdent | jroll: it was mostly lack of time and lack of clear specification. There was talk of it happening, but it was struck off as a not immediate requirement. If you've got cycles though, please have at :) | 11:55 |
| cdent | (seriously, not being snarky, your help would be very welcome) | 11:55 |
| sdague | cdent: https://github.com/openstack-dev/devstack/blob/ecfc8a77c203c7832b8b7542342cbb8e258a7765/lib/neutron#L292-L312 | 11:55 |
| jroll | cdent: okay cool, just curious | 11:55 |
| sdague | that's where neutron configures nova | 11:55 |
| cdent | thanks sdague | 11:55 |
| sdague | jroll: we also kind of need to synchronize the role downgrade to service at least across base iaas | 11:56 |
| sdague | otherwise it's going to be kick you in the knees confusing to ops | 11:56 |
| cdent | I've added notes to the review and -W'd myself, will get to that in a few minutes, currently adding the database tables. thanks sdague | 11:57 |
| *** rfolco has joined #openstack-nova | 11:57 | |
| jroll | sdague: fair enough, we did the move when we changed our policy to be more than admin or nothing | 11:58 |
| *** efried has joined #openstack-nova | 11:58 | |
| sdague | cdent: yep, no prob. I think I can sketch using the keystone auth correctly in parallel, and then hope it all comes together | 11:58 |
| *** mingyu has quit IRC | 12:00 | |
| *** pradk has joined #openstack-nova | 12:02 | |
| *** jaosorior has quit IRC | 12:04 | |
| *** mamandle has quit IRC | 12:04 | |
| *** jaosorior has joined #openstack-nova | 12:04 | |
| *** ducttape_ has joined #openstack-nova | 12:06 | |
| *** vladikr has joined #openstack-nova | 12:06 | |
| tdurakov | hi folks | 12:11 |
| *** ttrumm has joined #openstack-nova | 12:11 | |
| *** jaosorior has quit IRC | 12:11 | |
| *** jaosorior has joined #openstack-nova | 12:12 | |
| tdurakov | who could check one thing: on master/devstack with nova-network will try to boot instance without passing net-id | 12:12 |
| *** tonytan4ever has joined #openstack-nova | 12:13 | |
| *** clenimar has quit IRC | 12:15 | |
| *** gouthamr has joined #openstack-nova | 12:17 | |
| sdague | cdent - just so this is somewhere else than in my head - http://paste.openstack.org/show/564819/ is the minimal way to use keystoneauth as a rest client, which does all the token bits behind the scenes. That's on my devstack using the random admin user we setup by default. | 12:23 |
| *** kswiatek has joined #openstack-nova | 12:23 | |
| openstackgerrit | liuxiuli proposed openstack/nova: Modify workflow of rebuild_instance when schedule fails https://review.openstack.org/348111 | 12:24 |
| cdent | sdague: thanks was looking at when starting the python-placementclient code (which is far from done at this point) | 12:24 |
| cdent | I struggled to figure out "what's the bare minimum", so this is great | 12:25 |
| sdague | the keystone session interface is basically a wrapper on requests, with a couple of extra parameters to do selection of endpoint | 12:25 |
| cdent | yeah, this was another one of those situations where I was expecting some kind of reasonable defaulting (to environment variables in this case) | 12:29 |
| cdent | (and maybe it does, but the docs were not at all clear) | 12:29 |
| sdague | ok, going to get another cup of coffee, how is the devstack how is the devstack patch coming? I think I have a *should barely work* version of this locally now. But need some coffee / breakfast anyway. | 12:31 |
| *** mdrabe has joined #openstack-nova | 12:31 | |
| *** su_zhang has joined #openstack-nova | 12:31 | |
| *** ducttape_ has quit IRC | 12:32 | |
| cdent | sdague: I'm just testing it is okay, then ready to push | 12:32 |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Add volume status check when deleting instances https://review.openstack.org/260339 | 12:33 |
| *** tuanla_ has quit IRC | 12:34 | |
| cdent | where do tests for api migrations live? | 12:35 |
| cdent | s/api/api db/ | 12:36 |
| cdent | ah found em, in functional | 12:39 |
| *** lucas-hungry is now known as lucasagomes | 12:41 | |
| *** sudipto has joined #openstack-nova | 12:42 | |
| *** sudswas__ has joined #openstack-nova | 12:42 | |
| *** baoli has joined #openstack-nova | 12:44 | |
| *** EinstCrazy has joined #openstack-nova | 12:44 | |
| kswiatek | sdague, pkoniszewski, melwitt, mdbooth, could you review: https://review.openstack.org/#/c/314928/ | 12:45 |
| alaski | johnthetubaguy: bauzas responded on https://review.openstack.org/#/c/328491 | 12:47 |
| *** annegentle has joined #openstack-nova | 12:48 | |
| *** cdelatte has joined #openstack-nova | 12:49 | |
| *** gouthamr_ has joined #openstack-nova | 12:49 | |
| *** gouthamr has quit IRC | 12:51 | |
| cdent | sdague: just pushed new version of the devstack stuff | 12:52 |
| *** markvoelker has joined #openstack-nova | 12:52 | |
| bauzas | alaski: cool, will look into it asap | 12:53 |
| *** tbachman has quit IRC | 12:53 | |
| *** claudiub has joined #openstack-nova | 12:54 | |
| bauzas | alaski: okay, so basically, we have the same opinion | 12:54 |
| *** roeyc has quit IRC | 12:54 | |
| bauzas | alaski: I'm okay with merging this one for Newton, and then discussing for Ocata about providing better ways | 12:54 |
| *** roeyc has joined #openstack-nova | 12:55 | |
| bauzas | I mean, providing a better warning in case the operator runs it in the child cell, and also discuss about how it could be possible to leave a northbound call for computes adding automatically to the cell | 12:55 |
| bauzas | johnthetubaguy: you okay with that? | 12:55 |
| * johnthetubaguy is looking | 12:56 | |
| bauzas | johnthetubaguy: tbc, like I said, all our nova-manage cmds for cells v2 have the same issue if the operator runs them in a child cell host | 12:56 |
| openstackgerrit | Ludovic Beliveau proposed openstack/nova: Don't parse PCI whitelist eveytime neutron ports are created https://review.openstack.org/337236 | 12:56 |
| *** xinliang has quit IRC | 12:57 | |
| *** trisq has joined #openstack-nova | 12:57 | |
| johnthetubaguy | alaski: thanks for the response, I am +1 your comments on that patch | 12:58 |
| johnthetubaguy | alaski: just a slight docs tweak is good with me | 12:58 |
| *** madhaviy has quit IRC | 12:59 | |
| *** edand has joined #openstack-nova | 12:59 | |
| *** clenimar has joined #openstack-nova | 13:00 | |
| bauzas | johnthetubaguy: we can possibly add a follow-up reno change for adding like a comment saying 'please run all the nova-manage commands for cellsv2 at the API level" | 13:01 |
| bauzas | johnthetubaguy: but tbc, it's the same problem for all our already existing calls | 13:01 |
| johnthetubaguy | I guess most of them do the same, thats true | 13:01 |
| bauzas | so, maybe adding a prelude section is okay by me | 13:02 |
| *** sbezverk_ has quit IRC | 13:02 | |
| bauzas | like we did for Mitaka | 13:02 |
| johnthetubaguy | alaski: bauzas: how about we put all the tweaks in a follow up patch, to make them all consistent (somehow)? | 13:02 |
| bauzas | johnthetubaguy: you mean a doc thing ? | 13:02 |
| *** xyang1 has joined #openstack-nova | 13:02 | |
| bauzas | johnthetubaguy: or more something like an helper decorator (or something else) for all the commands, verifying whether we can access the API DB, and if not, fail gracefully? | 13:04 |
| *** gcb has quit IRC | 13:04 | |
| *** ttrumm has quit IRC | 13:04 | |
| *** GB21 has quit IRC | 13:04 | |
| johnthetubaguy | bauzas: I am fine with just a doc thing for this cycle | 13:04 |
| *** ttrumm has joined #openstack-nova | 13:04 | |
| *** mamandle has joined #openstack-nova | 13:05 | |
| *** abalutoiu has joined #openstack-nova | 13:06 | |
| *** su_zhang has quit IRC | 13:07 | |
| *** su_zhang has joined #openstack-nova | 13:07 | |
| sdague | cdent: ok, cool, just working out some of the expected exceptions from the keystone auth side, let me see if the happy path works now | 13:08 |
| sdague | cdent: resource reporting is every 60 seconds? | 13:09 |
| *** ttrumm has quit IRC | 13:09 | |
| cdent | sdague, that's the default yeah | 13:09 |
| *** rodrigods has quit IRC | 13:09 | |
| *** rodrigods has joined #openstack-nova | 13:09 | |
| *** david-lyle has quit IRC | 13:10 | |
| bauzas | sdague: cdent: well that depends | 13:10 |
| bauzas | sdague: cdent: resource reporting can be more often, tbc it's every time we claim for an instance | 13:10 |
| sdague | bauzas: ok, so if I trigger a server build, it will happen immediately? | 13:11 |
| bauzas | by default, we periodically make a whole new global update every 60 secs, that's true but we also amend the host every time an instance is claimed | 13:11 |
| bauzas | sdague: yup, if the claim is successful in the compute manager | 13:11 |
| *** adisky has quit IRC | 13:11 | |
| sdague | great | 13:11 |
| *** su_zhang has quit IRC | 13:11 | |
| *** antonyfm has quit IRC | 13:11 | |
| sdague | gives me enough to tickle this | 13:11 |
| *** ducttape_ has joined #openstack-nova | 13:12 | |
| sdague | cdent: fyi, keystone auth gives us some reasonable exceptions to distinguish the failure cases we cared about - http://paste.openstack.org/show/564830/ | 13:12 |
| cdent | excellent smithers | 13:13 |
| *** rlrossit has joined #openstack-nova | 13:13 | |
| sdague | as I was trying to handle the non happy path, which I did with glance instead, because with placement api in keystone, dropping it for 30 seconds isn't as simple in spot testing :) | 13:13 |
| *** sheel has quit IRC | 13:16 | |
| *** markvoelker has quit IRC | 13:17 | |
| alex_xu | johnthetubaguy: we didn't create migration obj for rebuild action, if we fail at that check, the rebuild action will be broken https://review.openstack.org/#/c/348111/9/nova/conductor/manager.py@648 | 13:17 |
| *** ducttape_ has quit IRC | 13:17 | |
| *** dinobot_ is now known as dakhmetov | 13:17 | |
| *** david-lyle has joined #openstack-nova | 13:18 | |
| *** mlavalle has joined #openstack-nova | 13:18 | |
| alex_xu | johnthetubaguy: from the comment at here https://review.openstack.org/#/c/200485/28/nova/conductor/manager.py@898, looks like Nikola's plan is passing migration from compute api. Maybe that was forgetten to change in the end. | 13:19 |
| *** jamesdenton has joined #openstack-nova | 13:20 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 13:20 |
| openstackgerrit | Claudiu Belu proposed openstack/nova: Adds Hyper-V OVS ViF driver https://review.openstack.org/140045 | 13:21 |
| johnthetubaguy | alex_xu: I just think the log message is bad, and we should add a comment why we don't error out there | 13:22 |
| johnthetubaguy | alex_xu: passing down the migrate object would be a much better idea log term, maybe we could add that as a TODO? | 13:22 |
| johnthetubaguy | alex_xu: my worry was really the save_and_reraise_exception decorator | 13:23 |
| *** bdeschenes has joined #openstack-nova | 13:23 | |
| *** GB21 has joined #openstack-nova | 13:23 | |
| *** daemontool has quit IRC | 13:25 | |
| openstackgerrit | Aline Bousquet proposed openstack/nova-specs: Spec for supporting authentication methods based on image metadata https://review.openstack.org/326073 | 13:26 |
| *** daemontool has joined #openstack-nova | 13:27 | |
| *** mriedem has joined #openstack-nova | 13:30 | |
| *** mmedvede_ has joined #openstack-nova | 13:30 | |
| *** dakhmetov has quit IRC | 13:30 | |
| *** dinobot_ has joined #openstack-nova | 13:30 | |
| *** mmedvede has quit IRC | 13:30 | |
| *** mmedvede_ has quit IRC | 13:30 | |
| *** dinobot_ is now known as dakhmetov | 13:30 | |
| mriedem | is danpb out? | 13:30 |
| *** luis5tb has joined #openstack-nova | 13:31 | |
| *** EinstCrazy has quit IRC | 13:31 | |
| *** dakhmetov is now known as dinobot_ | 13:31 | |
| *** dinobot_ is now known as dakhmetov | 13:31 | |
| *** su_zhang has joined #openstack-nova | 13:31 | |
| mriedem | moshele: are you aware of any changes in os-vif that need to be released for newton? | 13:31 |
| mriedem | anything needed for a pending feature? | 13:32 |
| mriedem | sahid: russellb: ^ | 13:32 |
| *** pwnall has joined #openstack-nova | 13:32 | |
| mriedem | https://review.openstack.org/#/q/project:openstack/os-vif+status:open | 13:32 |
| *** sbezverk has joined #openstack-nova | 13:33 | |
| mriedem | like maybe https://review.openstack.org/#/c/344433/ | 13:33 |
| *** abhishekk has joined #openstack-nova | 13:33 | |
| mriedem | i'm asking b/c the final release for os-vif in newton was supposed to be last week and the release team wants to cut the stable/newton branch | 13:34 |
| *** lixiaoy1 has quit IRC | 13:34 | |
| alex_xu | johnthetubaguy: yea, agree with a TODO | 13:34 |
| *** tbachman has joined #openstack-nova | 13:36 | |
| cdent | sdague: could you help me with something? This failure should only happen if placement_database.connection is not None. Is something strange about CONF in that environment: http://logs.openstack.org/66/362766/1/check/gate-nova-python27-db-ubuntu-xenial/e4aa915/console.html#_2016-08-30_10_44_42_866978 on https://review.openstack.org/#/c/362766/ | 13:38 |
| *** markvoelker has joined #openstack-nova | 13:40 | |
| *** pwnall has quit IRC | 13:40 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 13:40 |
| *** pwnall has joined #openstack-nova | 13:40 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 13:41 |
| *** tbachman has quit IRC | 13:41 | |
| *** mmedvede has joined #openstack-nova | 13:41 | |
| *** ManishD has joined #openstack-nova | 13:42 | |
| mriedem | cdent: my guess would be the issue is in https://review.openstack.org/#/c/362766/1/nova/db/sqlalchemy/migration.py@177 | 13:42 |
| mriedem | the api migrations are being run twice | 13:42 |
| mriedem | once for the api db and once for the placement db | 13:42 |
| openstackgerrit | Claudiu Belu proposed openstack/os-vif: Adds NoopPlugin https://review.openstack.org/362876 | 13:43 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Removed enum duplication from nova.compute https://review.openstack.org/335066 | 13:43 |
| cdent | mriedem: yeah, but why would the test fail on that job but not locally? | 13:43 |
| cdent | (with just tox -epy27 <the test> ?) | 13:44 |
| cdent | mriedem: and whatever is happening how to fix it? :) | 13:45 |
| *** mgoddard_ has joined #openstack-nova | 13:45 | |
| *** markvoelker has quit IRC | 13:45 | |
| mriedem | good question | 13:46 |
| * cdent does a full run | 13:46 | |
