| *** achanda has quit IRC | 00:00 | |
| *** tnovacik has quit IRC | 00:03 | |
| *** baoli has joined #openstack-infra | 00:04 | |
| *** baoli has quit IRC | 00:04 | |
| *** salv-orl_ has quit IRC | 00:06 | |
| *** armax has quit IRC | 00:07 | |
| *** masoz has joined #openstack-infra | 00:22 | |
| *** sdake has joined #openstack-infra | 00:23 | |
| *** sdake has quit IRC | 00:26 | |
| *** tiswanso has joined #openstack-infra | 00:26 | |
| *** annegentle has quit IRC | 00:26 | |
| *** sdake has joined #openstack-infra | 00:26 | |
| *** annegentle has joined #openstack-infra | 00:29 | |
| *** tiswanso has quit IRC | 00:30 | |
| openstackgerrit | Angus Lees proposed openstack-infra/system-config: Add libcap-dev in anticipation of python-prctl https://review.openstack.org/168750 | 00:35 |
|---|---|---|
| *** davideagnello has joined #openstack-infra | 00:37 | |
| *** gokrokve has joined #openstack-infra | 00:39 | |
| *** davideagnello has quit IRC | 00:43 | |
| *** ddieterl_ has joined #openstack-infra | 00:44 | |
| *** gokrokve has quit IRC | 00:44 | |
| *** ddieterly has quit IRC | 00:44 | |
| *** smccully has quit IRC | 00:48 | |
| *** yamamoto has joined #openstack-infra | 00:49 | |
| *** jhesketh has quit IRC | 00:50 | |
| *** jhesketh has joined #openstack-infra | 00:53 | |
| *** tiswanso has joined #openstack-infra | 00:55 | |
| *** hdd has quit IRC | 00:59 | |
| *** imcsk8 has quit IRC | 01:00 | |
| *** imcsk8 has joined #openstack-infra | 01:00 | |
| *** achanda has joined #openstack-infra | 01:01 | |
| *** eharney has quit IRC | 01:04 | |
| *** tiswanso has quit IRC | 01:06 | |
| *** achanda has quit IRC | 01:06 | |
| *** tiswanso has joined #openstack-infra | 01:06 | |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Add query for horizon integration test bug 1436903 https://review.openstack.org/168753 | 01:07 |
| openstack | bug 1436903 in OpenStack Dashboard (Horizon) "integration tests failing blocking gate" [Critical,Triaged] https://launchpad.net/bugs/1436903 - Assigned to David Lyle (david-lyle) | 01:07 |
| *** salv-orlando has joined #openstack-infra | 01:07 | |
| *** shakamunyi has joined #openstack-infra | 01:07 | |
| *** shakamunyi_ has joined #openstack-infra | 01:07 | |
| *** salv-orlando has quit IRC | 01:12 | |
| *** chris6131 has joined #openstack-infra | 01:14 | |
| *** Longgeek has joined #openstack-infra | 01:15 | |
| *** yamamoto has quit IRC | 01:15 | |
| *** ddieterly has joined #openstack-infra | 01:18 | |
| *** annegentle has quit IRC | 01:19 | |
| *** ddieterl_ has quit IRC | 01:21 | |
| *** masoz has quit IRC | 01:29 | |
| *** kaisers1 has joined #openstack-infra | 01:35 | |
| *** kaisers has quit IRC | 01:35 | |
| *** baoli has joined #openstack-infra | 01:37 | |
| *** gokrokve has joined #openstack-infra | 01:39 | |
| *** gokrokve has quit IRC | 01:44 | |
| *** stevemar has quit IRC | 01:45 | |
| *** stevemar has joined #openstack-infra | 01:46 | |
| *** yamamoto has joined #openstack-infra | 01:46 | |
| *** sdake has quit IRC | 01:51 | |
| *** ghostpl_ has joined #openstack-infra | 01:53 | |
| *** salv-orlando has joined #openstack-infra | 01:53 | |
| *** salv-orlando has quit IRC | 01:58 | |
| *** ghostpl_ has quit IRC | 02:00 | |
| *** baoli has quit IRC | 02:01 | |
| *** sputnik13 has quit IRC | 02:05 | |
| *** SumitNaiksatam has left #openstack-infra | 02:07 | |
| *** salv-orlando has joined #openstack-infra | 02:10 | |
| *** salv-orlando has quit IRC | 02:12 | |
| *** Longgeek has quit IRC | 02:12 | |
| *** SumitNaiksatam has joined #openstack-infra | 02:13 | |
| *** Longgeek has joined #openstack-infra | 02:16 | |
| *** Longgeek has quit IRC | 02:26 | |
| *** Longgeek has joined #openstack-infra | 02:26 | |
| openstackgerrit | Angus Lees proposed openstack/requirements: Add pyroute2 (python netlink bindings) https://review.openstack.org/168764 | 02:28 |
| *** julim has quit IRC | 02:29 | |
| *** baoli has joined #openstack-infra | 02:29 | |
| *** Somay has quit IRC | 02:32 | |
| *** tiswanso has quit IRC | 02:32 | |
| *** mriedem1 has joined #openstack-infra | 02:37 | |
| *** mriedem has quit IRC | 02:39 | |
| *** gokrokve has joined #openstack-infra | 02:39 | |
| *** baoli has quit IRC | 02:42 | |
| *** gokrokve has quit IRC | 02:44 | |
| *** baoli has joined #openstack-infra | 02:45 | |
| *** annegentle has joined #openstack-infra | 02:45 | |
| *** sputnik13 has joined #openstack-infra | 02:51 | |
| *** annegentle has quit IRC | 02:51 | |
| *** shakamunyi__ has joined #openstack-infra | 02:51 | |
| *** david-ly_ is now known as david-lyle | 02:52 | |
| openstackgerrit | Steve Kowalik proposed openstack-infra/puppet-zanata: Add a zanata::client manifest https://review.openstack.org/167081 | 02:53 |
| *** shakamunyi has quit IRC | 02:54 | |
| *** shakamunyi_ has quit IRC | 02:54 | |
| *** shakamunyi has joined #openstack-infra | 02:54 | |
| *** mriedem1 has quit IRC | 02:58 | |
| *** dims has quit IRC | 02:59 | |
| *** dims has joined #openstack-infra | 02:59 | |
| *** dims has quit IRC | 03:00 | |
| *** Longgeek_ has joined #openstack-infra | 03:07 | |
| *** Somay has joined #openstack-infra | 03:07 | |
| *** Longgeek has quit IRC | 03:10 | |
| *** jerryz_ has joined #openstack-infra | 03:10 | |
| *** sabeen has joined #openstack-infra | 03:11 | |
| *** ajmiller has quit IRC | 03:13 | |
| *** sabeen has quit IRC | 03:14 | |
| *** ayoung is now known as ayoung_ZZzz__ | 03:21 | |
| *** medieval1_ has quit IRC | 03:24 | |
| *** armax has joined #openstack-infra | 03:24 | |
| *** salv-orlando has joined #openstack-infra | 03:25 | |
| *** armax has quit IRC | 03:25 | |
| *** ddieterly has quit IRC | 03:27 | |
| *** sandywalsh has quit IRC | 03:28 | |
| *** salv-orlando has quit IRC | 03:30 | |
| *** tkelsey has joined #openstack-infra | 03:36 | |
| *** sandywalsh has joined #openstack-infra | 03:37 | |
| *** otter768 has quit IRC | 03:38 | |
| *** gokrokve has joined #openstack-infra | 03:39 | |
| *** tkelsey has quit IRC | 03:41 | |
| *** hdd has joined #openstack-infra | 03:42 | |
| *** ghostpl_ has joined #openstack-infra | 03:43 | |
| *** gokrokve has quit IRC | 03:43 | |
| *** ghostpl_ has quit IRC | 03:48 | |
| *** baoli has quit IRC | 03:49 | |
| *** sdake has joined #openstack-infra | 03:49 | |
| *** sabeen1 has joined #openstack-infra | 03:58 | |
| *** sabeen1 has quit IRC | 04:05 | |
| *** dims has joined #openstack-infra | 04:15 | |
| *** woodster_ has quit IRC | 04:20 | |
| *** prmtl has joined #openstack-infra | 04:20 | |
| *** dims has quit IRC | 04:21 | |
| *** pal_ has joined #openstack-infra | 04:22 | |
| *** achanda has joined #openstack-infra | 04:24 | |
| *** ddieterly has joined #openstack-infra | 04:28 | |
| *** ddieterly has quit IRC | 04:32 | |
| *** shakamunyi_ has joined #openstack-infra | 04:39 | |
| *** gokrokve has joined #openstack-infra | 04:39 | |
| *** shakamunyi__ has quit IRC | 04:41 | |
| *** shakamunyi has quit IRC | 04:41 | |
| *** tnovacik has joined #openstack-infra | 04:41 | |
| *** shakamunyi has joined #openstack-infra | 04:41 | |
| *** gokrokve has quit IRC | 04:44 | |
| *** sdake has quit IRC | 04:49 | |
| *** pal__ has joined #openstack-infra | 04:53 | |
| *** pal_ has quit IRC | 04:56 | |
| *** BharatK has joined #openstack-infra | 04:57 | |
| *** masoz has joined #openstack-infra | 04:59 | |
| *** shakamunyi_ has quit IRC | 05:04 | |
| *** jerryz_ has quit IRC | 05:05 | |
| *** pal__ has quit IRC | 05:10 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/project-config: Restore devstack centos7 to non-voting check job https://review.openstack.org/168780 | 05:11 |
| *** pal_ has joined #openstack-infra | 05:12 | |
| *** mrodden has quit IRC | 05:17 | |
| *** jerryz_ has joined #openstack-infra | 05:17 | |
| *** shakamunyi has quit IRC | 05:18 | |
| *** Alexandra_ has joined #openstack-infra | 05:19 | |
| *** unicell has quit IRC | 05:22 | |
| *** unicell has joined #openstack-infra | 05:22 | |
| *** achanda has quit IRC | 05:22 | |
| *** achanda_ has joined #openstack-infra | 05:22 | |
| *** shakamunyi has joined #openstack-infra | 05:22 | |
| *** jlibosva has joined #openstack-infra | 05:25 | |
| *** ildikov has quit IRC | 05:27 | |
| greghaynes | jhesketh: Reviewing your zuul connection patch, having a hard time figuring out why https://review.openstack.org/#/c/121528/22/zuul/model.py is needed | 05:27 |
| greghaynes | (the none -> [] changes) | 05:27 |
| greghaynes | makes me afraid were referencing those actions before the config is loaded, which seems like a possible race issue | 05:28 |
| *** ddieterly has joined #openstack-infra | 05:29 | |
| *** achanda has joined #openstack-infra | 05:29 | |
| *** achanda_ has quit IRC | 05:30 | |
| greghaynes | I removed it and tests still pass... | 05:30 |
| *** chris6131 has quit IRC | 05:30 | |
| *** mrodden has joined #openstack-infra | 05:31 | |
| *** pal__ has joined #openstack-infra | 05:32 | |
| *** ghostpl_ has joined #openstack-infra | 05:32 | |
| *** ddieterly has quit IRC | 05:34 | |
| *** pal_ has quit IRC | 05:34 | |
| *** woodster_ has joined #openstack-infra | 05:37 | |
| *** ghostpl_ has quit IRC | 05:37 | |
| *** otter768 has joined #openstack-infra | 05:39 | |
| *** hdd has quit IRC | 05:39 | |
| *** otter768 has quit IRC | 05:43 | |
| *** tnovacik has quit IRC | 05:47 | |
| *** Somay has quit IRC | 05:48 | |
| jhesketh | greghaynes: all the actions are looped over to send notifications... looping over an empty list is harmless and avoids the case where no merge_failure action types are set for example | 05:51 |
| greghaynes | jhesketh: Yes, but they should all be empty lists due to scheduler._parseConfig | 05:53 |
| greghaynes | as long as they are only being reference after the config is loaded, if they arent then for some reason were trying to perform actions before loading the config, which seems like it could be bad | 05:53 |
| jhesketh | it's true, changing them to a list isn't needed here | 05:57 |
| jhesketh | but it denotes what is expected in those values | 05:57 |
| jhesketh | happy to change it back though | 05:57 |
| greghaynes | ok, just making sure. I dont mind that they are I just wanted to make sure it wasnt a result of something bad | 05:57 |
| *** prmtl has quit IRC | 06:01 | |
| *** pal_ has joined #openstack-infra | 06:02 | |
| *** pal__ has quit IRC | 06:05 | |
| greghaynes | jhesketh: ok, replied, nothing huge but I think runtime imports are really important to avoid if at all possible | 06:07 |
| greghaynes | Need to go to bed now, see you on the other side :) | 06:09 |
| *** YorikSar has joined #openstack-infra | 06:10 | |
| *** prmtl has joined #openstack-infra | 06:10 | |
| *** e0ne has joined #openstack-infra | 06:19 | |
| *** jamielennox is now known as jamielennox|away | 06:20 | |
| *** deepakcs has joined #openstack-infra | 06:22 | |
| *** markvoelker has joined #openstack-infra | 06:22 | |
| *** e0ne has quit IRC | 06:23 | |
| *** shakamunyi has quit IRC | 06:25 | |
| *** yfried has joined #openstack-infra | 06:26 | |
| *** stevemar has quit IRC | 06:27 | |
| *** markvoelker has quit IRC | 06:27 | |
| *** stevemar has joined #openstack-infra | 06:27 | |
| *** e0ne has joined #openstack-infra | 06:28 | |
| openstackgerrit | Jan Provaznik proposed openstack-infra/project-config: Create os-cloud-management project under TripleO https://review.openstack.org/165433 | 06:29 |
| *** ddieterly has joined #openstack-infra | 06:29 | |
| *** ddieterly has quit IRC | 06:34 | |
| *** stevemar has quit IRC | 06:34 | |
| *** stevemar has joined #openstack-infra | 06:34 | |
| *** e0ne is now known as e0ne_ | 06:35 | |
| *** e0ne_ is now known as e0ne | 06:35 | |
| *** e0ne has quit IRC | 06:36 | |
| *** Abhi_ has joined #openstack-infra | 06:42 | |
| Abhi_ | while running CI for any patch, does the change is merged automatically?? | 06:44 |
| Abhi_ | like we do the cherrypick?? | 06:44 |
| *** ildikov has joined #openstack-infra | 06:47 | |
| *** belmoreira has joined #openstack-infra | 06:48 | |
| *** tnovacik has joined #openstack-infra | 06:48 | |
| *** scheuran has joined #openstack-infra | 06:51 | |
| *** hichihara has quit IRC | 06:54 | |
| *** salv-orlando has joined #openstack-infra | 06:54 | |
| *** Ala has joined #openstack-infra | 06:57 | |
| *** salv-orl_ has joined #openstack-infra | 06:58 | |
| *** salv-orlando has quit IRC | 06:58 | |
| *** doude has joined #openstack-infra | 06:59 | |
| *** afazekas has joined #openstack-infra | 07:02 | |
| *** pal__ has joined #openstack-infra | 07:02 | |
| *** tkelsey has joined #openstack-infra | 07:04 | |
| *** pal_ has quit IRC | 07:05 | |
| *** sergsh has joined #openstack-infra | 07:07 | |
| *** achanda has quit IRC | 07:08 | |
| *** tkelsey has quit IRC | 07:08 | |
| *** mrda is now known as mrda-away | 07:12 | |
| *** ominakov has joined #openstack-infra | 07:13 | |
| *** doude_ has joined #openstack-infra | 07:18 | |
| *** stevemar has quit IRC | 07:19 | |
| *** ghostpl_ has joined #openstack-infra | 07:21 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/storyboard-webclient: Initial proposal for integration tests https://review.openstack.org/156217 | 07:22 |
| *** ihrachyshka has joined #openstack-infra | 07:22 | |
| *** doude has quit IRC | 07:22 | |
| *** markvoelker has joined #openstack-infra | 07:23 | |
| *** ominakov has quit IRC | 07:24 | |
| *** ibiris_away is now known as ibiris | 07:26 | |
| *** markvoelker has quit IRC | 07:27 | |
| *** jlibosva has quit IRC | 07:27 | |
| *** ghostpl_ has quit IRC | 07:28 | |
| deepakcs | Abhi_, u mean whether the change is picked up automatically, then yes it is (I don't know whehter it does cherry pick or some other stuff tho') | 07:29 |
| deepakcs | Abhi_, but using 'merge' is the wrong way of putting it, since typically merge is done to master | 07:29 |
| *** ildikov has quit IRC | 07:30 | |
| *** ddieterly has joined #openstack-infra | 07:30 | |
| deepakcs | Abhi_, from a CI perspective, all you need to know is that the devstack job is setup with your patch included! | 07:30 |
| Abhi_ | deepakcs, ok so it must be doing cherry -pick for a patch, and we are testing that cherry-pick through our CI, right?? | 07:30 |
| *** unicell has quit IRC | 07:30 | |
| *** Alexandra_ has quit IRC | 07:30 | |
| *** unicell has joined #openstack-infra | 07:31 | |
| deepakcs | Abhi_, we are not testing cherry-pick, we are testing whether the change the patch introduced, is working fine with the vendor's storage backend using the CI infra | 07:31 |
| *** shardy has joined #openstack-infra | 07:32 | |
| Abhi_ | ok thanks for the info | 07:33 |
| *** ddieterly has quit IRC | 07:35 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/storyboard-webclient: Initial proposal for integration tests https://review.openstack.org/156217 | 07:37 |
| *** nmagnezi has joined #openstack-infra | 07:38 | |
| *** amotoki has joined #openstack-infra | 07:38 | |
| *** markus_z has joined #openstack-infra | 07:39 | |
| *** otter768 has joined #openstack-infra | 07:39 | |
| *** woodster_ has quit IRC | 07:40 | |
| *** arxcruz has joined #openstack-infra | 07:41 | |
| *** salv-orl_ has quit IRC | 07:42 | |
| openstackgerrit | Jordan Pittier proposed openstack-infra/system-config: Third party CI: document how to properly format message with Gerrit Trigger https://review.openstack.org/167716 | 07:42 |
| *** Somay has joined #openstack-infra | 07:42 | |
| *** jlanoux has joined #openstack-infra | 07:43 | |
| *** liusheng has quit IRC | 07:44 | |
| *** otter768 has quit IRC | 07:44 | |
| *** dulek has joined #openstack-infra | 07:44 | |
| *** ildikov has joined #openstack-infra | 07:45 | |
| *** jyuso1 has joined #openstack-infra | 07:46 | |
| *** cody-somerville has quit IRC | 07:47 | |
| *** hashar has joined #openstack-infra | 07:47 | |
| *** btully has quit IRC | 07:50 | |
| *** ominakov has joined #openstack-infra | 07:51 | |
| *** notnownikki has joined #openstack-infra | 07:51 | |
| *** ominakov has quit IRC | 07:52 | |
| *** mpaolino has joined #openstack-infra | 07:52 | |
| *** davideagnello has joined #openstack-infra | 07:53 | |
| *** dims has joined #openstack-infra | 07:53 | |
| *** tkelsey has joined #openstack-infra | 07:54 | |
| *** Somay has quit IRC | 07:54 | |
| *** pal_ has joined #openstack-infra | 07:55 | |
| *** yamamoto has quit IRC | 07:56 | |
| *** davideagnello has quit IRC | 07:57 | |
| *** pal__ has quit IRC | 07:57 | |
| *** dims has quit IRC | 07:58 | |
| *** tnovacik has quit IRC | 07:59 | |
| *** ssam2 has joined #openstack-infra | 08:00 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/storyboard-webclient: Initial proposal for integration tests https://review.openstack.org/156217 | 08:02 |
| *** cody-somerville has joined #openstack-infra | 08:03 | |
| *** HeOS has quit IRC | 08:04 | |
| *** vigneshvar has joined #openstack-infra | 08:05 | |
| *** vigneshvar_ has joined #openstack-infra | 08:05 | |
| *** _nadya_ has joined #openstack-infra | 08:05 | |
| *** tnovacik has joined #openstack-infra | 08:05 | |
| *** _nadya_ has quit IRC | 08:08 | |
| *** jistr has joined #openstack-infra | 08:09 | |
| *** jlibosva has joined #openstack-infra | 08:09 | |
| *** ominakov has joined #openstack-infra | 08:14 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/storyboard-webclient: Initial proposal for integration tests https://review.openstack.org/156217 | 08:17 |
| *** yamamoto has joined #openstack-infra | 08:19 | |
| *** ildikov has quit IRC | 08:23 | |
| *** pal__ has joined #openstack-infra | 08:25 | |
| *** yamamoto has quit IRC | 08:25 | |
| *** yfried is now known as yfried|afk | 08:25 | |
| *** pal_ has quit IRC | 08:28 | |
| *** ddieterly has joined #openstack-infra | 08:31 | |
| *** yfried|afk is now known as yfried | 08:31 | |
| *** MaxV has joined #openstack-infra | 08:33 | |
| *** kashyap has left #openstack-infra | 08:33 | |
| *** Abhi_ has quit IRC | 08:33 | |
| *** Somay has joined #openstack-infra | 08:34 | |
| *** teran has quit IRC | 08:34 | |
| *** ddieterly has quit IRC | 08:36 | |
| *** sergsh has quit IRC | 08:37 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/storyboard: Add tools to run integration tests https://review.openstack.org/155975 | 08:38 |
| *** ildikov has joined #openstack-infra | 08:38 | |
| *** ghostpl_ has joined #openstack-infra | 08:38 | |
| *** ihrachyshka has quit IRC | 08:39 | |
| *** Ala has quit IRC | 08:39 | |
| *** dtantsur|afk is now known as dtantsur | 08:40 | |
| *** salv-orlando has joined #openstack-infra | 08:43 | |
| *** e0ne has joined #openstack-infra | 08:43 | |
| *** btully has joined #openstack-infra | 08:46 | |
| *** ghostpl_ has quit IRC | 08:47 | |
| *** salv-orlando has quit IRC | 08:47 | |
| *** dhritishikhar_ has joined #openstack-infra | 08:48 | |
| *** sergsh has joined #openstack-infra | 08:49 | |
| *** btully has quit IRC | 08:50 | |
| *** dhritishikhar__ has joined #openstack-infra | 08:51 | |
| *** e0ne is now known as e0ne_ | 08:51 | |
| *** dhritishikhar_ has quit IRC | 08:51 | |
| *** Longgeek_ has quit IRC | 08:53 | |
| *** e0ne_ is now known as e0ne | 08:54 | |
| *** vigneshvar has quit IRC | 08:56 | |
| *** ZZelle has joined #openstack-infra | 08:56 | |
| *** vigneshvar_ has quit IRC | 08:56 | |
| *** jlibosva has quit IRC | 08:58 | |
| *** jlibosva has joined #openstack-infra | 09:02 | |
| *** jlibosva has quit IRC | 09:06 | |
| *** ominakov has quit IRC | 09:07 | |
| *** zz_johnthetubagu is now known as johnthetubaguy | 09:07 | |
| *** jlibosva has joined #openstack-infra | 09:09 | |
| *** jlibosva has quit IRC | 09:10 | |
| *** jlibosva has joined #openstack-infra | 09:11 | |
| *** ominakov has joined #openstack-infra | 09:12 | |
| *** HeOS has joined #openstack-infra | 09:17 | |
| *** ociuhandu has joined #openstack-infra | 09:17 | |
| *** dhritishikhar__ has quit IRC | 09:19 | |
| *** pal_ has joined #openstack-infra | 09:21 | |
| *** ihrachyshka has joined #openstack-infra | 09:21 | |
| *** pal__ has quit IRC | 09:23 | |
| *** markvoelker has joined #openstack-infra | 09:24 | |
| *** ajo has joined #openstack-infra | 09:25 | |
| *** mpaolino has quit IRC | 09:25 | |
| *** smccully has joined #openstack-infra | 09:26 | |
| *** ghostpl_ has joined #openstack-infra | 09:27 | |
| *** markvoelker has quit IRC | 09:29 | |
| *** ajo_ has joined #openstack-infra | 09:30 | |
| *** ddieterly has joined #openstack-infra | 09:32 | |
| *** jlibosva has quit IRC | 09:33 | |
| *** ajo has quit IRC | 09:33 | |
| *** flip214 has joined #openstack-infra | 09:36 | |
| *** ddieterly has quit IRC | 09:36 | |
| *** derekh has joined #openstack-infra | 09:36 | |
| flip214 | clarkb: ping | 09:37 |
| *** Somay has quit IRC | 09:37 | |
| *** dtantsur is now known as dtantsur|brb | 09:37 | |
| *** dizquierdo has joined #openstack-infra | 09:38 | |
| *** ajo_ is now known as ajo | 09:38 | |
| *** ajo has quit IRC | 09:40 | |
| *** ajo has joined #openstack-infra | 09:40 | |
| *** otter768 has joined #openstack-infra | 09:40 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Branches, milestones and projects ids validation in tasks https://review.openstack.org/162126 | 09:41 |
| *** otter768 has quit IRC | 09:45 | |
| *** skolekonov_ has joined #openstack-infra | 09:46 | |
| *** rlandy has joined #openstack-infra | 09:49 | |
| *** teran has joined #openstack-infra | 09:50 | |
| *** pal__ has joined #openstack-infra | 09:53 | |
| *** dims has joined #openstack-infra | 09:54 | |
| *** teran_ has joined #openstack-infra | 09:54 | |
| *** pal_ has quit IRC | 09:56 | |
| *** teran has quit IRC | 09:57 | |
| *** vigneshvar has joined #openstack-infra | 09:58 | |
| *** Longgeek has joined #openstack-infra | 09:58 | |
| *** vigneshvar_ has joined #openstack-infra | 09:59 | |
| *** ominakov has quit IRC | 10:01 | |
| openstackgerrit | Merged openstack-infra/storyboard: Refactor Tasks to work as a Subcontroller https://review.openstack.org/150052 | 10:02 |
| *** ominakov has joined #openstack-infra | 10:02 | |
| *** ominakov has quit IRC | 10:03 | |
| openstackgerrit | Merged openstack-infra/storyboard: Added events for tags https://review.openstack.org/165074 | 10:03 |
| *** ssam2 has quit IRC | 10:06 | |
| openstackgerrit | Merged openstack-infra/storyboard: Removed Nickname https://review.openstack.org/165194 | 10:07 |
| kaisers1 | Hi! Is anybody around for a devstack config question? | 10:10 |
| *** yamamoto has joined #openstack-infra | 10:12 | |
| *** flip214 has quit IRC | 10:17 | |
| *** flip214 has joined #openstack-infra | 10:17 | |
| openstackgerrit | Alex proposed openstack-infra/subunit2sql: Extend database schema of run_metadata https://review.openstack.org/168840 | 10:17 |
| *** Longgeek has quit IRC | 10:17 | |
| *** ssam2 has joined #openstack-infra | 10:20 | |
| *** _nadya_ has joined #openstack-infra | 10:20 | |
| *** samueldmq-away is now known as samueldmq | 10:22 | |
| *** pal_ has joined #openstack-infra | 10:23 | |
| *** salv-orlando has joined #openstack-infra | 10:23 | |
| *** markvoelker has joined #openstack-infra | 10:25 | |
| *** jlibosva has joined #openstack-infra | 10:26 | |
| *** pal__ has quit IRC | 10:26 | |
| *** markvoelker has quit IRC | 10:30 | |
| *** pblaho has joined #openstack-infra | 10:32 | |
| *** yfried is now known as yfried|afk | 10:32 | |
| *** ddieterly has joined #openstack-infra | 10:32 | |
| *** tnovacik has quit IRC | 10:33 | |
| *** hashar has quit IRC | 10:33 | |
| *** hashar has joined #openstack-infra | 10:34 | |
| *** yfried|afk is now known as yfried | 10:34 | |
| *** yamamoto has quit IRC | 10:35 | |
| *** pc_m has joined #openstack-infra | 10:36 | |
| *** ddieterly has quit IRC | 10:37 | |
| *** ominakov has joined #openstack-infra | 10:40 | |
| *** johnthetubaguy is now known as zz_johnthetubagu | 10:41 | |
| *** prmtl has quit IRC | 10:41 | |
| *** prmtl has joined #openstack-infra | 10:42 | |
