| *** mikeit_ has quit IRC | 00:00 | |
| *** Alexandra_ has joined #openstack-infra | 00:01 | |
| *** mikeit_ has joined #openstack-infra | 00:02 | |
| *** sdake_ has joined #openstack-infra | 00:07 | |
| *** woodster_ has quit IRC | 00:10 | |
| *** sdake has quit IRC | 00:11 | |
| *** akshai has joined #openstack-infra | 00:18 | |
| *** mikeit_ has quit IRC | 00:18 | |
| *** ashleighfarnham has joined #openstack-infra | 00:19 | |
| *** markvoelker has joined #openstack-infra | 00:21 | |
| *** akshai has quit IRC | 00:21 | |
| *** mikeitt has joined #openstack-infra | 00:21 | |
| *** armax has joined #openstack-infra | 00:22 | |
| *** markvoelker has quit IRC | 00:25 | |
| *** mikeitt has quit IRC | 00:31 | |
| *** mikeit has joined #openstack-infra | 00:32 | |
| *** mpaolino has joined #openstack-infra | 00:34 | |
| *** camunoz has quit IRC | 00:34 | |
| *** mpaolino has quit IRC | 00:36 | |
| *** zhiwei has joined #openstack-infra | 00:37 | |
| *** otter768 has joined #openstack-infra | 00:46 | |
| *** camunoz has joined #openstack-infra | 00:47 | |
| *** salv-orlando has joined #openstack-infra | 00:47 | |
| *** salv-orlando has quit IRC | 00:50 | |
| *** kiseok7_ has quit IRC | 00:53 | |
| *** ashleighfarnham has quit IRC | 00:53 | |
| *** cody-somerville has quit IRC | 00:54 | |
| *** marcusvrn has joined #openstack-infra | 00:57 | |
| *** salv-orlando has joined #openstack-infra | 01:05 | |
| *** armax has quit IRC | 01:08 | |
| *** salv-orlando has quit IRC | 01:12 | |
| *** marcusvrn has quit IRC | 01:19 | |
| *** markvoelker has joined #openstack-infra | 01:21 | |
| *** sdake has joined #openstack-infra | 01:24 | |
| *** sdake_ has quit IRC | 01:26 | |
| *** markvoelker has quit IRC | 01:26 | |
| *** sdake_ has joined #openstack-infra | 01:27 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Enhance the OperatorCloud constructor https://review.openstack.org/179647 | 01:29 |
|---|---|---|
| mordred | Shrews: ^^ there ya go | 01:30 |
| *** sdake has quit IRC | 01:30 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config: Split diskimage-builder built images out https://review.openstack.org/167837 | 01:40 |
| *** Alexandra_ has quit IRC | 01:41 | |
| *** sputnik13 has joined #openstack-infra | 01:41 | |
| *** sputnik13 has quit IRC | 01:43 | |
| *** sputnik13 has joined #openstack-infra | 01:43 | |
| *** camunoz has quit IRC | 01:44 | |
| *** kiseok7_ has joined #openstack-infra | 01:44 | |
| *** sputnik13 has quit IRC | 01:45 | |
| *** sputnik13 has joined #openstack-infra | 01:49 | |
| *** patrickeast has joined #openstack-infra | 01:49 | |
| *** zz_dimtruck is now known as dimtruck | 01:54 | |
| *** rkukura_ has quit IRC | 01:55 | |
| *** rkukura has joined #openstack-infra | 01:56 | |
| *** sarob has quit IRC | 01:57 | |
| *** camunoz has joined #openstack-infra | 01:57 | |
| *** gokrokve has joined #openstack-infra | 01:57 | |
| *** btully has quit IRC | 01:58 | |
| *** patrickeast has quit IRC | 02:00 | |
| *** Kennan has quit IRC | 02:01 | |
| *** dimsum__ has quit IRC | 02:04 | |
| *** Kennan has joined #openstack-infra | 02:06 | |
| openstackgerrit | Merged openstack-infra/system-config: Split diskimage-builder built images out https://review.openstack.org/167837 | 02:06 |
| *** Kennan has quit IRC | 02:08 | |
| *** emagana has joined #openstack-infra | 02:11 | |
| *** Kennan has joined #openstack-infra | 02:11 | |
| *** dimtruck is now known as zz_dimtruck | 02:15 | |
| *** sabeen1 has quit IRC | 02:17 | |
| *** Somay has joined #openstack-infra | 02:19 | |
| *** Somay has quit IRC | 02:22 | |
| *** markvoelker has joined #openstack-infra | 02:22 | |
| *** achanda has quit IRC | 02:25 | |
| *** markvoelker has quit IRC | 02:27 | |
| *** sputnik13 has quit IRC | 02:28 | |
| *** mikeit has quit IRC | 02:38 | |