| *** annegentle has quit IRC | 13:46 | |
| *** hoangcx2 has quit IRC | 13:47 | |
| *** burt has joined #openstack-nova | 13:48 | |
| *** zz_dimtruck is now known as dimtruck | 13:48 | |
| *** mgoddard has quit IRC | 13:48 | |
| *** thorst_ is now known as thorst | 13:48 | |
| openstackgerrit | Sean Dague proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/358797 | 13:49 |
| openstackgerrit | Sean Dague proposed openstack/nova: WIP: Totally hacky happy path for placement API working. https://review.openstack.org/362883 | 13:49 |
| sdague | mriedem: ok, so I basically didn't change jay's patch, though there is a rebase in there I guess | 13:49 |
| moshele | mriedem: I am not aware, but I wasn't much involved in the os-vif, so it better to ask danpb | 13:49 |
| tdurakov | cdent: I'd try to clear .tox/py27 for example | 13:49 |
| *** daemontool has quit IRC | 13:49 | |
| sdague | I was working on top of his patch to get some things working, and the happy path looks ok | 13:49 |
| mriedem | moshele: danpb appears to be out | 13:49 |
| mriedem | moshele: hence me pinging the entire os-vif core team | 13:50 |
| sdague | I actually got REST talking to placement API without horrible errors | 13:50 |
| *** daemontool has joined #openstack-nova | 13:50 | |
| tdurakov | paul-carlton2 eliqiao shaohe_feng pkoniszewski jlanoux yuntongjin mdbooth kashyap kmartin claudiub tobasco tdurakov andreas_s abhishekk mrhillsman raj_singh live-migration subteam meeting in 10 inutes on #openstack-meeting-3 | 13:51 |
| sdague | mriedem/cdent: I'm going to step away from computer for a bit and go for a bike ride while the weather is nice, and will dig back into this once I get back | 13:51 |
| kashyap | tdurakov: Back from long travel, and not feeling quite productive yet, I'll just lurk :-) | 13:51 |
| tdurakov | kashyap: sure no worries | 13:51 |
| cdent | sdague: that's cool, my battery is dead so I need to go home, and will try to get the loose ends together from there | 13:52 |
| *** takedakn has joined #openstack-nova | 13:52 | |
| * cdent waves | 13:52 | |
| *** cdent has quit IRC | 13:52 | |
| sdague | cdent: cool. The placement api doesn't seem to be logging much | 13:52 |
| sdague | I wonder if the wrong log file is being followed in devstack | 13:52 |
| *** ayoung has joined #openstack-nova | 13:52 | |
| sdague | or if there is something else that needs configuration | 13:52 |
| openstackgerrit | Andrey Volkov proposed openstack/nova: WIP Map instances to cell0 on failed schedule https://review.openstack.org/267827 | 13:52 |
| sdague | but I can look later | 13:52 |
| *** ducttape_ has joined #openstack-nova | 13:52 | |
| *** gouthamr_ is now known as gouthamr | 13:53 | |
| *** liverpooler has quit IRC | 13:53 | |
| johnthetubaguy | cdent: when you get back on, this might be an issue, but I am not totally sure: https://review.openstack.org/#/c/362766/1/nova/db/sqlalchemy/api.py@133 | 13:54 |
| openstackgerrit | viswesuwara nathan proposed openstack/python-novaclient: Fix incorrect output of "nova show" for long user data https://review.openstack.org/362303 | 13:54 |
| *** anush__ has joined #openstack-nova | 13:56 | |
| *** yamahata has joined #openstack-nova | 13:57 | |
| *** takedakn has quit IRC | 13:57 | |
| openstackgerrit | Andrey Volkov proposed openstack/nova: WIP Map instances to cell0 on failed schedule https://review.openstack.org/267827 | 13:59 |
| *** ratailor has quit IRC | 13:59 | |
| *** zzzeek has quit IRC | 14:00 | |
| *** sbezverk has quit IRC | 14:01 | |
| *** rmcall_ has joined #openstack-nova | 14:01 | |
| *** jdillaman has joined #openstack-nova | 14:02 | |
| *** eharney has joined #openstack-nova | 14:02 | |
| *** annegentle has joined #openstack-nova | 14:03 | |
| *** su_zhang has quit IRC | 14:03 | |
| *** rmcall has quit IRC | 14:03 | |
| *** rmcall_ is now known as rmcall | 14:03 | |
| *** zzzeek has joined #openstack-nova | 14:03 | |
| *** su_zhang has joined #openstack-nova | 14:04 | |
| *** dimtruck is now known as zz_dimtruck | 14:04 | |
| *** zzzeek has quit IRC | 14:04 | |
| *** zzzeek has joined #openstack-nova | 14:05 | |
| *** lixiaoy1 has joined #openstack-nova | 14:06 | |
| *** browne has joined #openstack-nova | 14:07 | |
| cmartin | markus_z: Hey I was told to talk to you about an issue with a nova bug | 14:07 |
| *** rama_y has joined #openstack-nova | 14:08 | |
| *** sdake has joined #openstack-nova | 14:08 | |
| markus_z | cmartin: sure, which one? | 14:08 |
| *** su_zhang has quit IRC | 14:08 | |
| cmartin | markus_z: It's this one https://bugs.launchpad.net/nova/+bug/1587108 I had a few questions because I am trying to figure out how to proceed | 14:09 |
| openstack | Launchpad bug 1587108 in OpenStack Compute (nova) "Booting an instance from snapshot (which is an instance booted from Cinder volume) fails with: "Volume is smaller than the minimum size specified in image metadata."" [Medium,In progress] - Assigned to Chris Martin (cm876n) | 14:09 |
| *** cleong has joined #openstack-nova | 14:09 | |
| * markus_z is reading the bug | 14:10 | |
| *** sbezverk has joined #openstack-nova | 14:10 | |
| johnthetubaguy | cmartin: that looks like the correct behaviour | 14:11 |
| johnthetubaguy | but that means I must be missing something | 14:11 |
| johnthetubaguy | if the disk is bigger than the volume, you can't boot it | 14:11 |
| *** garyk1 has quit IRC | 14:11 | |
| *** ljxiash has joined #openstack-nova | 14:11 | |
| *** sdake_ has joined #openstack-nova | 14:12 | |
| *** bdeschenes has quit IRC | 14:12 | |
| *** sdake has quit IRC | 14:13 | |
| cmartin | johnthetubaguy: Well here is what happens, the flavor size is 3, the volume size is 1, it lets you boot the intial instance but not the snapshot, because the exception that catches if the Volume is Larger than the disk passes when creating the initial instance because the size is 1, min_disk is 0 UNTIL the root_gb from the flavor is used to set the min_disk which is after the exception, which is why when you take a snapshot and reboot now the | 14:13 |
| cmartin | exception catches it | 14:13 |
| cmartin | johnthetubaguy: In this case I was using a volume of size 1 and flavor of disk size 3 | 14:14 |
| markus_z | Yeah, I would honestly expect that both cases fail or succeed | 14:14 |
| cmartin | Yeah exactly | 14:14 |
| johnthetubaguy | cmartin: so the initial is allowed | 14:14 |
| johnthetubaguy | this boils down to when we take the snapshot, we set the wrong min_disk value | 14:14 |
| cmartin | No the min_disk is set in the initial instance | 14:14 |
| cmartin | it is just set after the exception check | 14:15 |
| johnthetubaguy | we take it from the flavor, but in the BFV case, it needs to come from the volume | 14:15 |
| cmartin | And yes the min_disk is taken from the flavor | 14:15 |
| *** roeyc has quit IRC | 14:15 | |
| johnthetubaguy | right, thats the bug here, when BFV, the min_disk should be set from the size of the boot disk, not the flavor | 14:15 |
| cmartin | Yea that's right | 14:15 |
| johnthetubaguy | melwitt has a similar patch up to look at things around the resource tracker and BFV | 14:17 |
| cmartin | can you link it to me and i'll take a look | 14:17 |
| johnthetubaguy | I don't have a link to that handy though | 14:17 |
| *** pwnall has quit IRC | 14:17 | |
| cmartin | oh I'll look it up then | 14:17 |
| *** prateek has quit IRC | 14:17 | |
| johnthetubaguy | cmartin: I think this is it: https://review.openstack.org/355091 | 14:21 |
| cmartin | johnthetubaguy: yea I am looking at it right now | 14:21 |
| johnthetubaguy | cool | 14:21 |
| cmartin | It does seem to be trying to fix the overall larger problem | 14:21 |
| *** roeyc has joined #openstack-nova | 14:22 | |
| cmartin | But I'm not completely sure since most of my bug tracking has been done in nova/nova/compute/api.py and her patch doesn't touch those files | 14:22 |
| *** rlrossit has quit IRC | 14:23 | |
| *** zz_dimtruck is now known as dimtruck | 14:23 | |
| johnthetubaguy | cmartin: it will not fix your bug, AFAIK, but yeah, it fixes related issues | 14:25 |
| johnthetubaguy | cmartin: I just think the fix should follow in those footsteps | 14:25 |
| *** raunak has joined #openstack-nova | 14:26 | |
| cmartin | johnthetubaguy: alright thanks I'm still a pretty new contributor so I was just trying to figure out how to move forward and this helps | 14:26 |
| *** rlrossit has joined #openstack-nova | 14:26 | |
| johnthetubaguy | cmartin: no problem, happy to try and help | 14:27 |
| *** mriedem has quit IRC | 14:27 | |
| *** mriedem has joined #openstack-nova | 14:28 | |
| *** daemontool has quit IRC | 14:28 | |
| *** daemontool has joined #openstack-nova | 14:28 | |
| *** marst has quit IRC | 14:29 | |
| *** mamandle has quit IRC | 14:29 | |
| *** pbandark has quit IRC | 14:31 | |
| *** mgoddard_ has quit IRC | 14:32 | |
| *** mgoddard has joined #openstack-nova | 14:32 | |
| kashyap | johnthetubaguy: cmartin: What is "BFV" here? -- https://bugs.launchpad.net/nova/+bug/1587108/comments/5 | 14:33 |
| openstack | Launchpad bug 1587108 in OpenStack Compute (nova) "Booting an instance from snapshot (which is an instance booted from Cinder volume) fails with: "Volume is smaller than the minimum size specified in image metadata."" [Medium,In progress] - Assigned to Chris Martin (cm876n) | 14:33 |
| bhagyashri | vek: Hi, could you please review https://review.openstack.org/#/c/346026/ ? | 14:33 |
| johnthetubaguy | kashyap: boot from volume | 14:33 |
| *** ociuhandu has joined #openstack-nova | 14:33 | |
| johnthetubaguy | kashyap: at least thats what I meant | 14:33 |
| kashyap | johnthetubaguy: Err, okay, my brain is in hibernate mode | 14:33 |
| johnthetubaguy | kashyap: no worries | 14:34 |
| *** su_zhang has joined #openstack-nova | 14:34 | |
| *** dikonoor has joined #openstack-nova | 14:34 | |
| *** woodster_ has joined #openstack-nova | 14:34 | |
| *** pwnall has joined #openstack-nova | 14:35 | |
| *** sudswas__ has quit IRC | 14:35 | |
| *** sudipto has quit IRC | 14:35 | |
| *** oanson has quit IRC | 14:36 | |
| mriedem | so to run the placement db migrations we do it via the nova-manage api_db sync command? https://review.openstack.org/#/c/342362/10/lib/placement@156 | 14:36 |
| mriedem | that seems weird | 14:36 |
| *** su_zhang has quit IRC | 14:36 | |
| dansmith | mriedem: there are no placement db migrations | 14:36 |
| *** nkrinner has quit IRC | 14:36 | |
| bauzas | mriedem: yup, that's what we agreed | 14:36 |
| dansmith | mriedem: there are only api migrations, and we shouldn't have to run the db sync twice | 14:37 |
| bauzas | mriedem: no clear outstanding placement db | 14:37 |
| *** su_zhang has joined #openstack-nova | 14:37 | |
| *** kaisers_ has joined #openstack-nova | 14:37 | |
| bauzas | dansmith: mriedem: I was reviewing cdent's patch https://review.openstack.org/#/c/362766/ | 14:37 |
| *** tonytan_brb has joined #openstack-nova | 14:37 | |
| bauzas | dansmith: mriedem: I actually wonder if we should provide a separate contextmanager for the placement DB | 14:38 |
| dansmith | you mean for in the db_api? | 14:38 |
| dansmith | placement.writer? | 14:38 |
| *** nkrinner has joined #openstack-nova | 14:39 | |
| bauzas | dansmith: I mean https://review.openstack.org/#/c/362766/1/nova/objects/resource_provider.py | 14:39 |
| bauzas | I was thinking we were not yet using a separate engine facade, just using the same as for the API DB | 14:40 |
| *** marst has joined #openstack-nova | 14:40 | |
| bauzas | and just help operators to sync a new DB if really needed | 14:40 |
| *** tonytan4ever has quit IRC | 14:40 | |
| bauzas | but I could be wrong | 14:40 |
| dansmith | bauzas: we have to have a new context manger to honor the alternate connection url for people that want to pre-split, right? | 14:40 |
| *** su_zhang has quit IRC | 14:41 | |
| *** ekuris has quit IRC | 14:41 | |
| *** prateek has joined #openstack-nova | 14:42 | |
| *** tonytan_brb has quit IRC | 14:42 | |
| *** daemontool has quit IRC | 14:42 | |
| kashyap | Does anyone know top off their heads: if it's possible to change an existing flavour's public / private visibility setting to 'projects' (tenants)? | 14:42 |
| *** tonytan4ever has joined #openstack-nova | 14:42 | |
| mriedem | dansmith: bauzas: see the devstack change from cdent above | 14:43 |
| *** david-lyle has quit IRC | 14:43 | |
| kashyap | (I realize I could create a new flavour and decide about its visibility) | 14:43 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: do not merge https://review.openstack.org/362922 | 14:43 |
| mriedem | he has an optional flag to create the placement db and run schema migrations on it | 14:43 |
| bauzas | dansmith: well, indeed, we could need a separate connection string | 14:44 |
| bauzas | okay | 14:44 |
| dansmith | mriedem: right, so if we want a split one, then you run api sync against a new database and point to that with the placement connection url | 14:44 |
| *** jaosorior is now known as jaosorior_away | 14:45 | |
| *** diogogmt has joined #openstack-nova | 14:45 | |
| *** raunak has quit IRC | 14:45 | |
| *** nkrinner has quit IRC | 14:46 | |
| *** sdake_ has quit IRC | 14:47 | |
| *** lixiaoy1 has quit IRC | 14:48 | |
| *** david-lyle has joined #openstack-nova | 14:48 | |
| johnthetubaguy | with the current patch you only need to run the api sync once, buts thats OK for now maybe | 14:48 |
| *** lixiaoy1 has joined #openstack-nova | 14:48 | |
| mriedem | running nova-manage api_db sync twice shouldn't hurt, if you ran it the first time it will have the api db migrations at the latest version and is a noop | 14:49 |
| mriedem | and will then pick up the placement db migrations | 14:49 |
| johnthetubaguy | yup yup | 14:49 |
| openstackgerrit | Lucian Petrut proposed openstack/nova: Hyper-V: fix image handling when shared storage is being used https://review.openstack.org/347437 | 14:49 |
| dansmith | sure, I meant that you don't have to run it twice against two databases | 14:50 |
| melwitt | dansmith: I just commented on the mq switching patch. it's failing on the new side of grenade because the transport url for the cell mapping is missing the rabbit userid. I'm not yet sure how that is happening from the simple cell setup for grenade | 14:51 |
| *** lixiaoy1 has quit IRC | 14:52 | |
| dansmith | melwitt: okay I saw your comment.. do you need help or were you just explaining why the debug is still there? | 14:52 |