| *** yamamoto has joined #openstack-infra | 10:42 | |
| *** mpavone has joined #openstack-infra | 10:48 | |
| *** dtantsur|brb is now known as dtantsur | 10:50 | |
| *** pal_ has quit IRC | 10:52 | |
| *** mwagner_lap has quit IRC | 10:52 | |
| *** yamamoto has quit IRC | 10:52 | |
| *** pal_ has joined #openstack-infra | 10:52 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Branches, milestones and projects ids validation in tasks https://review.openstack.org/162126 | 10:55 |
| *** tnovacik has joined #openstack-infra | 10:57 | |
| *** pal_ has quit IRC | 11:01 | |
| *** ominakov has quit IRC | 11:02 | |
| *** ominakov has joined #openstack-infra | 11:03 | |
| *** vigneshvar_ has quit IRC | 11:04 | |
| *** vigneshvar has quit IRC | 11:04 | |
| *** vigneshvar has joined #openstack-infra | 11:05 | |
| *** vigneshvar_ has joined #openstack-infra | 11:05 | |
| *** ominakov has quit IRC | 11:06 | |
| *** marcusvrn has joined #openstack-infra | 11:08 | |
| *** marcusvrn1 has joined #openstack-infra | 11:08 | |
| *** ominakov has joined #openstack-infra | 11:11 | |
| *** sergsh has quit IRC | 11:11 | |
| *** marcusvrn has quit IRC | 11:13 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Added story types https://review.openstack.org/163412 | 11:15 |
| *** Longgeek has joined #openstack-infra | 11:15 | |
| *** ominakov has quit IRC | 11:15 | |
| *** jaypipes has joined #openstack-infra | 11:17 | |
| *** prmtl has quit IRC | 11:18 | |
| flip214 | hi, short questions about http://lists.openstack.org/pipermail/openstack-dev/2015-February/057585.html. | 11:21 |
| flip214 | in this ML post clarkb had two gerrit links, which I'm not sure whether they're the correct ones. | 11:21 |
| flip214 | does gerrit reuse its review numbers? | 11:22 |
| *** AJaeger has joined #openstack-infra | 11:22 | |
| *** AJaeger has joined #openstack-infra | 11:22 | |
| *** dhritishikhar_ has joined #openstack-infra | 11:23 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Branches, milestones and projects ids validation in tasks https://review.openstack.org/162126 | 11:25 |
| *** lucasagomes has joined #openstack-infra | 11:25 | |
| *** sergsh has joined #openstack-infra | 11:26 | |
| *** markvoelker has joined #openstack-infra | 11:26 | |
| lucasagomes | hi, if you have a time please take a look at https://review.openstack.org/#/c/163344/ , it add a non-voting check on the openstack/ironic-python-agent project to test IPA with the PXE ramdisks. It has been approved once but it had to be rebased. (Need one more +2/+A now). Tanks | 11:26 |
| lucasagomes | thanks* | 11:26 |
| *** markus_z has quit IRC | 11:27 | |
| bookwar | hi, we have hanging merge here https://review.openstack.org/#/c/167976/ and i don't see it in zuul, how can this be resolved? | 11:28 |
| *** lucasagomes is now known as lucas-hungry | 11:28 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Branches, milestones and projects ids validation in tasks https://review.openstack.org/162126 | 11:29 |
| *** markvoelker has quit IRC | 11:31 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Added story types https://review.openstack.org/163412 | 11:32 |
| *** BharatK has quit IRC | 11:33 | |
| *** miqui has quit IRC | 11:33 | |
| *** ddieterly has joined #openstack-infra | 11:33 | |
| *** btully has joined #openstack-infra | 11:35 | |
| *** doug-fish has joined #openstack-infra | 11:36 | |
| hughsaunders | How can I filter by subject/commit message in gerrit? I want to exclude WIPs (some commits are marked WIP in the subject, not in the workflow field) | 11:37 |
| *** Murad has joined #openstack-infra | 11:37 | |
| *** ddieterly has quit IRC | 11:38 | |
| Murad | can anybody help in creating html report for tempest? | 11:39 |
| *** btully has quit IRC | 11:40 | |
| *** Longgeek has quit IRC | 11:40 | |
| *** otter768 has joined #openstack-infra | 11:41 | |
| *** ldnunes has joined #openstack-infra | 11:42 | |
| *** otter768 has quit IRC | 11:46 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Added request statistic https://review.openstack.org/164696 | 11:46 |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard-webclient: Support for tags https://review.openstack.org/166847 | 11:48 |
| *** deepakcs has quit IRC | 11:49 | |
| *** vigneshvar has quit IRC | 11:49 | |
| *** vigneshvar_ has quit IRC | 11:50 | |
| *** yamamoto has joined #openstack-infra | 11:50 | |
| *** kgiusti has joined #openstack-infra | 11:51 | |
| *** dhritishikhar_ has quit IRC | 11:58 | |
| *** woodster_ has joined #openstack-infra | 12:00 | |
| *** dulek_ has joined #openstack-infra | 12:03 | |
| *** ominakov has joined #openstack-infra | 12:03 | |
| *** dulek has quit IRC | 12:06 | |
| *** jistr has quit IRC | 12:07 | |
| *** ominakov has quit IRC | 12:07 | |
| *** jistr has joined #openstack-infra | 12:08 | |
| *** cdent has joined #openstack-infra | 12:08 | |
| *** markvoelker has joined #openstack-infra | 12:10 | |
| *** dulek___ has joined #openstack-infra | 12:11 | |
| *** dulek_ has quit IRC | 12:14 | |
| *** prmtl has joined #openstack-infra | 12:14 | |
| *** prmtl has joined #openstack-infra | 12:14 | |
| *** ildikov has quit IRC | 12:17 | |
| *** dims has quit IRC | 12:17 | |
| *** dims has joined #openstack-infra | 12:18 | |
| *** dkranz has joined #openstack-infra | 12:18 | |
| *** krtaylor has quit IRC | 12:19 | |
| *** yfried is now known as yfried|afk | 12:20 | |
| *** Ala has joined #openstack-infra | 12:21 | |
| *** bswartz has quit IRC | 12:22 | |
| *** dulek___ is now known as dulek | 12:23 | |
| *** aysyd has joined #openstack-infra | 12:23 | |
| *** salv-orlando has quit IRC | 12:27 | |
| *** dkliban_afk is now known as dkliban | 12:27 | |
| *** gordc has joined #openstack-infra | 12:28 | |
| *** ayoung_ZZzz__ is now known as ayoung_short | 12:28 | |
| *** ayoung_short is now known as ayoung_cough | 12:28 | |
| *** ayoung_cough is now known as ayoung_yawn | 12:29 | |
| *** ayoung_yawn is now known as ayoung | 12:29 | |
| *** salv-orlando has joined #openstack-infra | 12:30 | |
| *** sputnik13 has quit IRC | 12:30 | |
| *** weshay has joined #openstack-infra | 12:30 | |
| *** smccully has quit IRC | 12:31 | |
| *** yfried|afk is now known as yfried | 12:31 | |
| *** pal_ has joined #openstack-infra | 12:32 | |
| *** rfolco has joined #openstack-infra | 12:32 | |
| *** lucas-hungry is now known as lucasagomes | 12:32 | |
| openstackgerrit | Merged openstack-infra/project-config: Ironic: Create and enable pxe_ipa-src-nv on ironic-python-agent https://review.openstack.org/163344 | 12:34 |
| *** ddieterly has joined #openstack-infra | 12:34 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config: Move gantt,python-gantclient to attic https://review.openstack.org/168883 | 12:35 |
| *** Longgeek has joined #openstack-infra | 12:36 | |
| *** sandywalsh has quit IRC | 12:37 | |
| *** tiswanso has joined #openstack-infra | 12:38 | |
| *** ddieterly has quit IRC | 12:39 | |
| *** sandywalsh has joined #openstack-infra | 12:39 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Added request statistic https://review.openstack.org/164696 | 12:40 |
| *** baoli has joined #openstack-infra | 12:42 | |
| *** ildikov has joined #openstack-infra | 12:43 | |
| *** ddieterly has joined #openstack-infra | 12:44 | |
| *** e0ne is now known as e0ne_ | 12:44 | |
| *** e0ne_ is now known as e0ne | 12:51 | |
| *** lascii is now known as alaski | 12:52 | |
| *** timcline has joined #openstack-infra | 12:52 | |
| *** timcline has quit IRC | 12:52 | |
| *** dkranz has quit IRC | 12:53 | |
| *** mwagner_lap has joined #openstack-infra | 12:53 | |
| *** timcline has joined #openstack-infra | 12:53 | |
| *** tnovacik has quit IRC | 12:54 | |
| *** dboik has quit IRC | 12:56 | |
| *** ddieterly has quit IRC | 12:56 | |
| *** shakamunyi has joined #openstack-infra | 12:57 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config: Move gantt,python-gantclient to attic https://review.openstack.org/168883 | 12:58 |
| *** timcline has quit IRC | 12:59 | |
| *** julim has joined #openstack-infra | 12:59 | |
| *** krtaylor has joined #openstack-infra | 13:00 | |
| *** julim has quit IRC | 13:01 | |
| anteaya | flip214: gerrit doesn't reuse its review numbers | 13:01 |
| *** miqui has joined #openstack-infra | 13:02 | |
| flip214 | anteaya: thank you for your answer! | 13:03 |
| flip214 | the link in the ML post https://review.openstack.org/#/c/141530/ seems to have different contents, though... | 13:03 |
| *** e0ne is now known as e0ne_ | 13:03 | |
| flip214 | I remember that the gerrit data matched the textual description "current state of multinode work" some weeks ago, | 13:04 |
| flip214 | but now it's title is "Set custom cpu_model for live_migrate" | 13:04 |
| anteaya | bookwar: I don't see any Jenkins reports on that patch so I don't know what it is doing | 13:04 |
| *** dkranz has joined #openstack-infra | 13:04 | |
| *** bswartz has joined #openstack-infra | 13:05 | |
| anteaya | hughsaunders: if you use label:Workflow=0 that removes all wip patches | 13:05 |
| *** bknudson has quit IRC | 13:05 | |
| anteaya | and also any approved patches | 13:05 |
| hughsaunders | anteaya: problem is that some people use "WIP" in the subject rather than the workflow column | 13:06 |
| *** hrubi_ has quit IRC | 13:06 | |
| *** lucasagomes has left #openstack-infra | 13:06 | |
| anteaya | you can additionally use label:Verified=+1,jenkins to filter on patches that don't have +1 returned from Jenkins, so either patches currently running check tests or patches that have failed check | 13:06 |
| bookwar | anteaya: the jenkins is configured to do noop check and the merge itself, here is for example it works: https://review.openstack.org/#/c/167662/ but the latest patch didn't trigger Jenkins task | 13:06 |
| hughsaunders | anteaya: is it possible to filter on commit message? | 13:07 |
| anteaya | hughsaunders: I have no ability to filter out people who don't use gerrit at it is intended to be used | 13:07 |
| anteaya | flip214: I confirm your findings | 13:07 |
| *** Hal has joined #openstack-infra | 13:07 | |
| anteaya | bookwar: yes I confirm what you are seeing | 13:08 |
| *** Hal has quit IRC | 13:08 | |
| hughsaunders | anteaya: gerrit does have a message: query, but it doesn't seem to match the commit message, is there another way to do that? | 13:08 |
| *** annegentle has joined #openstack-infra | 13:08 | |
| bookwar | i'll ask devs to reset Workflow and CR mark one's again, maybe it'll help | 13:09 |
| *** Longgeek has quit IRC | 13:10 | |
| *** Hal has joined #openstack-infra | 13:11 | |
| anteaya | hughsaunders: I personally haven't had good luck with the message query | 13:11 |
| *** Hal is now known as Guest30324 | 13:11 | |
| boris-42 | anteaya: hi | 13:11 |
| anteaya | hughsaunders: I have found that for me it returns all sorts of results with have nothing to do with with my query | 13:12 |
| hughsaunders | anteaya: I have also experienced that | 13:12 |
| *** julim has joined #openstack-infra | 13:12 | |
| anteaya | hughsaunders: I also have been unsuccessful in getting it to take more than one word for the query | 13:12 |
| anteaya | hughsaunders: which is why I say that I have no ability to filter for people who don't use gerrit as it was intended | 13:12 |
| *** shakamunyi has quit IRC | 13:13 | |
| anteaya | bookwar: it might | 13:13 |
| hughsaunders | anteaya: I wonder if thats a gerrit upstream issue or our implementation of gerrit? | 13:13 |
| anteaya | hughsaunders: well our implementation of gerrit draws from upstream | 13:13 |
| anteaya | we haven't done any customization of the gerrit query functionality | 13:14 |
| anteaya | to the best of my knowledge | 13:14 |
| anteaya | which is why I stick to the labels | 13:14 |
| *** e0ne_ has quit IRC | 13:14 | |
| hughsaunders | is it possible to get workflow -1 to stick across patchsets? | 13:15 |
| *** esker has quit IRC | 13:15 | |
| anteaya | boris-42: how about you just post to the channel, my average of questions I can answer compared to questions asked today is not very high | 13:15 |
| *** ghostpl_ has quit IRC | 13:15 | |
| anteaya | boris-42: I have no confidence that I can help you boris-42 | 13:15 |
| *** dboik has joined #openstack-infra | 13:16 | |
| anteaya | hughsaunders: well, the only vote that sticks across patchsets is a -2 from a reviewer | 13:16 |
| *** e0ne has joined #openstack-infra | 13:16 | |
| anteaya | hughsaunders: since having workflow -1 stick across patchsets interferes with the functionality of workflow -1 | 13:16 |
| openstackgerrit | Mehdi Abaakouk proposed openstack-infra/project-config: Update gnocchi repository url https://review.openstack.org/168895 | 13:17 |
| anteaya | hughsaunders: you seem to be dealing with something rather specific but asking general questions to get there | 13:17 |
| anteaya | hughsaunders: are you going to tell me what you are actually dealing with? | 13:17 |
| hughsaunders | anteaya: XY | 13:17 |
| anteaya | what does that mean? | 13:17 |
| *** dprince has joined #openstack-infra | 13:18 | |
| hughsaunders | anteaya: because workflow -1 doesn't stick across patchsets, reviews get labelled with WIP in the subject, rather than using the workflow field. Which means I can't filter them :( | 13:18 |
| anteaya | right | 13:18 |
| hughsaunders | xy problem: http://xyproblem.info/ | 13:18 |
| anteaya | because that is developer behaviour | 13:18 |
| sileht | Hi, does anyone can review https://review.openstack.org/#/c/168895/, it blocks the gate for gnocchi since we have moved the repository to the openstack namespace. | 13:19 |
| anteaya | hughsaunders: I suggest you bring it up at the weekly project meeting of whatever project you are working on | 13:19 |
| anteaya | hughsaunders: and present your perspective | 13:20 |
| anteaya | personally I just reset workflow -1 on a wip patch | 13:20 |
| anteaya | it isn't that hard | 13:20 |
| *** peristeri has joined #openstack-infra | 13:20 | |
| anteaya | and yes, the link describes it beautifully, thank you | 13:21 |
| *** zz_johnthetubagu is now known as johnthetubaguy | 13:21 | |
| hughsaunders | anteaya: I'll check if git-review has a flag for w-1, maybe a useful feature if not there already | 13:21 |
| anteaya | sileht: hi, hopefully you can get another +2 before too long | 13:22 |
| anteaya | hughsaunders: no argument from me | 13:22 |
| sileht | anteaya, thanks | 13:22 |
| anteaya | sileht: welcome, sorry I didn't catch that in review | 13:22 |
| openstackgerrit | Mehdi Abaakouk proposed openstack-infra/project-config: Update gnocchi repository url https://review.openstack.org/168895 | 13:23 |
| flip214 | anteaya: well, so what can I do about that? Is #infra interested in taking a look what happened there? | 13:23 |
| sileht | anteaya, I have updated the review (I have missed 1 occurence of stackforge) | 13:24 |
| *** dustins has joined #openstack-infra | 13:24 | |
| anteaya | flip214: I can only speak for myself, you asked a few questions, I was able to answer one of them and answered the one I could | 13:25 |
| anteaya | flip214: I can't give you information I don't have | 13:25 |
| *** sdake has joined #openstack-infra | 13:25 | |
| anteaya | sileht: I will review again | 13:25 |
| sileht | anteaya, thanks | 13:25 |
| flip214 | anteaya: well, who can I ask about gerrit issues like these? | 13:27 |
| *** sdake_ has joined #openstack-infra | 13:28 | |
| anteaya | sileht: okay I grepped this time and found a total of two entries that need to be changed, looks like you found them both | 13:28 |
| *** tnovacik has joined #openstack-infra | 13:28 | |
| *** hdd has joined #openstack-infra | 13:29 | |
| *** ssam2 has quit IRC | 13:30 | |
| *** sdake has quit IRC | 13:31 | |
| AJaeger | anteaya, Good morning! Could you review https://review.openstack.org/#/c/168022/ , please? | 13:31 |
| flip214 | anteaya: thanks a lot, btw! | 13:32 |
| openstackgerrit | Cyril Roelandt proposed openstack-infra/project-config: saharaclient: enable Python3 jobs https://review.openstack.org/168899 | 13:33 |
| *** johnthetubaguy is now known as zz_johnthetubagu | 13:34 | |
| anteaya | AJaeger: done | 13:35 |
| anteaya | flip214: you're welcome | 13:35 |
| AJaeger | thanks, anteaya | 13:35 |
| anteaya | AJaeger: thanks for reviewing sileht's patch | 13:36 |
| AJaeger | anteaya, sure | 13:36 |
| *** chlong has joined #openstack-infra | 13:36 | |
| sileht | AJaeger, thx | 13:36 |
| * anteaya goes to find some breakfast | 13:36 | |
| AJaeger | anteaya, enjoy! | 13:37 |
| *** ddieterly has joined #openstack-infra | 13:37 | |
| *** hdd has quit IRC | 13:38 | |
| *** deepakcs has joined #openstack-infra | 13:39 | |
| *** rfolco has quit IRC | 13:40 | |
| *** jlanoux has quit IRC | 13:40 | |
| *** eharney has joined #openstack-infra | 13:41 | |
| *** otter768 has joined #openstack-infra | 13:42 | |
| *** rfolco has joined #openstack-infra | 13:42 | |
| *** mriedem has joined #openstack-infra | 13:42 | |
| *** markus_z has joined #openstack-infra | 13:43 | |
| *** e0ne is now known as e0ne_ | 13:43 | |
| *** ssam2 has joined #openstack-infra | 13:44 | |
| fungi | hughsaunders: workflow -1 doesn't persist on new patchsets because not persisting on new patchsets was a desired behavior of workflow -1 | 13:44 |
| *** jgrimm is now known as zz_jgrimm | 13:45 | |
| *** bknudson has joined #openstack-infra | 13:45 | |
| fungi | hughsaunders: it's currently implemented as an individual vote, not a change state, so the only ways to clear it are either to reverse your own workflow -1 vote or to upload a new patchset to a change. if workflow -1 persisted on new patchsets then the only way to remove a workflow -1 would be to undo the vote yourself. not great if you're collaborating on a change with the rest of your team | 13:46 |
| openstackgerrit | Merged openstack-infra/project-config: Enable PyPi push for the stackforge/faafo project https://review.openstack.org/168022 | 13:46 |
| *** e0ne_ is now known as e0ne | 13:46 | |
| hughsaunders | fungi: thanks, I hadn't twigged that people could disagree on the workflow state. | 13:46 |
| openstackgerrit | Merged openstack-infra/project-config: Update gnocchi repository url https://review.openstack.org/168895 | 13:47 |
| *** rfolco has quit IRC | 13:47 | |
| *** otter768 has quit IRC | 13:47 | |
| *** rfolco has joined #openstack-infra | 13:48 | |
| *** claudiupopa has joined #openstack-infra | 13:49 | |
| *** sdake has joined #openstack-infra | 13:49 | |
| jogo | clarkb: oh you got 64 bit live migration working, nice | 13:50 |
| *** ghostpl_ has joined #openstack-infra | 13:51 | |
| *** stevemar has joined #openstack-infra | 13:52 | |
| *** sdake_ has quit IRC | 13:53 | |
| *** tonytan4ever has joined #openstack-infra | 13:53 | |
| *** xyang1 has joined #openstack-infra | 13:53 | |
| openstackgerrit | Merged openstack-infra/project-config: Add new project nerd-reviewer https://review.openstack.org/166945 | 13:53 |
| openstackgerrit | Merged openstack-infra/project-config: Add new project nerd-reviewer https://review.openstack.org/166945 | 13:53 |
| *** alexpilotti has joined #openstack-infra | 13:53 | |
| *** spzala has joined #openstack-infra | 13:54 | |
| *** superdan is now known as dansmith | 13:56 | |
| *** sabeen1 has joined #openstack-infra | 13:56 | |
| fungi | hughsaunders: when we upgrade to gerrit 2.9 i think we'll have the ability to start using the wip plugin which has been under development for our use case. it implements a custom change state rather than using a vote, and so can be switched on and off by different users | 13:57 |
| hughsaunders | fungi: sounds good | 13:58 |
| *** baoli has quit IRC | 13:59 | |
| *** baoli_ has joined #openstack-infra | 14:00 | |
| *** dtantsur is now known as dtantsur|brb | 14:01 | |
| *** pal_ has quit IRC | 14:01 | |
| *** gokrokve has joined #openstack-infra | 14:02 | |
| *** yamamoto has quit IRC | 14:03 | |
| *** yamamoto has joined #openstack-infra | 14:03 | |
| *** dboik_ has joined #openstack-infra | 14:03 | |
| fungi | flip214: you'd probably have to ask clarkb what he meant when linking https://review.openstack.org/141530 in http://lists.openstack.org/pipermail/openstack-dev/2015-February/057585.html | 14:04 |
| fungi | flip214: as far as i can tell the change has always had that title all the way back to december when it was first uploaded | 14:05 |
| *** ominakov has joined #openstack-infra | 14:05 | |
| *** jyuso has joined #openstack-infra | 14:05 | |
| *** jyuso1 has quit IRC | 14:05 | |
| fungi | flip214: it's possible clarkb meant to link a different change | 14:05 |
| *** dboik has quit IRC | 14:06 | |
| *** nelsnelson has joined #openstack-infra | 14:09 | |
| *** doude_ has quit IRC | 14:09 | |
| *** doude has joined #openstack-infra | 14:09 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:09 | |
| *** yamamoto has quit IRC | 14:10 | |
| *** tkelsey has quit IRC | 14:10 | |
| *** btully has joined #openstack-infra | 14:11 | |
| *** dkranz has quit IRC | 14:11 | |
| *** ominakov has quit IRC | 14:12 | |
| fbo | Hi, I saw there is lot of errors in the system-config integration test "apply-test.sh" logs run by the gate. So I'm wondering if is it still relevant ? | 14:12 |
| anteaya | fbo: have you a link to the errors you are seeing? | 14:13 |
| fbo | anteaya, yep http://logs.openstack.org/52/166252/4/check/gate-infra-puppet-apply-precise/be3a209/console.html | 14:14 |
| fbo | anteaya, I see this kind of error "Skipping because of failed dependencies" for all patches and for all tests on precise/trusty/centos6 platform | 14:15 |
| anteaya | okay looking at the tests for this patch, the apply tests failed one time and all three apply tests failed | 14:16 |
| fbo | But it seems the script reports a success to gerrit | 14:16 |
| anteaya | every other test run on that patch had passing apply tests | 14:16 |
| fungi | fbo: right now it's not safe/easy to do a full apply of all our manifests together on one machine, so we use --noop | 14:17 |
| fungi | fbo: unfortunately since there are many operations puppet won't perform in --noop mode it sees a lot of dependent actions as unsatisfied, hence the error messages | 14:17 |
| fbo | fungi, yes I agree not saw easy I guess | 14:17 |
| *** jcoufal has joined #openstack-infra | 14:18 | |
| *** tonytan_brb has joined #openstack-infra | 14:18 | |
| *** mtanino has joined #openstack-infra | 14:18 | |
| anteaya | fbo: so based on my initial findings so far I'm going with yes it is still relevant | 14:19 |
| fbo | fungi, and do you have some ideas to improve the integration testing for the system-config manifests ? | 14:20 |
| *** bhunter71 has joined #openstack-infra | 14:20 | |
| *** timcline has joined #openstack-infra | 14:20 | |
| kaisers1 | anteaya: Hi! is there a thirdparty mentoring today? | 14:21 |
| *** tonytan4ever has quit IRC | 14:22 | |
| fungi | fbo: probably best to talk to nibalizer since i think he was working on something to address that | 14:22 |
| *** mattfarina has joined #openstack-infra | 14:23 | |
| anteaya | kaisers1: there is a third-party meeting today | 14:24 |
| fbo | anteaya, ok because I sent some patches last week on system-config but I was a bit lost with all that log output and wasn't able to check if my patches pass the integration test or not :) | 14:24 |
| anteaya | kaisers1: I have no idea why you call it mentoring | 14:25 |
| anteaya | fbo: did jenkins report success or failure? | 14:25 |
| kaisers1 | anteaya: sorry, i seemed to remember it was called that. | 14:26 |
| *** jlanoux has joined #openstack-infra | 14:26 | |
| *** dkranz has joined #openstack-infra | 14:27 | |
| fbo | fungi, Ok I'll do that, I've some experience with this kind of complex testing using LXC containers and I'll be happy to help if I can :) | 14:27 |
| fbo | anteaya, yes jenkins reports a success but it should not as manifest apply fail | 14:28 |