| *** patrickeast has joined #openstack-infra | 02:41 | |
| *** patrickeast has quit IRC | 02:42 | |
| openstackgerrit | Merged openstack-infra/bindep: Allow mixed-case package names https://review.openstack.org/179111 | 02:44 |
| *** sdake_ is now known as sdake | 02:44 | |
| openstackgerrit | Merged openstack-infra/bindep: Add positive/negative tests exercising the parser https://review.openstack.org/178378 | 02:46 |
| openstackgerrit | Jamie Lennox proposed openstack-infra/project-config: Add django-openstack-auth-kerberos project https://review.openstack.org/172803 | 02:51 |
| *** salv-orlando has joined #openstack-infra | 02:56 | |
| *** sergsh has quit IRC | 02:56 | |
| *** sergsh has joined #openstack-infra | 02:56 | |
| *** salv-orlando has quit IRC | 03:01 | |
| *** dimsum__ has joined #openstack-infra | 03:05 | |
| openstackgerrit | Merged openstack-infra/bindep: Allow hyphens in profile strings https://review.openstack.org/178379 | 03:09 |
| *** dimsum__ has quit IRC | 03:10 | |
| *** salv-orlando has joined #openstack-infra | 03:14 | |
| *** sputnik13 has joined #openstack-infra | 03:16 | |
| *** yfried|afk is now known as yfried | 03:19 | |
| *** salv-orlando has quit IRC | 03:19 | |
| *** coolsvapl is now known as coolsvap | 03:22 | |
| *** SumitNaiksatam has joined #openstack-infra | 03:22 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Look for yum instead of rpm when caching devstack https://review.openstack.org/179687 | 03:22 |
| *** markvoelker has joined #openstack-infra | 03:23 | |
| *** achanda has joined #openstack-infra | 03:26 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Import subprocess in cache devstack element https://review.openstack.org/179688 | 03:27 |
| *** markvoelker has quit IRC | 03:27 | |
| *** unicell has joined #openstack-infra | 03:30 | |
| *** achanda has quit IRC | 03:31 | |
| *** unicell1 has quit IRC | 03:32 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/devstack-gate: [WIP] run centos7 with neutron https://review.openstack.org/179689 | 03:34 |
| *** reed_ has joined #openstack-infra | 03:35 | |
| *** reed_ has quit IRC | 03:37 | |
| *** jyuso1 has joined #openstack-infra | 03:42 | |
| *** Somay has joined #openstack-infra | 03:43 | |
| openstackgerrit | Merged openstack-infra/project-config: Import subprocess in cache devstack element https://review.openstack.org/179688 | 03:48 |
| *** bswartz has quit IRC | 03:50 | |
| *** bswartz has joined #openstack-infra | 03:52 | |
| lifeless | tchaypo: I've closed off the pr's you had open on my github tripleo-end-to-end repo: I'm not hacking on that, nor maintaining it atm | 04:01 |
| *** ivar-laz_ has joined #openstack-infra | 04:02 | |
| *** ivar-laz_ has quit IRC | 04:03 | |
| *** otter768 has quit IRC | 04:04 | |
| *** fedexo has joined #openstack-infra | 04:04 | |
| *** ivar-laz_ has joined #openstack-infra | 04:04 | |
| *** ivar-lazzaro has quit IRC | 04:04 | |
| tchaypo | thanks | 04:07 |
| *** sarob has joined #openstack-infra | 04:11 | |
| *** sputnik13 has quit IRC | 04:12 | |
| *** Hal has joined #openstack-infra | 04:14 | |
| *** Hal is now known as Guest17598 | 04:14 | |
| *** AlecTaylor has joined #openstack-infra | 04:18 | |
| AlecTaylor | hi | 04:18 |
| *** jyuso has joined #openstack-infra | 04:20 | |
| *** Guest17598 has quit IRC | 04:21 | |
| *** sputnik13 has joined #openstack-infra | 04:23 | |
| *** jyuso1 has quit IRC | 04:23 | |
| *** markvoelker has joined #openstack-infra | 04:24 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Do a better job of identifying CentOS 6 https://review.openstack.org/179694 | 04:25 |
| *** markvoelker has quit IRC | 04:28 | |
| *** heyongli has quit IRC | 04:30 | |
| *** heyongli has joined #openstack-infra | 04:31 | |
| *** tjones1 has left #openstack-infra | 04:31 | |