| openstackgerrit | Maciej Szankin proposed openstack/nova: WIP conf opt: added check for help msg line lenght https://review.openstack.org/362454 | 14:52 |
| *** roeyc has quit IRC | 14:52 | |
| *** jlanoux has quit IRC | 14:53 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements https://review.openstack.org/362595 | 14:53 |
| melwitt | dansmith: might need help. I'm going to start trying to find why the rabbit userid isn't being set in grenade | 14:54 |
| *** ManishD has quit IRC | 14:54 | |
| *** tbachman has joined #openstack-nova | 14:54 | |
| kashyap | Okay, disregard me above -- for my question, `nova flavour-access-add` should help | 14:54 |
| *** ManishD has joined #openstack-nova | 14:54 | |
| melwitt | dansmith: or the rabbit userid variable. I'm not sure how this is happening yet | 14:54 |
| dansmith | melwitt: so this would be that we're doing the cell setup in grenade but passing an empty user, and you're fetching that broken thing out of the db... | 14:55 |
| dansmith | I'm just using the get_transport_url thing from devstack | 14:55 |
| dansmith | maybe I need to source something else to make that whole | 14:55 |
| dansmith | sdague: ^ | 14:55 |
| openstackgerrit | Maciej Szankin proposed openstack/nova: WIP conf opt: added check for help msg line lenght https://review.openstack.org/362454 | 14:55 |
| melwitt | dansmith: oh, passing empty user deliberately? because the rest of the data is there in the transport_url correctly. it pulls them from devstack variables | 14:56 |
| dansmith | melwitt: no not intentionally, | 14:56 |
| *** jlanoux has joined #openstack-nova | 14:56 | |
| melwitt | it should have Just Worked(tm) I thought. because it's just stuff that ends up being written to nova.conf anyhow | 14:56 |
| dansmith | but if I haven't sourced something that actually defaults the rabbit userid or something, then this would happen I think | 14:56 |
| *** cdent has joined #openstack-nova | 14:56 | |
| melwitt | oh, gotcha | 14:56 |
| dansmith | yeah, it's defaulted in stack.sh, | 14:57 |
| dansmith | which we don't run | 14:57 |
| dansmith | in grenade | 14:57 |
| melwitt | a-ha | 14:57 |
| dansmith | let me hack something up to test | 14:57 |
| melwitt | I also noticed grenade is using the old school deprecated rpc backend variables. not critical but something we'll need to change later | 14:57 |
| melwitt | (in nova.conf) | 14:58 |
| dansmith | melwitt: depends-on this and try again: https://review.openstack.org/362936 | 14:59 |
| melwitt | dansmith: k, thanks | 14:59 |
| *** nkrinner has joined #openstack-nova | 14:59 | |
| *** amotoki has quit IRC | 15:01 | |
| *** mingyu has joined #openstack-nova | 15:01 | |
| *** david-lyle has quit IRC | 15:02 | |
| openstackgerrit | melanie witt proposed openstack/nova: Move cell message queue switching and add caching https://review.openstack.org/326906 | 15:02 |
| cdent | sdague: on [t 38wp]. Yeah, I was never too clear on what the requirements were there so have been assuming since it has never come up in review that the web log was sufficient. Certainly possible to add more. In a not-quite-related thing there's a pending WIP for starting notifications from resource provider objects. | 15:02 |
| purplerbot | <sdague> cdent: cool. The placement api doesn't seem to be logging much [2016-08-30 13:52:22] [n 38wp] | 15:02 |
| *** moshele has quit IRC | 15:03 | |
| dansmith | melwitt: I think this is the actual proper fix: https://review.openstack.org/362941 | 15:05 |
| dansmith | unless sdague knows a better way | 15:05 |
| *** diogogmt has quit IRC | 15:06 | |
| *** mingyu has quit IRC | 15:06 | |
| bauzas | melwitt: FWIW, I'm currently rebasing https://review.openstack.org/355091 | 15:06 |
| *** tbachman has quit IRC | 15:06 | |
| melwitt | dansmith: okay, cool | 15:06 |
| bauzas | melwitt: I had a question tho, about why we shouldn't update the original instance | 15:06 |
| bauzas | any reason I missed ? | 15:06 |
| melwitt | bauzas: I was just about to reply on that one. I think there's still one last problem with it in that resize is still saving instances with root_gb = 0 in the db | 15:06 |
| *** coolsvap is now known as coolsvap_ | 15:07 | |
| *** mjura_ has quit IRC | 15:07 | |
| bauzas | melwitt: but if an instance is BFV, then it would still be root_gb=0 even when moving ? | 15:07 |
| melwitt | bauzas: oh, and make a copy? to avoid the possibility of the instance getting saved to the db with a modified flavor | 15:07 |
| *** tonytan_brb has joined #openstack-nova | 15:07 | |
| *** diogogmt has joined #openstack-nova | 15:08 | |
| ducttape_ | question about default policy in nova - novaclient(request).quotas.get(tenant_id, detail=True) causes failures for _member_ type users (detail=False works). this seems like a strange default, since I can find all the details via making the 35 other api calls to get this info. is there a reason for this ? | 15:09 |
| *** sdake has joined #openstack-nova | 15:09 | |
| *** armax has joined #openstack-nova | 15:09 | |
| melwitt | bauzas: yeah we want it to be considered as root_gb=0 for calculations for moving it, but we were trying to make sure it doesn't get persisted that way. and currently it is being persisted with root_gb=0 in the resize case, abhishekk noticed | 15:10 |
| *** rmart04 has quit IRC | 15:10 | |
| melwitt | I had been meaning to figure out how to avoid that but hadn't gotten to it yet | 15:10 |
| dansmith | OR, we don't try to hack this up in the face of the real solution.. :) | 15:11 |
| *** tbachman has joined #openstack-nova | 15:11 | |
| *** tonytan4ever has quit IRC | 15:11 | |
| *** nkrinner is now known as nkrinner_afk | 15:11 | |
| *** xwizard has quit IRC | 15:13 | |
| *** tbachman_ has joined #openstack-nova | 15:15 | |
| *** praneshp has joined #openstack-nova | 15:15 | |
| bauzas | melwitt: sorry, otp | 15:16 |
| *** tbachman has quit IRC | 15:17 | |
| *** tbachman_ is now known as tbachman | 15:17 | |
| bauzas | melwitt: so, just to make it sure, take one BFV instance, and then moving it | 15:18 |
| *** sudswas__ has joined #openstack-nova | 15:18 | |
| *** sudipto has joined #openstack-nova | 15:18 | |
| *** zzzeek has quit IRC | 15:19 | |
| *** kaisers_ has quit IRC | 15:19 | |
| mriedem | sdague: looks like the re-enabled auto-allocate tempest tests are still ok http://status.openstack.org//elastic-recheck/index.html#1616498 | 15:20 |
| mriedem | the blip there is a single change failing in the check queue | 15:21 |
| mriedem | so probably a broken change | 15:21 |
| *** abalutoiu has quit IRC | 15:21 | |
| mriedem | sdague: so can we get https://review.openstack.org/#/c/353018/ approved so i can release 6.0.0 today or tomorrow | 15:21 |
| sdague | mriedem: yes | 15:22 |
| *** david-lyle has joined #openstack-nova | 15:22 | |
| *** pwnall has quit IRC | 15:22 | |
| melwitt | bauzas: yeah, resize/move one BFV instance. when considering the root_gb want to consider it as 0 | 15:23 |
| openstackgerrit | Anusha Unnam proposed openstack/nova: Add a new release note https://review.openstack.org/362448 | 15:23 |
| *** mdrabe has quit IRC | 15:23 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Optional separate database for placement API https://review.openstack.org/362766 | 15:23 |
| sdague | dansmith: that d-g change is probably fine | 15:23 |
| *** mdrabe has joined #openstack-nova | 15:24 | |
| dansmith | sdague: cool, thanks | 15:24 |
| *** sridharg has quit IRC | 15:24 | |
| sdague | mriedem: +A | 15:24 |
| dansmith | sdague: the grenade canary will let melwitt prove that it fixes the problem and then if we can land the d-g fix, we can drop the grenade one | 15:25 |
| sdague | ok, is there any sign of the follow on jaypipes patches for other things the placement api should do? | 15:25 |
| *** mingyu has joined #openstack-nova | 15:26 | |
| dansmith | he's not around I guess? | 15:26 |
| *** artom has quit IRC | 15:27 | |
| mriedem | i didn't get a reply to my email | 15:27 |
| dansmith | no reply from him overnight, which isn't a good sign | 15:27 |
| *** links has quit IRC | 15:28 | |
| *** kaisers_ has joined #openstack-nova | 15:29 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 15:30 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 15:30 |
| *** sdake has quit IRC | 15:31 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 15:31 |
| *** Jeffrey4l_ has quit IRC | 15:32 | |
| *** andreas_s has quit IRC | 15:33 | |
| *** TravT has quit IRC | 15:34 | |
| sdague | cdent: do you know what the follow on reporting patches would look like? | 15:37 |
| dansmith | cdent: we had a discussion yesterday about the missing reporting functionality that we expect jay has in a queue | 15:37 |
| dansmith | cdent: do you know that those are done and just not submitted or do you think they still need writing? | 15:37 |
| dansmith | sdague: heh, jinx | 15:38 |
| *** rama_y has quit IRC | 15:38 | |
| *** TravT has joined #openstack-nova | 15:38 | |
| dansmith | cdent: i.e. reporting allocations and inventory from the RT through the reporting client... because the patch up that we all thought did that only creates resource providers | 15:39 |
| *** zzzeek has joined #openstack-nova | 15:40 | |
| cdent | dansmith, sdague: I'm sorry to say that I don't know. Jay's currently as unavailable to me as he is to you | 15:40 |
| *** salv-orlando has joined #openstack-nova | 15:40 | |
| dansmith | yeah, but I was hoping you'd have more context about the big picture here | 15:40 |
| *** david-lyle has quit IRC | 15:40 | |
| *** pwnall has joined #openstack-nova | 15:41 | |
| *** ManishD has quit IRC | 15:41 | |
| *** zzzeek has quit IRC | 15:41 | |
| sdague | right, I can get the current patch he has into merge state, however, it seems we need some other stuff to hit the finish line, and I definitely don't have context on what additional reporting should be in there | 15:42 |
| dansmith | we need to be reporting compute node inventories | 15:42 |
| cdent | dansmith: I bet I was hoping for that more... | 15:42 |
| dansmith | and instance allocations | 15:42 |
| *** unicell has quit IRC | 15:42 | |
| cdent | in broad strokes what I understood was that instead of using the get_inventory stuff that he added in one of his branches he was going to do things raw in one of the existing methods and then change it over later. I don't know if some/none/all of that is done. | 15:43 |
| *** zzzeek has joined #openstack-nova | 15:43 | |
| cdent | Writing allocations to the api is currently not possible because although the api code is written, the object code is not, Jay said he was working on that but I've seen no concrete evidence | 15:43 |
| *** paul-carlton2 has quit IRC | 15:44 | |
| cdent | I have a pretty good picture of how that object code is supposed to work though if a) someone wants to do or b) I should go ahead and do it | 15:44 |
| dansmith | um, what? | 15:44 |
| dansmith | there's not full API support for allocations? | 15:44 |
| dansmith | that's pretty much going to kill our ocata plans | 15:44 |
| *** klkumar has quit IRC | 15:44 | |
| cdent | there's been pending api code for allocations for plural weeks but jay's been saying he's going to write the object code to wire it up | 15:45 |
| cdent | AllocationList.{create,delete}_all() | 15:45 |
| cdent | was the plan, as I recall | 15:45 |
| *** cfriesen has quit IRC | 15:45 | |
| * dansmith headdesks | 15:46 | |
| sdague | could we do a hangout in 15 minutes to figure out where we stand, and what's reasonable for this week. Because I'm happy to have this be my week, if it means we get to a useful place. But if we can't get there, there are other things to go poke at. | 15:46 |
| * tbachman hands dansmith a headdesk cushion | 15:46 | |
| dansmith | sdague: yeah | 15:46 |
| *** derekh has quit IRC | 15:46 | |
| cdent | dansmith: a) there's not _that_ much to do on the api side for allocations b) I feel like I've been meekly waving my hands about this for a long time so please let me know what I need to do be less meek next time | 15:46 |
| cdent | c) happy to do a hangout whenever | 15:47 |
| dansmith | cdent: there's also not that much time left to get this done... | 15:47 |
| cdent | dansmith: yes, I know | 15:48 |
| *** lpetrut has quit IRC | 15:48 | |
| *** unicell has joined #openstack-nova | 15:50 | |
| *** david-lyle has joined #openstack-nova | 15:50 | |
| cdent | sdague, dansmith say top of the hour for the hangout? | 15:50 |
| dansmith | wfm | 15:51 |
| *** markus_z has quit IRC | 15:51 | |
| *** jlanoux has quit IRC | 15:52 | |
| *** ihrachys has quit IRC | 15:54 | |
| *** ducttape_ has quit IRC | 15:55 | |
| *** ducttape_ has joined #openstack-nova | 15:56 | |
| *** vishwanathj_zzz is now known as vishwanathj | 15:56 | |
| *** artom has joined #openstack-nova | 15:56 | |
| *** Sukhdev has joined #openstack-nova | 15:58 | |
| cdent | sdague, dansmith, edleafe give this a try: https://hangouts.google.com/call/ikfdxn75dfbf7masd3rknl5ndye | 15:59 |
| *** Hieu has joined #openstack-nova | 16:00 | |
| *** abalutoiu has joined #openstack-nova | 16:01 | |
| *** tbachman has quit IRC | 16:01 | |
| auggy | good morning nova! | 16:02 |
| *** devananda is now known as devananda|MOVING | 16:04 | |
| *** nic has joined #openstack-nova | 16:04 | |
| *** mriedem1 has joined #openstack-nova | 16:04 | |
| bauzas | cdent: dansmith: edleafe: sdague: I need to bail out :/ | 16:05 |
| *** moshele has joined #openstack-nova | 16:05 | |
| *** mriedem has quit IRC | 16:06 | |
| sfinucan | moshele: Sorry for missing the sriov meeting again today | 16:06 |
| sfinucan | I've been using Fedora since my move to Red Hat, and I only just figured out how to enable Outlook-like notifications in Evolution. Shan't happen again | 16:06 |
| mriedem1 | cdent: we end here? https://review.openstack.org/#/c/339874/ | 16:06 |
| *** mriedem1 is now known as mriedem | 16:07 | |
| sfinucan | I'll review the logs now | 16:07 |
| moshele | sfinucan: np :) | 16:07 |
| sdague | mriedem: https://etherpad.openstack.org/p/placement-next | 16:08 |
| *** klkumar has joined #openstack-nova | 16:10 | |
| *** ljxiash has quit IRC | 16:13 | |
| *** raunak has joined #openstack-nova | 16:15 | |
| *** ociuhandu has quit IRC | 16:15 | |
| *** armax has quit IRC | 16:16 | |
| *** sudswas__ has quit IRC | 16:16 | |
| *** sudipto has quit IRC | 16:16 | |