| *** ghostpl_ has quit IRC | 14:28 | |
| *** ajo has quit IRC | 14:28 | |
| *** ominakov has joined #openstack-infra | 14:29 | |
| *** ghostpl_ has joined #openstack-infra | 14:29 | |
| *** garyh has joined #openstack-infra | 14:29 | |
| anteaya | kaisers1: I never called it that | 14:29 |
| fungi | fbo: puppet "failures" aren't black and white. it returns a variety of different exit codes for different conditions, so we catch specific ones we know indicate a problem | 14:29 |
| *** ChuckC has quit IRC | 14:29 | |
| fungi | dboik_: it's mostly there to catch things like parsing errors | 14:30 |
| fungi | er, fbo ^ | 14:30 |
| fungi | (sorrt dboik_) | 14:30 |
| *** jlanoux_ has joined #openstack-infra | 14:31 | |
| kaisers1 | anteaya: apologies, it was called like that in some cinder CI related mails, that's where my impression came from. Just looked it up. I won't address it like that no more! :) | 14:31 |
| *** wenlock has joined #openstack-infra | 14:32 | |
| *** achanda has joined #openstack-infra | 14:33 | |
| *** jlanoux has quit IRC | 14:34 | |
| anteaya | kaisers1: thanks | 14:34 |
| anteaya | I'll talk to cinder folks | 14:34 |
| *** ghostpl_ has quit IRC | 14:34 | |
| anteaya | and yes 1500 is meeting time | 14:35 |
| *** ghostpl_ has joined #openstack-infra | 14:35 | |
| *** prmtl_ has joined #openstack-infra | 14:36 | |
| *** tonytan_brb is now known as tonytan4ever | 14:36 | |
| *** jlibosva has quit IRC | 14:37 | |
| *** ghostpl_ has quit IRC | 14:37 | |
| *** asselin has quit IRC | 14:37 | |
| *** tkelsey has joined #openstack-infra | 14:37 | |
| *** asselin has joined #openstack-infra | 14:37 | |
| *** ildikov has quit IRC | 14:38 | |
| *** ilyashakhat has joined #openstack-infra | 14:38 | |
| *** ghostpl_ has joined #openstack-infra | 14:38 | |
| *** zz_dimtruck is now known as dimtruck | 14:38 | |
| *** prmtl has quit IRC | 14:40 | |
| claudiupopa | mordred: ping | 14:40 |
| *** asselin has quit IRC | 14:40 | |
| *** asselin has joined #openstack-infra | 14:41 | |
| *** ghostpl_ has quit IRC | 14:41 | |
| *** asselin has quit IRC | 14:41 | |
| *** tkelsey has quit IRC | 14:41 | |
| *** asselin_ has joined #openstack-infra | 14:41 | |
| *** asselin_ has quit IRC | 14:42 | |
| *** patrickeast has joined #openstack-infra | 14:42 | |
| *** asselin_ has joined #openstack-infra | 14:43 | |
| *** ghostpl_ has joined #openstack-infra | 14:43 | |
| *** asselin_ has quit IRC | 14:44 | |
| *** achanda has quit IRC | 14:44 | |
| *** asselin_ has joined #openstack-infra | 14:44 | |
| *** zz_jgrimm is now known as jgrimm | 14:44 | |
| *** asselin_ has quit IRC | 14:45 | |
| *** asselin_ has joined #openstack-infra | 14:45 | |
| *** ominakov has quit IRC | 14:46 | |
| *** asselin_ has quit IRC | 14:46 | |
| *** asselin has joined #openstack-infra | 14:47 | |
| *** asselin has quit IRC | 14:47 | |
| *** zul has quit IRC | 14:47 | |
| *** asselin has joined #openstack-infra | 14:48 | |
| *** nmagnezi has quit IRC | 14:48 | |
| *** yfried has quit IRC | 14:48 | |
| *** asselin has quit IRC | 14:49 | |
| *** asselin_ has joined #openstack-infra | 14:49 | |
| *** armax has joined #openstack-infra | 14:49 | |
| *** timcline has quit IRC | 14:49 | |
| mordred | claudiupopa: hi | 14:50 |
| *** asselin_ has quit IRC | 14:50 | |
| *** asselin_ has joined #openstack-infra | 14:50 | |
| *** asselin_ has quit IRC | 14:51 | |
| *** thingee has quit IRC | 14:51 | |
| *** asselin_ has joined #openstack-infra | 14:51 | |
| *** skolekonov_ has quit IRC | 14:51 | |
| *** dtantsur|brb is now known as dtantsur | 14:52 | |
| clarkb | fungi flip214 thats where multinode work was. setting cpu models to enable nova live migration | 14:52 |
| *** asselin_ has quit IRC | 14:52 | |
| *** asselin has joined #openstack-infra | 14:52 | |
| *** asselin_ has joined #openstack-infra | 14:52 | |
| fungi | clarkb: which seems more like something you're testing on multinode (live migration) rather than something necessary to make multinode work (which seemed more like the point of the ml reply) so that's probably the confusion | 14:53 |
| *** zul has joined #openstack-infra | 14:53 | |
| clarkb | maybe, it is about workong multinode because libvirt | 14:53 |
| *** MarkAtwood has joined #openstack-infra | 14:53 | |
| fungi | e.g. are cpu model settings relevant to testing drbd backend? | 14:54 |
| claudiupopa | mordred: any idea why this change isn't in pbr 0.10? https://review.openstack.org/#/c/136707/ | 14:54 |
| clarkb | fungi I have no idea because I dont know what specialsauce they are doing | 14:54 |
| fungi | i guess multi-node devstack is broken without the cpu settings? | 14:54 |
| clarkb | fungi yes | 14:54 |
| claudiupopa | it's non on pypi either. | 14:54 |
| mordred | claudiupopa: that's right | 14:54 |
| mordred | claudiupopa: currently master is not releasable because there is a bug in the semver support there | 14:55 |
| mordred | so bugfixes have been going out to the 0.10 when necessary on a stable/0.10 branch which we're not particularly happy about | 14:55 |
| *** Longgeek has joined #openstack-infra | 14:55 | |
| mordred | however, I think we're pretty fed up with waiting for the semver stuff to be fixed | 14:55 |
| fungi | claudiupopa: master in pbr is essentially a development branch right now, and the releases have been coming from a separate "feature/0.10" release branch lately | 14:55 |
| *** thedodd has joined #openstack-infra | 14:56 | |
| claudiupopa | Oh I see. | 14:56 |
| fbo | Thanks fungi for the explanation I understand this cannot be really called integration tests for now. | 14:56 |
| yolanda | hi mordred, wanted to share with you that change :https://review.openstack.org/#/c/168686/ | 14:56 |
| yolanda | what do you think about it? | 14:56 |
| clarkb | fungi anyways I was pointing at general work for multinode. It had nothing to do directly with drbd other than indicating where to go for multinode stuff | 14:56 |
| *** tkelsey has joined #openstack-infra | 14:57 | |
| clarkb | I encouraged them to start witha single node test which has not happened | 14:57 |
| mordred | fungi: perhaps we should send a message to the the list that if someone doesn't step up to finish the semver patches, we're going to consider them abandoned, move them to a feature branch and re-promote 0.10 to master | 14:57 |
| *** timcline has joined #openstack-infra | 14:57 | |
| mordred | fungi: with a timeframe - maybe a week? | 14:57 |
| claudiupopa | mordred: Well, couldnt this change be cherrypicked? ;-) We were believing it's there already and we just had one bad release on cloudbaseinit because to it. | 14:57 |
| *** Guest96413 is now known as redrobot | 14:58 | |
| mordred | claudiupopa: you are more than welcome to propose a cherry-pick of the patch to feature/0.10! :) | 14:58 |
| jogo | fungi: question related to what to do with gantt | 14:58 |
| jogo | fungi: what is the stance on changing the history of a repo (equivalent of git push --force)? | 14:58 |
| jogo | fungi: because that is what will happen if we re-use the existing gantt repo | 14:59 |
| anteaya | jogo: why? | 14:59 |
| jeblair | jogo: we have two choices -- we could do what we did with keystone and replace it while keeping history (basically, new history into a branch and then merge that branch into master) | 15:00 |
| *** timcline_ has joined #openstack-infra | 15:00 | |
| jeblair | jogo: or we could rename gantt to something else and rename the new thing to gantt | 15:00 |
| jeblair | jogo: it kind of depends on the degree of relation between the two | 15:00 |
| *** reed has joined #openstack-infra | 15:01 | |
| *** rwsu has joined #openstack-infra | 15:01 | |
| jeblair | jogo: but we will never erase history with "push --force" * | 15:01 |
| jogo | jeblair: the contents of the existing gantt repo are worthless and confusing | 15:01 |
| jeblair | * (except for really minor things like an incorrect commit on a repo import or something technical like that) | 15:01 |
| mordred | jeblair: ++ | 15:01 |
| jeblair | jogo: that's not a value judgement any of us get to make :) | 15:01 |
| jogo | and there will be no relationship between the existing content and what the new scheduler split would contain | 15:02 |
| fungi | jogo: might be a good idea to merge a commit which deletes all the files and replaces them with an explanatory readme about the situation if you're concerned about it being confusing in its current state? | 15:02 |
| jeblair | jogo: that is a more compelling argument to do a rename swap :) | 15:02 |
| jogo | https://github.com/openstack/gantt#disclaimer | 15:02 |
| *** zul has quit IRC | 15:02 | |
| fungi | i see it as a very compelling argument to use a new name for the project | 15:02 |
| jogo | jeblair: rename swap? | 15:02 |
| *** zul has joined #openstack-infra | 15:03 | |
| mordred | jogo: 15:00:01 jeblair | jogo: we have two choices -- we could do what we did with keystone and replace it while keeping history (basically, new history into a | 15:03 |
| mordred | | branch and then merge that branch into master) | 15:03 |
| jogo | fungi: ++, but I will defer to jaypipes on what makes the most sense | 15:03 |
| *** timcline has quit IRC | 15:03 | |
| jogo | ahh | 15:03 |
| jeblair | fungi: ah, yes, the zeroth option. i did forget to mention that. it's a good one. :) | 15:04 |
| mordred | jogo: we're never going to delete git history, but we can make the repo look like what whatever people want | 15:04 |
| *** dboik_ has quit IRC | 15:04 | |
| jogo | mordred: ack | 15:04 |
| jeblair | jogo: are you going to pull code out of nova and put it into gantt? | 15:04 |
| jeblair | * scheduler code | 15:04 |
| jogo | jeblair: honestly don't actually know at this point | 15:04 |
| *** dboik has joined #openstack-infra | 15:04 | |
| jogo | jeblair: that was the idea, but not sure if that is still the idea. Once again I defer to jaypipes | 15:05 |
| *** SumitNaiksatam has quit IRC | 15:05 | |
| mordred | yolanda: that's an interesting idea ... it makes https://review.openstack.org/#/c/168665/ a bit different to reason about | 15:05 |
| *** SumitNaiksatam has joined #openstack-infra | 15:05 | |
| *** arxcruz has quit IRC | 15:05 | |
| mordred | yolanda: in terms of where a priority would be expressed, because I don't think any of those actions would have inherent priorities that would make sense to include in an API | 15:06 |
| anteaya | jogo: if I can suggest, why don't you and jaypipes add an item to the infra meeting and then we call all be there at the same time and make a decision? | 15:06 |
| yolanda | mm, shade with out namanger? | 15:06 |
| yolanda | manager? | 15:06 |
| yolanda | do we want to go at that level? | 15:06 |
| *** carl_baldwin_ has joined #openstack-infra | 15:06 | |
| mordred | yolanda: I believe that we have to | 15:06 |
| mordred | yolanda: because otherwise once we move to shade we will lose the ability to have rate limiting context | 15:06 |
| jogo | anteaya: well hopefully jaypipes will just chime on on the patch itself | 15:07 |
| anteaya | jogo: that works too | 15:07 |
| mordred | yolanda: since shade necessarily hides whether a given action is 0, 1, 2 or 10 REST API calls | 15:07 |
| *** jlanoux_ has quit IRC | 15:07 | |
| mordred | otherwise it's not very useful | 15:07 |
| yolanda | mordred, i see... but actually i'm testing and that priority concept is proving to be useful | 15:07 |
| jeblair | jogo: if i thought these were completely different ideas, i would go with fungi's option -- new name. if i thought that we were going to make a second attempt to do what we did the first time, i would keep it in repo and commit an rm -fr. if i thought the previous effort was a mistake and so do the people working on the previous effort, i'd rename it out of the way and use the name on a new repo. | 15:08 |
| jeblair | jogo: hth | 15:08 |
| yolanda | how could we port that to shade if needed? | 15:08 |
| jeblair | jogo: what patch? | 15:08 |
| mordred | yolanda: well, we'll need to think about that | 15:08 |
| *** zul has quit IRC | 15:08 | |
| jogo | jeblair: https://review.openstack.org/#/c/168883/ | 15:08 |
| *** zul has joined #openstack-infra | 15:08 | |
| *** jlanoux has joined #openstack-infra | 15:08 | |
| yolanda | going to stress hpcloud a bit more with that, i hope to don't break it :) | 15:09 |
| mordred | yolanda: good lukc! | 15:09 |
| *** mrmartin has joined #openstack-infra | 15:09 | |
| fungi | jogo: great summary | 15:10 |
| *** smccully has joined #openstack-infra | 15:11 | |
| *** ChuckC has joined #openstack-infra | 15:12 | |
| flip214 | fungi: thanks, but IIRC both referenced commits were relevant when I read his answer. Guess I'm wrong, though. | 15:12 |
| fungi | wow... jenkins thinks check-pbr-installation-dsvm is averaging 2 hours lately? | 15:12 |
| *** harlowja_at_home has joined #openstack-infra | 15:13 | |
| clarkb | flip214 it is relevant... | 15:13 |
| *** Longgeek has quit IRC | 15:13 | |
| flip214 | clarkb: hi. it is? sorry, couldn't tell from the commit message. | 15:13 |
| flip214 | all right, thanks. | 15:13 |
| clarkb | flip214 yes setting the guest cpu model is reauired for a working multinode cloud | 15:13 |
| flip214 | I was afraid that the needed features got merged when I wasn't looking ;) | 15:13 |
| *** radez_g0n3 is now known as radez | 15:13 | |
| clarkb | however, please start with a single node test | 15:14 |
| clarkb | it will make everything simpler to start | 15:14 |
| *** tonytan4ever has quit IRC | 15:15 | |
| *** afazekas has quit IRC | 15:16 | |
| *** tonytan4ever has joined #openstack-infra | 15:16 | |
| *** dannywilson has joined #openstack-infra | 15:16 | |
| claudiupopa | mordred: just to be sure, do you want this cherry pick to go through git review? | 15:18 |
| *** baoli_ has quit IRC | 15:18 | |
| *** annegentle has quit IRC | 15:19 | |
| *** annegentle has joined #openstack-infra | 15:19 | |
| mordred | claudiupopa: yes. all changes must go throguh git review | 15:22 |
| *** baoli_ has joined #openstack-infra | 15:22 | |
| *** dkranz has quit IRC | 15:22 | |
| claudiupopa | mordred: here it is https://review.openstack.org/#/c/168953/1 | 15:22 |
| *** sputnik13 has joined #openstack-infra | 15:23 | |
| *** dulek has quit IRC | 15:24 | |
| *** shakamunyi has joined #openstack-infra | 15:25 | |
| *** chlong has quit IRC | 15:26 | |
| *** Ala has quit IRC | 15:26 | |
| *** sdake_ has joined #openstack-infra | 15:26 | |
| *** belmoreira has quit IRC | 15:27 | |
| *** hrubi has joined #openstack-infra | 15:29 | |
| *** tiswanso has quit IRC | 15:29 | |
| *** dboik_ has joined #openstack-infra | 15:30 | |
| *** baoli_ has quit IRC | 15:30 | |
| *** sdake has quit IRC | 15:30 | |
| *** baoli_ has joined #openstack-infra | 15:31 | |
| *** tiswanso_ has joined #openstack-infra | 15:31 | |
| *** MaxV has quit IRC | 15:32 | |
| *** e0ne is now known as e0ne_ | 15:32 | |
| *** dboik has quit IRC | 15:32 | |
| *** ghostpl_ has quit IRC | 15:35 | |
| *** ghostpl_ has joined #openstack-infra | 15:36 | |
| *** dizquierdo has quit IRC | 15:37 | |
| *** e0ne_ is now known as e0ne | 15:38 | |
| *** dboik_ has quit IRC | 15:42 | |
| *** dboik has joined #openstack-infra | 15:43 | |
| *** otter768 has joined #openstack-infra | 15:43 | |
| *** deepakcs has quit IRC | 15:43 | |
| *** ddieterly has quit IRC | 15:43 | |
| clarkb | mordred: I am not sure how 168634 addresses your concern over the weekend. Is it relying on the shade internal rate limiting that you linked earlier? | 15:46 |
| *** zz_johnthetubagu is now known as johnthetubaguy | 15:46 | |
| clarkb | mordred: the only problem with that is the rate limit won't know about any of the limiting that nodepool is doing internally but that is likely ok while the transition to shade happens | 15:46 |
| openstackgerrit | Cyril Roelandt proposed openstack-infra/project-config: saharaclient: enable Python3 jobs https://review.openstack.org/168899 | 15:47 |
| openstackgerrit | Monty Taylor proposed openstack-infra/project-config: Add elements for Infra servers https://review.openstack.org/140840 | 15:47 |
| mordred | clarkb: you have to look at the two previous patches | 15:48 |
| mordred | clarkb: nodepool will pass in the nodepool providermanager as a TaskManager parameter | 15:48 |
| *** otter768 has quit IRC | 15:48 | |
| mordred | that means that shade will use nodepool's taskmanager instead of its own task manager | 15:48 |
| mordred | clarkb: which means that any API tasks it runs will get done in the context of the nodepool ratelimiting | 15:48 |
| mordred | clarkb: but nodepool doesn't need to know the specifics about which shade API call run how many REST API calls | 15:49 |
| clarkb | I see | 15:49 |
| zaro | morning | 15:49 |
| clarkb | mordred: does that formalize the task manager in a way that may change (thinking about jeblair's change that made hpcloud have a sad) | 15:50 |
| mordred | clarkb: it would mean we need to define an interface for sure | 15:50 |
| anteaya | morning zaro | 15:50 |
| mordred | and there are additional design questions we'll want to think about - such as how would we handle things like what yolanda is trying to do | 15:50 |
| *** jcoufal has quit IRC | 15:51 | |
| mordred | but basically, I would hope that at the very least it shows that we can have nodepool own the knowledge about how often it wants to do things, and still have shade own and hide the logic about which cloud things are needed for a particular action | 15:52 |
| yolanda | so at first tries i like what i see there, it's regularly triggering the polling tasks and adding nodes, not waiting for the whole queue of create and deletes to finish | 15:52 |
| clarkb | mordred: yup the general idea seems good. I am just a bit concerend because we just tried to change this particular interface | 15:52 |
| yolanda | mordred, you are right, shade should be dumb and nodepool should have the knowledge | 15:52 |
| clarkb | mordred: and trying to coordinate that with a library that is much more concerend about bcakward compatibility will only make it harder | 15:52 |
| boris-42 | clarkb: hi | 15:52 |
| *** shakamunyi has quit IRC | 15:52 | |
| clarkb | boris-42: hello | 15:52 |
| boris-42 | clarkb: could you please add me to nerdreviewer-cores https://review.openstack.org/#/admin/groups/654,members ?) | 15:53 |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Add query for horizon integration test bug 1436903 https://review.openstack.org/168753 | 15:53 |
| openstack | bug 1436903 in OpenStack Dashboard (Horizon) "integration tests failing blocking gate" [Critical,Triaged] https://launchpad.net/bugs/1436903 - Assigned to David Lyle (david-lyle) | 15:53 |
| clarkb | boris-42: done | 15:54 |
| boris-42 | clarkb: thank you) | 15:55 |
| mordred | clarkb: well, actually, I do not think the TaskManager interface is one that we would need to make one we'd care about - it shoudl be an impl detail that shade users do not think about except in extreme cases like nodepool | 15:55 |
| mordred | clarkb: in fact, we could make it a private function to add a taskmanager and put big warning signs "this is onlly to be used by nodepool, don't use it if you don't know what you're doing" | 15:56 |
| *** jcoufal has joined #openstack-infra | 15:56 | |
| *** amitgandhinz has joined #openstack-infra | 15:56 | |
| jogo | have all the nodepool zombie instances been cleaned out and we are actually using both clouds now? | 15:56 |
| mordred | clarkb: (just thinking out loud) | 15:56 |
| *** jcoufal has quit IRC | 15:56 | |
| mordred | clarkb: the important parts of the API for shade and backwards compat woudl be the parts we expect people to program to - which is definitely not the internal task interface | 15:56 |
| sileht | I have a review where I have used 'Depends-On' the dependencies have been merged, but this one is never processed by zuul: https://review.openstack.org/#/c/166763/ | 15:57 |
| clarkb | mordred: I am just thinking in the nodepool case. eg if we have to backpedal on a task manager change like that again it will require either a downgrade of shade too or a roll forward of shade | 15:57 |
| *** ddieterly has joined #openstack-infra | 15:57 | |
| sileht | 'recheck' can help in this case ? | 15:57 |
| *** scheuran has quit IRC | 15:57 | |
| *** markus_z has quit IRC | 15:57 | |
| mordred | clarkb: totally. it's definitely a think we'll need a plan on | 15:58 |
| clarkb | sileht: if that change was approved before its deps then zuul would not have processed it once the deps did merge | 15:59 |
| clarkb | sileht: a recheck should get the ball rolling | 15:59 |
| clarkb | sileht: actually hrm | 15:59 |
| sileht | clarkb, thanks I try | 15:59 |
| clarkb | because the verified vote is already a +1 that may not work | 15:59 |
| anteaya | clarkb: yeah I was looking at that too | 15:59 |
| clarkb | zaro: did you ever get traction on your gerrit change to fix ^ | 15:59 |
| clarkb | sileht: instead a reapproval may be best | 15:59 |
| anteaya | clarkb: I'm inclined to say a new +A | 15:59 |
| clarkb | anteaya: yup | 15:59 |
| marcusvrn1 | asselin kaisers1: hi! I was out of office, then I couldn't attend to the meeting...but have read the log and I have also a problem with test_minimum_basic_scenario | 16:00 |
| *** tonytan4ever has quit IRC | 16:00 | |
| *** dhritishikhar_ has joined #openstack-infra | 16:00 | |
| sileht | clarkb, I will ping a core dev, so | 16:00 |
| anteaya | marcusvrn1: can you post a stacktrace? | 16:00 |
| *** gokrokve_ has joined #openstack-infra | 16:00 | |
| * asselin looks up his log of that too | 16:00 | |
| kaisers1 | marcusvrn1: ah, ok | 16:00 |
| *** Guest30324 has quit IRC | 16:00 | |
| marcusvrn1 | anteaya: sure, one sec... | 16:01 |
| anteaya | thanks | 16:01 |
| nibalizer | fbo: hi | 16:01 |
| kaisers1 | marcusvrn1: asselin: Are these similar issues or different issues with the same test? | 16:01 |
| mordred | clarkb: although - actually, the only thing in the change we just made and reverted that would have been shade-side would have been in Task - and even that was an impl detail change - the real guts f the change was TaskManager's logic - so that really could have been done almost totally with a change to nodepool | 16:01 |
| *** YorikSar has quit IRC | 16:01 | |
| nibalizer | if you have some ideas on how to improve the apply-test im all ears | 16:01 |
| mordred | even in the new patches | 16:01 |
| *** Hal has joined #openstack-infra | 16:02 | |
| *** tqtran has joined #openstack-infra | 16:02 | |
| anteaya | kaisers1: hence the request for some stacktraces | 16:02 |
| clarkb | mordred: I think the task interface changed too | 16:02 |
| zaro | clarkb: no, only got one review. maybe you can post a review?. | 16:02 |
| mordred | clarkb: all we need to do is _add_ a Task api that lets a task manager indicate it wants the other behavior - and nodepool's TaskManager could use either | 16:02 |
| *** Hal is now known as Guest17829 | 16:02 | |
| anteaya | zaro: link? | 16:02 |
| clarkb | zaro: sure, have a link handy? | 16:02 |
| mordred | clarkb: it did - but the Task interface change can be done in a way that can do both things | 16:02 |
| mordred | clarkb: (was just looking at the code) | 16:02 |
| *** gokrokv__ has joined #openstack-infra | 16:02 | |
| *** gokrokve_ has quit IRC | 16:02 | |
| zaro | clarkb: https://gerrit-review.googlesource.com/65530 | 16:03 |
| marcusvrn1 | kaisers1: The error that I got is related to the SSH connection timeout | 16:03 |
| Shrews | mordred: you were busy this weekend | 16:03 |
| marcusvrn1 | anteaya: kaisers1: http://paste.openstack.org/show/197584/ | 16:03 |
| asselin | marcusvrn1, anteaya kaisers1 I have same issue as marcusvrn1 http://paste.openstack.org/show/197586/ | 16:04 |
| *** gokrokve has quit IRC | 16:04 | |
| anteaya | awesome | 16:04 |