| *** sks has joined #openstack-infra | 04:31 | |
| *** cryptarium has quit IRC | 04:31 | |
| *** yfried has quit IRC | 04:39 | |
| *** unicell has quit IRC | 04:41 | |
| *** unicell has joined #openstack-infra | 04:43 | |
| *** tjones1 has joined #openstack-infra | 04:51 | |
| *** abregman has joined #openstack-infra | 04:54 | |
| AlecTaylor | PaaS cloud, considering: libcloud+consul+etcd+mesos+deis+drone, thoughts? | 04:56 |
| *** woodster_ has joined #openstack-infra | 04:57 | |
| *** achanda has joined #openstack-infra | 05:02 | |
| *** chhavi has joined #openstack-infra | 05:04 | |
| *** btully has joined #openstack-infra | 05:04 | |
| chhavi | Hi all | 05:04 |
| chhavi | I am facing issues while commit the changes to openstack | 05:04 |
| chhavi | fatal: The remote end hung up unexpectedly | 05:05 |
| chhavi | chhavi@chhavi-KVM:~/chhavi/nova$ git push gerrit HEAD:refs/for/master/fix-bug-1451030 | 05:05 |
| chhavi | fatal: ICLA contributor agreement requires current contact information. | 05:05 |
| chhavi | Please review your contact information: | 05:05 |
| chhavi | https://review.openstack.org/#/settings/contact | 05:05 |
| chhavi | fatal: The remote end hung up unexpectedly | 05:05 |
| jhesketh | chhavi: have you updated your contact information as per the instructions? | 05:21 |
| chhavi | yes, the issue just got resolved. there was issue while running git remote | 05:22 |
| chhavi | I am able to commit the change now | 05:22 |
| jhesketh | excellent | 05:24 |
| *** markvoelker has joined #openstack-infra | 05:24 | |
| *** ivar-laz_ has quit IRC | 05:27 | |
| *** markvoelker has quit IRC | 05:29 | |
| *** ashleighfarnham has joined #openstack-infra | 05:34 | |
| tchaypo | dhellmann: . | 05:36 |
| *** gokrokve has quit IRC | 05:37 | |
| *** chhavi has left #openstack-infra | 05:37 | |
| *** gokrokve has joined #openstack-infra | 05:37 | |
| *** ashleighfarnham has quit IRC | 05:39 | |
| *** gokrokve has quit IRC | 05:42 | |
| *** hdd has joined #openstack-infra | 05:42 | |
| *** BharatK has joined #openstack-infra | 05:43 | |
| *** sarob has quit IRC | 05:43 | |
| tchaypo | well done me. | 05:49 |
| *** ibiris_away is now known as ibiris | 05:50 | |
| *** salv-orlando has joined #openstack-infra | 05:53 | |
| *** Somay has quit IRC | 05:55 | |
| *** Somay has joined #openstack-infra | 05:56 | |
| *** salv-orlando has quit IRC | 05:56 | |
| *** Darkwan has quit IRC | 05:58 | |
| *** salv-orlando has joined #openstack-infra | 05:59 | |
| *** salv-orlando has quit IRC | 05:59 | |
| *** tnovacik has joined #openstack-infra | 06:03 | |
| *** otter768 has joined #openstack-infra | 06:05 | |
| *** AlecTaylor has quit IRC | 06:05 | |
| *** gokrokve has joined #openstack-infra | 06:08 | |
| *** gokrokve has quit IRC | 06:09 | |
| *** gokrokve has joined #openstack-infra | 06:09 | |
| *** otter768 has quit IRC | 06:10 | |
| *** salv-orlando has joined #openstack-infra | 06:11 | |
| *** deepakcs has joined #openstack-infra | 06:14 | |
| *** salv-orlando has quit IRC | 06:15 | |
| *** mrunge has joined #openstack-infra | 06:16 | |
| *** vlaza has joined #openstack-infra | 06:18 | |
| *** sdake_ has joined #openstack-infra | 06:22 | |
| *** fedexo has quit IRC | 06:22 | |
| openstackgerrit | Andreas Jaeger proposed openstack/requirements: Rename keystoneclient-federation to saml2 https://review.openstack.org/179706 | 06:23 |
| *** sergsh has quit IRC | 06:25 | |
| *** sergsh_ has quit IRC | 06:25 | |
| *** yfried has joined #openstack-infra | 06:25 | |
| *** sdake has quit IRC | 06:25 | |
| *** sdake_ is now known as sdake | 06:35 | |
| *** yfried is now known as yfried|afk | 06:36 | |
| *** yfried|afk is now known as yfried | 06:37 | |
| *** dimsum__ has joined #openstack-infra | 06:42 | |
| *** ibiris is now known as ibiris_away | 06:43 | |