| *** tonytan_brb has quit IRC | 16:17 | |
| *** matrohon has quit IRC | 16:19 | |
| mriedem | the good news is devstack is at least standing up with the latest devstack placement service change | 16:19 |
| mriedem | http://logs.openstack.org/41/362441/1/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/669af17/logs/apache/placement-api.txt.gz | 16:20 |
| *** lpetrut has joined #openstack-nova | 16:20 | |
| *** david-lyle has quit IRC | 16:21 | |
| mriedem | just updated the d-g change to point at sean's hack on top of jay's change | 16:22 |
| *** david-lyle has joined #openstack-nova | 16:23 | |
| cdent | edleafe: are you savvy to this logging change that sdague says needs to happen? | 16:23 |
| edleafe | cdent: no, not really | 16:24 |
| cdent | do you want to become so? :) | 16:24 |
| sdague | it's this, about wsgi requests - https://specs.openstack.org/openstack/openstack-specs/specs/log-guidelines.html#specific-event-types | 16:26 |
| mriedem | cdent: edleafe: https://github.com/openstack/keystone/blob/master/httpd/wsgi-keystone.conf | 16:26 |
| mriedem | CustomLog /var/log/apache2/keystone_access.log combined | 16:26 |
| cdent | mriedem: so yeah, placement is already logging via apache, and sdague says that does't have enough info | 16:26 |
| cdent | (and maybe is not going somewhere good enough_ | 16:26 |
| mriedem | well, we don't have anything hitting that api yet, at least not in those d-g results i have up | 16:27 |
| cdent | dansmith: I'm going to send some "oh hey, jay, we made this etherpad, just so you know" email. Unless you want that gig. | 16:27 |
| dansmith | cdent: all yours | 16:27 |
| cdent | dansmith: ✔ | 16:27 |
| *** kaisers_ has quit IRC | 16:28 | |
| dansmith | cdent: so placement has no api samples which makes it kinda hard to know exactly what to look for | 16:28 |
| dansmith | cdent: and there is no api-ref stuff for placement right? | 16:28 |
| dansmith | it something else being generated I can look at? | 16:28 |
| cdent | dansmith: what do you mean by "look for"? What do you want to see? | 16:28 |
| *** moshele has quit IRC | 16:29 | |
| mriedem | sample api requests | 16:29 |
| *** Swami has joined #openstack-nova | 16:29 | |
| mriedem | like the api samples in nova's functional test tree and docs | 16:29 |
| mriedem | does gabbi provide that? | 16:29 |
| dansmith | cdent: s/know what to look for/know what to PUT/ | 16:29 |
| dansmith | right, what mriedem said | 16:29 |
| edleafe | sdague: so is something like this what's needed to set up oslo logging? https://github.com/openstack/nova/blob/master/nova/wsgi/nova-api.py#L33 | 16:29 |
| sdague | edleafe: well, at the end of the day we want lines like this in the keystone log - http://logs.openstack.org/61/360761/1/check/gate-grenade-dsvm-neutron-ubuntu-trusty/08cb274/logs/apache/keystone.txt.gz#_2016-08-25_20_28_31_196 | 16:30 |
| cdent | ah, okay: either here https://github.com/openstack/nova/tree/master/nova/tests/functional/api/openstack/placement/gabbits or http://specs.openstack.org/openstack/nova-specs/specs/newton/approved/generic-resource-pools.html#rest-api-impact | 16:30 |
| sdague | especially because they contain request-id and context info | 16:30 |
| *** tbachman has joined #openstack-nova | 16:30 | |
| dansmith | yeah, I guess the gabbits are readable if you eye-parse out the data bits | 16:31 |
| cdent | dansmith: the gabbits show it in yaml, the spec shows it in json strings | 16:31 |
| *** unicell has quit IRC | 16:31 | |
| cdent | dansmith: there's a yaml file that I used to exercise the devstack set up here, which is very plain: https://gist.github.com/cdent/a9590764fbc7402d450fa36df14f35e0 | 16:32 |
| *** nic has quit IRC | 16:33 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Make api-ref bug link point to nova https://review.openstack.org/362170 | 16:33 |
| *** code-R_ has quit IRC | 16:34 | |
| *** david-lyle_ has joined #openstack-nova | 16:36 | |
| *** trisq has quit IRC | 16:37 | |
| sdague | edleafe: are you looking into logging? | 16:38 |
| *** mamandle has joined #openstack-nova | 16:38 | |
| *** Sukhdev has quit IRC | 16:39 | |
| edleafe | sdague: yes, but I'm not sure how to enable it. You mentioned middleware, but I don't see any | 16:39 |
| *** yamahata has quit IRC | 16:39 | |
| sdague | no, I don't think there is any middleware | 16:39 |
| *** david-lyle has quit IRC | 16:39 | |
| sdague | I think it's just about putting the request logging in there | 16:39 |
| edleafe | sdague: maybe that's why I didn't find it :) | 16:39 |
| edleafe | sdague: so just import logging from oslo? | 16:39 |
| *** zenoway has quit IRC | 16:40 | |
| sdague | well, to start, but I assume that's already there. The important thing is to actually log requests. Also, errors. My current stack has 500 errors, and I have no idea why (probably auth issue) | 16:40 |
| sdague | there are zero things in the placement-api.log | 16:40 |
| *** zenoway has joined #openstack-nova | 16:41 | |
| cdent | sdague: one sec | 16:41 |
| sdague | well, not quite zero things - http://paste.openstack.org/show/564859/ :) | 16:41 |
| edleafe | sdague: ok, digging | 16:41 |
| dansmith | cdent: hmm, we can't post multiple inventories in one go? | 16:41 |
| cdent | dansmith: you can | 16:42 |
| cdent | that was added later | 16:42 |
| lbeliveau | I raised a new jira on bugs.launchpad.net, proposed a fix with "Closes-bug:", but the review is not getting logged in the bug, anybody knows how to fix this ? | 16:42 |
| cdent | PUT /resource_providers/{uuid}/inventories | 16:42 |
| dansmith | cdent: https://github.com/openstack/nova/blob/master/nova/tests/functional/api/openstack/placement/gabbits/inventory.yaml#L254 ? | 16:42 |
| dansmith | okay good | 16:42 |
| dansmith | (whew) | 16:42 |
| cdent | yeah that one | 16:42 |
| cdent | sdague: have you looked to be sure that apache isn't putting it somewhere else? I ask because the site-templates are weird and the weirdness has been telephone gamed from project to project as they have mod-wsgi'd themselves | 16:44 |
| *** zenoway has quit IRC | 16:45 | |
| melwitt | lbeliveau: it will only update the bug on PS1 so if you added Closes-Bug later than that, you have to go comment and add a link manually for the review in progress. upon merge, the bug will be updated automatically to close it | 16:45 |
| cdent | it could be that 500s in the auth middleware are not being properly trapped | 16:45 |
| cdent | sdague: ^ . In fact that is fairly likely. | 16:45 |
| sdague | right | 16:45 |
| sdague | so, my guess, we should have a LOG.info here - https://github.com/openstack/nova/blob/bc344458e7603886509a5a80ca7dca628b203c93/nova/api/openstack/placement/handler.py#L145 for success, and log the failures as well in a blanket except handler | 16:46 |
| lbeliveau | melwitt: thanks, usually it was just working for me (meaning didn't needed to manual link the review) | 16:46 |
| edleafe | sdague: yeah, that's where I was looking | 16:46 |
| cdent | sdague: so one option would be a derivation of the faultwrap middleware as the outer most piece of middleware that traps all thigns | 16:46 |
| edleafe | sdague: also the webob exceptions | 16:47 |
| cdent | sdague, edleafe that wouldn't be the best place for it, as that's inside the stack of the middleware | 16:47 |
| melwitt | lbeliveau: it should work automatic if you had the Closes-Bug in the commit message on the first revision of the patch | 16:47 |
| cdent | you want something that is outermost | 16:47 |
| cdent | so something that is added in deploy.py | 16:47 |
| *** Oku_OS is now known as Oku_OS-away | 16:47 | |
| cdent | doing it at the linked line will only catch errors in the placment handlers, and nothing to do with any of the middlewares | 16:48 |
| edleafe | cdent: don't those middlewares already log their errors? | 16:48 |
| sdague | so FaultWrapper is fine for the 500 case | 16:48 |
| cdent | edleafe: I'm guess not if sdague is having untrapped 500s | 16:48 |
| *** devkulkarni has joined #openstack-nova | 16:49 | |
| sdague | edleafe: no, a lot of stuff rolls into FaultWrapper iirc. | 16:49 |
| *** mamandle has quit IRC | 16:49 | |
| sdague | because that's the thing that catches random exceptions | 16:49 |
| *** liverpooler has joined #openstack-nova | 16:50 | |
| *** fragatina has joined #openstack-nova | 16:50 | |
| sdague | edleafe: so I think there are 2 things. Integrate FaultWrapper (I think as is), then handle the non fault cases as well | 16:50 |
| dansmith | cdent: I see no gabbits that validate/expose the resource provider's generation field, only ones that pass it during inventory CRUD | 16:51 |
| *** fragatin_ has joined #openstack-nova | 16:51 | |
| *** mingyu has quit IRC | 16:51 | |
| *** jaosorior_away has quit IRC | 16:51 | |
| cdent | dansmith: what do you mean by validate? | 16:51 |
| cdent | this one tries with a bad one and fails: https://github.com/openstack/nova/blob/master/nova/tests/functional/api/openstack/placement/gabbits/inventory.yaml#L283 | 16:52 |
| *** rlrossit has quit IRC | 16:52 | |
| dansmith | cdent: check that it's returned? I just want to know what key from the result to use, and I don't see any gabbits that use it | 16:52 |
| dansmith | cdent: that's putting it during inventory create | 16:52 |
| *** diogogmt has quit IRC | 16:52 | |
| dansmith | cdent: the client has to know what the current generation is though | 16:52 |
| dansmith | cdent: surely it's something inside of the actual GET /resource_providers/uuid or something right? | 16:52 |
| *** pradk has quit IRC | 16:53 | |
| *** pradk- has joined #openstack-nova | 16:53 | |
| cdent | dansmith: when you get a resource provider it comes with a 'generation' field: https://github.com/openstack/nova/blob/master/nova/tests/functional/api/openstack/placement/gabbits/resource-provider.yaml#L69 | 16:53 |
| dansmith | cdent: that's what I was looking for, yeah, thanks | 16:54 |
| *** fragatina has quit IRC | 16:55 | |
| cdent | sdague: so to restate what your'e after, just to be clear: 1) trap and log (at the outermost middleware layer) uncaught exceptions that are of the 500 variety, which can be done like FaultWrapper 2) at the PlacementHandler layer log, but reraise, exceptions which are <500 status. Is that in the realm of correct? | 16:56 |
| dansmith | sdague: that caching of the rp in the base patch you're working on will not work for us, unless the scheduler client is recreated each time | 16:56 |
| dansmith | caching the generation will not work | 16:56 |
| sdague | dansmith: ok, explain | 16:56 |
| dansmith | sdague: I guess we just go forward with what is there for the moment, but it'll break | 16:56 |
| dansmith | sdague: we pull the generation once at create and if we're ever out of sync, we'll die | 16:57 |
| sdague | what is the issue with a long standing keystone token? | 16:57 |
| *** erlon has quit IRC | 16:57 | |
| *** unicell has joined #openstack-nova | 16:57 | |
| *** lucasagomes is now known as lucas-dinner | 16:58 | |
| sdague | dansmith: can you point me at the bit of code that does that? | 16:58 |
| dansmith | sdague: https://review.openstack.org/#/c/362883/1/nova/scheduler/client/report.py@169 | 16:59 |
| dansmith | sdague: I guess we can just invalidate the cache if and when we get a failed update so I'll just handle it downstream of you | 16:59 |
| *** devkulkarni has quit IRC | 17:00 | |
| *** sdague has quit IRC | 17:00 | |
| *** Hieu has quit IRC | 17:01 | |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] remove a comment that is no longer a todo https://review.openstack.org/363039 | 17:01 |
| *** rama_y has joined #openstack-nova | 17:01 | |
| cdent | biab | 17:01 |
| *** cdent has quit IRC | 17:01 | |
| *** sdague has joined #openstack-nova | 17:01 | |
| *** raj_singh_ has joined #openstack-nova | 17:01 | |
| *** su_zhang has joined #openstack-nova | 17:01 | |
| *** mingyu has joined #openstack-nova | 17:02 | |
| *** hieulq_ has joined #openstack-nova | 17:02 | |
| *** marst has quit IRC | 17:03 | |
| *** pradk- is now known as pradk | 17:03 | |
| *** praneshp has quit IRC | 17:03 | |
| *** su_zhang has quit IRC | 17:03 | |
| *** Apoorva has joined #openstack-nova | 17:03 | |
| *** su_zhang has joined #openstack-nova | 17:04 | |
| *** su_zhang has quit IRC | 17:04 | |
| *** su_zhang has joined #openstack-nova | 17:05 | |
| *** cdent has joined #openstack-nova | 17:06 | |
| *** unicell has quit IRC | 17:06 | |
| *** mgoddard_ has joined #openstack-nova | 17:06 | |
| *** diogogmt has joined #openstack-nova | 17:07 | |
| mriedem | hmm, someone must have changed an api error response message | 17:07 |
| *** devkulkarni has joined #openstack-nova | 17:07 | |
| sdague | dansmith: oh, that bit, yeh that's part of jay's patch that I didn't touch | 17:08 |
| dansmith | sdague: I know | 17:08 |
| sdague | so, sure, we can tweak there | 17:08 |
| sdague | hmmm.... ok, placement-api isn't being installed with pbr? | 17:08 |
| cdent | sdague: i just copied what nova was doing | 17:09 |
| *** mgoddard has quit IRC | 17:09 | |
| cdent | was hoping it would come out clean in the review process... | 17:10 |
| cdent | which I guess is what is happening now | 17:10 |
| mriedem | nova does this https://github.com/openstack/nova/blob/master/setup.cfg#L49 | 17:10 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Use more specific asserts in tests https://review.openstack.org/363043 | 17:10 |
| cdent | mriedem: I thought sdague meant in devstack | 17:10 |
| mriedem | keystone does this https://github.com/openstack/keystone/blob/master/setup.cfg#L81 | 17:10 |
| mriedem | i'm assuming sdague meant an entry point is missing from setup.cfg so that pbr would install it | 17:10 |
| sdague | right | 17:10 |
| cdent | mriedem: yeah, I realize that now | 17:10 |
| cdent | but yes, it does not exist yet | 17:10 |
| *** su_zhang has quit IRC | 17:11 | |
| *** ijw has joined #openstack-nova | 17:11 | |
| *** su_zhang has joined #openstack-nova | 17:11 | |
| edleafe | cdent: we're not using paste for the placement API, right? | 17:11 |
| cdent | but same comment about coming out clean still applies | 17:11 |
| cdent | edleafe: correct, it does its own thing in nova/api/openstack/placement/deploy.py | 17:12 |
| *** rcernin has quit IRC | 17:12 | |
| *** klkumar has quit IRC | 17:12 | |
| cdent | because we didn't want paste to be config | 17:12 |