| anteaya | you may have found a bug | 16:04 |
| anteaya | well done | 16:04 |
| marcusvrn1 | asselin: the same error here | 16:05 |
| kaisers1 | Mine seems to be different, I'll dig deeper... :) | 16:05 |
| kaisers1 | probably i'll hit the same once i'm done with my issue... | 16:05 |
| marcusvrn1 | kaisers1: yes | 16:05 |
| clarkb | zaro: I really dislike the gerrit overriding ^F in my browser behavior... | 16:05 |
| mordred | clarkb: anyway -I'm nt saying that change would have worked, or that we don't need to do some deep thinking - but looking at the change itself I can see how we could make it so that the forward/back changes were contained within nodepool still | 16:05 |
| anteaya | so now the question is, is it a bug with the test or has the test actually turned up a bug in cinder code | 16:05 |
| anteaya | marcusvrn1 asselin kaisers1 do one of you want to start an etherpad to keep track of links so that you can work together on this? | 16:06 |
| *** annegentle has quit IRC | 16:07 | |
| asselin | anteaya, the test runs as part of cinder in the gate | 16:07 |
| *** unicell has quit IRC | 16:07 | |
| anteaya | asselin: and it is passing in the gate? | 16:07 |
| anteaya | so then why is it failing for two ci systems? | 16:07 |
| marcusvrn1 | anteaya: asselin: another test returns the same error, but it's intermittent... | 16:08 |
| marcusvrn1 | this one: tempest.thirdparty.boto.test_ec2_instance_run.InstanceRunTest.test_compute_with_volumes | 16:08 |
| fungi | jogo: it looks like once our utilization of hpcloud picked back up for the day, we've got 180+ nodes which have been in a "building" state for more than half an hour and all are in hpcloud | 16:08 |
| anteaya | marcusvrn1: which again could be a bug, but a race condition | 16:08 |
| *** ssam2 has quit IRC | 16:09 | |
| jogo | fungi: too bad thanks. | 16:09 |
| asselin | anteaya, drivers exposes other issues such as timing. Default driver (gate) is lvm with is run locally on the same instance. | 16:09 |
| *** shakamunyi has joined #openstack-infra | 16:10 | |
| *** thedodd has quit IRC | 16:10 | |
| anteaya | asselin: ah | 16:10 |
| anteaya | asselin: and yes timing issues are worth looking at | 16:10 |
| *** yfried has joined #openstack-infra | 16:13 | |
| marcusvrn1 | asselin: hmmm may be that's why gate doesn't fail | 16:13 |
| *** ssam2 has joined #openstack-infra | 16:13 | |
| * asselin will be back in 45 mins | 16:13 | |
| fungi | jogo: though it looks like in the last few minutes hpcloud has started returning more ready nodes, so maybe it's better behaved than last week | 16:14 |
| clarkb | zaro: commented, one inline thing but not familiar enough with the code to know if its a real -1 | 16:14 |
| fbo | nibalizer, hi yes to be short I think we can use lxc on virtual node to spawn some precise,trusty,centos containers and run the manifests based on a custom manifests/site.pp and hiera db filled with fake config/key pairs in order to have an almost complete CI up and running | 16:14 |
| fbo | nibalizer, then we can add some smoke testing with tools like serverspec to check services are correctly setup | 16:15 |
| clarkb | mordred: shade still needs to know what the Task interface is though right? eg it has to pick one of the options that may be available to it? in any case I am sure this is doable, just making sure we think about the backward compat requirements around it | 16:16 |
| nibalizer | fbo: yea i've done some stuff like that | 16:16 |
| clarkb | mordred: I don't think its too far fetched to say if nodepool requires this then any full features shade consumer will likely need it too | 16:17 |
| nibalizer | (outside of openstack testing i should say) | 16:17 |
| jogo | fungi: good to hear | 16:17 |
| nibalizer | there are some things that we manage that don't or sometimes dont work in lxc containers/docker containers like iptables | 16:17 |
| nibalizer | but we can get there when we get there | 16:17 |
| nibalizer | i was thinking to use envassert instead of serverspec since it is python | 16:17 |
| *** asselin has quit IRC | 16:19 | |
| fbo | nibalizer, same for me I have some experience with LXC to test deployment based on puppet for tool like gerrit/jenkins/zuul but not with the system-config manifest | 16:19 |
| *** armax has quit IRC | 16:19 | |
| yolanda | so mordred, the stress test was quite good, trying to generate some charts about that | 16:19 |
| fbo | Yes if we can skip ruby stuff I'll be happy, I'll look at envassert then :) | 16:20 |
| mordred | clarkb: I think it's less likely. ansible, for instance, wants to have nothing to do with knowing about this. (and I had to make zero changes to ansible for it to continue working) BUT - I'm on the phone, so I'll sync back up on it in about an hour | 16:20 |
| fbo | nibalizer, so what do you think if we merge efforts to bring something in openstack-infra to perform more in depth testing of the manifests ? | 16:21 |
| *** ssam2 has quit IRC | 16:24 | |
| clarkb | yolanda: for the priority queue work shouldn't most of those hit the internal cache instead? | 16:26 |
| yolanda | well, the request is only generated when the cache is missing | 16:26 |
| clarkb | yolanda: right, so 99% of the time the cache will be hit and we keep moving? | 16:26 |
| yolanda | so it just generates that api call and puts that as high priority, so the polling is run on time | 16:26 |
| yolanda | well, it's not what happens really... the cache interval is quite short | 16:26 |
| yolanda | and under load, the create, delete nodes tasks are just queued and preventing the polling tasks to run | 16:27 |
| yolanda | i've seen cases of 15-20 minutes without this task being executed | 16:27 |
| nibalizer | fbo: i dont really have an efforts in flight | 16:27 |
| clarkb | yolanda: the polling tasts are? | 16:27 |
| nibalizer | there is one patch but it doesn't go close to what you're talking about | 16:27 |
| yolanda | getserverlists, getfloatingiplists... | 16:27 |
| *** ajmiller has joined #openstack-infra | 16:28 | |
| *** dkranz has joined #openstack-infra | 16:28 | |
| yolanda | clarkb, so from our side, the problem was that nodes were created, but as the getserverlists wasn't called, system wasn't able to asign a floating ip . Even when that floating ip was assigned, the floatingiplist task was taking 10 minutes to run, so the node was idle there | 16:28 |
| yolanda | clarkb, it comes a bit with a conversation we had with mordred and jeblair on friday, talking about having 2 different managers: one for lists tasks, one for the create/delete... but i think that this priority makes thing simpler and it works | 16:29 |
| *** baoli_ has quit IRC | 16:29 | |
| clarkb | yolanda: basically the cache update period is > than the cache timeout period | 16:29 |
| fbo | nibalizer ok I'm writing a spec to describe that so maybe it can be a good starting point to go deeper in order to discuss and defined something ? | 16:29 |
| yolanda | much greater, yes | 16:30 |
| *** prmtl_ has quit IRC | 16:30 | |
| yolanda | clarkb, but it is nothing we can control the way nodepool it's now, because it depends on the rate and the number of api calls queued | 16:30 |
| *** e0ne has quit IRC | 16:30 | |
| yolanda | so that's my reason for the priority, so we make sure that this interval is accomplished | 16:30 |
| clarkb | yolanda: right and thats a FIFO with expensive tasks inline | 16:30 |
| *** ssam2 has joined #openstack-infra | 16:30 | |
| yolanda | clarkb, that should be implemented as a heapq in python? | 16:31 |
| nibalizer | fbo: good idea | 16:31 |
| nibalizer | fbo: also I think the openstackci module is a good place to jump off our testing | 16:32 |
| clarkb | yolanda: maybe? my only concern here is that the list servers/fips/images/etc may starve the creates and deletes | 16:32 |
| nibalizer | since that will be a known excpected state with a lot of complexity | 16:32 |
| yolanda | clarkb, no, because they rely on cache | 16:32 |
| yolanda | so you won't be duplicated calls, only once at a time that cache expires | 16:32 |
| *** pblaho has quit IRC | 16:32 | |
| *** hashar has quit IRC | 16:32 | |
| clarkb | yolanda: and without updating the priorities of any tasks based on age (or some other criteria) we have no way of ensuring that the other tasks are serviced | 16:32 |
| *** aysyd has quit IRC | 16:32 | |
| clarkb | yolanda: thats only true if the api responds within a cache timeout period right? | 16:33 |
| clarkb | yolanda: and I think we are already saying that this may not be the case? though I guess that is not the case or other api calls | 16:33 |
| clarkb | since the listings don't starve themselves, the deletes starve them | 16:33 |
| jeblair | yolanda, clarkb: what priority thing? | 16:33 |
| *** marun has joined #openstack-infra | 16:33 | |
| clarkb | jeblair: 168686 | 16:33 |
| yolanda | jeblair https://review.openstack.org/#/c/168686/ | 16:33 |
| *** esker has joined #openstack-infra | 16:33 | |
| *** esker has quit IRC | 16:34 | |
| yolanda | clarkb, let me check about the caching implementation | 16:34 |
| openstackgerrit | Spencer Krum proposed openstack-infra/system-config: Print the node definition before apply test https://review.openstack.org/167346 | 16:34 |
| yolanda | although this will be moved to shade so it will be different | 16:34 |
| *** samueldmq has quit IRC | 16:34 | |
| *** emagana has joined #openstack-infra | 16:34 | |
| *** amitgandhinz has quit IRC | 16:35 | |
| *** amitgandhinz has joined #openstack-infra | 16:35 | |
| yolanda | clarkb, so server lists and floating ip lists have an interval of 5 secons | 16:35 |
| jeblair | yolanda, clarkb: the default cache interval is 5 seconds -- hpcloud was taking 20-30 seconds for a create 2 weeks ago, and 40-60 seconds for a create last week | 16:35 |
| yolanda | seconds | 16:35 |
| *** marcusvrn1 has quit IRC | 16:35 | |
| *** marcusvrn1 has joined #openstack-infra | 16:35 | |
| yolanda | jeblair, so as per our friday conversation, i thought into a different approach, that was that priorityqueue thing | 16:35 |
| *** aysyd has joined #openstack-infra | 16:35 | |
| fbo | nibalizer, yes you are right ! I'll mention in the spec. | 16:36 |
| jeblair | yolanda, clarkb: i believe that means that create and deletes will be starved until all existing create/deletes are complete | 16:36 |
| yolanda | i've been testing that on my hpcloud account and results were better | 16:36 |
| nibalizer | fbo: 167346 is my current set of changes to the apply test, just trying to split it up and make it easier to read | 16:36 |
| mordred | I believe the create/delete speed is going to be dependent on the tenant | 16:36 |
| jeblair | mordred: yes, and change | 16:36 |
| mordred | because of soft deletes and quota calculation | 16:36 |
| yolanda | as it's getting the created servers and floating ips as they come | 16:36 |
| jeblair | yolanda: how do you define better? i believe it _will_ make a single bringup faster. but it will lower overall throughput. | 16:36 |
| fbo | nibalizer, thanks I'll have a look | 16:36 |
| jeblair | yolanda: i think we care more about overall throughput | 16:36 |
| yolanda | jeblair, i'm working on some carts | 16:36 |
| yolanda | charts | 16:36 |
| fungi | it's sort of like measuring resistance across a circuit with a trickle current rather than under load. you can get entirely different readings | 16:37 |
| jeblair | ya | 16:37 |
| yolanda | jeblair, well, my thought is: you have ready nodes without a floating ip, because the serverlists task hasn't come. That means idle nodes , increasing quota and doing nothing | 16:37 |
| fungi | i've seen hpcloud perform just fine over the weekend when we ask very little of it, then crumble again when the week starts | 16:37 |
| jeblair | for our workload, i think we want to know how many servers we can create and delete over a multi-hour period | 16:37 |
| yolanda | if you are able to get that nodes working as they come, you will gain time | 16:38 |
| *** harlowja_at_home has quit IRC | 16:38 | |
| *** baoli has joined #openstack-infra | 16:38 | |
| yolanda | we've had cases of serverlists not executing on 15 minutes, you said that as well on friday. So it means 15 minutes for a node waiting there without activity and wasting quota | 16:38 |
| jeblair | yolanda: yes that is true when you look at that single node. but when you look at hundreds of them, you see that you can only create and delete them at a fixed, slow, rate. so the overall throughput of the system is defined by that rate. | 16:38 |
| *** unicell has joined #openstack-infra | 16:39 | |
| clarkb | yolanda: I think thats more of an issue around atomicity or requests and less about priority | 16:39 |
| clarkb | basically we want atomic work units that result in "working node" "node removed" | 16:39 |
| clarkb | instead of 5 api calls for each thing all dependent on each other | 16:39 |
| jeblair | yolanda: yeah, which is why i suggested separating the providers so that you maximize the throughput of the slow calls and the fast ones. | 16:40 |
| *** jlvillal_ has joined #openstack-infra | 16:40 | |
| jeblair | yolanda: however, if we go with the priority queue as written, we will actually just reverse the situation | 16:40 |
| yolanda | jeblair, but the lists task come on a controlled rate due to caching | 16:41 |
| jeblair | yolanda: also, splitting the providers means separating calls that require the quota lock and ones that don't. | 16:41 |
| *** sergsh has quit IRC | 16:41 | |
| yolanda | jeblair, do we have calls that don't need a rate? | 16:41 |
| *** jlvillal_ has left #openstack-infra | 16:41 | |
| jeblair | yolanda: not to my knowledge (though implementation my vary by provider) | 16:42 |
| *** ghostpl_ has quit IRC | 16:42 | |
| jeblair | yolanda: hpcloud is behaving pathologically bad | 16:42 |
| jeblair | i am not certain we should try to engineer a complete solution for this behavior | 16:42 |
| greghaynes | clarkb: Is what youre getting at that essentially that we dont mind starving list if its not going to actually result in getting a resource sooner? | 16:43 |
| *** MaxV has joined #openstack-infra | 16:43 | |
| yolanda | ok, let me work on the provider approach as well, it's not complicated | 16:43 |
| yolanda | but wanted to show and test that as well | 16:43 |
| yolanda | we have some tooling to analyze it so we could test both | 16:43 |
| jeblair | yolanda: okay. i'd love to see data from this approach too | 16:43 |
| *** Somay has joined #openstack-infra | 16:43 | |
| clarkb | greghaynes: yes | 16:43 |
| jeblair | yolanda: yeah, i think there's stuff we can learn from here, but i have concerns about it :) | 16:44 |
| clarkb | greghaynes: because without atomic work unit api calls you have a bunch of calls scattered through that queue and the list* calls are only a portion of that | 16:44 |
| greghaynes | ok, I think thats key - and preserving insertion ordering should inherently do that | 16:44 |
| jeblair | yolanda: but i think we will need to know how it performs at scale -- with at least 100 nodes over several hours | 16:44 |
| *** crc32 has joined #openstack-infra | 16:44 | |
| yolanda | yes, i just wanted to leave that running overnight | 16:44 |
| openstackgerrit | Merged openstack-infra/nodepool: Correct diskimage-builder name usage https://review.openstack.org/167771 | 16:44 |
| jeblair | yolanda: if you don't have the quota for that, you should be able to create a simulation by using the timing data in the nodepool log | 16:44 |
| *** yamahata has joined #openstack-infra | 16:44 | |
| *** jistr has quit IRC | 16:44 | |
| clarkb | greghaynes: instead we need to either process work units together internal to nodepool or fix nova/neutron to give us apis that represent logical changes | 16:45 |
| yolanda | jeblair, i got a magical quota increase this afternoon :) | 16:45 |
| jeblair | yolanda: since you know that under load it takes X seconds for a create call, and Y seconds for a list ip call | 16:45 |
| jeblair | yolanda: can you use 100 nodes without affecting infra or gozer? i thought hpcloud was under a lot of capacity pressure? | 16:45 |
| *** isviridov has joined #openstack-infra | 16:45 | |
| yolanda | jeblair, well, at least i've been doing that for half an hour | 16:45 |
| yolanda | and it was doing the job | 16:46 |
| nibalizer | yolanda: in us east or a different region? | 16:46 |
| greghaynes | clarkb: Yep, thats the basis of a lot of modern scheduling problems ;) | 16:46 |
| yolanda | us east | 16:46 |
| jeblair | clarkb: i would love atomic work units from openstack, but by not making them atomic in nodepool we are able to be much more efficient | 16:46 |
| greghaynes | clarkb: inserting into work queue in order is a quick way to get that though | 16:46 |
| *** ghostpl_ has joined #openstack-infra | 16:46 | |
| yolanda | i have over 100 instances creted now | 16:46 |
| clarkb | greghaynes: it is a queue so it already is a fifo | 16:46 |
| clarkb | greghaynes: the problem is you have more than one input | 16:47 |
| greghaynes | oh, its mutiprocess | 16:47 |
| yolanda | so i'll let it run until night, and will grab the logs, and tomorrow i can test with the independent provider approach | 16:47 |
| *** MaxV has quit IRC | 16:47 | |
| yolanda | don't want to let them run on the night so i crash something and i'm not awake... | 16:47 |
| clarkb | jeblair: ya, I don't think there is a major problem with what we do now. Its easier for me to think about this problem in the context of scheduling logical work units than it is "prioritize list servers" | 16:47 |
| jeblair | yolanda: okay. while you're waiting, you could try writing a program to simulate it much faster :) | 16:48 |
| clarkb | greghaynes: its multi threaded | 16:48 |
| greghaynes | actually, I think what you have still gets what youre describing - something is necessarially blocking on the task in the queue | 16:48 |
| greghaynes | it just might not be the most important task, which is a different problem | 16:48 |
| jeblair | yolanda, clarkb: so i think the starvation potential with the priority queue is more nuanced than what i described before... | 16:48 |
| *** jaypipes has quit IRC | 16:49 | |
| yolanda | what do you mean? | 16:50 |
| *** MaxV has joined #openstack-infra | 16:50 | |
| jeblair | yolanda, clarkb: i think what would be required for it to happen is for a list servers and list floating ips task to get into the queue around the same time. if the second one takes more than 5 seconds, then the next list <something> task will get promoted. and that's how we starve create and delete. | 16:50 |
| yolanda | you are right with that | 16:51 |
| clarkb | jeblair: yes that is what I was trying to describe earlier without specific cache timeout numbers | 16:51 |
| jeblair | yolanda, clarkb: so the requirement for starvation is that ether list servers or list floating ips takes more than the cache interval. | 16:51 |
| *** Swami has joined #openstack-infra | 16:51 | |
| jeblair | clarkb: sorry, i'm catching up :) | 16:51 |
| *** ildikov has joined #openstack-infra | 16:52 | |
| jeblair | yolanda: so anyway, let's collect data and think about this, but even if it looks better, we need to carefully consider it because if listing servers currently takes 4.9 seconds, it could mean that performance drastically changes next week when it takes 5.1 seconds. | 16:52 |
| yolanda | jeblair, and we could prevent in some way to launch a request if one was already launched before? | 16:53 |
| jeblair | yolanda: well, that's what the cache does. we could increase the timeout, but that will have ripple effects (making the whole system slower again) | 16:54 |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard: Fixed import https://review.openstack.org/168467 | 16:54 |
| yolanda | that's a complicated balance :) | 16:54 |
| jeblair | ya | 16:54 |
| yolanda | it we had callbacks of some way we could just stop the polling ones... | 16:54 |
| mordred | we don't | 16:54 |
| mordred | sadly | 16:55 |
| mordred | but yes | 16:55 |
| mordred | it would be so great if openstack had a user-facing callback mechanism | 16:55 |
| yolanda | modred, ask for it :) | 16:55 |
| *** dizquierdo has joined #openstack-infra | 16:55 | |
| mordred | yolanda: heh. I'm just focusing on getting ssh host keys out of nova for next cycle - I think a generalized callback mechanism might take a while :) | 16:56 |
| *** dizquierdo_ has joined #openstack-infra | 16:57 | |
| *** derekh has quit IRC | 16:57 | |
| *** wicope has joined #openstack-infra | 16:58 | |
| *** AJaeger_ has joined #openstack-infra | 16:58 | |
| *** AJaeger has quit IRC | 16:58 | |
| *** erlon has joined #openstack-infra | 16:59 | |
| *** YorikSar has joined #openstack-infra | 16:59 | |
| *** esker has joined #openstack-infra | 16:59 | |
| *** dizquierdo has quit IRC | 17:00 | |
| *** dkranz has quit IRC | 17:01 | |
| *** ssam2 has quit IRC | 17:01 | |
| *** btran has joined #openstack-infra | 17:02 | |
| *** AJaeger has joined #openstack-infra | 17:03 | |
| *** esker has quit IRC | 17:03 | |
| *** AJaeger has quit IRC | 17:03 | |
| *** AJaeger has joined #openstack-infra | 17:03 | |
| *** Guest17829 has quit IRC | 17:03 | |
| *** sarob has joined #openstack-infra | 17:03 | |
| *** adalbas has joined #openstack-infra | 17:04 | |
| *** mpavone has quit IRC | 17:04 | |
| *** Hal has joined #openstack-infra | 17:04 | |
| *** bknudson has quit IRC | 17:04 | |
| *** Hal is now known as Guest94647 | 17:05 | |
| *** tiswanso_ has quit IRC | 17:05 | |
| *** sarob_ has joined #openstack-infra | 17:06 | |
| *** davideagnello has joined #openstack-infra | 17:06 | |
| *** hdd has joined #openstack-infra | 17:06 | |
| AJaeger_ | jogo, jeblair: REgarding gantt: Is then https://review.openstack.org/#/c/168682/ the right approach? Or should I abandon both for now and we leave status quo? | 17:08 |
| *** MaxV has quit IRC | 17:08 | |
| *** sarob has quit IRC | 17:10 | |
| *** Somay has quit IRC | 17:11 | |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard: Remove page_size_maximum https://review.openstack.org/167633 | 17:11 |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard: Remove page_size_default https://review.openstack.org/167652 | 17:11 |
| *** HeOS has quit IRC | 17:11 | |
| openstackgerrit | Spencer Krum proposed openstack-infra/system-config: Add a helpful friend to gerrit reviews https://review.openstack.org/169010 | 17:13 |
| *** pal_ has joined #openstack-infra | 17:13 | |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard: Fixed import https://review.openstack.org/168467 | 17:13 |
| *** ssam2 has joined #openstack-infra | 17:14 | |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard: Remove page_size_maximum https://review.openstack.org/167633 | 17:14 |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard: Remove page_size_default https://review.openstack.org/167652 | 17:14 |
| *** armax has joined #openstack-infra | 17:16 | |
| *** tsg_ has joined #openstack-infra | 17:16 | |
| *** dkranz has joined #openstack-infra | 17:16 | |
| *** ssam2 has quit IRC | 17:17 | |
| *** sarob_ is now known as sarob | 17:17 | |
| *** yamahata has quit IRC | 17:18 | |
| *** dtantsur is now known as dtantsur|afk | 17:18 | |
| *** Somay has joined #openstack-infra | 17:20 | |
| *** MarkAtwood has quit IRC | 17:20 | |
| *** peristeri has quit IRC | 17:21 | |
| *** Somay has quit IRC | 17:22 | |
| *** prmtl has joined #openstack-infra | 17:22 | |
| *** jgrimm is now known as zz_jgrimm | 17:22 | |
| *** sarob_ has joined #openstack-infra | 17:24 | |
| *** achanda has joined #openstack-infra | 17:24 | |
| *** sarob has quit IRC | 17:24 | |
| *** melwitt has joined #openstack-infra | 17:25 | |
| *** harlowja_away is now known as harlowja | 17:25 | |
| *** tnovacik has quit IRC | 17:25 | |