| *** xylan_kong has left #openstack-infra | 06:44 | |
| *** salv-orl_ has joined #openstack-infra | 06:45 | |
| *** ibiris_away is now known as ibiris | 06:47 | |
| *** dimsum__ has quit IRC | 06:47 | |
| *** jcoufal has joined #openstack-infra | 06:55 | |
| *** sushilkm has left #openstack-infra | 06:55 | |
| ttx | lifeless: doing it now | 06:58 |
| *** hdd has quit IRC | 06:59 | |
| *** hdd has joined #openstack-infra | 07:00 | |
| ttx | mtreinish, morganfainberg: stable-maint-core owns $PROJECT-stable-maint because we are supposed to check that proposed members know and will follow the stable policy | 07:02 |
| ttx | You don't want each $PROJECT to come up with their own culture of what is an acceptable stable branch backport -- consistency is needed there | 07:03 |
| *** Ala has joined #openstack-infra | 07:04 | |
| ttx | So project teams propose members and stable-maint-core lecture the proposed member with Stable Branch Policy, then we add them | 07:05 |
| ttx | Reference: https://wiki.openstack.org/wiki/StableBranch#Project-specific_teams | 07:06 |
| *** scheuran has joined #openstack-infra | 07:09 | |
| lifeless | ttx: thanks | 07:12 |
| *** mrmartin has joined #openstack-infra | 07:12 | |
| *** yamahata has quit IRC | 07:14 | |
| *** sdake has quit IRC | 07:21 | |
| *** doude has joined #openstack-infra | 07:21 | |
| openstackgerrit | Steve Kowalik proposed openstack-infra/infra-specs: Add a specification about building packages https://review.openstack.org/179713 | 07:22 |
| *** achanda has quit IRC | 07:23 | |
| *** ashleighfarnham has joined #openstack-infra | 07:24 | |
| openstackgerrit | Sebastian Kalinowski proposed openstack-infra/project-config: Remove old Fuel plugins https://review.openstack.org/179714 | 07:24 |
| *** btully has quit IRC | 07:26 | |
| *** markvoelker has joined #openstack-infra | 07:26 | |
| *** achanda has joined #openstack-infra | 07:26 | |
| *** ashleighfarnham has quit IRC | 07:28 | |
| *** ildikov has joined #openstack-infra | 07:29 | |
| *** markvoelker has quit IRC | 07:30 | |
| *** Kennan has quit IRC | 07:33 | |
| *** Kennan has joined #openstack-infra | 07:34 | |
| *** markus_z has joined #openstack-infra | 07:36 | |
| *** dizquierdo has joined #openstack-infra | 07:36 | |
| *** achanda has quit IRC | 07:37 | |
| *** chlong has quit IRC | 07:37 | |
| *** mpavone has joined #openstack-infra | 07:40 | |
| *** woodster_ has quit IRC | 07:40 | |
| *** Hal has joined #openstack-infra | 07:43 | |
| *** Hal is now known as Guest36998 | 07:44 | |
| *** swat30 has quit IRC | 07:49 | |
| *** begal-sop_ has quit IRC | 07:50 | |
| *** funzo_ has joined #openstack-infra | 07:50 | |
| *** aswadr has joined #openstack-infra | 07:50 | |
| *** arxcruz has joined #openstack-infra | 07:50 | |
| *** funzo has quit IRC | 07:50 | |
| *** swat30_ has joined #openstack-infra | 07:50 | |
| *** swat30_ is now known as swat30 | 07:51 | |
| *** begal-sop__ has joined #openstack-infra | 07:51 | |
| *** tnovacik has quit IRC | 07:52 | |
| openstackgerrit | Jan Klare proposed openstack-infra/project-config: added chef-rake-jobs for openstack-chef-repo https://review.openstack.org/179721 | 07:56 |
| openstackgerrit | Fabien Boucher proposed openstack-infra/infra-specs: Specification proposal about system-config testing using containers https://review.openstack.org/172833 | 07:56 |
| *** hashar has joined #openstack-infra | 07:56 | |
| *** hashar_ has joined #openstack-infra | 07:57 | |
| *** jistr has joined #openstack-infra | 07:59 | |
| *** hashar has quit IRC | 08:02 | |
| *** AJaeger has joined #openstack-infra | 08:04 | |
| *** otter768 has joined #openstack-infra | 08:05 | |
| GheRiver1 | hi all | 08:08 |
| *** GheRiver1 is now known as GheRivero | 08:08 | |
| *** pblaho has joined #openstack-infra | 08:08 | |