| edleafe | cdent: ok, that's what I thought | 17:12 |
| edleafe | cdent: so I'll add FaultWrapper there | 17:12 |
| cdent | edleafe: to preserve the separate you might want to copy and trim it to its bare essentials (that's what was done with the other middlewares) | 17:13 |
| edleafe | cdent: ah, I see. OK, I'll take a whack at that | 17:13 |
| *** tesseract- has quit IRC | 17:14 | |
| mriedem | shit well novaclient is busted in the gate, that's blocking 2.37 from landing in there, so i'll be working on that now | 17:15 |
| *** su_zhang has quit IRC | 17:16 | |
| cdent | sdague, mriedem I've add the wsgi bit for setup.cfg to the etherpad, perhaps we can dump things which are not immediate there as we find them? | 17:16 |
| mriedem | cdent: a TODO section sounds good | 17:16 |
| cdent | done | 17:17 |
| *** tonytan4ever has joined #openstack-nova | 17:18 | |
| sdague | yeh, so logging never is setup anywhere in this stack I don't think | 17:18 |
| sdague | ok, I need to grab some food, I'm going to need to unwind this part a bit before I can figure out why the devstack patch doesn't work | 17:19 |
| *** hieulq_ has quit IRC | 17:19 | |
| cdent | sdague: ping me later if you want a pseudo pair | 17:20 |
| *** sahid has quit IRC | 17:21 | |
| *** marst has joined #openstack-nova | 17:21 | |
| dansmith | sounds like he wants a pseudo sandwich as much as anything | 17:21 |
| mriedem | https://www.youtube.com/watch?v=YLpUatutZDw | 17:23 |
| *** tonytan4ever has quit IRC | 17:23 | |
| *** ijw has quit IRC | 17:24 | |
| *** tesseract- has joined #openstack-nova | 17:24 | |
| *** pcaruana has quit IRC | 17:25 | |
| *** ijw has joined #openstack-nova | 17:25 | |
| mriedem | oh yes http://instantrimshot.com/index.php?sound=rimshot&play=true | 17:26 |
| *** sdake has joined #openstack-nova | 17:27 | |
| *** tesseract- has quit IRC | 17:28 | |
| *** praneshp has joined #openstack-nova | 17:28 | |
| *** tonytan4ever has joined #openstack-nova | 17:29 | |
| *** su_zhang has joined #openstack-nova | 17:32 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Report compute node inventories through placement https://review.openstack.org/363061 | 17:33 |
| dansmith | cdent: sdague mriedem ^ untested unworking stub of the inventory stuff I think | 17:34 |
| dansmith | when sdague updates the base and devstack patches, I'll fix it up and actually try it | 17:34 |
| openstackgerrit | Matt Riedemann proposed openstack/python-novaclient: Fix test_trigger_crash_dump_in_locked_state_nonadmin test https://review.openstack.org/363062 | 17:34 |
| mriedem | ^ unblocks novaclient | 17:34 |
| dansmith | melwitt: ^ ready for +W | 17:35 |
| melwitt | on it | 17:35 |
| *** su_zhang has quit IRC | 17:35 | |
| cdent | dansmith: at a glance that looks reasonable, but haven't looked very closely | 17:36 |
| *** su_zhang has joined #openstack-nova | 17:36 | |
| dansmith | cdent: yep, not much point until I can see if it even works :) | 17:36 |
| dansmith | just FYI | 17:36 |
| mriedem | dansmith: melwitt: thanks | 17:37 |
| cdent | dansmith: if you're feeling like you want to be implicit you can leave out reserved, {min,max}_unit and step_size, they will be default to what you've got there | 17:39 |
| cdent | actually, that's not true | 17:40 |
| cdent | the defaults are different from what you've got, never mind | 17:40 |
| dansmith | okay | 17:40 |
| * cdent tries to stay focused on what's he actually doing | 17:41 | |
| dansmith | I think being explicit there is useful anyway, both for future proofness and also just for being an easy place to see what exactly is being reported | 17:41 |
| dansmith | it's not that verbose | 17:41 |
| * cdent nods | 17:41 | |
| dansmith | cdent: one thing.. am I supposed to report those strings as the resource classes? or do I need to resolve them to indexes? | 17:42 |
| dansmith | DISK_GB etc | 17:42 |
| cdent | the api expects strings | 17:42 |
| dansmith | haven't followed where the dynamic stuff is going | 17:42 |
| dansmith | okay cool | 17:42 |
| *** unicell has joined #openstack-nova | 17:42 | |
| *** prateek_ has joined #openstack-nova | 17:42 | |
| *** devkulkarni has quit IRC | 17:44 | |
| *** cfriesen has joined #openstack-nova | 17:45 | |
| *** prateek has quit IRC | 17:46 | |
| *** lpetrut has quit IRC | 17:46 | |
| *** raj_singh_ has quit IRC | 17:47 | |
| *** rlrossit has joined #openstack-nova | 17:48 | |
| cdent | dansmith: since the placement api doesn't do rpc (and ought to scale in the future via horizontally having more api servers without needing a conductor) is there a distance future where the methods on the resource provider objects won't need to be remotable? | 17:49 |
| dansmith | cdent: they probably don't need to be remotable now technically, right? | 17:50 |
| dansmith | aside from the fact that we have had code in the past that has used them as remotable | 17:50 |
| cdent | dansmith: correct, but it seemed odd to break the pattern now | 17:50 |
| cdent | and correct on that too | 17:50 |
| dansmith | yeah, can't undo it now but eventually we probably can yeah | 17:50 |
| cdent | cool, just checking my understanding, thanks | 17:50 |
| *** mlavalle has quit IRC | 17:53 | |
| *** gouthamr has quit IRC | 17:54 | |
| *** markvoelker has joined #openstack-nova | 17:54 | |
| *** gouthamr has joined #openstack-nova | 17:54 | |
| *** devkulkarni has joined #openstack-nova | 17:55 | |
| *** baoli has quit IRC | 17:59 | |
| *** baoli has joined #openstack-nova | 18:00 | |
| *** su_zhang has quit IRC | 18:00 | |
| *** Sukhdev has joined #openstack-nova | 18:01 | |
| *** rmcall has quit IRC | 18:03 | |
| *** rmcall has joined #openstack-nova | 18:03 | |
| *** Sukhdev has quit IRC | 18:04 | |
| *** tesseract- has joined #openstack-nova | 18:05 | |
| *** tesseract- has quit IRC | 18:05 | |
| *** tesseract- has joined #openstack-nova | 18:06 | |
| *** salv-orl_ has joined #openstack-nova | 18:08 | |
| *** tesseract- has quit IRC | 18:08 | |
| *** mnestratov has quit IRC | 18:08 | |
| *** Apoorva has quit IRC | 18:10 | |
| openstackgerrit | Arun yerra proposed openstack/nova-specs: Proposal for standardizing hardware NIC capabilities. https://review.openstack.org/341341 | 18:10 |
| *** yamahata has joined #openstack-nova | 18:10 | |
| *** hemanthm is now known as hemanthm|lunch | 18:10 | |
| *** Apoorva has joined #openstack-nova | 18:10 | |
| *** tesseract- has joined #openstack-nova | 18:10 | |
| *** salv-orlando has quit IRC | 18:11 | |
| *** tonytan4ever has quit IRC | 18:13 | |
| *** parora has joined #openstack-nova | 18:14 | |
| *** tonytan4ever has joined #openstack-nova | 18:14 | |
| *** dikonoor has quit IRC | 18:15 | |
| *** dikonoor has joined #openstack-nova | 18:16 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Add a TODO and add info to a releasenote https://review.openstack.org/362919 | 18:16 |
| *** prateek_ has quit IRC | 18:17 | |
| sdague | dansmith: humph, I'm getting a method not allowed here which is my crazy 500 error | 18:17 |
| dansmith | sdague: from the placement api you mean? | 18:17 |
| sdague | yeh | 18:17 |
| dansmith | POST or PUT? | 18:18 |
| sdague | POST | 18:18 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add more vd2 unit tests https://review.openstack.org/349298 | 18:18 |
| *** ducttape_ has quit IRC | 18:18 | |
| sdague | 'REQUEST_METHOD': 'POST', 'PATH_INFO': '/resource_providers/5c42a360-cd71-4586-a65e-471933757434' | 18:18 |
| openstackgerrit | Vu Tran proposed openstack/nova: libvirt: Enable instance resizing for LVM backed instances https://review.openstack.org/360764 | 18:18 |
| sdague | is that supposed to be a PUT? | 18:18 |
| dansmith | sdague: the code says you can't POST to /uuid | 18:18 |
| sdague | it was a post in jay's patch | 18:19 |
| dansmith | sdague: but jay's code seems to be doing that | 18:19 |
| dansmith | so it has to be POST /resource_providers or PUT /resource_providers/uuid | 18:19 |
| dansmith | I think right? | 18:19 |
| sdague | yeh, seems right. | 18:19 |
| dansmith | sorry, jay's code seems to be doing ti wrong I mean | 18:20 |
| dansmith | you need to snip the /uuid part on create | 18:20 |
| sdague | yeh, let me try that bit | 18:21 |
| sdague | that will at least maybe, let me know if the devstack patch is ok | 18:21 |
| *** markvoelker has quit IRC | 18:22 | |
| *** gszasz has quit IRC | 18:23 | |
| sdague | hmmm... GET is no good either? | 18:25 |
| sdague | 'REQUEST_METHOD': 'GET', 'PATH_INFO': '/resource_providers/5c42a360-cd71-4586-a65e-471933757434' | 18:25 |
| *** su_zhang has joined #openstack-nova | 18:25 | |
| dansmith | uh | 18:25 |
| *** su_zhang has quit IRC | 18:26 | |
| dansmith | 404 or no method? | 18:26 |
| *** rrecio has joined #openstack-nova | 18:26 | |
| sdague | no, 405 | 18:27 |
| dansmith | are you sure you're hitting the right API? | 18:28 |
| *** tesseract- has quit IRC | 18:30 | |
| sdague | yes, let me get this logging tweaked, then I'll have a better thing to say here | 18:30 |
| *** tongli has joined #openstack-nova | 18:30 | |
| *** tesseract- has joined #openstack-nova | 18:30 | |
| *** tesseract- has quit IRC | 18:30 | |
| *** tesseract- has joined #openstack-nova | 18:31 | |
| *** mgoddard_ has quit IRC | 18:31 | |
| sdague | it might have been me overrunning the logs on the GET request, and it turning into something stupid | 18:32 |
| sdague | the GETs are working for me now, the POST is failing on media type enforcement | 18:32 |
| sdague | 2016-08-30 14:34:34.686 ERROR nova.scheduler.client.report [req-4b6fcb8a-6437-4050-a96c-763422f8abed None None] Failed to create resource provider record in placement API for UUID 5c42a360-cd71-4586-a65e-471933757434. Got 201: . | 18:34 |
| sdague | well progress | 18:34 |
| cdent | sdague: I think jay's code is pretty much completely untested so I wouldn't take what it is doing as canonical in any way | 18:36 |
| openstackgerrit | Doug Hellmann proposed openstack/os-vif: Update reno for stable/newton https://review.openstack.org/363115 | 18:36 |
| *** yamahata has quit IRC | 18:37 | |
| *** tesseract- has quit IRC | 18:37 | |
| *** fragatin_ has quit IRC | 18:38 | |
| *** nic has joined #openstack-nova | 18:39 | |
| *** mtanino has joined #openstack-nova | 18:41 | |
| *** zengine has joined #openstack-nova | 18:41 | |
| *** ducttape_ has joined #openstack-nova | 18:41 | |
| *** tonytan_brb has joined #openstack-nova | 18:42 | |
| *** Apoorva has quit IRC | 18:43 | |
| *** GB21 has quit IRC | 18:43 | |
| *** su_zhang has joined #openstack-nova | 18:43 | |
| sdague | cdent: yes, definitely not tested :) | 18:44 |
| *** lpetrut has joined #openstack-nova | 18:45 | |
| *** su_zhang has quit IRC | 18:45 | |
| *** tonytan4ever has quit IRC | 18:45 | |
| *** Apoorva has joined #openstack-nova | 18:45 | |
| *** tonytan4ever has joined #openstack-nova | 18:47 | |
| sdague | ok, so the devstack patch is validated now, let me get someone else to +2 that https://review.openstack.org/#/c/342362 | 18:47 |
| sdague | mtreinish: how about you? | 18:47 |
| *** yamahata has joined #openstack-nova | 18:48 | |
| sdague | and I have another round of fixes here I need to bring back into the fold | 18:48 |
| *** delattec has joined #openstack-nova | 18:48 | |
| *** mdrabe_ has joined #openstack-nova | 18:49 | |
| *** salv-orl_ has quit IRC | 18:49 | |
| *** hemanthm|lunch is now known as hemanthm | 18:50 | |
| *** liangy_ has joined #openstack-nova | 18:50 | |
| *** tonytan_brb has quit IRC | 18:50 | |
| *** aignatov has quit IRC | 18:50 | |
| *** dstepanenko_ has joined #openstack-nova | 18:51 | |
| * dansmith netboots | 18:51 | |
| *** cdelatte has quit IRC | 18:51 | |
| *** lxsli has quit IRC | 18:51 | |
| *** dstepanenko has quit IRC | 18:51 | |
| *** ianw has quit IRC | 18:51 | |
| *** mdrabe has quit IRC | 18:52 | |
| *** lennyb has quit IRC | 18:52 | |
| *** liangy has quit IRC | 18:52 | |
| *** NobodyCam has quit IRC | 18:52 | |
| *** sayalilunkad has quit IRC | 18:52 | |
| *** scsnow has quit IRC | 18:52 | |
| *** scsnow has joined #openstack-nova | 18:53 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: WIP - Added FaultWrapper to the placement API https://review.openstack.org/363125 | 18:53 |
| edleafe | cdent: sdague: ^^ am I on the right track? | 18:53 |
| *** dimtruck is now known as zz_dimtruck | 18:54 | |
| *** NobodyCam has joined #openstack-nova | 18:54 | |
| * cdent looks | 18:54 | |
| *** cn28h has quit IRC | 18:54 | |
| edleafe | crap -never mind | 18:55 |
| edleafe | missed a git add | 18:55 |
| edleafe | one minute | 18:55 |
| cdent | yeah, was just gonna say, the bit in deploy.py seems fine though :) | 18:55 |
| sdague | edleafe: yeh, though honestly I don't know why we'd dup the fault wrapper, just use it in place | 18:55 |
| *** prateek_ has joined #openstack-nova | 18:56 | |
| cdent | dansmith: if you're around, I could do with a small borrow of your brain: When writing allocations, do we need to process generations | 18:56 |
| sdague | we can dup it out when we split | 18:56 |
| cdent | sdague: everything else has already been duped, and when it was duped it was made much smaller | 18:56 |
| edleafe | sdague: that was cdent's suggestion to make it similar to the rest | 18:56 |
| dansmith | cdent: okay | 18:56 |
| cdent | but yeah, it's not strictly necessary | 18:56 |
| *** aignatov has joined #openstack-nova | 18:56 | |
| *** edand has quit IRC | 18:57 | |
| cdent | dansmith: the api doesn't have the generation in the input when we write allocations, so I'm not quite sure... | 18:57 |
| *** annegentle has quit IRC | 18:57 | |
| sdague | cdent: at this point in the cycle, I don't want to manage any new differences | 18:57 |
| cdent | dansmith: http://specs.openstack.org/openstack/nova-specs/specs/newton/approved/generic-resource-pools.html#put-allocations-consumer-uuid and http://specs.openstack.org/openstack/nova-specs/specs/newton/approved/resource-providers-allocations.html | 18:57 |
| cdent | sdague: makes sense | 18:58 |
| openstackgerrit | Sean Dague proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/358797 | 18:58 |
| openstackgerrit | Sean Dague proposed openstack/nova: WIP: bang in some logging https://review.openstack.org/363130 | 18:58 |
| sdague | edleafe: also, yes, that should work. Though, it turns out that wasn't really the issue | 18:58 |
| *** intr1nsic has quit IRC | 18:58 | |