| *** gokrokve has joined #openstack-infra | 17:25 | |
| *** sarob has joined #openstack-infra | 17:25 | |
| *** amitgandhinz has quit IRC | 17:25 | |
| *** samueldmq has joined #openstack-infra | 17:25 | |
| zaro | clarkb: replied to your question. i agree with you about the ^f i'm guessing that it's because it works well on macs because there's the command-f key. | 17:26 |
| *** tiswanso has joined #openstack-infra | 17:26 | |
| arosen | quick question. to push a git tag the wiki says: git push gerrit tag version; what is the remote of gerrit? | 17:27 |
| clarkb | zaro: fwiw I don't think they should be doing different things, they should do the same exact thing so that the behavior is always the same for stream events | 17:27 |
| clarkb | arosen: if you use git-review it is 'gerrit' | 17:27 |
| SpamapS | mordred: regarding user facing callback mechanisms: Zaqar was supposed to be the facilitator for that. But it got pounded on. :-P | 17:27 |
| arosen | clarkb: i do use git review but I guess that dosen't set my .git/config to gerrit ? | 17:27 |
| *** MarkAtwood has joined #openstack-infra | 17:27 | |
| clarkb | arosen: do a `git remote -v` | 17:28 |
| *** dkranz has quit IRC | 17:28 | |
| *** e0ne has joined #openstack-infra | 17:28 | |
| clarkb | SpamapS: I thought zaqar specifically didn't want to be in cloud | 17:28 |
| arosen | clarkb: doh i see. | 17:28 |
| arosen | Creating a git remote called "gerrit" that maps to: | 17:28 |
| arosen | ssh://arosen@review.openstack.org:29418/stackforge/python-congressclient.git | 17:28 |
| *** gokrokv__ has quit IRC | 17:28 | |
| arosen | thanks | 17:29 |
| clarkb | SpamapS: that was a huge point of confusion and discussion | 17:29 |
| *** sarob_ has quit IRC | 17:29 | |
| *** ivar-lazzaro has joined #openstack-infra | 17:29 | |
| *** gokrokve has quit IRC | 17:30 | |
| *** ivar-lazzaro has quit IRC | 17:30 | |
| clarkb | jogo: I think the tests that failed on my 64bit guest cpu change may have done so because that cpu isn't "real" enough to actually log into the VM | 17:30 |
| prmtl | folks, I kindly ask for a review (and hopefully merge): https://review.openstack.org/#/c/167539/ | 17:30 |
| *** ivar-lazzaro has joined #openstack-infra | 17:30 | |
| clarkb | jogo: but knowing what I know now we can do a core2duo - monitor - pse36 and that *should* work so going to try cleaning that up this morning | 17:31 |
| *** julim has quit IRC | 17:31 | |
| *** hodos has joined #openstack-infra | 17:31 | |
| SpamapS | clarkb: it was supposed to be available to in and out | 17:31 |
| SpamapS | clarkb: so that was supposed to be where things like Heat and Nova would put messages about events users had requested. | 17:32 |
| SpamapS | I believe Heat still wants to do that. | 17:33 |
| SpamapS | and may even drive the effort into the others. | 17:33 |
| *** julim has joined #openstack-infra | 17:33 | |
| *** jerryz_ has quit IRC | 17:33 | |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard-webclient: Removed grunt:server https://review.openstack.org/163165 | 17:34 |
| *** mriedem has quit IRC | 17:34 | |
| fungi | arosen: yep, see the first note at http://docs.openstack.org/infra/manual/drivers.html#tagging-a-release | 17:34 |
| zaro | clarkb: probably agree with you there, but didn't want to change the existing flow too much for that change. | 17:34 |
| *** ghostpl_ has quit IRC | 17:34 | |
| clarkb | zaro: ok | 17:35 |
| *** dboik has quit IRC | 17:35 | |
| *** e0ne is now known as e0ne_ | 17:35 | |
| *** dboik has joined #openstack-infra | 17:35 | |
| *** crc32 has quit IRC | 17:36 | |
| *** tonytan4ever has joined #openstack-infra | 17:36 | |
| zaro | clarkb: too bad the shortcut keys are not confiugrable. that would be a nice feature | 17:36 |
| *** amotoki has quit IRC | 17:36 | |
| *** annegentle has joined #openstack-infra | 17:37 | |
| *** Guest94647 has quit IRC | 17:38 | |
| yolanda | SpamapS, that looks as an interesting feature, didn't know about that effort | 17:38 |
| *** yamahata has joined #openstack-infra | 17:39 | |
| *** amitgandhinz has joined #openstack-infra | 17:39 | |
| *** e0ne_ is now known as e0ne | 17:39 | |
| zaro | clarkb: looks like someone created a patch for that but, it's on the android fork of gerrit but I can't access it. Argg! probably in draft mode. | 17:39 |
| *** aduarte has joined #openstack-infra | 17:41 | |
| *** ZZelle_ has joined #openstack-infra | 17:41 | |
| *** dkranz has joined #openstack-infra | 17:41 | |
| *** jaypipes has joined #openstack-infra | 17:42 | |
| *** annegentle has quit IRC | 17:43 | |
| *** otter768 has joined #openstack-infra | 17:44 | |
| *** annegentle has joined #openstack-infra | 17:44 | |
| *** MaxV has joined #openstack-infra | 17:44 | |
| openstackgerrit | Merged openstack-infra/project-config: Add CI jobs for fuel-tasklib https://review.openstack.org/167539 | 17:46 |
| *** wicope has joined #openstack-infra | 17:46 | |
| *** ihrachyshka has quit IRC | 17:47 | |
| *** otter768 has quit IRC | 17:48 | |
| *** Sukhdev has joined #openstack-infra | 17:48 | |
| *** e0ne is now known as e0ne_ | 17:49 | |
| SpamapS | yolanda: I think it would be massively useful if a user could subscribe to status changes for an individual instance. | 17:51 |
| *** mriedem has joined #openstack-infra | 17:51 | |
| SpamapS | yolanda: and just open up a longpoll http request to receive notifications. | 17:51 |
| *** prmtl has quit IRC | 17:51 | |
| SpamapS | or a gearman connection. :) | 17:51 |
| *** peristeri has joined #openstack-infra | 17:52 | |
| openstackgerrit | Andrey Pavlov proposed openstack-infra/project-config: Improve ec2-api gating jobs for rally testing https://review.openstack.org/166648 | 17:53 |
| *** e0ne_ has quit IRC | 17:54 | |
| *** pal_ has quit IRC | 17:55 | |
| *** dustins_ has joined #openstack-infra | 17:56 | |
| *** dustins_ has quit IRC | 17:57 | |
| jogo | AJaeger_: not sure myself actually | 17:57 |
| jogo | Jay's trademark concern over gantt make me thing the attic is a good solution | 17:58 |
| *** sarob has quit IRC | 17:58 | |
| AJaeger_ | jogo: we don't need to force a solution yet, I was just trying to clean this up,... | 17:59 |
| *** dustins has quit IRC | 17:59 | |
| jogo | AJaeger_: yeah thanks for doing this. It will be good to figure out an answer | 18:00 |
| *** dustins has joined #openstack-infra | 18:01 | |
| clarkb | jogo: trademark concern? A quick uspto search doesn't really bring anything up | 18:02 |
| jogo | clarkb: I was taking jaypipes comment on face value, not sure what he was referring to | 18:02 |
| fungi | more like contention over the term "gantt" (because gantt charts are a thing) | 18:02 |
| clarkb | jogo: oh is there a comment? /me digs through sb for change link | 18:03 |
| *** hodos|2 has joined #openstack-infra | 18:03 | |
| *** ghostpl_ has joined #openstack-infra | 18:03 | |
| clarkb | neat uspto shows you liveness and abandonment dates | 18:04 |
| clarkb | I wonder how up to date that info is | 18:04 |
| jaypipes | clarkb: I commented that I would be surprised if the project name "gantt" didn't get messy, legally | 18:04 |
| *** teran_ has quit IRC | 18:04 | |
| *** teran has joined #openstack-infra | 18:04 | |
| *** hodos has quit IRC | 18:05 | |
| *** tsg_ has quit IRC | 18:05 | |
| *** tsg has joined #openstack-infra | 18:05 | |
| *** greghaynes has quit IRC | 18:07 | |
| yolanda | SpamapS, yes, i'd love to have some way of callbacks, but this doesn't sound like an easy task | 18:09 |
| *** jlanoux has quit IRC | 18:09 | |
| *** teran has quit IRC | 18:09 | |
| *** bknudson has joined #openstack-infra | 18:10 | |
| *** dhritishikhar_ has quit IRC | 18:11 | |
| *** thingee has joined #openstack-infra | 18:11 | |
| *** bknudson has left #openstack-infra | 18:12 | |
| *** greghaynes has joined #openstack-infra | 18:13 | |
| *** dizquierdo_ has quit IRC | 18:14 | |
| *** bknudson has joined #openstack-infra | 18:14 | |
| *** svarnau has joined #openstack-infra | 18:18 | |
| svarnau | Is bug-tracking for zuul still on storyboard? I thought I filed a bug, but I see no trace of it. | 18:22 |
| *** esker has joined #openstack-infra | 18:22 | |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard-webclient: Typo in Project Groups. https://review.openstack.org/169037 | 18:23 |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard-webclient: Project Group detail page improvements. https://review.openstack.org/169038 | 18:23 |
| *** johnthetubaguy is now known as zz_johnthetubagu | 18:23 | |
| *** dizquierdo has joined #openstack-infra | 18:23 | |
| fungi | svarnau: default storyboard page size is small and the ui currently doesn't give you any indication it's exceeded (pagination support hasn't been completed yet). check your storyboard account prefs and increase your minimum page size | 18:25 |
| svarnau | fungi: Thanks! that was indeed the problem | 18:26 |
| *** gokrokve has joined #openstack-infra | 18:26 | |
| *** crc32 has joined #openstack-infra | 18:26 | |
| *** _nadya_ has quit IRC | 18:26 | |
| *** ddieterly has quit IRC | 18:27 | |
| *** Swami has quit IRC | 18:28 | |
| *** Swami has joined #openstack-infra | 18:28 | |
| *** svarnau has quit IRC | 18:30 | |
| *** MaxV has quit IRC | 18:31 | |
| *** e0ne has joined #openstack-infra | 18:34 | |
| *** ddieterly has joined #openstack-infra | 18:36 | |
| *** achanda has quit IRC | 18:38 | |
| *** HeOS has joined #openstack-infra | 18:39 | |
| openstackgerrit | Matthew Treinish proposed openstack-infra/project-config: Use localrc passthrough for tempest no admin job https://review.openstack.org/169042 | 18:39 |
| mtreinish | sdague: ^^^ will that do the right thing? | 18:39 |
| *** pal_ has joined #openstack-infra | 18:40 | |
| sdague | mtreinish: nope | 18:40 |
| openstackgerrit | Joe Gordon proposed openstack-infra/project-config: Move aiopcpu to nova check queue https://review.openstack.org/169043 | 18:40 |
| mtreinish | heh, figured | 18:40 |
| jogo | armax sdague: ^ | 18:41 |
| jogo | dansmith: ^ | 18:41 |
| sdague | mtreinish: DEVSTACK_LOCAL_CONFIG | 18:41 |
| mtreinish | ah, ok thanks | 18:41 |
| *** carl_baldwin_ has quit IRC | 18:41 | |
| sdague | jogo: is that non-voting? | 18:41 |
| *** dustins has quit IRC | 18:42 | |
| armax | jogo: thanks for the heads-up | 18:42 |
| sdague | anyway, got to run for a bit | 18:42 |
| jogo | sdague: yes, but let me double check | 18:43 |
| *** dhritishikhar_ has joined #openstack-infra | 18:43 | |
| jogo | sdague: yes it is non voting (line 694) | 18:44 |
| openstackgerrit | Matthew Treinish proposed openstack-infra/project-config: Add os-testr repo https://review.openstack.org/167243 | 18:44 |
| openstackgerrit | Matthew Treinish proposed openstack-infra/project-config: Use localrc passthrough for tempest no admin job https://review.openstack.org/169042 | 18:45 |
| *** ivar-lazzaro has quit IRC | 18:46 | |
| *** spzala has quit IRC | 18:46 | |
| *** xyang1 has quit IRC | 18:46 | |
| *** dustins has joined #openstack-infra | 18:48 | |
| *** pal_ has quit IRC | 18:51 | |
| *** Hal has joined #openstack-infra | 18:52 | |
| *** Hal is now known as Guest65061 | 18:53 | |
| *** Rockyg has joined #openstack-infra | 18:57 | |
| *** achanda has joined #openstack-infra | 18:57 | |
| *** achanda has quit IRC | 18:57 | |
| *** achanda has joined #openstack-infra | 18:58 | |
| *** dizquierdo has quit IRC | 19:00 | |
| *** baoli has quit IRC | 19:00 | |
| *** xyang1 has joined #openstack-infra | 19:01 | |
| *** tkelsey has quit IRC | 19:01 | |
| openstackgerrit | Matthew Treinish proposed openstack-infra/project-config: Add os-testr repo https://review.openstack.org/167243 | 19:02 |
| clarkb | mordred: for 168601 why wouldn't nodepool provide that info to shade using its existing config? | 19:04 |
| clarkb | mordred: hopefully the use of shade should be transparent to nodepool users | 19:05 |
| *** tnovacik has joined #openstack-infra | 19:06 | |
| *** claudiub has joined #openstack-infra | 19:07 | |
| openstackgerrit | Armando Migliaccio proposed openstack-infra/project-config: Enable DVR for Neutron aiopcpu job https://review.openstack.org/158415 | 19:08 |
| armax | clarkb, jogo: ^^^ | 19:08 |
| *** spzala has joined #openstack-infra | 19:08 | |
| armax | clarkb: that was fast! | 19:08 |
| mordred | clarkb: it can continue to do that - nothing will break | 19:09 |
| clarkb | mordred: then can we abandon that change stack? | 19:10 |
| mordred | clarkb: however, there are additional things that we can get by starting to use the clouds.yaml file | 19:10 |
| mordred | no | 19:10 |
| mordred | we want it | 19:10 |
| mordred | but it's not going to break other people | 19:10 |
| clarkb | why do we want it? can you make sure that is explained in the commit messages (I am pretty sure it is not) | 19:10 |
| clarkb | mordred: I was hoping to avoid this philisophica ldiscussion but I think that users should never need to configure libraries directly when using libs in applications | 19:11 |
| clarkb | mordred: this is a major issue I have with how oslo.config works | 19:11 |
| mordred | sure. I'll go rework the config file - tl;dr - if we have the file, then when the openstackclient patch lands we can use with openstackclient or ansible on that host to perform any ad-hoc maint tasks without having to duplicate auth config | 19:11 |
| mordred | clarkb: what? | 19:11 |
| clarkb | mordred: as a nodepool user that wants features you shouldn't need to write multiple config files | 19:12 |
| mordred | I agree | 19:12 |
| mordred | and you don't have to | 19:12 |
| clarkb | and more importantl you should need to know that shade ever exists and has special config options | 19:12 |
| clarkb | *shouldn't need to know | 19:12 |
| mordred | you don't have to | 19:12 |
| mordred | HOWEVER - if you want to _Also_ take advantage of teh fact that there is a config file that configures several other openstack utilities as well, you can opt in to doing so | 19:12 |
| clarkb | mordred: so this is purely for convenience and has nothing to do with nodepool? | 19:13 |
| mordred | which is why we want it | 19:13 |
| mordred | yes | 19:13 |
| mordred | nodepool does not need to require a clouds.yaml | 19:13 |
| clarkb | mordred: ok that isn't clear at all given that 168602 exists | 19:13 |
| mordred | if we, or anyone else, already has named configured clouds on their system beacuse they want to use those features, they can use them | 19:13 |
| mordred | clarkb: fair. I'll rework the commit message | 19:14 |
| clarkb | mordred: and yes oslo.confg has us generating environment specific config files because each time *client changes your config changes too | 19:14 |
| clarkb | mordred: its a bit crazy | 19:15 |
| mordred | clarkb: I totally agree with that | 19:15 |
| *** hashar has joined #openstack-infra | 19:15 | |
| mordred | clarkb: this is more "there are 5 pieces of information about your clouds that multiple tools need to know to operate, so if we put those informations in a sane place and have all the tools be able to opt-in to referring to a named cloud, it reduces the duplicative places we need to express those informations" | 19:15 |
| mordred | but I think opt-in is key | 19:15 |
| clarkb | yes optin is good | 19:16 |
| mordred | the openstackclient patch, for instance, allows you do say "openstack --os-cloud=foo servers list" | 19:16 |
| mordred | and will read the clouds.yaml file to figure out which things you want | 19:16 |
| mordred | but also still just works with OS_ env vars or command line options | 19:16 |
| mordred | if doing that isnt' somethign you're interested in | 19:16 |
| clarkb | gotcha | 19:17 |
| stevemar | mordred, i'm hoping to take a deeper look at that in a weeks time, when all this RC craziness is behind us | 19:18 |
| openstackgerrit | Monty Taylor proposed openstack-infra/system-config: Add a clouds.yaml file in anticipation of shade https://review.openstack.org/168601 | 19:19 |
| mordred | stevemar: ++ | 19:19 |
| mordred | clarkb: I updated the commit message on the base one to indicate it's not _just_ about nodepool | 19:20 |
| *** Sukhdev has quit IRC | 19:20 | |
| mordred | clarkb: (fwiw, I'll also want to install one of these on puppetmaster but with a different set of clouds in ) | 19:20 |
| clarkb | mordred: thanks, looking | 19:20 |
| *** esheffield_ has quit IRC | 19:20 | |
| *** notnownikki has quit IRC | 19:21 | |
| *** SumitNaiksatam has quit IRC | 19:21 | |
| *** SumitNaiksatam has joined #openstack-infra | 19:21 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/system-config: Switch to consuming cloud auth from clouds.yaml https://review.openstack.org/168602 | 19:21 |
| clarkb | mordred: 168602 needs a reparenting now but 168601 helps quite a bit | 19:21 |
| clarkb | beat me to it :) | 19:22 |
| *** dhritishikhar_ has quit IRC | 19:22 | |
| *** SumitNaiksatam has quit IRC | 19:23 | |
| *** Guest65061 has quit IRC | 19:23 | |
| clarkb | mordred: stevemar one suggestion about clouds.yaml. the region_name should be a yaml list rather than csv | 19:23 |
| *** rkukura_ has joined #openstack-infra | 19:24 | |
| *** rkukura has quit IRC | 19:25 | |
| *** rkukura_ is now known as rkukura | 19:25 | |
| clarkb | mordred: stevemar and the expectation is that service type and name and auth url are all known via os-client-config? | 19:26 |
| mordred | clarkb, stevemar: that's a good idea | 19:26 |
| mordred | clarkb: yes. they all are | 19:26 |
| mordred | clarkb: or can be | 19:26 |
| clarkb | cool | 19:26 |
| mordred | they are characteristics of a cloud | 19:26 |
| mordred | clarkb: btw - I'm fixing a review comment from ianw on the nodepool side of adding shade support ... so if you're reviewing it I'm about to push up new patches | 19:27 |
| mordred | (he suggested adding validation for the new config option, which I agree with) | 19:27 |
| clarkb | I haven't gotten that far yet but thanks for the heads up | 19:27 |
| stevemar | mordred, does each cloud entry need a 'cloud' value? | 19:28 |
| mordred | stevemar: in nodepool.yaml or clouds.yaml ? | 19:28 |
| mordred | ah - clouds.yaml - it does not need one, no | 19:28 |
| stevemar | looking at the example in the README | 19:28 |
| mordred | it's an optional shorthand to specify a known cloud | 19:28 |
| stevemar | ah okay | 19:29 |
| stevemar | is there a list of valid values? | 19:29 |
| mordred | so if you're using rackspace, you can say "cloud: rackspace" and you'll get some default values | 19:29 |
| stevemar | gotcha | 19:29 |
| mordred | the list is "hp,rackspace" right now | 19:29 |
| mordred | although I'm going to add dreamhost soon - we should update the README | 19:29 |
| stevemar | makes sense | 19:29 |
| openstackgerrit | Merged openstack-infra/project-config: Create core groups for main fuel-* projects https://review.openstack.org/168182 | 19:29 |
| *** Somay has joined #openstack-infra | 19:29 | |
| mordred | clarkb: oh - btw - I explained in the nodepool series in the commit message that it was optional :) | 19:30 |
| *** cdent has quit IRC | 19:30 | |
| mordred | clarkb: so I apparently knew you needed to know that - but just added it a different commit :) | 19:30 |
| clarkb | nibalizer: can you confirm my comment on 168572? it sa puppetism | 19:30 |
| *** adalbas has quit IRC | 19:31 | |
| * nibalizer looks | 19:31 | |
| *** carl_baldwin has joined #openstack-infra | 19:32 | |
| clarkb | danke | 19:32 |
| openstackgerrit | Matthew Treinish proposed openstack-infra/project-config: Add os-testr repo https://review.openstack.org/167243 | 19:33 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Pass TaskManager in to shade https://review.openstack.org/168663 | 19:35 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Use shade for image uploads https://review.openstack.org/168633 | 19:35 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Use shade to get the client objects https://review.openstack.org/168603 | 19:35 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Use shade API for flavors https://review.openstack.org/168634 | 19:35 |
| clarkb | nibalizer: basically undef like that will override a default right? | 19:35 |
| mordred | clarkb: I fleshed out the discussion of opt-in in the nodepool support for it there | 19:35 |
| nibalizer | clarkb: yea if you pass in undef... wtf happens idunno i think you get to pick up the default | 19:35 |
| nibalizer | anyways im gonna -1 | 19:35 |
| nibalizer | for kinda related reasons | 19:35 |
| nibalizer | more testing needed :P | 19:36 |
| clarkb | mordred: thanks, going to continue reviewing changes after lunch | 19:37 |
| clarkb | but EHUNGRY now | 19:37 |
| *** ivar-lazzaro has joined #openstack-infra | 19:37 | |
| mordred | clarkb: cool - I actually made a nodepool-dib etherpad | 19:37 |
| mordred | clarkb: to help deal with teh current onslaught | 19:37 |
| mordred | clarkb: https://etherpad.openstack.org/p/dib-nodepool-merges | 19:37 |
| mordred | jeblair, fungi, pleia2, jhesketh, SergeyLukjanov: ^^ | 19:37 |
| *** carl_baldwin has quit IRC | 19:39 | |
| annegentle | hey infra mordred jeblair etc, is it generally correct that the big tent is now open? | 19:40 |
| greghaynes | mordred: nice | 19:40 |
| *** rkukura has quit IRC | 19:40 | |
| annegentle | what I mean is, can all projects now use docs.openstack.org for docs, using our known frameworks. | 19:40 |
| annegentle | I can bring up the topic at the TC meeting, but looking for "generally, this is what we all think happens now" | 19:41 |
| *** timcline_ has quit IRC | 19:41 | |
| AJaeger_ | annegentle: right now I only let projects throught that are in the openstack namespace | 19:42 |
| annegentle | AJaeger_: yes, was thinking that's our new line, what is the timing then? "by kilo"? | 19:42 |
| annegentle | AJaeger_: I mean, I think I know, but wanted to vet with others :) | 19:42 |
| AJaeger_ | annegentle: magnum for example was just approved by the TC and therefore can use docs.openstack.org IMHO. | 19:42 |
| annegentle | AJaeger_: yep. | 19:43 |
| annegentle | AJaeger_: same for murano | 19:43 |
| AJaeger_ | yep | 19:43 |
| annegentle | AJaeger_: ok | 19:43 |
| *** claudiupopa has left #openstack-infra | 19:43 | |
| AJaeger_ | I'm not aware of a timeline - if the TC approves, they can use it is my practical motto ;) | 19:43 |
| annegentle | AJaeger_: :) | 19:44 |
| *** otter768 has joined #openstack-infra | 19:45 | |
| *** adalbas has joined #openstack-infra | 19:45 | |
| AJaeger_ | annegentle: yeah, the big tent needs some documentation discussion... | 19:46 |
| annegentle | AJaeger_: I'd like to write up guidelines. We can continue in #openstack-docs. But it will help me think about data-driven tags for docs. | 19:47 |
| annegentle | AJaeger_: like "has CLI Ref, check" "Is in config-ref, check" "is in Install Guide, nope" | 19:47 |
| annegentle | and so on | 19:47 |
| annegentle | seems data driven anyway | 19:47 |
| AJaeger_ | yep | 19:47 |
| AJaeger_ | let's discuss this in #openstack-docs - but timelines sounds like a TC discussion | 19:48 |
| nibalizer | clarkb: so i tested it and you were wrong (but could easily have been right) | 19:49 |
| *** otter768 has quit IRC | 19:49 | |
| clarkb | nibalizer the default is used? | 19:50 |
| clarkb | I dont even follow how that is sane | 19:50 |
| clarkb | we have specifically overridden the default with undef | 19:50 |
| anteaya | annegentle: do you have any stipulations about what projects use docs.o.o? | 19:50 |
| clarkb | byt the default is applied | 19:50 |
| *** mpavone has joined #openstack-infra | 19:50 | |
| nibalizer | !shrug | 19:51 |
| openstack | nibalizer: Error: "shrug" is not a valid command. | 19:51 |
| anteaya | annegentle: in my mind that is your territory and would be telling me what you want to see | 19:51 |
| clarkb | ha | 19:51 |
| crinkle | passing param => undef is the same as not setting the param, so the default is used | 19:51 |