| *** ildikov has quit IRC | 08:10 | |
| *** otter768 has quit IRC | 08:10 | |
| *** fhubik has joined #openstack-infra | 08:13 | |
| *** sushilkm has joined #openstack-infra | 08:14 | |
| *** sushilkm has left #openstack-infra | 08:14 | |
| *** vlaza has quit IRC | 08:17 | |
| *** vlaza1 has joined #openstack-infra | 08:17 | |
| *** ajo has joined #openstack-infra | 08:26 | |
| *** ildikov has joined #openstack-infra | 08:26 | |
| *** markvoelker has joined #openstack-infra | 08:27 | |
| *** sergsh_ has joined #openstack-infra | 08:27 | |
| *** vlaza has joined #openstack-infra | 08:29 | |
| *** vlaza1 has quit IRC | 08:29 | |
| *** Hal has joined #openstack-infra | 08:30 | |
| *** SumitNaiksatam has quit IRC | 08:30 | |
| *** Hal is now known as Guest64971 | 08:31 | |
| *** markvoelker has quit IRC | 08:31 | |
| *** fhubik is now known as fhubik_afk | 08:31 | |
| *** sergsh has joined #openstack-infra | 08:31 | |
| openstackgerrit | Huang Rui proposed openstack-infra/project-config: enable python-jobs for networking-zvm https://review.openstack.org/179730 | 08:32 |
| *** Guest36998 has quit IRC | 08:33 | |
| *** tnovacik has joined #openstack-infra | 08:33 | |
| *** SumitNaiksatam has joined #openstack-infra | 08:34 | |
| *** Guest64971 has quit IRC | 08:35 | |
| *** mrunge has quit IRC | 08:36 | |
| *** vlaza has quit IRC | 08:38 | |
| *** vlaza has joined #openstack-infra | 08:38 | |
| openstackgerrit | Merged openstack-infra/project-config: Puppet/Beaker: drop puppet-{gnocchi,tuskar} https://review.openstack.org/179601 | 08:39 |
| *** Somay has quit IRC | 08:41 | |
| *** Somay has joined #openstack-infra | 08:42 | |
| *** alexpilotti has quit IRC | 08:45 | |
| *** ociuhandu has joined #openstack-infra | 08:48 | |
| *** vlaza has quit IRC | 08:50 | |
| *** vlaza has joined #openstack-infra | 08:50 | |
| *** alexpilotti has joined #openstack-infra | 08:54 | |
| openstackgerrit | Sirushti Murugesan proposed openstack-infra/project-config: Add grenade jobs for Heat https://review.openstack.org/178352 | 08:56 |
| openstackgerrit | Huang Rui proposed openstack-infra/project-config: Move zvm-driver to stackforge-attic https://review.openstack.org/179738 | 09:02 |
| *** e0ne has joined #openstack-infra | 09:04 | |
| *** hashar_ is now known as hasharOut | 09:06 | |
| *** yfried is now known as yfried|afk | 09:10 | |
| *** e0ne has quit IRC | 09:10 | |
| *** asrangne has joined #openstack-infra | 09:11 | |
| *** salv-orlando has joined #openstack-infra | 09:13 | |
| *** salv-orl_ has quit IRC | 09:13 | |
| *** ashleighfarnham has joined #openstack-infra | 09:13 | |
| *** aswadr has quit IRC | 09:13 | |
| *** hdd has quit IRC | 09:16 | |
| *** _dguerri is now known as dguerri | 09:17 | |
| *** ashleighfarnham has quit IRC | 09:17 | |
| *** hdd has joined #openstack-infra | 09:18 | |
| *** sergsh has quit IRC | 09:22 | |
| *** sergsh_ has quit IRC | 09:22 | |
| *** fhubik_afk is now known as fhubik | 09:23 | |
| *** markvoelker has joined #openstack-infra | 09:27 | |
| *** jyuso has quit IRC | 09:27 | |
| *** dguerri is now known as _dguerri | 09:28 | |
| *** mpaolino has joined #openstack-infra | 09:29 | |
| *** vlaza has left #openstack-infra | 09:30 | |
| *** markvoelker has quit IRC | 09:32 | |
| openstackgerrit | Giulio Fidente proposed openstack-infra/tripleo-ci: Bump up delorean pinning https://review.openstack.org/177176 | 09:32 |
| openstackgerrit | Julian Taylor proposed openstack-infra/jenkins-job-builder: add support for measurement plots plugin to junit publisher https://review.openstack.org/179743 | 09:33 |
| *** Somay_ has joined #openstack-infra | 09:34 | |
| *** Somay has quit IRC | 09:35 | |
| *** sergsh has joined #openstack-infra | 09:35 | |
| *** sergsh_ has joined #openstack-infra | 09:35 | |