| sdague | I put something similar in my loggint patch to figure out what is going on | 18:58 |
| *** annegentle has joined #openstack-nova | 18:58 | |
| cdent | sdague other reason it happened is that the calling strategy for the middlewares is a bit different from how nova likes things | 18:58 |
| *** cn28h has joined #openstack-nova | 18:58 | |
| edleafe | sdague: the extra logging will come next | 18:58 |
| openstackgerrit | Ed Leafe proposed openstack/nova: WIP - Added FaultWrapper to the placement API https://review.openstack.org/363125 | 18:59 |
| *** parora has quit IRC | 18:59 | |
| edleafe | ^^ should be better now | 18:59 |
| *** salv-orlando has joined #openstack-nova | 18:59 | |
| *** erlon has joined #openstack-nova | 18:59 | |
| dansmith | cdent: I think we had a convo about this at some point, but are you sure you need a generation when updating an allocation? | 19:00 |
| *** devkulkarni has quit IRC | 19:00 | |
| dansmith | cdent: you're refreshing all the allocations for a given consumer every time you update, right? | 19:00 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Pull from cell0 and build_requests for instance list https://review.openstack.org/356138 | 19:00 |
| openstackgerrit | Andrew Laski proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | 19:00 |
| cdent | dansmith: not refreshing, setting. if there's an allocation for that consumer already it is considered an error | 19:00 |
| sdague | dansmith: do we have an existing decorator for a method example somewhere that traps exceptions and sets attributes on self? | 19:01 |
| *** intr1nsic has joined #openstack-nova | 19:01 | |
| dansmith | cdent: really? how are we going to do the heal then? | 19:01 |
| sdague | just to speed up me writing this error handler | 19:01 |
| cdent | dansmith: your statement "but are you sure you need a generation when updating an allocation?" is my question | 19:01 |
| dansmith | sdague: we probably have decorators for those two things, but not sure all in one | 19:01 |
| dansmith | sdague: which part do you need to copy? | 19:01 |
| cdent | dansmith: that's a great question, but currently the spec is: "409 Conflict if there is already an allocation record for the specified consumer against a specified resource provider. " | 19:02 |
| sdague | dansmith: I just need refreshers, I do decorators so infrequently | 19:02 |
| edleafe | sdague: your WIP logging patch looks similar to what I was adding. Can I just add the faultwrapper copy to it and abandon mine? | 19:02 |
| cdent | (so one option would be a delete/put pair) | 19:02 |
| dansmith | cdent: not sure how resizes are going to work then | 19:02 |
| sdague | edleafe: sure | 19:02 |
| dansmith | sdague: just a sec | 19:02 |
| cdent | dansmith: alaski had that same comment at some point and I'm not sure how it was resolved :( | 19:02 |
| sdague | edleafe: actually, please don't dup faultwrapper | 19:02 |
| sdague | I don't want to manage extra dups right now, we'll use the existing one | 19:03 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Documentation for the vendordata reboot. https://review.openstack.org/353793 | 19:03 |
| dansmith | sdague: http://pastebin.com/wCQtnqKc | 19:03 |
| dansmith | is that what you want? | 19:03 |
| *** lxsli has joined #openstack-nova | 19:04 | |
| sdague | dansmith: yep, perfect | 19:04 |
| sdague | thank you | 19:04 |
| *** sayalilunkad has joined #openstack-nova | 19:04 | |
| dansmith | heh, np | 19:04 |
| *** lennyb has joined #openstack-nova | 19:04 | |
| *** david-lyle_ is now known as david-lyle | 19:05 | |
| *** raunak has quit IRC | 19:05 | |
| cdent | dansmith: I don't think there's anything about the data model that makes doing the heal impossible, so maybe it's just okay and the spec is "wrong" | 19:05 |
| *** zz_dimtruck is now known as dimtruck | 19:06 | |
| dansmith | cdent: well, it means a lot of polling to check that the current allocation matches what it should be, and if we need to correct it, a delete/post with a race window of not consuming resources | 19:06 |
| * cdent nods | 19:06 | |
| dansmith | cdent: not sure why we wouldn't be able to PUT a specific allocation to change it | 19:07 |
| dansmith | a same-host resize is going to have to either have two allocations for the same host or change one | 19:07 |
| dansmith | in fact, two allocations is the more correct case, IMHO | 19:07 |
| dansmith | because during resize_confirm you're using two of several things | 19:07 |
| cdent | i'm entirely willing to accept that | 19:07 |
| cdent | and can code it that way; but I'm still unsure about generation | 19:08 |
| dansmith | yeah, well, not sure why generation would be needed during the allocation bit | 19:08 |
| cdent | at some point during the allocation we need to check if there's capacity for it | 19:09 |
| cdent | which means its view is dependent on a particular view of a resource provider (at a generation when inventory was set). Is that right? | 19:10 |
| * cdent ow | 19:10 | |
| dansmith | um | 19:10 |
| *** baoli has quit IRC | 19:10 | |
| *** lpetrut has quit IRC | 19:10 | |
| dansmith | we don't check for capacity, we POST what we need and it either succeeds or doesn't right? | 19:10 |
| cmartin | Hey I am working on a bugfix and before I do something very tedious I was wondering if I could get a quick answer here | 19:10 |
| cdent | dansmith: that's what I mean the: server checks for capacity | 19:11 |
| *** coolsvap_ has quit IRC | 19:12 | |
| dansmith | the server can't check against anything other than the current inventory so I'm not sure why generation gives you anything, | 19:12 |
| dansmith | unless you want to fail if the generation has changed since you submitted the request, | 19:12 |
| dansmith | but the client requesting the allocation doesn't really care, right? | 19:12 |
| mriedem | sdague: cdent: dansmith: got a clean run with the devstack change and sean's fix on top of jay's change http://logs.openstack.org/41/362441/2/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/4b43ad8/console.html | 19:13 |
| dansmith | it doesn't need to examine the capacity before it decides to submit a request, right? | 19:13 |
| cdent | dansmith: correct | 19:13 |
| alaski | cdent: dansmith just catching up. but I think I convinced jaypipes that it should be possible to update allocations | 19:13 |
| mriedem | there is sadness here, but it sounds like that's a known issue http://logs.openstack.org/41/362441/2/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/4b43ad8/logs/screen-n-cpu.txt.gz?level=TRACE#_2016-08-30_17_38_00_797 | 19:13 |
| dansmith | alaski: I think we have to be able to do that | 19:13 |
| cdent | alaski: ah, good, I guess the spec never made it up to date | 19:13 |
| alaski | dansmith: right, I had the same concerns as you on that | 19:14 |
| dansmith | \o/ | 19:14 |
| cdent | alaski: do you remember about generations and allocations. the reason I'm fussing over this is because I seem to remember a discussion about it in sunnyvale but not enough details | 19:14 |
| alaski | cdent: seems like it. I discussed it in person with him and then I guess it never made it to writing | 19:14 |
| *** armax has joined #openstack-nova | 19:15 | |
| openstackgerrit | Sean Dague proposed openstack/nova: WIP: bang in some logging https://review.openstack.org/363130 | 19:15 |
| openstackgerrit | Sean Dague proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/363148 | 19:15 |
| alaski | cdent: I need to keep reading to get full context as that's not ringing a bell | 19:15 |
| *** sbezverk has quit IRC | 19:16 | |
| * cdent is here all night, tip your server | 19:16 | |
| sdague | mriedem: right, https://review.openstack.org/363148 should be a thing now | 19:16 |
| *** ijw has quit IRC | 19:16 | |
| mriedem | yeah you squashed the changes, lost the old change id | 19:16 |
| mriedem | was that intentional? | 19:16 |
| sdague | nope | 19:17 |
| sdague | fixing | 19:17 |
| openstackgerrit | Sean Dague proposed openstack/nova: WIP: bang in some logging https://review.openstack.org/363130 | 19:17 |
| openstackgerrit | Sean Dague proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/358797 | 19:17 |
| sdague | too fast on the last squash | 19:17 |
| mriedem | can we abandon https://review.openstack.org/#/c/362883/ ? | 19:17 |
| mriedem | to keep things straight | 19:17 |
| sdague | yep | 19:17 |
| mriedem | ok, and https://review.openstack.org/#/c/363061/ will have to be rebased | 19:18 |
| sdague | yep | 19:18 |
| alaski | cdent: I see what you mean. I don't recall anything about having generations with allocations. But you'd want to check allocations against the current inventory right? passing a generation would potentially check against a previous inventory which would be odd | 19:18 |
| cdent | alaski: it's not about passing the generation, it's making sure that the generation hasn't changed out from under you while making changes | 19:18 |
| sdague | ok, I think that take on 358797 should be basically sound | 19:18 |
| sdague | but, all unit tests are pointless, let me go and work on those now | 19:19 |
| dansmith | I'm rebasing on that and stacking | 19:19 |
| alaski | cdent: a generation for the allocation? I'm not sure I can see a situation where you're reporting an allocation with potentially outdated data | 19:19 |
| dansmith | alaski: he means RP generation with the allocation | 19:19 |
| tdurakov | alaski: ping | 19:19 |
| mriedem | ok, i'll update the d-g change to pick up the latest and pull in dan's change and the log jammin' change | 19:19 |
| dansmith | which I don't think makes sense | 19:19 |
| cdent | writing an allocation against a changed inventory | 19:19 |
| alaski | cdent: dansmith | 19:20 |
| alaski | gotcha | 19:20 |
| sdague | mriedem: so the log jammin' change is all kinds of gross | 19:20 |
| dansmith | cdent: you don't write allocations against an inventory though, from the client perspective | 19:20 |
| tdurakov | alaski: question on rpodolyaka's patch that you've reviewed | 19:20 |
| * mriedem hopes sdague gets the big lebowski reference | 19:20 | |
| cdent | dansmith: but if there's no room in the inventory the write of the allocation must fail | 19:20 |
| alaski | cdent: dansmith yeah, I'm not seeing the usefulness. if sum(allocation) < inventory then proceed. the inventory changing shouldn't matter as long as there's capacity | 19:20 |
| *** armax has quit IRC | 19:20 | |
| tdurakov | alaski: why we use admin_context for host_manager, and shouldn't for ironic_host_manager https://review.openstack.org/#/c/347948/3/nova/scheduler/host_manager.py | 19:20 |
| dansmith | cdent: right, but that's the point | 19:20 |
| sdague | we really need to retool how the placement api gets setup as a wsgi script, including log initialization. That's going to be on my stack after I make the unit tests sound on 358797 | 19:21 |
| dansmith | cdent: you fail the SQL update in that case, but nothing about the generation means anything to that | 19:21 |
| alaski | tdurakov: looking | 19:21 |
| sdague | mriedem: apparently not entirely :) | 19:21 |
| cdent | okay, that all sounds sane to me, I wish I could shake the niggle that says jay's going to come along and disagree, but I guess if he wants to do that he needs to be here | 19:22 |
| dansmith | sdague: https://www.youtube.com/watch?v=FiG7z1lrs70 | 19:22 |
| mriedem | https://s-media-cache-ak0.pinimg.com/736x/86/f9/59/86f959ffac58dc752f9359111dd1fb35.jpg | 19:22 |
| alaski | tdurakov: usually get_admin_context is used because it's the only utility method there so people favor it over just doing RequestContext(). If there was a get_context method it would probably be used | 19:22 |
| mriedem | he's here to fix your cable | 19:22 |
| sdague | dansmith: ah, man ... I had forgotten that | 19:23 |
| tdurakov | alaski: should I change context to non-admin one for host_manager? | 19:23 |
| alaski | tdurakov: yes. A few cycles ago it may have been necessary because there were checks for admin in the db layer. But those were all moved to policy a while back | 19:23 |
| tdurakov | acked, also point about objects is valid to both manager, I assume | 19:24 |
| tdurakov | alaski: ^ | 19:24 |
| *** annegentle has quit IRC | 19:24 | |
| alaski | tdurakov: it may not be. HostManager does len(compute_nodes) which may not work without accessing the .objects | 19:25 |
| alaski | or the slice operation | 19:26 |
| sdague | hmmm... our version detection for qemu version is totally bonkers | 19:26 |
| tdurakov | alaski: ok, working on that one right now | 19:26 |
| alaski | tdurakov: leaving the .objects is fine for consistency. I didn't realize it was being copied from hostmanager | 19:26 |
| sdague | http://logs.openstack.org/61/360761/1/check/gate-tempest-dsvm-full-ubuntu-xenial/d847714/logs/screen-n-cpu.txt.gz#_2016-08-25_20_15_57_163 | 19:27 |
| sdague | xenial has 2.5.0 | 19:27 |
| tdurakov | alaski: tbh both works for me, well, will leave object for consistency:) | 19:27 |
| alaski | tdurakov: cool. either works for me | 19:28 |
| *** ianw_ has joined #openstack-nova | 19:28 | |
| *** rama_y has quit IRC | 19:28 | |
| *** Sukhdev has joined #openstack-nova | 19:28 | |
| *** baoli has joined #openstack-nova | 19:28 | |
| cdent | can I rely on methods with database context manager decorators to nest properly and rollback as expected when things go wrong? | 19:29 |
| *** parora has joined #openstack-nova | 19:30 | |
| alaski | cdent: I think that's a rpodolyaka question, who unfortunately is not here | 19:31 |
| cdent | i'll try to test it | 19:31 |
| cdent | thanks | 19:32 |
| *** prateek_ has quit IRC | 19:33 | |
| *** rtheis_ has joined #openstack-nova | 19:33 | |
| *** armax has joined #openstack-nova | 19:35 | |
| *** rtheis has quit IRC | 19:35 | |
| *** rcernin has joined #openstack-nova | 19:39 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Fix qemu version check. https://review.openstack.org/363160 | 19:39 |
| sdague | mriedem et al ^^^ quick thing that I noticed when looking at logs | 19:40 |
| *** bvanhav has quit IRC | 19:41 | |
| *** fragatina has joined #openstack-nova | 19:43 | |
| *** su_zhang has joined #openstack-nova | 19:43 | |
| mriedem | looking | 19:45 |
| mriedem | heh, "Running Nova with a QEMU version less than 1.5.3 is deprecated. The required minimum version of QEMU will be raised to 1.5.3 in the next release." | 19:45 |