| *** rkukura has joined #openstack-infra | 19:51 | |
| nibalizer | clarkb: fwiw i totes see where you're comming from | 19:51 |
| annegentle | anteaya: what we're moving towards is "in the github namespace qualifies you for docs.openstack.org publishing privileges" I suppose? That's sort of what I want to talk through. I'm envisioning the PTL sets quality bars, etc, but is there any legal consideration I should ask the Foundation about? Hm. | 19:52 |
| nibalizer | this is how you end up passing things like 'EMPTY VALUE' into things | 19:52 |
| mordred | anteaya, annegentle: my understanding is similar to what AJaeger_ said. nothing has changed in terms of supporting stackforge projects, so I would not expect a change in terms of stackforge projects using docs.o.o. what has changed is that the barrier to being in the openstack namespace is much lower | 19:52 |
| clarkb | I mean if I pass None as a kwarg in python I get None | 19:52 |
| clarkb | I dont get $random default | 19:52 |
| AJaeger_ | annegentle: for formal defintion, use "git.openstack" instead of github ;) | 19:52 |
| anteaya | annegentle: I'm assuming you meant in the openstack namespace | 19:53 |
| anteaya | annegentle: my point is that you are the one managing what is on docs.o.o in terms of usage and load | 19:53 |
| crinkle | clarkb: None is different from undef | 19:53 |
| annegentle | AJaeger_:woops, yes of course :) | 19:53 |
| anteaya | annegentle: so if you feel you need any qualifiers to what uses it, I'd like to hear your thoughts | 19:54 |
| annegentle | anteaya: but the infra team has done a fantastic job of "self-service" so docs should follow. | 19:54 |
| anteaya | annegentle: if that is your call, I'm fine with that | 19:54 |
| *** carl_baldwin_ has joined #openstack-infra | 19:54 | |
| AJaeger_ | annegentle: if you want infra to not approve publishing to docs.o.o, please give us a policy/guidelines | 19:55 |
| anteaya | annegentle: my point is I don't think there is a default decision here, it is whatever you say works for you | 19:55 |
| anteaya | annegentle: in the absence of any guidelines from you, any project in the openstack namespace can publish to docs.o.o | 19:56 |
| clarkb | crinkle they are both the absence of value values | 19:56 |
| anteaya | annegentle: but anything from you that overrides that supercedes that | 19:56 |
| annegentle | anteaya: ok, that's helpful, thanks! | 19:57 |
| * AJaeger_ agrees with anteaya | 19:57 | |
| anteaya | annegentle: thanks so if we need to do anything other than what we are doing now, be sure to let us know | 19:58 |
| crinkle | clarkb: you're right, my mistake. puppet just has different behavior than python | 19:58 |
| crinkle | clarkb: they're changing the behavior of undef in puppet 4 since this is obviously confusing | 19:58 |
| *** Swami has quit IRC | 19:59 | |
| *** dprince has quit IRC | 19:59 | |
| clarkb | crinkle thats good. maybe we should respect puppet 4s change for future proofing then? | 19:59 |
| clarkb | nibalizer ^ | 19:59 |
| *** AJaeger_ has quit IRC | 20:00 | |
| *** julim has quit IRC | 20:02 | |
| *** annegentle has quit IRC | 20:04 | |
| *** peristeri has quit IRC | 20:04 | |
| nibalizer | uh, i think thats a bigger conversation kinda? | 20:05 |
| nibalizer | I'd like to get some minimal commitment that we want to upgrade to puppet 4 before choming off a bunch of work to make infra puppet4 go | 20:05 |
| bookwar | AJaeger: hi, thank you for merging this thing https://review.openstack.org/168182 I am worried about how gerrit changes are applied. To not block current development process, while we are populating these new groups, we need fuel-core to be added to all groups which were created by this patch | 20:07 |
| fungi | bookwar: i can do that once puppet finishes updating gerrit to create them | 20:08 |
| anteaya | nibalizer: what are the advantages/disadvantages of moving to puppet 4? | 20:09 |
| mordred | nibalizer: I'd like to get on to puppet apply before we start puppet 4-ing | 20:09 |
| bookwar | fungi: thanks, so i'll wait then | 20:10 |
| mordred | nibalizer: because I think that shoudl make it "easier" due to being able to do it piecemeal, yeah? | 20:10 |
| *** peristeri has joined #openstack-infra | 20:10 | |
| nibalizer | well... first some background, puppet4 isnt out yet | 20:10 |
| nibalizer | however, its sorta bundled with puppet 3.7, add future_parser = true to puppet.conf and you get puppet4 | 20:11 |
| mordred | cool | 20:11 |
| nibalizer | mordred: yes puppetapply will make it easier | 20:11 |
| mordred | that's a nice testing step | 20:11 |
| nibalizer | mordred: yea they learned from last time | 20:11 |
| mordred | ++ | 20:11 |
| nibalizer | they've had that flag available for like a year, so people have been kicking the tires | 20:11 |
| nibalizer | anteaya: some advantages we see above, its a better language with better handling of undef and basic types | 20:12 |
| anteaya | nibalizer: do we need that | 20:12 |
| nibalizer | we can throw away the 'validate_*' crap that we've had to do, it should be faster | 20:12 |
| anteaya | are we blocked in what we do without having that? | 20:12 |
| nibalizer | the disadvantage is that its work to upgrade | 20:12 |
| nibalizer | anteaya: not blocked at all | 20:12 |
| *** mwhudson has quit IRC | 20:12 | |
| nibalizer | really upgrading to puppet4 isn't even on my radar really | 20:13 |
| nibalizer | i have other things i want to fix in infra more | 20:13 |
| anteaya | so yes, remembering the move to puppet 3 was a lot of work, mostly on your shoulders | 20:13 |
| anteaya | nibalizer: so given that, I would be on the side of letting you fix the stuff in infra you think is more important | 20:13 |
| *** carl_baldwin_ is now known as carl_baldwin | 20:13 | |
| *** mwhudson has joined #openstack-infra | 20:14 | |
| *** carl_baldwin has quit IRC | 20:14 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Pass socket timeout to swiftclient https://review.openstack.org/168672 | 20:14 |
| jeblair | mordred: i left feedback in the etherpad -- i'd like to either have a real plan for a build infrastructure, or documentation for the manual steps to use the ppa, or both... | 20:16 |
| jeblair | mordred: i think i'm going to -1 167625 based on that, even though i don't think it is something that can actually be fixed in the patch; does that sound okay? | 20:16 |
| mordred | jeblair: yup! | 20:17 |
| mordred | jeblair: do you think I should upload to an infra-core-available PPA so that any infra-core can upload a new pacakge to it? | 20:17 |
| jeblair | mordred: that would probably be a good idea. it should make the 'document the manual process' easier for now and not be too difficult, right? | 20:18 |
| mordred | yup | 20:18 |
| mordred | jeblair: of of https://launchpad.net/~openstack-ci-core perhaps? | 20:18 |
| mordred | jeblair: also - we shoudl add pleia2 and jhesketh there? | 20:19 |
| jeblair | wow that's dusty | 20:19 |
| jeblair | and yep. | 20:19 |
| mordred | jeblair: where should I document taht? system-config docs? | 20:20 |
| jeblair | mordred: yep, probably nodepool section for now, but i think it would be more fundamental puppet stuff if it becomes required for infra server launching | 20:21 |
| *** timcline has joined #openstack-infra | 20:22 | |
| mordred | jeblair: we can move it if/when it does | 20:22 |
| *** timcline has quit IRC | 20:22 | |
| jeblair | ++ | 20:22 |
| *** timcline has joined #openstack-infra | 20:23 | |
| nibalizer | etherpad where? | 20:23 |
| jeblair | nibalizer: https://etherpad.openstack.org/p/dib-nodepool-merges | 20:23 |
| * clarkb does a happy dance https://jenkins07.openstack.org/job/check-tempest-dsvm-aiopcpu-full/29/console ran succussfully using 64bit guest oses | 20:23 | |
| clarkb | jogo: ^ | 20:23 |
| clarkb | and live migration ran | 20:24 |
| *** annegentle has joined #openstack-infra | 20:25 | |
| *** tiswanso has quit IRC | 20:25 | |
| anteaya | clarkb: congratulations | 20:25 |
| *** dizquierdo has joined #openstack-infra | 20:26 | |
| *** MaxV has joined #openstack-infra | 20:27 | |
| *** samueldmq has quit IRC | 20:27 | |
| *** afazekas has joined #openstack-infra | 20:27 | |
| jeblair | clarkb: see followup comment on 168601 | 20:28 |
| *** bswartz has quit IRC | 20:28 | |
| openstackgerrit | Stephanie Miller proposed openstack-infra/system-config: Add infra setup for zanata https://review.openstack.org/169074 | 20:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/system-config: Document the vhd-util ppa https://review.openstack.org/169075 | 20:31 |
| *** tsg has quit IRC | 20:31 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/puppet-nodepool: Install additional depends for dib https://review.openstack.org/167625 | 20:31 |
| mordred | jeblair: ^^ | 20:31 |
| mordred | clarkb: woot! | 20:32 |
| *** melwitt has quit IRC | 20:33 | |
| *** melwitt_ has joined #openstack-infra | 20:33 | |
| *** afazekas has quit IRC | 20:34 | |
| *** ociuhandu has quit IRC | 20:34 | |
| *** tsg has joined #openstack-infra | 20:35 | |
| fungi | bookwar: i've added fuel-core as an included group of the 10 groups created by that change... fuel-astute-core, fuel-devops-core, fuel-docs-core, fuel-library-core, fuel-main-core, fuel-ostf-core, fuel-plugins-core, fuel-qa-core, fuel-stats-core and fuel-web-core | 20:37 |
| *** afazekas has joined #openstack-infra | 20:37 | |
| bookwar | fungi: exactly, thanks | 20:39 |
| *** mwagner_lap has quit IRC | 20:40 | |
| fungi | you're welcome! | 20:41 |
| *** annegentle has quit IRC | 20:41 | |
| *** carl_baldwin has joined #openstack-infra | 20:42 | |
| *** mrda-away is now known as mrda | 20:43 | |
| clarkb | do you guys see my edits on the etherpad as being in a different font? I have no idea why that is happening | 20:43 |
| *** afazekas has quit IRC | 20:44 | |
| jeblair | clarkb: nope, same font, different color | 20:44 |
| *** dizquierdo has quit IRC | 20:45 | |
| *** e0ne has quit IRC | 20:46 | |
| *** Sukhdev has joined #openstack-infra | 20:46 | |
| mordred | clarkb: maybe it's a portland thing? | 20:47 |
| anteaya | clarkb: I see your edits in the same font as the rest of the etherpad in a lovely hot pink | 20:48 |
| clarkb | could be, my fonts are cool before I even know about it | 20:48 |
| mordred | clarkb: :) | 20:48 |
| fungi | looks like the same font, but with glasses and a handlebar mustache | 20:48 |
| *** Rockyg has quit IRC | 20:48 | |
| anteaya | who is neon turquoise? | 20:50 |
| anteaya | mordred: that you? | 20:50 |
| mordred | anteaya: yeah | 20:50 |
| anteaya | must be you | 20:50 |
| anteaya | nice | 20:50 |
| *** annegentle has joined #openstack-infra | 20:52 | |
| dansmith | clarkb: awesome | 20:54 |
| clarkb | mordred: https://review.openstack.org/#/c/168673/2 has a comment for clarification which you may not see since its over on swiftlcient | 20:54 |
| *** rfolco has quit IRC | 20:54 | |
| *** adalbas has quit IRC | 20:56 | |
| openstackgerrit | Stephanie Miller proposed openstack-infra/system-config: Miscellaneous linter cleanups https://review.openstack.org/169085 | 20:56 |
| *** sarob has joined #openstack-infra | 20:58 | |
| *** signed8bit has joined #openstack-infra | 20:58 | |
| mordred | clarkb: thanks | 20:58 |
| *** yamamoto has joined #openstack-infra | 20:58 | |
| mordred | clarkb: and he's right, right? it is a socket read timeout | 20:58 |
| clarkb | mordred: yup I think so | 20:59 |
| clarkb | mordred: also given the two dib changes you linked, do I need to resurrect my nodepool change to support multiple output formats | 20:59 |
| *** amotoki has joined #openstack-infra | 20:59 | |
| clarkb | mordred: I *think* it should mostly just work since shade is handling uploads and dib will use vhd-util | 20:59 |
| mordred | clarkb: yah - I believe so | 20:59 |
| mordred | clarkb: to both | 20:59 |
| clarkb | ok will clean that up soon | 21:00 |
| *** kgiusti has quit IRC | 21:00 | |
| clarkb | I wasn't sure if nodepool would have to grow the vhd-util stuff internally | 21:00 |
| mordred | clarkb: yeah - thankfully not | 21:00 |
| *** melwitt_ has quit IRC | 21:01 | |
| *** melwitt has joined #openstack-infra | 21:01 | |
| *** dkliban is now known as dkliban_afk | 21:01 | |
| *** melwitt has quit IRC | 21:02 | |
| *** ociuhandu has joined #openstack-infra | 21:02 | |
| *** tsg has quit IRC | 21:04 | |
| *** radez is now known as radez_g0n3 | 21:06 | |
| *** afazekas has joined #openstack-infra | 21:06 | |
| clarkb | mordred: greghaynes question for you on https://review.openstack.org/#/c/156433/ too | 21:07 |
| fungi | mordred: just out of curiosity because i haven't looked very closely at it yet... vhd-util is basically a fork of the xen codebase? | 21:08 |
| *** melwitt has joined #openstack-infra | 21:09 | |
| mordred | fungi: maybe? | 21:09 |
| clarkb | I think its the xen source with one patch over the top | 21:09 |
| mordred | yah | 21:09 |
| clarkb | http://blogs.citrix.com/2012/10/04/convert-a-raw-image-to-xenserver-vhd/ | 21:10 |
| fungi | mordred: is the end-game for that utility to get it shipped as part of xen itself? | 21:10 |
| clarkb | fungi: its already shipped by xen | 21:10 |
| clarkb | fungi: the problem is the patch that is getting applied has not gotten merged. Its not the greatest code | 21:11 |
| fungi | okay, so what we really need is for the distro xen packages to split that utility out? | 21:11 |
| *** mpavone has quit IRC | 21:11 | |
| *** dustins has quit IRC | 21:11 | |
| clarkb | fungi: and we need upstream (whether xen or distro) to either take the patch as is or fix it | 21:11 |
| mordred | greghaynes: ImportError: No module named '__builtin__' :) | 21:11 |
| clarkb | I personally think it would be nice to not use -t 1 and -t2 to describe image types | 21:11 |
| mordred | fungi: yes, split the utility out, and also apply that patch | 21:12 |
| clarkb | or have to manually do two passes for an image conversion | 21:12 |
| mordred | I dont' think we shoudl have to convert our image at all | 21:12 |
| clarkb | that too :) | 21:12 |
| fungi | okay, but there is some endgame where we're (openstack infra) not maintaining some heavyweight fork of the entire xen codebase just to get access to one utility shade needs | 21:12 |
| mordred | I think the glance API should always accept a RAW upload | 21:12 |
| mordred | fungi: I do not believe that I believe in that endgame, but I do think there is at least a theory of that endgame | 21:13 |
| greghaynes | mordred: orly | 21:13 |
| *** isviridov has quit IRC | 21:13 | |
| clarkb | fungi: I don't think anyone is working on that. The blog above is ancient | 21:13 |
| *** sarob has quit IRC | 21:13 | |
| *** mpavone has joined #openstack-infra | 21:13 | |
| mordred | greghaynes: but I'm goign to upload a new version in response to slagle's comment - so don't fix your thing yet | 21:13 |
| *** tsg has joined #openstack-infra | 21:13 | |
| clarkb | greghaynes: yes, glance should do glancy things and make me not care about hypervisor | 21:13 |
| * fungi sighs at... just... everyone and everything in the world | 21:13 | |
| greghaynes | clarkb: hah, sorry, that was re: the no module named builtin | 21:14 |
| clarkb | greghaynes: oh derp | 21:14 |
| jeblair | fungi, clarkb, mordred: i actually think it would be a fun weekend project to reimplement the vhd conversion from scratch in python. or rust. can't be that hard, right? :) | 21:15 |
| clarkb | I know little about image formats but vhd has been nothing but trouble. It would also be awesome if Xen/hyperv supported image formats that were compressed and all that fun stuff | 21:15 |
| greghaynes | clarkb: yea, I think your comment will work about not doing the extra copy, need to test | 21:15 |
| jeblair | otoh, it is spring, so i won't spend my weekends doing that. | 21:15 |
| *** tnovacik has quit IRC | 21:15 | |
| clarkb | greghaynes: my concern there is the -i and -o being the same | 21:16 |
| greghaynes | clarkb: oh, well thats fine | 21:16 |
| clarkb | greghaynes: does vhd-util use a temporary file name then rename when its done? | 21:16 |
| clarkb | greghaynes: otherwise it likely isn't fine | 21:16 |
| *** ldnunes has quit IRC | 21:16 | |
| fungi | jeblair: yeah, seems like it would make a fine lib for glance (or even nova) to load, and include a cli endpoint so it can also be used as a general conversion tool | 21:17 |
| clarkb | especially if in the course of a conversion you can write ahead of your read pointer | 21:17 |
| greghaynes | clarkb: I believe so, which is why the .bak needs to be rm'd | 21:17 |
| greghaynes | clarkb: but that is tested an dworks | 21:17 |
| clarkb | greghaynes: where does intermediate.bak come from? that chunk of code just reads as buggy | 21:18 |
| *** sdake has joined #openstack-infra | 21:18 | |
| *** boris-42 has quit IRC | 21:18 | |
| greghaynes | clarkb: the vhd-util conversion creates it | 21:18 |
| fungi | that blktap2 patch doesn't look terrible. i wonder why xen upstream is allergic | 21:18 |
| clarkb | greghaynes: does it create it if -i and -o are different? if not then I extra vote for different -i and -o | 21:18 |
| *** afazekas has quit IRC | 21:18 | |
| fungi | i mean, more than half the patch is comments | 21:19 |
| marun | neutron is wedged until this merges: https://review.openstack.org/#/c/169092/ | 21:19 |
| mtreinish | jeblair, clarkb, fungi: I was wondering if there was an interest in setting up a submodule repo for all the puppet things | 21:19 |
| marun | can we get it bumped if/when a reset occurs? | 21:19 |
| fungi | some are even usefully commenting the not-yet-commented code which was already there | 21:19 |
| mtreinish | because a common problem I have now is finding things post split | 21:19 |
| clarkb | fungi its incredibly user unfriendly | 21:20 |
| mordred | clarkb, fungi, jeblair: in theory, qemu-image _should_ support doing this but does not | 21:20 |
| mordred | so honestly, if someone were going to hack on things, just fixing qemu-image would be the best chocie | 21:20 |
| *** mpavone has quit IRC | 21:20 | |
| greghaynes | clarkb: so are you worried that theres a potential issue where it works sometimes and not others? Its a pretty obtuse tool and I have no idea what the internals that are going on are | 21:20 |
| clarkb | mtreinish we have a module installer in system-config | 21:20 |
| jeblair | marun: did the devstack up/down job not catch that? is that fixable? | 21:20 |
| greghaynes | that setup does work for booting rax images though | 21:21 |
| fungi | clarkb: oh, so it's less about code quality and more that it's just a terrible idea to implement it that way? | 21:21 |
| clarkb | greghaynes yes exactly and the dib change reads as wrong | 21:21 |
| clarkb | fungi yes :) | 21:21 |
| mordred | clarkb, greghaynes: also, the raw to first .vhd happens in place - the second conversion makes a backup file :) | 21:21 |
| *** SumitNaiksatam has joined #openstack-infra | 21:21 | |
| *** annegentle has quit IRC | 21:21 | |
| marun | jeblair: it's only the neutron functional job that's broken, and there is nothing stopping ds from merging breaking fixes | 21:21 |
| greghaynes | ok, so if it happens in place why is there a -o at all? | 21:21 |
| greghaynes | clarkb: what reads as wrong | 21:21 |
| mordred | greghaynes: it does the move | 21:21 |
| *** sdake_ has quit IRC | 21:21 | |
| clarkb | fungi: see current discussion about inputs and outputs and backup files | 21:21 |
| marun | jeblair: sdague kindly tagged a release that we could pin to but I hadn't done so yet. will make it a priority now | 21:21 |
| mordred | greghaynes: I think | 21:22 |
| mordred | greghaynes: it's a terrible tool | 21:22 |
| clarkb | fungi: also the -s and -t flags take integer args which is lolwut | 21:22 |
| jeblair | marun, dtroyer: yeah, but i thought there was a job on devstack to test the parts used by functional tests | 21:22 |
| marun | sdague: could you add a kilo3 tag? | 21:22 |
| mordred | clarkb: it's amazingly bad | 21:22 |
| jeblair | marun: my concern is that seems not to have worked | 21:22 |
| clarkb | greghaynes: deleting a file you never create (the .bak), using an input and output that are the same | 21:22 |
| jeblair | marun: tagged a release of what? | 21:22 |
| clarkb | greghaynes: those are both smells in shell scripts | 21:22 |
| marun | jeblair: they might test the parts that are used, but in this case it was a 'source lib/neutron' that broke | 21:22 |
| fungi | clarkb: agreed | 21:22 |
| marun | jeblair: The intent was to tag a milestone release of devstack we could pin to | 21:23 |
| mtreinish | clarkb: it was more for grepping and keeping things up to date all at once, not for necessarily running things | 21:23 |
| jeblair | marun: how would you pin to it? i don't think that does (or should) work... | 21:23 |
| clarkb | mtreinish: it should do that | 21:23 |
| marun | jeblair: we'd checkout the tag in the configure script before we sourced anything | 21:24 |
| greghaynes | clarkb: so you want me to use different outputs and not rm the .bak? | 21:24 |
| jeblair | marun: please don't do that, you will defeat cross-project testing | 21:24 |
| marun | jeblair: ?? | 21:24 |
| greghaynes | sounds like the tool does an in place modify and then maybe moves | 21:24 |
| clarkb | greghaynes: if possible. If the .bak is created even when inputs and outputs are different then maybe we need a comment to explain the sillyness | 21:24 |
| greghaynes | so its all the same | 21:24 |
| marun | jeblair: we're only talking about pinning devstack for a script that installs os deps | 21:24 |
| greghaynes | sure | 21:24 |
| marun | jeblair: it doesn't have anything to do with cross-project | 21:24 |
| marun | https://github.com/openstack/neutron/blob/master/tools/configure_for_func_testing.sh | 21:25 |
| jeblair | marun: you're using devstack to set all of this up though, right? | 21:25 |
| marun | jeblair: I'm using devstack functions, yes. | 21:25 |
| marun | jeblair: I don't invoke stack.sh | 21:25 |
| mtreinish | clarkb: sure, but it's different than my normal git workflow which is kinda my issue | 21:26 |
| clarkb | greghaynes: I prefer code to be readable without knowing silly magic in broken tools | 21:26 |
| clarkb | least surprise etc | 21:26 |
| greghaynes | clarkb: sure, unfortunately were using a magic black box of a tool though | 21:26 |
| clarkb | ya :/ | 21:26 |
| mordred | greghaynes: also - I'm out of that dib change if you want the lock back to work on that test | 21:26 |
| jeblair | marun: so what should have happened is that the change to devstack should not have merged because it broke a job that depended on devstack. i believe the best way to fix that is to ensure devstack has adequate testing of this... | 21:27 |
| anteaya | clarkb: I'm curious what patch is under discussion, I've looked through the backscroll and if it is there I haven't seen it yet | 21:27 |
| greghaynes | mordred: the vhd or config drive process one? | 21:27 |
| greghaynes | or both | 21:27 |
| *** ibiris is now known as ibiris_away | 21:28 | |
| clarkb | anteaya: https://review.openstack.org/#/c/156433/10 that one | 21:28 |
| anteaya | thanks | 21:28 |
| marun | jeblair: testing of what? the fix is s+lib/neutron+lib/neutron-legacy+ | 21:28 |
| clarkb | mtreinish: typically we have resisted submodules because people trip on them always | 21:28 |
| anteaya | this created the fungi sigh, a memorable event | 21:28 |
| mordred | greghaynes: the config drive one | 21:28 |
| mordred | greghaynes: both actuall - I'm done mucking with both | 21:28 |
| mordred | greghaynes: whichever one is broken | 21:28 |
| greghaynes | ok, I think both ;) | 21:28 |
| marun | jeblair: how do we track something like the name of an import across everyone that imports it? | 21:28 |