| *** mpaolino has quit IRC | 09:35 | |
| *** abregman is now known as abregmansd | 09:38 | |
| *** abregmansd is now known as abregman|food | 09:38 | |
| *** abregman|food is now known as abregman|food-ti | 09:39 | |
| *** abregman|food-ti is now known as abregman|launch | 09:39 | |
| *** abregman|launch is now known as abregman|lunch | 09:39 | |
| *** _dguerri is now known as dguerri | 09:42 | |
| *** salv-orlando has quit IRC | 09:43 | |
| *** ociuhandu has quit IRC | 09:44 | |
| *** salv-orl_ has joined #openstack-infra | 09:46 | |
| *** mrunge has joined #openstack-infra | 09:46 | |
| *** fhubik is now known as fhubik_afk | 09:47 | |
| openstackgerrit | Giulio Fidente proposed openstack-infra/tripleo-ci: Bump up delorean pinning https://review.openstack.org/177176 | 09:49 |
| *** luqas has joined #openstack-infra | 09:49 | |
| *** abregman|lunch is now known as abregman | 09:49 | |
| *** dguerri is now known as _dguerri | 09:50 | |
| *** fhubik_afk is now known as fhubik | 09:50 | |
| *** salv-orlando has joined #openstack-infra | 09:50 | |
| openstackgerrit | Victor Stinner proposed openstack/requirements: Add mysqlclient dependency https://review.openstack.org/179745 | 09:52 |
| *** salv-orl_ has quit IRC | 09:53 | |
| *** dimsum__ has joined #openstack-infra | 09:53 | |
| openstackgerrit | Sirushti Murugesan proposed openstack-infra/project-config: Add grenade jobs for Heat https://review.openstack.org/178352 | 09:55 |
| *** devvesa has joined #openstack-infra | 09:55 | |
| *** Hal has joined #openstack-infra | 09:55 | |
| *** Hal is now known as Guest84883 | 09:56 | |
| *** zhiwei has quit IRC | 10:00 | |
| *** rlandy has joined #openstack-infra | 10:04 | |
| *** dguitarbite has quit IRC | 10:05 | |
| *** otter768 has joined #openstack-infra | 10:06 | |
| *** abregman is now known as abregman|lunch | 10:08 | |
| *** vlaza has joined #openstack-infra | 10:09 | |
| *** otter768 has quit IRC | 10:11 | |
| lifeless | a few hours late, but just FYI, new releases of testscenarios and fixtures were cut earlier today | 10:14 |
| lifeless | not expecting any fallout,but just in case.... | 10:14 |
| *** armax has joined #openstack-infra | 10:15 | |
| *** yfried|afk is now known as yfried | 10:16 | |
| *** pc_m has joined #openstack-infra | 10:17 | |
| *** dguitarbite has joined #openstack-infra | 10:17 | |
| *** Somay_ has quit IRC | 10:22 | |
| *** macjack has joined #openstack-infra | 10:23 | |
| *** fhubik is now known as fhubik_afk | 10:26 | |
| *** _dguerri is now known as dguerri | 10:27 | |
| *** tnovacik has quit IRC | 10:28 | |
| *** sushilkm has joined #openstack-infra | 10:28 | |
| *** sushilkm has left #openstack-infra | 10:28 | |
| *** fhubik_afk is now known as fhubik | 10:28 | |
| *** salv-orl_ has joined #openstack-infra | 10:31 | |
| *** e0ne has joined #openstack-infra | 10:31 | |
| *** salv-orlando has quit IRC | 10:34 | |
| *** weshay has joined #openstack-infra | 10:36 | |
| *** mrmartin has quit IRC | 10:40 | |
| *** samueldmq has joined #openstack-infra | 10:40 | |
| samueldmq | morning :) | 10:40 |
| *** asrangne has quit IRC | 10:43 | |
| *** dguerri is now known as _dguerri | 10:43 | |
| *** armax has quit IRC | 10:44 | |
| *** salv-orl_ has quit IRC | 10:46 | |
| *** salv-orlando has joined #openstack-infra | 10:47 | |
| openstackgerrit | Julian Taylor proposed openstack-infra/jenkins-job-builder: add support for measurement plots plugin to junit publisher https://review.openstack.org/179743 | 10:49 |
| *** ociuhandu has joined #openstack-infra | 10:49 | |
| *** fhubik is now known as fhubik_afk | 10:50 | |
| *** fhubik_afk is now known as fhubik | 10:52 | |
| *** tjones2 has joined #openstack-infra | 10:54 | |
| *** tjones1 has quit IRC | 10:57 | |
| *** luqas has quit IRC | 10:57 | |