| mriedem | and we have: | 19:46 |
| mriedem | ii qemu-kvm 1:2.5+dfsg-5ubuntu10.4 amd64 QEMU Full virtualization | 19:46 |
| sdague | mriedem: yep | 19:46 |
| sdague | also, yay for error messages that don't tell you what it thinks we have | 19:47 |
| sdague | because, it would have been easy to see it was wrong | 19:47 |
| mriedem | open a low-hanging-fruit bug to clean that up and someone will pick it up fast | 19:48 |
| *** su_zhang has quit IRC | 19:48 | |
| *** jamesdenton has quit IRC | 19:48 | |
| dansmith | sdague: I get a nasty trace with your patch on top of a so-enabled devstack | 19:51 |
| dansmith | sdague: 2016-08-30 12:51:21.869 TRACE nova.compute.manager MissingAuthPlugin: An auth plugin is required to determine endpoint URL | 19:51 |
| melwitt | mriedem: what's the difference between terminating a volume connection and detaching it? I was checking out the recent code you added that checks the stashed connector to determine whether to terminate, | 19:52 |
| mriedem | if you'll excuse me, i have to head to church and start praying | 19:52 |
| mriedem | http://www.cbssports.com/nfl/news/vikings-stop-practice-after-teddy-bridgewater-suffers-apparent-knee-injury/ | 19:52 |
| sdague | dansmith: hmmmmm | 19:52 |
| sdague | dansmith: you have the devstack patch? | 19:52 |
| melwitt | mriedem: and I'm seeing it not seem to matter whether terminate connection is called or not. (if I have instance in error state with no host set) the detach call seems to free the volume and it's then available for attaching again | 19:53 |
| dansmith | aye | 19:53 |
| cmartin | mriedem: Dang that's unfortunate, if it makes you feel any better my favorite team is starting a 2nd year 7th round pick QB | 19:53 |
| dansmith | sdague: I'm on 4d6017566a2fd550b418609c8452e6cf35dd29a7 | 19:53 |
| mriedem | melwitt: that's probably a better question for the cinder channel people | 19:53 |
| sdague | dansmith: what does your placement section look like? | 19:53 |
| *** akshai has joined #openstack-nova | 19:53 | |
| dansmith | sdague: placement section of ... | 19:54 |
| dansmith | ? | 19:54 |
| sdague | in /etc/nova/nova.conf | 19:54 |
| melwitt | mriedem: okay, thanks | 19:54 |
| dansmith | sdague: grep placement -i /etc/nova/nova.conf | 19:54 |
| dansmith | sdague: nothing | 19:54 |
| cdent | dansmith: did you enable_service placement-api in your local.conf? | 19:54 |
| * dansmith looks sheepish | 19:55 | |
| mriedem | melwitt: i imagine it means different things to different volume backends | 19:55 |
| sdague | heh | 19:55 |
| dansmith | cdent: sdague: just enable_service placement-api ? | 19:55 |
| mriedem | yeah, latest version of the devstack change doesn't enable the placement-api by default | 19:55 |
| sdague | dansmith: yes | 19:55 |
| dansmith | sdague: regardless, we need to handle that gracefully as everyone will be in that boat on day zero of newton | 19:55 |
| cdent | dansmith: before today it was automatic, but sdague made me change it :) | 19:55 |
| sdague | dansmith: you have the full stack trace? | 19:56 |
| sdague | dansmith: I agree | 19:56 |
| dansmith | sdague: lemme try to get it before devstack kills it | 19:56 |
| openstackgerrit | Pushkar Umaranikar proposed openstack/nova: Expand feature matrix with server admin actions https://review.openstack.org/362408 | 19:56 |
| dansmith | sdague: http://pastebin.com/raw/azJusvqw | 19:57 |
| sdague | dansmith: ok, I can put that into the wrapper | 19:57 |
| *** harlowja has quit IRC | 19:58 | |
| dansmith | okay | 19:58 |
| *** rama_y has joined #openstack-nova | 19:58 | |
| *** su_zhang has joined #openstack-nova | 19:58 | |
| *** Apoorva has quit IRC | 20:00 | |
| *** Sukhdev has quit IRC | 20:00 | |
| openstackgerrit | Sean Dague proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/358797 | 20:00 |
| sdague | dansmith: I think that will fix it | 20:00 |
| dansmith | okay well, I am hoping just enabling the thing will let me test my part and then I will rebase | 20:01 |
| sdague | sure | 20:01 |
| dansmith | I have a question | 20:01 |
| sdague | shoot | 20:01 |
| dansmith | when you people are watching devstack run, | 20:02 |
| dansmith | you're shouting vile obscenities at the screen too, right? | 20:02 |
| *** su_zhang has quit IRC | 20:02 | |
| sdague | not usually, why? | 20:02 |
| dansmith | oh, hmm | 20:02 |
| dansmith | I thought that was a thing | 20:02 |
| sdague | because of the copious output? | 20:02 |
| dansmith | no | 20:02 |
| sdague | is that how you summon the gods of testing? | 20:02 |
| dansmith | haha | 20:03 |
| sdague | usually goat sacrifice works better | 20:03 |
| * dansmith looks around for a goat, finds only a cat | 20:03 | |
| cdent | dansmith: yes | 20:04 |
| mriedem | i usually look elsewhere whilst stacking | 20:04 |
| dansmith | I just expand the window to a full 27" and mirror the window on all screens | 20:04 |
| dansmith | in hopes someone who has ever seen a movie will come visit and be impressed | 20:05 |
| mriedem | and start swearing? | 20:05 |
| dansmith | hah | 20:05 |
| mriedem | oh i see, like a 90s hacker movie? | 20:05 |
| dansmith | s/a 90s/every/ | 20:05 |
| mriedem | where people are flying through some cyber space thing while coding | 20:05 |
| sdague | we should totally add some extra matrix bits into it | 20:05 |
| dansmith | and furious typing as if I'm causing all the output | 20:05 |
| mriedem | https://www.youtube.com/watch?v=u1Ds9CeG-VY | 20:05 |
| mriedem | ^ is dan at home | 20:06 |
| dansmith | Yes, I "assemble port scan objects" regularly | 20:06 |
| melwitt | lol ... I feel embarrassed watching that | 20:09 |
| *** dikonoor has quit IRC | 20:10 | |
| *** su_zhang has joined #openstack-nova | 20:10 | |
| cdent | hugh jackman and dansmith do look a bit alike | 20:10 |
| * dansmith feels sorry for hugh jackman | 20:11 | |
| *** sdake has quit IRC | 20:11 | |
| *** pradk has quit IRC | 20:11 | |
| *** sdake has joined #openstack-nova | 20:12 | |
| *** mdrabe_ is now known as mdrabe | 20:13 | |
| *** raunak has joined #openstack-nova | 20:14 | |
| tdurakov | folks, do we have any issues with py27 unit-tests right now? | 20:16 |
| dansmith | sdague: now I get this: 2016-08-30 13:15:40.043 ERROR nova.scheduler.client.report [req-990b33be-ecc5-4ea6-bcef-7d1cc5859876 None None] Failed to create resource provider record in placement API for UUID 5cdba15c-1409-422c-8274-746795e6c595. Got 400: 400 Bad Request | 20:16 |
| *** sbezverk has joined #openstack-nova | 20:16 | |
| *** pradk has joined #openstack-nova | 20:16 | |
| dansmith | sdague: because: JSON does not validate | 20:16 |
| dansmith | sdague: http://pastebin.com/Wpugnpdh | 20:17 |
| sdague | json inbound or outbound | 20:17 |
| dansmith | presumably means you don't match the schema that the server side requires | 20:17 |
| mriedem | it's not nested in some object | 20:17 |
| mriedem | does it need to be in a 'resource' dict or something? | 20:17 |
| sdague | ok, I was pretty sure I saw that working here, but I can do a full clean / start again | 20:18 |
| dansmith | okay | 20:18 |
| sdague | that's definitely server side json enforcement | 20:19 |
| *** ducttape_ has quit IRC | 20:19 | |
| dansmith | I'm going to run off for a bit but will circle back after you've had a chance to look | 20:19 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/api/openstack/placement/handlers/resource_provider.py#L27 | 20:19 |
| mriedem | seems right | 20:19 |
| sdague | yeh, I'm really confused honestly, this is exactly an object | 20:20 |
| mriedem | here is the related gabbit https://github.com/openstack/nova/blob/master/nova/tests/functional/api/openstack/placement/gabbits/resource-provider.yaml#L34 | 20:20 |
| cdent | is the content-type header being sent? | 20:21 |
| sdague | if not, it would be a 415 | 20:21 |
| cdent | yeah | 20:21 |
| * cdent hmmms | 20:21 | |
| cdent | oh, hang on | 20:21 |
| *** ducttape_ has joined #openstack-nova | 20:22 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: ironic_host_manager: fix population of instances info on start https://review.openstack.org/347948 | 20:22 |
| cdent | nevermind, I thought perhaps this has already merged and was causing confusion: https://review.openstack.org/#/c/361422/ | 20:22 |
| cdent | dansmith: is there any chance you have an older jsonschema? | 20:23 |
| cdent | (python module) | 20:23 |
| sdague | cdent: in devstack we should be using whatever is in upper constraints | 20:24 |
| cdent | but unless pip is being told to force update and he's got old cruft laying around? (speculating) | 20:24 |
| sdague | no, devstack is pretty explicit | 20:25 |
| sdague | so 2.5.1 should be what we are on | 20:25 |
| *** gyee has joined #openstack-nova | 20:25 | |
| *** rtheis_ is now known as rtheis | 20:26 | |
| *** kevinz has quit IRC | 20:27 | |
| *** sbezverk_ has joined #openstack-nova | 20:29 | |
| *** sbezverk has quit IRC | 20:29 | |
| *** FL1SK has quit IRC | 20:30 | |
| *** hemanthm is now known as hemanthm|bbl | 20:32 | |
| *** rcernin has quit IRC | 20:33 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: WIP - Added FaultWrapper to the placement API https://review.openstack.org/363125 | 20:33 |
| *** rcernin has joined #openstack-nova | 20:33 | |
| *** Apoorva has joined #openstack-nova | 20:34 | |
| *** rstarmer has joined #openstack-nova | 20:34 | |
| *** ianw_ is now known as ianw | 20:39 | |
| *** ijw has joined #openstack-nova | 20:40 | |
| mriedem | sdague: dansmith: cdent: we now have a job in the experimental queue for running the placement-api | 20:42 |
| mriedem | gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv | 20:42 |
| cdent | mriedem++ | 20:42 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: do not merge https://review.openstack.org/362922 | 20:42 |
| sdague | mriedem: awesome | 20:44 |
| bauzas | mriedem: just catching up, bravo | 20:46 |
| mriedem | oh but it takes awhile to be deployed i think | 20:46 |
| mriedem | like job changes happen every 30 minutes or something? | 20:47 |
| openstackgerrit | Andrew Laski proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | 20:49 |
| bauzas | dansmith: sdague: cdent: I briefly followed your hangout and the etherpad as I was/am mostly off for tonight, but I can help continuing work on https://review.openstack.org/#/c/363061/ and https://review.openstack.org/#/c/358797 on my EU morning | 20:52 |
| bauzas | dansmith: just make sure to upload your latest branch before leaving today so I can take your work and continue | 20:53 |
| sdague | mriedem: something like that | 20:53 |
| tdurakov | g | 20:53 |
| tdurakov | ops | 20:53 |
| tdurakov | sdague: hi, do we have a plan to move ceph ci jobs to xenial? | 20:54 |
| *** lpetrut has joined #openstack-nova | 20:54 | |
| tdurakov | sdague: if we do who is working on that? | 20:54 |
| sdague | tdurakov: probably just needs to be worked with infra | 20:54 |
| sdague | tdurakov: probably no one | 20:54 |
| tdurakov | sdague: at least we need a patch to devstack | 20:54 |
| sdague | do we? | 20:55 |
| mriedem | tdurakov: any changes would be in the devstack ceph plugin repo | 20:55 |
| mriedem | to probably make it straddle both xenial and trusty | 20:55 |
| mriedem | but i'm not sure what changes would be needed | 20:55 |
| tdurakov | sdague: at least change this one https://github.com/openstack-dev/devstack/blob/master/lib/ceph#L356 | 20:56 |
| tdurakov | sdague: systemd instead of upstart | 20:56 |
| mriedem | tdurakov: i think https://github.com/openstack-dev/devstack/blob/master/lib/ceph needs to just be dropped | 20:57 |
| mriedem | for one | 20:57 |
| mriedem | i don't think any jobs are using it | 20:57 |
| mriedem | tdurakov: we use this https://github.com/openstack/devstack-plugin-ceph | 20:57 |
| sdague | tdurakov: if you have patches to review, let me know | 20:57 |
| tdurakov | mriedem: ha, just wanted to ask about that | 20:57 |
| mriedem | https://github.com/openstack/devstack-plugin-ceph/blob/master/devstack/lib/ceph#L335 | 20:57 |
| sdague | otherwise, I'm not thinking about it until post freeze | 20:58 |
| mriedem | looks like it already handles this | 20:58 |
| mriedem | https://github.com/openstack/devstack-plugin-ceph/commit/19051aa4cd779d9f6265a1a0a9fd00617ac72dab | 20:58 |
| tdurakov | mriedem: right, thank you for pointing to the right source | 20:59 |
| mriedem | tdurakov: it's been awhile since i've done an audit in project-config of the ceph jobs out there, but i think ceilometer was the last using a ceph job with the devstack stuff, and that was moving to use the plugin repo | 21:00 |
| *** baoli has quit IRC | 21:00 | |
| mriedem | i'd have to double check | 21:00 |
| mriedem | sorry i mean gnocchi | 21:00 |
| tdurakov | mriedem: hm, when I worked on ceph part for live-migration jobs there we several differences, that forces me to duplicate some of that ceph functions, I'll do some checks, and then try to reuse that plugin instead | 21:01 |
| mriedem | https://review.openstack.org/#/c/315994/ | 21:03 |
| mriedem | tdurakov: there is an experimental queue job for running devstack-plugin-ceph with xenial | 21:05 |
| mriedem | but that's on the devstack-plugin-ceph repo | 21:06 |
| *** jamielennox is now known as jamielennox|away | 21:06 | |
| tdurakov | mriedem: it looks like I could clone that in post_hook, we don't need ceph to be initially installed, I'd just reuse plugin lib functions | 21:08 |
| mriedem | yeah i think that's what gnocchi does | 21:09 |
| scottda | mriedem: You had comments on this patch and I think I've addressed them. Please review when you have time: https://review.openstack.org/#/c/326681/ | 21:12 |
| scottda | Tempest Migrate volume while attached to an instance to test Nova swap_volume, etc. | 21:12 |
| *** thorst has quit IRC | 21:13 | |
| mikal | . | 21:14 |
| *** raunak has quit IRC | 21:14 | |
| mriedem | hmm, i'm not sure if we can remove this https://github.com/openstack-dev/devstack/blob/master/files/debs/ceph | 21:16 |
| mriedem | since https://github.com/openstack/devstack-plugin-ceph/blob/master/devstack/lib/ceph doesn't install xfsprogs | 21:17 |
| *** harlowja has joined #openstack-nova | 21:17 | |
| mriedem | but it's used here https://github.com/openstack-dev/devstack/blob/master/functions#L630 | 21:17 |
| mriedem | mikal: what's going on with this? https://review.openstack.org/#/c/348066/ | 21:18 |