| jeblair | marun: part of devstack's api is now "lib/neutron" -- it broke that api because no jobs running on devstack test that. i think that should be corrected. | 21:28 |
| mordred | hehe | 21:29 |
| marun | jeblair: I guess I'm not sure that the devstack folks agreed to keep it stable. | 21:29 |
| jeblair | marun: eek. | 21:29 |
| mordred | can we reverse that trend? | 21:29 |
| mordred | I'm a little tired of people thinking they can opt in to supporting an api | 21:29 |
| marun | jeblair: I asked to have a setup-only neutron job added to the devstack gate pipeline to prevent this, but that was deemed unacceptable. | 21:29 |
| mordred | if it's a detail, mark it that way | 21:29 |
| mordred | in python, we use leading _'s for those things | 21:30 |
| *** boris-42 has joined #openstack-infra | 21:30 | |
| jeblair | marun: i thought there was one | 21:30 |
| marun | jeblair: the compromise was to tag a stable devstack release and pin | 21:30 |
| marun | jeblair: and it's my fault for not pinning | 21:30 |
| marun | jeblair: a pin isn't intended to be a get-out-of-jail-free card. | 21:30 |
| marun | jeblair: it would be my responsibility to keep current | 21:30 |
| clarkb | mtreinish: we do use them for our gerrit repo beause that is how gerrit does plugins and for the submodule track openstack openstack/openstack repo | 21:30 |
| jeblair | marun: i don't think projects should have pins on devstack | 21:30 |
| clarkb | mtreinish: so in theory its possible but I think in general we would try to avoid it | 21:31 |
| marun | jeblair: and fix breakage as it occurs. but at least that wouldn't block other developers | 21:31 |
| marun | jeblair: What are you concerns with a pinning-based strategy? | 21:31 |
| mtreinish | clarkb: yeah openstack/openstack was what I was thinking of | 21:31 |
| jeblair | marun: because _one_ job will have neutron set up in a certain way, and other jobs will have it set up in a different way | 21:31 |
| marun | jeblair: our use of devstack doesn't suggest any kind of stable api | 21:31 |
| marun | jeblair: it's not neutron that' being set up, though. | 21:32 |
| jeblair | marun: i don't see how you can evolve what devstack does to set up neutron for integration testing separately from what it does for functional testing | 21:32 |
| marun | jeblair: the functional job manages its own fixture | 21:32 |
| mtreinish | clarkb: sure I can understand that, submodules can get really weird. I just miss the one big repo for everything, it worked well for my infrequent puppet contribs | 21:32 |
| marun | jeblair: it's only os deps that the script installs | 21:32 |
| jeblair | marun: what's the 'source lib/neutron' for then? | 21:32 |
| marun | jeblair: to allow the installation of agent os deps | 21:32 |
| *** mattfarina has quit IRC | 21:32 | |
| marun | jeblair: that's it | 21:32 |
| marun | https://github.com/openstack/neutron/blob/master/tools/configure_for_func_testing.sh#L166 | 21:33 |
| jeblair | marun: shouldn't those still evolve together? | 21:33 |
| jeblair | marun: if you add/change an os dep for integration testing, shouldn't that also be reflected and tested with functional testing? | 21:33 |
| marun | jeblair: sure, but the addition that tries to add the os dep can be responsible for updating the pin | 21:33 |
| *** dkranz has quit IRC | 21:34 | |
| *** yamahata has quit IRC | 21:34 | |
| *** carl_baldwin has quit IRC | 21:34 | |
| jeblair | marun: how? it's a change to devstack, right? | 21:34 |
| marun | jeblair: we shouldn't be merging code without functional tests, and if that requires devstack updates, then the code depending on a newer devstack can't merge until the pin gets updated once devstack has thechange | 21:34 |
| jeblair | marun: meanwhile, the integration tests are using the new thing while the functional tests are not. | 21:35 |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Add query for ironic sideways bug 1438415 https://review.openstack.org/169099 | 21:35 |
| openstack | bug 1438415 in devstack "Some network tests failing with IndexError" [Undecided,In progress] https://launchpad.net/bugs/1438415 | 21:35 |
| marun | jeblair: no | 21:35 |
| marun | jeblair: er | 21:35 |
| marun | jeblair: we're talking os deps here... | 21:35 |
| marun | jeblair: So the only change would be 'I'm adding something new' | 21:35 |
| *** jamielennox|away is now known as jamielennox | 21:36 | |
| marun | jeblair: and nothing would be using that new thing until a change to neutron merged | 21:36 |
| marun | jeblair: if we were talking python dep that would be different - the versions might change | 21:36 |
| *** annegentle has joined #openstack-infra | 21:36 | |
| clarkb | greghaynes: I updated my comment on the vhd-util dib change, does that make sense? | 21:36 |
| marun | jeblair: but for a given distro the dep version is going to be common across all installs | 21:37 |
| clarkb | greghaynes: unfortunately vhd-util doesn't seem to have a manpage or other documentation I can refer to | 21:37 |
| greghaynes | Yep, I am very aware of that fact ;) | 21:38 |
| jeblair | marun: i guess i'm not as certain that the only changes that will ever affect neutron functional tests in devstack are os dependency additions. surely there's other stuff going on, right? configurations of certain things? if os deps is really all it is, why use devstack at all? | 21:38 |
| greghaynes | It also wouldnt apply since were using a patched version of it | 21:38 |
| greghaynes | basically, this is "we found an incantation that seems to work so were doing it" | 21:38 |
| marun | jeblair: service configuration could potentially change | 21:38 |
| marun | jeblair: mysql/pg and rabbit custom and distro-specific fixups | 21:38 |
| jeblair | marun: wouldn't it be good for that to stay in sync between functional and integration tests? | 21:39 |
| marun | jeblair: I don't think it's as critical as not blocking developers | 21:39 |
| marun | jeblair: the fact that mysql or rabbit behave a bit differently is unlikely to break tests | 21:39 |
| jeblair | marun: if you are saying that my suggestion is to block developers, we have a big communication problem here | 21:39 |
| jeblair | marun: you _are_ blocked now, and i'm trying to help you fix it. | 21:39 |
| marun | jeblair: I guess I'm not seeing a good alternative to pinning that doesn't involve more work for little gain | 21:40 |
| clarkb | greghaynes: magic | 21:40 |
| marun | jeblair: in an ideal world we have devstack-lib that knows how to install things in distro-specific and openstack-specific ways | 21:40 |
| jeblair | marun: simple -- we fix the devstack job that is _supposed_ to test this so that it does, and if it turns out it isn't really supposed to test it, then we add the job you suggested earlier. | 21:40 |
| jeblair | marun: because we need to accept that if devstack is going to be used by functional tests, it needs to have a tested api | 21:41 |
| jeblair | dtroyer, mtreinish: so what can we do to the devstack up/down job to make sure this part of devstack's api is tested in the future? | 21:41 |
| mtreinish | jeblair: there is a test job which does stack than unstack | 21:42 |
| jeblair | mtreinish: so it doesn't actually use the part of the devstack api that a functional test job might use? | 21:42 |
| mtreinish | jeblair: http://logs.openstack.org/07/168407/7/check/check-devstack-dsvm-updown/385a90c/console.html | 21:43 |
| dtroyer | jeblair: the project include file are not part of any imagined API that DevStack may or may not have. The only thing that has 'permission' to use them directly is Grenade and those two pojects often have to have coordinated changes to make things work. | 21:43 |
| *** timcline has quit IRC | 21:43 | |
| greghaynes | clarkb: also, this just makes me want to make jobs to test all the image creation formats more | 21:43 |
| greghaynes | clarkb: is adding a custom ppa for a job like that doable? | 21:44 |
| jeblair | marun: i read that as dtroyer saying that functional test jobs should not use those libs | 21:44 |
| marun | jeblair: hah | 21:44 |
| marun | jeblair: well, I could stop using devstack and roll my own | 21:44 |
| dtroyer | what marun is doing is playing with sharp objects and he got cut. if there needs to be some devstack-lib-like api, then we need to define those, until then I do not accept that anything in devstack is open for other use just because it exists | 21:44 |
| marun | jeblair: or I could use that time to do useful work and pin | 21:44 |
| jeblair | marun: let's actually discuss that seriously for a minute | 21:44 |
| jeblair | marun: what are you actually using from devstack? os dependencies i got, anything else? | 21:45 |
| dtroyer | if installing os deps beyond what we pulled out into tools is required, then that needs to be defined and pulled out somewhere | 21:45 |
| marun | jeblair: no. os deps is all | 21:45 |
| *** mrmartin has quit IRC | 21:45 | |
| *** ajmiller_ has joined #openstack-infra | 21:45 | |
| marun | jeblair: not just package installation, but post-installation fixup, sometimes distro-specific | 21:45 |
| *** mriedem is now known as mriedem_away | 21:45 | |
| dtroyer | that's one reason I keep moving things into tools. those are available | 21:45 |
| mordred | I believe the bindep work should help this, no? | 21:45 |
| jeblair | mordred: yeah, that's where i'm going with this | 21:45 |
| mordred | jeblair: sorry, I'll butt out :) | 21:46 |
| jeblair | mordred: but trying to figure out if it's feasible | 21:46 |
| *** otter768 has joined #openstack-infra | 21:46 | |
| jeblair | mordred: no problem, please stay butted in :) | 21:46 |
| *** jamielennox is now known as jamielennox|away | 21:46 | |
| jeblair | marun: what about mysql installation/config/etc? or rabbit? or any other service startup and configuration? | 21:46 |
| marun | jeblair: we do the same thing devstack does. package install, distro and openstack-specific post-install fixup/config | 21:46 |
| marun | jeblair: it's all there in the script | 21:47 |
| marun | rpc: https://github.com/openstack/neutron/blob/master/tools/configure_for_func_testing.sh#L96 | 21:47 |
| marun | db: https://github.com/openstack/neutron/blob/master/tools/configure_for_func_testing.sh#L111 | 21:47 |
| *** MaxV has quit IRC | 21:47 | |
| marun | base deps: https://github.com/openstack/neutron/blob/master/tools/configure_for_func_testing.sh#L82 | 21:47 |
| marun | agent deps: https://github.com/openstack/neutron/blob/master/tools/configure_for_func_testing.sh#L163 | 21:47 |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Add query for ironic sideways bug 1438415 https://review.openstack.org/169099 | 21:48 |
| openstack | bug 1438415 in devstack "Some network tests failing with IndexError" [Undecided,In progress] https://launchpad.net/bugs/1438415 | 21:48 |
| marun | it would be great to simply have a function call for each of these | 21:48 |
| marun | and have the resulting api supported | 21:48 |
| *** ajmiller has quit IRC | 21:48 | |
| jeblair | marun, dtroyer: it looks like there's some copying from devstack into there, but also invocation of other things from devstack/lib/* (not just neutron). is that correct? | 21:49 |
| dtroyer | there is, and I didn't know it was going that far. I would suggest that is dangerous… | 21:49 |
| marun | jeblair: correct. we need to install db and rpc | 21:49 |
| jeblair | dtroyer, marun: so could devstack evolve an API that does things like "install_and_setup_db" ...? and would that be useful? | 21:50 |
| *** julim has joined #openstack-infra | 21:50 | |
| marun | jeblair: absolutely. none of what I'm doing couldn't be done better in devstack. | 21:50 |
| dtroyer | it could. I'm accepting gifts of days with which to do so… | 21:51 |
| *** otter768 has quit IRC | 21:51 | |
| * anteaya hands dtroyer a Friday | 21:51 | |
| dtroyer | although it'll be pretty low on the priority list, ie we didn't even talk about this last week when sorting out the priorities | 21:51 |
| marun | which is why I went ahead and just did it | 21:52 |
| marun | and I'm still of the opinion that pinning is the best way to do this until more projects need the capability | 21:52 |
| marun | neutron is it for now as far as I know | 21:52 |
| jeblair | marun, dtroyer: so how about we start to define that api, and we set up a job that makes sure it is all tested? perhaps what's in neutron could be moved into devstack until the neutron func setup script is just a series of devstack api invocations? | 21:52 |
| *** jamielennox|away is now known as jamielennox | 21:53 | |
| jeblair | marun, mordred: to finish out the thought from ealier, because of the setup involved, i don't think bindep is the answer here; it would be a good one if we only cared about installing deps, but there's actual configuration going on which i think everyone at least agrees ideally should be in devstack. | 21:54 |
| dtroyer | I'll think about it after venvs are implemented and backported and a sensible neutron config is running as our default. at the current rate we'll talk about it in tokyo | 21:54 |
| mordred | jeblair: nod | 21:55 |
| *** aysyd has quit IRC | 21:55 | |
| marun | jeblair: I'd be happy to have things simpler. It doesn't sound like there are the cycles ot make it happen soon, though. | 21:56 |
| *** rbradfor has joined #openstack-infra | 21:56 | |
| jeblair | marun: okay. so under the circumstances, i think either pinning or just copying the bits of devstack you need into your repo are the options on the table. devstack has no public api for what you want to do, so testing that wouldn't seem to be the solution. | 21:56 |
| openstackgerrit | Aaron Rosen proposed openstack-infra/devstack-gate: Allow SERVICE_TIMEOUT to be configured https://review.openstack.org/169103 | 21:56 |
| jhesketh | Morning | 21:57 |
| jeblair | marun: honestly, i figure you've done much of the work already, so i'm happy to try to volunteer you for shaping at least part of that api, but it doesn't sound like there'd be a lot of bandwidth on the receiving side, so i won't push that. :) | 21:57 |
| anteaya | morning jhesketh | 21:57 |
| marun | jeblair: I'd be happy to help when the cycles are available. | 21:57 |
| jeblair | marun, dtroyer: our thought from a while back was "let's see a project implement functional testing and use that as a template for the rest" | 21:57 |
| jeblair | marun, dtroyer: unfortunately, it seems like what we ended up with is not a good template | 21:58 |
| jeblair | so i feel that we should not accept new functional jobs that involve devstack bits | 21:58 |
| marun | jeblair: :/ | 21:58 |
| marun | jeblair: I don't think that's a good solution | 21:58 |
| jeblair | i don't either | 21:59 |
| anteaya | jeblair: oh, how should folks set up functional jobs? | 21:59 |
| anteaya | what should we give them as guidance? | 21:59 |
| clarkb | I'm a fan of the zuul/nodepool functional teting model though it probably doesn't map onto testing things with many system requirements very well | 22:00 |
| jeblair | anteaya: i think the thing we have discovered is that functional jobs can't use certain parts of devstack. | 22:00 |
| clarkb | (they basically set everything up in test) | 22:00 |
| jeblair | anteaya: and it will be hard for us to determine whether a given job does so or not | 22:00 |
| marun | jeblair: I don't see why what we're doing now isn't viable with pinning | 22:00 |
| anteaya | jeblair: yes, I am reading that | 22:00 |
| marun | jeblair: until an api comes along | 22:00 |
| anteaya | jeblair: ah | 22:00 |
| marun | jeblair: I want to get stuff done, not block until we have some ideal | 22:00 |
| jeblair | marun: i don't think i'm suggesting anything that will inhibit your work? | 22:01 |
| jeblair | marun: i'm suggesting that the _next_ functional job that comes along should take what we've learned from this attempt and come up with a better one | 22:01 |
| marun | jeblair: fair enough | 22:02 |
| *** unicell has quit IRC | 22:02 | |
| *** unicell has joined #openstack-infra | 22:02 | |
| marun | jeblair: I think fungi was actually working towards similar ends in his work to allow unit tests to run on devstack nodes | 22:02 |
| anteaya | marun: yes I too am reading that a _new_ functional job needs more scrutinty | 22:02 |
| jeblair | marun: yeah, that's the bindep thing i was talking about earlier | 22:02 |
| *** nelsnelson has quit IRC | 22:02 | |
| marun | jeblair: yeah, but as you say, bindep doesn't manage distro-specific fixup or openstack-specific post-install config | 22:03 |
| *** YorikSar has quit IRC | 22:03 | |
| jeblair | yep | 22:03 |
| marun | jeblair: I'd be happy to see it take that on, frankly. | 22:03 |
| marun | jeblair: devstack could orchestrate and bindep could manage package install | 22:03 |
| jeblair | marun: oh yeah, we're definitely doing that :) | 22:03 |
| * fungi is working on it as we speak in fact. cross-checking package sets on held workers of divverent types | 22:03 | |
| *** hashar has quit IRC | 22:04 | |
| fungi | trying to make sure https://review.openstack.org/164447 covers all our current bases | 22:04 |
| anteaya | jeblair fungi clarkb sorry to change the topic but on my action item to come up with a post to the infra list about election tooling, I'm inclined to delay until after this election season is over | 22:04 |
| anteaya | so about 4 weeks, so as not to add to the confusion that I initially was trying to avoid | 22:04 |
| marun | fungi: any thoughts about expanding the scope of bindep? | 22:05 |
| anteaya | thoughts? | 22:05 |
| jeblair | anteaya: i don't see a huge reason not to start the thought process now, but i'm happy either way. :) | 22:05 |
| fungi | marun: er, depends on what scope expansion you're considering. it's aimed at being a simple tool to parse a list of packages and let you know what you're missing | 22:05 |
| fungi | marun: with the expectation that you can feed the result into apt-get install or yum install or whatever | 22:06 |
| anteaya | jeblair: you don't think a thread on infra with a topic of elections might confuse folks? | 22:06 |
| jeblair | anteaya: only usefully :) | 22:07 |
| anteaya | jeblair: ha ha ha | 22:07 |
| anteaya | well here's hoping | 22:07 |
| marun | fungi: :/ | 22:08 |
| jeblair | marun: to go waaay back to your original question -- i don't think we need to promote that change because it only affects neutron functional jobs which are only on the neutron repo, so any job behind that change in the gate queue will use it anyway. if you want validate jobs in check with that in place without waiting for it to merge, you can set "Depends-On: I5eb4b4052da4b0db128feb42feae50a8bc59f373" | 22:08 |
| fungi | jeblair: marun: also it's already in the gate with only two changes ahead of it in its shared change queue anyway | 22:09 |
| jeblair | and that | 22:09 |
| marun | yeah, looking good | 22:09 |
| jeblair | dtroyer: is there someplace where devstack's "api" is described? if not, it would probably be good to have that in writing, particularly what you think is acceptable for a functional test job to use. | 22:10 |
| *** carl_baldwin has joined #openstack-infra | 22:10 | |
| *** amotoki has quit IRC | 22:10 | |
| marun | fungi: so devstack/shell scripting/puppet still get to do post-install stuff... | 22:11 |
| *** miqui has quit IRC | 22:11 | |
| marun | fungi: I think the duplication of effort and lack of consumability of this status quo is unfortunate. | 22:11 |
| *** paul-- has quit IRC | 22:11 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Migrate API calls to task management https://review.openstack.org/168665 | 22:11 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Add task management framework https://review.openstack.org/168664 | 22:11 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Fix a use of in where it should be equality https://review.openstack.org/169105 | 22:11 |
| *** paul-- has joined #openstack-infra | 22:12 | |
| dtroyer | jeblair: not really… other than Grenade, the only things meant for external use is functions* (they are supposed to stand alone) and tools/* | 22:12 |
| marun | which disallows installation of rpc db | 22:12 |
| dtroyer | add inc/* as we're breaking up functions-common and adding tests | 22:12 |
| marun | and anything service-specific | 22:13 |
| fungi | marun: well, i'm not opposed to a tool scoped to perform configuration of things, but that sounds like something better decomposed out of devstack. i picked up bindep because it already existed, just needed a bit of fleshing out, but i think it's trying to be a traditional unix "do one thing and do it well" sort of utility | 22:13 |
| jeblair | dtroyer: can you write that down somewhere? | 22:13 |
| dtroyer | things like the rpc and databse that marun needs need to be pulled out somehow if they are to be generally used, again, other than Grenade | 22:13 |
| dtroyer | jeblair: I'll add it to the list | 22:13 |
| jeblair | marun: i agree, but i think we need to address that by resetting expectations around the proper tools. bindep isn't the place for that, but devstack or something like it is. in fact, a benefit of keeping bindep simple is that we can use it in devstack, unit test jobs, and functional test jobs. all of which need "some packages installed". | 22:13 |
| marun | jeblair: the hard part is not installing a list of deps | 22:14 |
| jeblair | on top of that, _some_ (but not all) of those need configuration done. | 22:14 |
| jeblair | marun: i agree ^ that is the hard part | 22:14 |
| *** dboik_ has joined #openstack-infra | 22:15 | |
| anteaya | jeblair fungi clarkb tristanC mordred anyone else who cares about elections: https://etherpad.openstack.org/p/XU4qRouXIH | 22:15 |
| jeblair | marun: and i don't have a solution to that right now other than devstack. so we either need to work with dtroyer to evolve it there (once he gets over the shock we're putting him through now, and we have time to plan this out properly) or decide that a different tool would be better. and hey, maybe devstack uses that. | 22:15 |
| anteaya | note not pinging pleia2 as she is away this week | 22:15 |
| *** asselin has joined #openstack-infra | 22:15 | |
| *** yamamoto has quit IRC | 22:15 | |
| anteaya | mordred: I recognize your colour now | 22:16 |
| marun | jeblair: So long as we can run functional tests I'm not wedded to how we configure the environment. | 22:16 |
| fungi | jeblair: marun: for example, something to take on the tasks performed by the .*sql-{db,prep} macros at http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/macros.yaml#n439 would be helpful | 22:16 |
| marun | jeblair: I don't like the idea of stalling testing effort in other projects because 'devstack isn't ready' though - by that logic we shouldn't have done what we've done. | 22:16 |
| fungi | just looking at things from a decomposability perspective | 22:16 |
| marun | jeblair: and I think making progress > doing things 'right' | 22:17 |
| marun | jeblair: we can always fix things that break. I'd rather risk breakage than stall. | 22:17 |
| fungi | though sometimes "making progress" really means "accumulating technical debt we can never hope to pay back" | 22:17 |
| *** dboik has quit IRC | 22:18 | |
| marun | fungi: we have to consider that possiblity, but I don't think it makes sense to consider things in that light by default. | 22:18 |
| jeblair | marun: i understand, but i also don't think that a bunch of other projects asking that devstack tag a new release with a new dep updated, but please also don't break anything that might make it harder to release a new tag for that other project later is actually a scalable solution | 22:18 |
| marun | fungi: it's always risk vs reward | 22:18 |
| anteaya | fungi: ++ | 22:18 |
| marun | jeblair: we don't need a tag, though | 22:18 |