| *** mrmartin has joined #openstack-infra | 10:57 | |
| *** tnovacik has joined #openstack-infra | 10:59 | |
| *** ashleighfarnham has joined #openstack-infra | 11:01 | |
| openstackgerrit | Michal Kovarik proposed openstack-infra/jenkins-job-builder: Adding support for Beaker Builder Plugin https://review.openstack.org/179756 | 11:02 |
| *** claudiub has joined #openstack-infra | 11:03 | |
| *** baoli has joined #openstack-infra | 11:04 | |
| *** baoli has quit IRC | 11:05 | |
| *** abregman|lunch is now known as abregman | 11:06 | |
| *** ashleighfarnham has quit IRC | 11:06 | |
| sdague | ls | 11:07 |
| sdague | doh | 11:07 |
| openstackgerrit | Sirushti Murugesan proposed openstack-infra/project-config: Add grenade jobs for Heat https://review.openstack.org/178352 | 11:09 |
| *** vlaza has left #openstack-infra | 11:09 | |
| samueldmq | sdague, echo "hehe"; ./good-morning.sh | 11:11 |
| AJaeger | morning, sdague! The requirements post job propose-requirements-update fails due to a project rename - https://review.openstack.org/#/c/179706/ should fix it... | 11:11 |
| samueldmq | AJaeger, and this also should be renamed in the openstack/requirements .. which was failing to find python-keystoneclient-federation | 11:14 |
| samueldmq | right ? | 11:14 |
| AJaeger | samueldmq, yep | 11:14 |
| samueldmq | AJaeger, nice | 11:15 |
| *** Longgeek has joined #openstack-infra | 11:15 | |
| AJaeger | samueldmq, https://jenkins.openstack.org/job/propose-requirements-updates/539/console | 11:15 |
| samueldmq | AJaeger, cool .. I just added a reference to your patch on the previous one (which renamed the project) | 11:17 |
| AJaeger | thanks, samueldmq. | 11:19 |
| sdague | AJaeger: could you make propose-requirements-update more robust as well? | 11:19 |
| sdague | it seems like 1 project not existing shouldn't make it go fatal | 11:19 |
| *** luqas has joined #openstack-infra | 11:19 | |
| AJaeger | sdague, would be a nice idea. I doubt I have time quickly for this, so if anybody else wants to take it... | 11:20 |
| *** salv-orlando has quit IRC | 11:23 | |
| openstackgerrit | Gal Sagie proposed openstack-infra/project-config: Enable a check job for stackforge/dragonflow https://review.openstack.org/175152 | 11:23 |
| samueldmq | sdague, AJaeger how would be that change ? I'd probably be able to help, but I am not familiar with this code | 11:26 |
| samueldmq | any entry point to the code ? | 11:26 |
| AJaeger | samueldmq, jenkins/scripts/propose_update.sh in project-config | 11:26 |
| *** ildikov has quit IRC | 11:27 | |
| samueldmq | AJaeger, sdague what should be the expected behavior when it cannot fetch a project ? | 11:27 |
| AJaeger | http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/scripts/propose_update.sh#n43 is the line that currently breaks it | 11:28 |
| *** eharney has quit IRC | 11:28 | |
| AJaeger | samueldmq, setting ALL_SUCCESS to 1 and moving on to the next project | 11:28 |
| AJaeger | Together with some error message to look out for | 11:29 |
| *** luqas has quit IRC | 11:29 | |
| sdague | samueldmq: yeh, so it should accumulate errors somewhere to report at the end, but definitely keep looping through all projects | 11:30 |
| *** wznoinsk has quit IRC | 11:30 | |
| *** marcusvrn has joined #openstack-infra | 11:30 | |
| pc_m | Hi, I'm getting a failure in gate-neutron-vpnaas-python27 that is quite odd. It passes all the tests and then fails in the script after. Anyone hear of this? | 11:30 |
| pc_m | Ref: http://logs.openstack.org/15/168115/32/check/gate-neutron-vpnaas-python27/5d9ecf1/console.html#_2015-05-02_16_28_39_906 | 11:30 |
| AJaeger | sdague, there are more projects in requirements that were renamed... | 11:30 |
| * AJaeger does a second patch... | 11:31 | |
| sdague | AJaeger: cool, thanks | 11:31 |