| mriedem | it's the last unmerged patch for the vd2 stuff | 21:18 |
| *** rstarmer has quit IRC | 21:18 | |
| mikal | mriedem: please hold | 21:19 |
| mikal | Oh that one | 21:19 |
| mikal | So I think we've had a brain fart there | 21:19 |
| mikal | There was an openstack-dev thread about it | 21:19 |
| *** raunak has joined #openstack-nova | 21:19 | |
| *** ayoung has quit IRC | 21:19 | |
| *** gouthamr has quit IRC | 21:19 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Make req spec and RT aware of volume-backed instances https://review.openstack.org/355091 | 21:20 |
| mikal | Oh, maybe not | 21:20 |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add create_all and delete_all to AllocationList https://review.openstack.org/363209 | 21:20 |
| mikal | mriedem: its a little complicated. Do you have a moment to work through the thought process? | 21:20 |
| *** rtheis has quit IRC | 21:20 | |
| cdent | dansmith, mriedem, sdague 363209 there ^^ is a start, but a lame one. I'll continue on it tomorrow, but need to stop now, getting stupid and sore. | 21:21 |
| sdague | cdent: no worries | 21:21 |
| mikal | It goes a bit like this... | 21:21 |
| cdent | sdague: any review commentary prior to tomorrow is most welcome, in case I've gone off in the weeds | 21:21 |
| * cdent waves goodnight | 21:22 | |
| *** cdent has quit IRC | 21:22 | |
| mikal | There are two cases for a vendordata request. The request is caused by a nova operation like boot, where we know the authentication details of the user doing the operation and we can pass those through to the external rest api | 21:22 |
| openstackgerrit | Sean Dague proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/358797 | 21:22 |
| mikal | The second case is the instance requesting metadata, where we don't know the auth details of the user that created the instance, and so we can't pass those through | 21:22 |
| mikal | So that patch passed through the auth details for the nova rest api iniated case | 21:23 |
| mikal | But leaves no auth for the instance initiated case | 21:23 |
| mikal | The suggestion on openstack-dev was to use a keystone account instead of the user creds | 21:23 |
| mikal | And I had presumed we'd do that for all the calls to the vendordata rest server things | 21:23 |
| mikal | But maybe I'm wrong? Should we just use a service account for those where we don't know the user details? | 21:24 |
| sdague | mriedem: I unfuxored a thing on https://review.openstack.org/358797 which unblocked dan | 21:25 |
| sdague | but it still needs unit tests, if anyone wants to pile in there, cool. Otherwise I'll probably get the placement-api doing logging tomorrow to make debugging the partial patches easier | 21:25 |
| mriedem | sdague: ah ok i see how that would mess up the validation | 21:25 |
| sdague | mriedem: yep | 21:25 |
| *** sdake has quit IRC | 21:26 | |
| sdague | we got on a hangout and I said "well the only reason this might not work is if we didn't need this encoding" and dan looked and said "oh, have you never used requests before" :P so, sorted | 21:26 |
| mriedem | mikal: hmm, if we have a service account, it seems we'd want to just use that always | 21:26 |
| *** salv-orlando has quit IRC | 21:26 | |
| mikal | mriedem: that's kind of where I was going, hence not wanting to land all that driver change for the other thing | 21:27 |
| sdague | signing off now for dinner | 21:27 |
| mikal | mriedem: then again, the user details when available tell you a lot more than a service account does | 21:27 |
| *** zengine has quit IRC | 21:27 | |
| mikal | mriedem: which is why I am conflicted | 21:27 |
| mikal | mriedem: also, I don't have the service account thing written because I've been pulled into something internal at work for the last couple of weeks | 21:27 |
| mriedem | mikal: yeah, so i was going to say, does not having the creds for when the instance is requesting vendordata a blocker to using it? | 21:28 |
| mriedem | or does that just depend on the impl of the actual vd provider? | 21:28 |
| mikal | So, I don't think its a blocker | 21:28 |
| mriedem | i mean, you don't have this with config drive either | 21:28 |
| mriedem | since it doesn't change | 21:28 |
| mikal | I am unsure | 21:29 |
| mikal | I think you need the user creds for things like AD to know which user to add the instance to | 21:29 |
| mriedem | well, i mean, with config drive you're just going to get the initial vendor data | 21:29 |
| *** mingyu has quit IRC | 21:29 | |
| *** rfolco has quit IRC | 21:29 | |
| mikal | And you'd have those after the first call if creating an instance caused a call to metadata | 21:29 |
| mikal | But I think it only does that for config drive | 21:29 |
| mikal | The external thing could cache it after that first call | 21:30 |
| mriedem | so if the vd provider blocks requests that don't provide auth, that's going to basically make this unusable for any non-libvirt driver out of the gate right? | 21:30 |
| mikal | Without that patch, yes | 21:30 |
| mriedem | so, it seems we want to get that patch in | 21:30 |
| mikal | And then add a service account on top later if we need to? | 21:31 |
| mriedem | and work a service account thing in as a later future enhancement | 21:31 |
| mriedem | yeah | 21:31 |
| mikal | I'd be ok with that | 21:31 |
| mriedem | ok | 21:31 |
| mikal | Ok, I can remove my -W then | 21:31 |
| mriedem | cool, was just going to ask you to :) | 21:31 |
| mikal | We should reply to the dev thread as well though | 21:31 |
| mriedem | was this a nova thread? | 21:31 |
| mikal | Let me find it | 21:31 |
| openstackgerrit | Andrew Laski proposed openstack/nova: DNM logs for debugging https://review.openstack.org/362442 | 21:31 |
| *** sdake has joined #openstack-nova | 21:32 | |
| *** armax has quit IRC | 21:32 | |
| mikal | mriedem: Subject line is "[openstack-dev] [nova][keystone] auth for new metadata plugins" | 21:32 |
| mikal | Oh, there's a middleware problem I hadn't twigged to | 21:33 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: DNM: test rm -f devstack/lib/ceph https://review.openstack.org/363216 | 21:34 |
| *** claudiub has quit IRC | 21:35 | |
| *** cleong has quit IRC | 21:35 | |
| mikal | mriedem: I replied to the thread, which should make it easier to find | 21:36 |
| mriedem | yup reading it | 21:36 |
| *** rfolco has joined #openstack-nova | 21:37 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: tests: verify pci passthrough with numa https://review.openstack.org/356077 | 21:37 |
| openstackgerrit | melanie witt proposed openstack/nova: Detach volume after deleting instance with no host https://review.openstack.org/340614 | 21:38 |
| *** rfolco has quit IRC | 21:39 | |
| mriedem | melwitt: whomg | 21:39 |
| mriedem | you're still trying to make that work huh | 21:39 |
| mriedem | i love the list of reverts in that commit message | 21:40 |
| melwitt | mriedem: yeah :( I had to spend some time figuring out what else was needed since the code evolved quite a bit since I last had it working | 21:40 |
| melwitt | and yeah, the commit message is ... a saga haha | 21:42 |
| *** abalutoiu has quit IRC | 21:42 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Add a hacking rule for string interpolation at logging https://review.openstack.org/334844 | 21:42 |
| *** Swami has quit IRC | 21:44 | |
| *** _alastor__ has joined #openstack-nova | 21:45 | |
| _alastor__ | Hey folks, quick question. Say I have a flat external network netA1. On this network I have a bunch of subnets, eg netA1-sub1, netA1-sub2 etc. I want to boot a VM that uses a specific subnet, but nova boot's `nic` argument only allows me to specify a net-id, not a subnet-id How do I accomplish this? | 21:45 |
| *** tblakes has quit IRC | 21:47 | |
| *** eharney has quit IRC | 21:48 | |
| *** mingyu has joined #openstack-nova | 21:49 | |
| *** chrichip has joined #openstack-nova | 21:49 | |
| *** salv-orlando has joined #openstack-nova | 21:49 | |
| *** marst has quit IRC | 21:50 | |
| mriedem | mikal: replied | 21:54 |
| *** tonytan4ever has quit IRC | 21:54 | |
| mriedem | just for the record | 21:54 |
| *** mlavalle has joined #openstack-nova | 21:55 | |
| mikal | Ok | 21:57 |
| mriedem | big ol +W on that, i'll mark the bp complete tomorrow | 21:57 |
| mriedem | thanks for getting those follow up changes done | 21:57 |
| mikal | Its ok, sorry for being pulled away for a couple of weeks recently | 21:57 |
| mikal | Its a thing I don't enjoy, but don't have a lot of control over | 21:58 |
| mriedem | np. just go merge all of the cells v2 stuff today :) | 21:58 |
| mikal | The service user thing doesn't look too bad to do. I might try and whip up a patch to do that thing to see how bad it is... | 21:58 |
| mikal | mriedem: are they on the priority etherpad? If they are I'll take a look today. | 21:58 |
| mriedem | i don't think alaski was updating that anymore | 21:58 |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Remove context object in oslo.log method https://review.openstack.org/317109 | 21:59 |
| mikal | mriedem: ok, so how do I determine what his review priorities are? | 22:00 |
| dansmith | sdague: I have it working now | 22:00 |
| mriedem | might have to ask him, or dansmith, or melwitt | 22:00 |
| mriedem | time for me to run | 22:01 |
| openstackgerrit | Dan Smith proposed openstack/nova: Report compute node inventories through placement https://review.openstack.org/363061 | 22:01 |
| *** marst has joined #openstack-nova | 22:01 | |
| mriedem | dansmith: run check experimental on that too | 22:01 |
| dansmith | mriedem: okay | 22:01 |
| *** mriedem has quit IRC | 22:01 | |
| *** burt has quit IRC | 22:02 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: WIP: Use admin context to create neutron ports https://review.openstack.org/357540 | 22:05 |
| *** chrichip has quit IRC | 22:06 | |
| *** rstarmer has joined #openstack-nova | 22:07 | |
| auggy | i've been having a couple folks repeatedly running intel-nfv-ci recheck on my WIP patches | 22:07 |
| auggy | is someone trying to fix an issue that anyone knows of? i'm more worried about gate load | 22:08 |
| *** rlrossit has quit IRC | 22:09 | |
| *** mdrabe has quit IRC | 22:11 | |
| *** mingyu has quit IRC | 22:11 | |
| *** lpetrut has quit IRC | 22:12 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: ironic_host_manager: fix population of instances info on start https://review.openstack.org/347948 | 22:13 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: WIP: Assume neutron port_binding extensions enabled https://review.openstack.org/357726 | 22:17 |
| *** diogogmt has quit IRC | 22:20 | |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Using get() method to prevent KeyError https://review.openstack.org/359424 | 22:22 |
| *** su_zhang has quit IRC | 22:22 | |
| *** cn28h has quit IRC | 22:23 | |
| *** cn28h has joined #openstack-nova | 22:26 | |
| *** ianw has quit IRC | 22:26 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: WIP: Assume neutron port_binding extensions enabled https://review.openstack.org/357726 | 22:31 |
| *** diogogmt has joined #openstack-nova | 22:32 | |
| *** ducttape_ has quit IRC | 22:34 | |
| *** su_zhang has joined #openstack-nova | 22:34 | |
| *** jamesdenton has joined #openstack-nova | 22:36 | |
| *** sdake has quit IRC | 22:36 | |
| *** nic has quit IRC | 22:38 | |
| *** hemna is now known as hemnafk | 22:39 | |
| *** FL1SK has joined #openstack-nova | 22:39 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: WIP: Use admin context to create neutron ports https://review.openstack.org/357540 | 22:41 |
| *** tomhambleton has joined #openstack-nova | 22:44 | |
| *** Swami has joined #openstack-nova | 22:47 | |
| *** Swami_ has joined #openstack-nova | 22:47 | |
| *** jamesdenton has quit IRC | 22:48 | |
| *** erhudy has quit IRC | 22:52 | |
| dansmith | bauzas: if you want to write unit tests on my patch for inventory, and maybe see if any more need writing for the base patch tomorrow before I get up that would rock | 22:56 |
| dansmith | bauzas: it works for me as it is now, so I think the functional code is good | 22:56 |
| *** ianw has joined #openstack-nova | 22:57 | |
| *** ayoung has joined #openstack-nova | 22:59 | |
| *** armax has joined #openstack-nova | 23:00 | |
| *** pradk has quit IRC | 23:00 | |
| melwitt | dansmith: I put up this patch for the devstack rabbit userid, we'll see if that's good enough https://review.openstack.org/#/c/363295/ | 23:00 |
| dansmith | melwitt: cool | 23:00 |
| *** mingyu has joined #openstack-nova | 23:01 | |
| *** ducttape_ has joined #openstack-nova | 23:03 | |
| *** zengine has joined #openstack-nova | 23:06 | |
| *** mingyu has quit IRC | 23:08 | |
| *** praneshp has quit IRC | 23:08 | |
| *** ducttape_ has quit IRC | 23:09 | |
| *** Apoorva has quit IRC | 23:09 | |
| *** takashin has joined #openstack-nova | 23:11 | |
| *** rkrum has joined #openstack-nova | 23:12 | |
| *** praneshp has joined #openstack-nova | 23:13 | |
| *** Apoorva has joined #openstack-nova | 23:14 | |
| *** jamesdenton has joined #openstack-nova | 23:16 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: WIP: Enable cold migration with target host https://review.openstack.org/357105 | 23:19 |
| *** praneshp has quit IRC | 23:19 | |
| *** mingyu has joined #openstack-nova | 23:20 | |
| *** salv-orlando has quit IRC | 23:21 | |
| cfriesen | dansmith: when you were switching from InstanceV1 to InstanceV2, how did you handle newer code receiving the InstanceV1 object since the major version is different? | 23:22 |
| *** su_zhang has quit IRC | 23:23 | |
| alaski | mikal: the priorities etherpad is actually up to date on cells stuff, outside of devstack/grenade patches | 23:31 |
| *** sdague has quit IRC | 23:31 | |
| mikal | alaski: oh cool. Thanks. | 23:33 |
| *** jamesdenton has quit IRC | 23:34 | |
| alaski | mikal: not at all. thank you | 23:35 |
| *** rcernin has quit IRC | 23:35 | |
| cfriesen | dansmith: never mind, procedural error in my testing | 23:36 |
| *** mingyu has quit IRC | 23:39 | |
| *** praneshp has joined #openstack-nova | 23:42 | |
| *** rama_y has quit IRC | 23:46 | |
| *** mingyu has joined #openstack-nova | 23:51 | |
| *** sameo has quit IRC | 23:52 | |
| *** mriedem has joined #openstack-nova | 23:57 | |
| *** tonytan4ever has joined #openstack-nova | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