| *** sabeen1 has quit IRC | 22:18 | |
| *** amitgandhinz has quit IRC | 22:18 | |
| marun | jeblair: We can always just pin to something arbitrary | 22:18 |
| marun | jeblair: if that's a concern, maybe that's best. | 22:19 |
| *** annegentle has quit IRC | 22:19 | |
| marun | jeblair: we're already running into limitations around our current practice of global deps, no pinning | 22:19 |
| *** dboik_ has quit IRC | 22:19 | |
| *** YorikSar has joined #openstack-infra | 22:19 | |
| marun | jeblair: I get the motivation, but one-size-fits-all breaks down at the scale we're seeing. | 22:20 |
| jeblair | marun: i don't think having projects say they need to use another project at this particular random git sha is a solution either. that's well below the level of engineering this project is capable of. | 22:20 |
| *** asselin has quit IRC | 22:20 | |
| jeblair | marun: no, this is not a breakdown of anything except thet fact that a proper inter-project api was not defined | 22:20 |
| marun | jeblair: I think disallowing a viable short-term approach on principle is self-defeating. | 22:20 |
| jeblair | but used regardless | 22:20 |
| marun | jeblair: we need to do what we need to do | 22:20 |
| *** gokrokve has quit IRC | 22:21 | |
| marun | jeblair: so long as the end-goal is quality software that evolves to meet user demand, what, really, should be off the table? | 22:21 |
| jeblair | marun: establishing an API when two modules depend on each other is pretty fundamental software engineering. it needs to be no more complicated than that. | 22:21 |
| marun | jeblair: it takes time | 22:21 |
| marun | jeblair: and the pain during that time is real | 22:21 |
| marun | jeblair: again, why deny usage of a short-term fix to make that pain more bearable? | 22:22 |
| marun | jeblair: nobody is saying we want to do the wrong thing forever. | 22:22 |
| fungi | the term "throwing good money after bad" comes to mind. development effort spent implementing something you already have plans to rewrite | 22:23 |
| marun | jeblair: honestly, why is this something that infra needs or wants to be involved in? The projects are responsible for converging on sanity every milestone. | 22:23 |
| marun | fungi: are you working on the project you're trying to dictate terms to? | 22:23 |
| fungi | marun: i don't personally have any inclination to work on that. but by the same token you shouldn't ask for my help implementing it | 22:24 |
| marun | fungi: I'm not. | 22:24 |
| * fungi is speaking figuratively, on the assumption you're doing the same | 22:24 | |
| jeblair | i think this would be a good time to end this conversation. | 22:24 |
| *** tsg has quit IRC | 22:24 | |
| marun | fungi: I'm asking to not be hamstrung by seemingly arbitrary decisions based more on principle than pragmatism. | 22:24 |
| dtroyer | marun: you are more than welcome to help define and build those interfaces. I have less time that you do available for this, if you can't wait, then run along ahead and don't say we're using quality practices here. when we talked about using devstack bits as a library, I made the apparently bad assumption that we would actually define the interface and not ust use whatever could be found | 22:24 |
| mordred | marun: I do not think you're being hamstrung by anything | 22:25 |
| mordred | marun: you came in here asking fora human override on a system | 22:25 |
| marun | mordred: I think we'll have to disagree, then. | 22:25 |
| mordred | taht is a cost | 22:25 |
| mordred | so we're going to ask to have that cost be justified | 22:25 |
| mordred | you did not end up doing that | 22:25 |
| mordred | you are still free to do whatever you feel like in neutron | 22:25 |
| mordred | but if you want us to pefrom work, then you're going to have to convince us that the work is a good idea | 22:26 |
| marun | mordred: Are you suggesting that the functional testing we're doing doesn't have value, then? | 22:26 |
| mordred | no | 22:26 |
| mordred | I am not | 22:26 |
| mordred | I am suggesting that us bumping a change in the queue is the work you asked for | 22:26 |
| marun | mordred: Then why do you think that the cost here isn't justified? | 22:26 |
| mordred | outside of that, I believe we're offering suggestions that we think that doing a pin of something in a neutorn repo is a bad idea - but none of us can stop you or are interested in doing so | 22:27 |
| *** YorikSar has quit IRC | 22:27 | |
| marun | dtroyer: I'm not against defining an API. It doesn't sound like helping me do so is a priority for you right now, though. | 22:27 |
| mordred | however, if it does break and causes more carnage and you need to come and ask for another bump of a change through a gate queue | 22:27 |
| mordred | you're probably not going to get much sympathy | 22:27 |
| marun | mordred: I find it troubling that I need sympathy to do my job. | 22:28 |
| mordred | you will need sympathy if you want to convince us to bypass the system that runs the automated testing for you | 22:28 |
| mordred | there is a reason we don't do that lightly | 22:28 |
| mordred | but, luckily, the system keeps chugging | 22:29 |
| mordred | and your change will land when it lands, like all changes do, and you will be happy, and the world will be good | 22:29 |
| *** tsg has joined #openstack-infra | 22:31 | |
| *** YorikSar has joined #openstack-infra | 22:33 | |
| mtreinish | clarkb, jeblair, fungi, mordred: would you be ok with me stealing subunit2html from the slave scripts and packaging that in os-testr? | 22:33 |
| anteaya | in case this was missed, my first draft of an email to the infra list about discussion election tooling: https://etherpad.openstack.org/p/XU4qRouXIH | 22:34 |
| openstackgerrit | Clark Boylan proposed openstack-infra/nodepool: Support multiple image formats in a diskimage https://review.openstack.org/126747 | 22:34 |
| *** pc_m has quit IRC | 22:34 | |
| anteaya | discussing | 22:34 |
| clarkb | mtreinish: yes please | 22:34 |
| mordred | mtreinish: fine by me | 22:34 |
| jeblair | mtreinish: i'm pretty sure the answer is yes. | 22:34 |
| clarkb | mordred: ^ 126747 should now be ready for rebasing and stuff | 22:34 |
| fungi | mtreinish: seems fine, though we should figure out how to stop using the copy in the slave scripts | 22:34 |
| mordred | clarkb: woot! | 22:34 |
| clarkb | mordred: not sure what order you want that to go in with shade image uploads | 22:35 |
| jeblair | mtreinish, fungi: will we need to install os-testr on nodes like we do zuul? | 22:35 |
| mordred | clarkb: I'll poke at it later tonight and try to add it into the stack | 22:35 |
| clarkb | mordred: I even added a test though in my haste did not run it locally >_> | 22:35 |
| mtreinish | cool, then I'll start doing that | 22:35 |
| jeblair | mtreinish, fungi: or will it be a tox dep? | 22:35 |
| mordred | jeblair: I _think_ it goes into teh venvs with tox, yeah? | 22:35 |
| mtreinish | jeblair: not yet especially since I haven't pushed a release to pypi yet :) | 22:35 |
| *** claudiub has quit IRC | 22:35 | |
| mtreinish | I was waiting for it to be in gerrit before I did that | 22:35 |
| mordred | mtreinish: but the theory is that it will go into tox/test-requirements.txt yeah? | 22:35 |
| fungi | becomes a tox dep preferably? because preinstalling things in our images is something we should strive to avoid where possible | 22:36 |
| mtreinish | mordred: yeah I was going to add it as the test requirement everywhere instead of testrepository, etc. | 22:36 |
| jeblair | mtreinish, fungi, mordred: yeah, i think a tox dep is preferable | 22:36 |
| mtreinish | kinda like how we do it with hacking for flake8 and friends | 22:36 |
| jeblair | mtreinish: so far so good | 22:36 |
| clarkb | anteaya: just one question on the etherpad, otherwise it looks pretty good | 22:36 |
| fungi | e.g. the more things we can fix live without needing to rebuild worker images, the better | 22:37 |
| jeblair | mtreinish: then what runs it -- will you add it to the commands in tox.ini so it runs after testr? | 22:37 |
| jeblair | mtreinish, fungi: so then the slave script does not need to run it separately? or is that not possible because testr might exit with error and therefore not run it? | 22:37 |
| anteaya | clarkb: thank you | 22:38 |
| mtreinish | jeblair: well at least to start, it'll have to be installed globally in dsvm nodes because it's only used in d-g functions right now | 22:38 |
| jeblair | mtreinish: isn't it used in unit tests? | 22:38 |
| mtreinish | jeblair: oh yeah there too | 22:38 |
| mtreinish | jeblair: but ideally I'll eventually add a --html flag to os-testr to just dump a testr_results.html in the cwd | 22:40 |
| jeblair | mtreinish, fungi, mordred: this is probably where it gets tricky -- should this be in tox.ini in every project, or in jjb macros? | 22:40 |
| jeblair | mtreinish: oh you're replacing the testr invocation with an os-testr invocation? | 22:40 |
| clarkb | oh that I would not do | 22:40 |
| mtreinish | jeblair: that's the eventual plan. It started as a replacement of my pretty_tox.sh tumors that were being spread to all the projects | 22:41 |
| clarkb | we alrady have enough confusion using the test entrypoint from pbr | 22:41 |
| fungi | a wrapper, huh? | 22:41 |
| mtreinish | clarkb: this would not use that at all | 22:41 |
| *** annegentle has joined #openstack-infra | 22:41 | |
| clarkb | mtreinish: yes I know, I am saying from experience with the pbr entrpoint using another wrapper to run testr is sadness | 22:41 |
| clarkb | but maybe this ship has sailed | 22:41 |
| clarkb | I guess thats what pretty_tox.sh was (that name trips me up every time) | 22:42 |
| mtreinish | clarkb: I think it has, because right now we have a bash wrapper to get output people like, calling the pbr wrapper to default things in parallel, calling testr which is itself a wrapper | 22:42 |
| clarkb | mtreinish: yes, I got tricked by the name again | 22:42 |
| *** smccully has quit IRC | 22:43 | |
| *** dboik_ has joined #openstack-infra | 22:43 | |
| *** eharney has quit IRC | 22:43 | |
| clarkb | mordred: oh! I think I need to add image-types to the nodepool config validator stuff | 22:44 |
| clarkb | mordred: so my change isn't quite ready, working on that now | 22:44 |
| *** YorikSar has quit IRC | 22:44 | |
| *** harlowja has quit IRC | 22:45 | |
| *** harlowja_ has joined #openstack-infra | 22:45 | |
| jeblair | mtreinish: anyway, please do steal it, but let's keep talking about the best way to use it. in general, having tox handle the dependency is good. if it needs to be more than one step, we may want to do those in jjb macros for easy applicability across all the projects. if you end up doing it in one step, maybe that's not an issue. | 22:45 |
| *** peristeri has quit IRC | 22:45 | |
| jeblair | mtreinish: also, consider whether the PTI should be updated for this (i don't know the answer right now) | 22:46 |
| *** YorikSar has joined #openstack-infra | 22:46 | |
| jeblair | mtreinish: http://governance.openstack.org/reference/project-testing-interface.html | 22:46 |
| *** Sukhdev has quit IRC | 22:46 | |
| openstackgerrit | Clark Boylan proposed openstack-infra/nodepool: Support multiple image formats in a diskimage https://review.openstack.org/126747 | 22:46 |
| anteaya | the syrup got away on me, it was further along than I thought :( | 22:48 |
| mtreinish | jeblair: yeah I was trying to do this within all the constraints listed there. If we want to we can try and force using os-testr as the testr interface in that, but I'm not sure that's needed | 22:48 |
| *** rlucio has quit IRC | 22:52 | |
| ianw | fungi: so https://github.com/pyca/pyopenssl/issues/202 is about the final wrap-up of the immediate centos7 memory issues I think. ubuntu is just "lucky" it install python-openssl in its base image, which is why it hasn't ever noticed this in ci | 22:52 |
| *** YorikSar has quit IRC | 22:53 | |
| clarkb | ianw: thanks for digging into that. Any ideas on how we can deal with it in the mean time? | 22:55 |
| fungi | ianw: makes sense. so we're effectively testing much newer pyopenssl on centos7 than on trusty? | 22:55 |
| *** YorikSar has joined #openstack-infra | 22:55 | |
| *** rlandy has quit IRC | 22:55 | |
| *** annegentle has quit IRC | 22:55 | |
| ianw | clarkb: for centos, we can pre-install the packaged 0.13 version -> https://github.com/pyca/pyopenssl/issues/202 | 22:55 |
| fungi | i mean, we _want_ to not be installing the distro-packaged version on ubuntu as well... i expect swift is going to asplode all teh memories there too if we do that? | 22:55 |
| ianw | but now i'm actually thinking that pin in global-requirements is better | 22:56 |
| ianw | <0.14 ... which is the re-write to use cryptography rather than direct C bindings | 22:56 |
| *** esker has quit IRC | 22:57 | |
| ianw | ^ i mean https://review.openstack.org/#/c/168217/ | 22:57 |
| *** markvoelker has quit IRC | 22:57 | |
| *** dimtruck is now known as zz_dimtruck | 22:58 | |
| *** gordc has quit IRC | 22:58 | |
| clarkb | might also be good to try a throwaway devstack change that installs latest pyOpenSSL everywhere to check that ubuntu falls over too? | 22:59 |
| ianw | clarkb: i can fire one off ... Ubuntu seems to have a little more headroom than centos | 22:59 |
| ianw | clarkb: but if you can log in once it's going, we can easily check relative usage v another vm | 23:00 |
| clarkb | I should be able to do that for the next hour or so, I have to afk after that though | 23:00 |
| *** unicell has quit IRC | 23:02 | |
| *** unicell has joined #openstack-infra | 23:03 | |
| jeblair | ianw: also, are you aware of the dstat log saved on devstack runs? | 23:03 |
| mtreinish | anteaya: if you get a sec https://review.openstack.org/169042 should be pretty quick | 23:05 |
| ianw | jeblair: yes, but it's not really enough detail to debug issues like this. https://review.openstack.org/#/c/160232/ is something, it needs a few tweaks | 23:05 |
| mtreinish | clarkb: ^^^ I think you reviewd the d-g patch which that supersedes (I still need to abandon that) | 23:05 |
| clarkb | mtreinish: why the switch? | 23:06 |
| jeblair | ianw: cool, you're ahead of me then. :) | 23:06 |
| ianw | clarkb: https://review.openstack.org/#/c/169112 | 23:06 |
| mtreinish | clarkb: because sdague said so and -1'd the d-g change | 23:07 |
| mtreinish | it doesn't really bug me one way or the other | 23:07 |
| *** patrickeast_ has joined #openstack-infra | 23:08 | |
| clarkb | ok, that was my first suggestion, but ifgured if we were going to test with that opiton upstream it was probably fine | 23:08 |
| clarkb | we should probably figure out what needs a DEVSTACK_GATE flag and what doesn't then clean the leftovers up | 23:08 |
| *** patrickeast has quit IRC | 23:08 | |
| *** patrickeast_ is now known as patrickeast | 23:08 | |
| *** MarkAtwood has quit IRC | 23:08 | |
| jeblair | yeah, i think the rule should be "things that alter devstack-gate's behavior" | 23:09 |
| mtreinish | clarkb: fwiw, the adding the flag approach was the only way I knew how to do it. It's been a while since I had to add a job | 23:09 |
| jeblair | that _might_ be a little more subtle than it seems when you take branches into account, but maybe not. | 23:09 |
| jeblair | mtreinish: it's recent | 23:09 |
| mtreinish | yeah, from when we added devstack plugins iirc | 23:10 |
| clarkb | jeblair: so anything that d-g needs to operate on rather than straight pass through? sounds good to me | 23:10 |
| jeblair | yeah | 23:10 |
| jeblair | i mean, that's a starting place at least. :) | 23:11 |
| jeblair | we'll see what that doesn't work for | 23:11 |
| anteaya | mtreinish: done | 23:11 |
| anteaya | I took this batch too far, it is on it's way to sugar | 23:11 |
| * anteaya goes out to gather while there is still some daylight | 23:12 | |
| mtreinish | anteaya: cool thanks | 23:12 |
| anteaya | mtreinish: welcome | 23:12 |
| jeblair | Shrews: mordred got both of us with 168671 | 23:13 |
| Shrews | jeblair: hehe | 23:14 |
| clarkb | ianw: what are the magical incantations I should run on the ubuntu nodes running tests for 169112? | 23:15 |
| *** tsg has quit IRC | 23:15 | |
| ianw | clarkb: if you can just pastebin "ps --sort -pmem -eo pid:10,pmem:6,cputime:10,nlwp:8,wchan:25,comm:25,args:100" once devstack has run | 23:15 |
| clarkb | ianw: thanks will get that up when devstack is done running | 23:16 |
| *** melwitt has quit IRC | 23:17 | |
| *** Somay has quit IRC | 23:18 | |
| openstackgerrit | Merged openstack-infra/project-config: Use localrc passthrough for tempest no admin job https://review.openstack.org/169042 | 23:18 |
| openstackgerrit | Merged openstack-infra/project-config: Use localrc passthrough for tempest no admin job https://review.openstack.org/169042 | 23:18 |
| jeblair | oh dear that looks weird. | 23:18 |
| *** stevemar has quit IRC | 23:18 | |
| clarkb | wow | 23:19 |
| mtreinish | ooh, a double merge :) | 23:19 |
| mtreinish | do I get two stackalytic points! | 23:19 |
| *** baoli has joined #openstack-infra | 23:20 | |
| jeblair | the bot claims to have received two gerrit events... | 23:22 |
| *** tsg has joined #openstack-infra | 23:22 | |
| *** ivar-lazzaro has quit IRC | 23:22 | |
| *** ddieterly has quit IRC | 23:23 | |
| mtreinish | jeblair: fwiw, I got 2 emails too | 23:23 |
| jeblair | and zuul also got 2 notifications | 23:23 |
| *** Somay has joined #openstack-infra | 23:24 | |
| *** alexpilotti has quit IRC | 23:24 | |
| *** baoli_ has joined #openstack-infra | 23:25 | |
| *** bswartz has joined #openstack-infra | 23:25 | |
| *** alexpilotti has joined #openstack-infra | 23:26 | |
| *** alexpilotti has quit IRC | 23:26 | |
| *** SumitNaiksatam has quit IRC | 23:28 | |
| *** baoli has quit IRC | 23:28 | |
| clarkb | nice | 23:28 |
| *** mattoliverau has quit IRC | 23:29 | |
| *** wenlock has quit IRC | 23:29 | |
| *** mattoliverau has joined #openstack-infra | 23:29 | |
| *** ghostpl_ has quit IRC | 23:30 | |
| jeblair | clarkb, mordred: i think i can guess the answer to this... but is there a way for nodepool or shade to know what image formats to create/upload? | 23:33 |
| clarkb | jeblair: https://review.openstack.org/#/c/126747/7 | 23:34 |
| jeblair | clarkb: well, that's the change i'm trying to review | 23:34 |
| jeblair | clarkb: so i probably didn't ask the question well | 23:34 |
| clarkb | jeblair: do you mean know what image format based on the remote cloud in use? | 23:34 |
| jeblair | clarkb: particularly, i'm looking at the fact that we specify what image types to build. i think we shouldn't have to... yes that. | 23:35 |
| clarkb | jeblair: that is possibly something that can go in os-client-config | 23:35 |
| clarkb | then shade/nodepool can be smart about it when talking to dib | 23:35 |
| jeblair | clarkb: well, i don't think that os-client-config should have that information either | 23:35 |
| jeblair | clarkb: the more stuff that we put in os-client-config, the more broken openstack is with private clouds | 23:35 |
| clarkb | then no, openstack does not tell you this info anywhere | 23:35 |
| *** carl_baldwin has quit IRC | 23:36 | |
| jeblair | clarkb: (and the more that os-client-config becomes the first-level support desk for all openstack clouds everywhere) | 23:36 |
| clarkb | we even have to tell dib to use qcow2 version 0.10 for compat with hpcloud's qemu-img | 23:36 |
| jeblair | we have failed | 23:36 |
| *** asettle has quit IRC | 23:37 | |
| fungi | seems like something glance should at least expose (if it can't just convert on the fly) | 23:37 |
| jeblair | clarkb: what do you think about instead of specifying that on the diskimage dict, we specify it for the provider, and then take the union of image types in providers and use that for all dib calls? | 23:38 |
| *** hodos|2 has quit IRC | 23:38 | |
| *** dims has quit IRC | 23:38 | |
| jeblair | it feels like it's really an attribute of the provider, and adding it to multiple diskimage sections is repetitive and potentially error prone ("i forgot to add qcow 0.10 to this image, so oops") | 23:39 |
| jeblair | clarkb: i should say, it's already on the provider in your change, so i guess i'm only proposing removing it from diskimage | 23:39 |
| clarkb | on the flip side its an image attribute | 23:39 |
| clarkb | hrm my change is diskimage based I thought | 23:40 |
| jeblair | clarkb: it is in both places on your change | 23:40 |
| jeblair | clarkb: so we would need to specify it in two places and make sure the things matched up. i'm saying let the computer do it for us. :) | 23:40 |
| *** SumitNaiksatam has joined #openstack-infra | 23:40 | |
| clarkb | ianw: http://paste.openstack.org/show/197633/ http://paste.openstack.org/show/197635/ | 23:41 |
| *** Somay has quit IRC | 23:42 | |
| clarkb | jeblair: I see, ya I can do that | 23:42 |
| clarkb | jeblair: I do not know how to model that for the qemu-img option | 23:42 |
| jeblair | clarkb: cool, i'll leave a message with a suggestion about that too | 23:43 |
| clarkb | possibly we do a union of those options between image in provider and diskimage | 23:43 |
| clarkb | since it is possible that you would want a special flag there that is provider independent | 23:43 |
| clarkb | whereas the image type is more closely related to what the provider supports in glance | 23:44 |
| ianw | clarkb: yeah, so see any swift-* line there; compare with line 135 of https://etherpad.openstack.org/p/oom-in-rax-centos7-CI-job ... it usually takes up around 0.2% of a ci host's memory, it's showing up as ~0.6% | 23:44 |
| clarkb | ianw: fun stuff | 23:44 |
| ianw | clarkb: interesting to see if it ooms ... even with centos it was a sometimes thing | 23:45 |
| clarkb | ya, it wasn't 100% | 23:45 |
| ianw | digging into whatever the heck cryptography is doing ... how could parsing a 1.4mb C file into an in-memory tree ever go wrong :) | 23:46 |
| *** ujuc has quit IRC | 23:46 | |
| *** otter768 has joined #openstack-infra | 23:46 | |
| *** hichihara has joined #openstack-infra | 23:50 | |
| *** ZZelle_ has quit IRC | 23:50 | |
| *** otter768 has quit IRC | 23:51 | |
| *** ghostpl_ has joined #openstack-infra | 23:53 | |
| *** markvoelker has joined #openstack-infra | 23:53 | |
| jeblair | clarkb, mordred: i have written a bit on 126747 that you both should probably read. | 23:53 |
| *** carl_baldwin has joined #openstack-infra | 23:54 | |
| *** shakamunyi has quit IRC | 23:55 | |
| * clarkb reads | 23:55 | |
| *** annegentle has joined #openstack-infra | 23:56 | |
| *** Somay has joined #openstack-infra | 23:57 | |
| *** dannywilson has quit IRC | 23:58 | |
| *** weshay has quit IRC | 23:58 | |
| clarkb | jeblair: I think the second option is much better. We specifically do not want to have several hundred images anymore | 23:58 |
| *** ghostpl_ has quit IRC | 23:58 | |
| clarkb | jeblair: basically image with name "foo" should be built once and be the same everywhere | 23:58 |
| jeblair | clarkb: yes, except we have to have 2 formats of every image | 23:58 |
| openstackgerrit | Ramy Asselin proposed openstack-infra/puppet-openstackci: WIP: Attempt to test log server https://review.openstack.org/169117 | 23:58 |
| clarkb | jeblair: ya but thats just a format not a different image | 23:59 |
| jeblair | clarkb: that's all i'm suggesting; what did i say wrong? | 23:59 |
| asselin_ | nibalizer, clarkb ^^ quick script to test log server. There's probably a better way, but there's a start. | 23:59 |
| jeblair | clarkb: it's only a 'problem' because of the tripleo cloud -- we would create devstack-tripleo images for hp and rax, and devstack-trusty for tripleo | 23:59 |
| anteaya | jeblair: I updated the agenda for tomorrow's meeting | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!