| openstackgerrit | Michal Kovarik proposed openstack-infra/jenkins-job-builder: Adding support for Beaker Builder Plugin https://review.openstack.org/179756 | 11:31 |
| *** luqas has joined #openstack-infra | 11:31 | |
| samueldmq | AJaeger, sdague should be something similar to what we do when a project fails to sync .. http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/scripts/propose_update.sh#n114 | 11:32 |
| samueldmq | right ? | 11:33 |
| AJaeger | yes | 11:33 |
| samueldmq | k got it | 11:33 |
| *** deepakcs has quit IRC | 11:36 | |
| openstackgerrit | Andreas Jaeger proposed openstack/requirements: Follow project renames https://review.openstack.org/179706 | 11:38 |
| AJaeger | sdague, samueldmq : Updated my patch ^ | 11:38 |
| openstackgerrit | Andreas Jaeger proposed openstack/requirements: Follow project renames https://review.openstack.org/179706 | 11:39 |
| AJaeger | and sorted projects again ^ | 11:39 |
| *** mase_x200 has joined #openstack-infra | 11:40 | |
| *** dimsum__ is now known as dims | 11:41 | |
| samueldmq | AJaeger, and congress/rally changed from stackforge to opesntack, right? | 11:41 |
| AJaeger | samueldmq, they did | 11:41 |
| samueldmq | AJaeger, k looks good, thanks | 11:42 |
| AJaeger | samueldmq, let me enhance the commit message... | 11:42 |
| *** ildikov has joined #openstack-infra | 11:43 | |
| samueldmq | AJaeger, oh, already +1'ed, the commit message already says rally and congress were renamed | 11:43 |
| samueldmq | AJaeger, but feel free to enhance it as you want :) | 11:43 |
| openstackgerrit | Andreas Jaeger proposed openstack/requirements: Follow project renames https://review.openstack.org/179706 | 11:43 |
| AJaeger | samueldmq, if you ask, somebody else wonders as well - so let's be clear here... | 11:43 |
| AJaeger | samueldmq, thanks for the reviews | 11:43 |
| samueldmq | AJaeger, you added links to the commits on each project, nice | 11:44 |
| *** claudiub_ has joined #openstack-infra | 11:44 | |
| samueldmq | AJaeger, np, just trying to get things moving, my reviews aren't still that valuable in #infra though | 11:45 |
| AJaeger | samueldmq, every review is important - and you learn as well with them ;) | 11:45 |
| *** claudiub has quit IRC | 11:45 | |
| samueldmq | AJaeger, :) | 11:46 |
| *** ldnunes has joined #openstack-infra | 11:47 | |
| *** esker has quit IRC | 11:47 | |
| *** markvoelker has joined #openstack-infra | 11:50 | |
| *** alexpilotti has quit IRC | 11:51 | |
| *** luqas has quit IRC | 11:51 | |
| *** dkranz has joined #openstack-infra | 11:52 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/project-config: Don't short circuit when projects fails to clone https://review.openstack.org/179760 | 11:53 |
| samueldmq | AJaeger, sdague ^ | 11:53 |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/project-config: Don't short circuit when projects fail to clone https://review.openstack.org/179760 | 11:53 |
| samueldmq | well, just fixed the commit message | 11:53 |
| fungi | pc_m: i've commented on your https://review.openstack.org/168115 change pointing out where you're introducing that issue | 11:54 |
| pc_m | fungi: Thanks. I was wondering if it was the copy of rootwrap.conf. will look at your comment. | 11:55 |
| pc_m | fungi: Ah that is what you saw. Thanks | 11:56 |
| AJaeger | thanks, samueldmq! | 11:56 |
| pc_m | fungi: I'm not too familiar with tox.ini. Can I have a command for a target, depend on another target's command? | 11:57 |
| claudiub_ | sdague: hello. Could I have your +2 on a 1 line commit? It already has a +2. https://review.openstack.org/#/c/178993/ Thanks! | 11:57 |
| samueldmq | AJaeger, np, anytime :) | 11:57 |
| *** rfolco has joined #openstack-infra | 11:58 | |
| pc_m | fungi: IOW, have the command on that and the previous two lines in a "functional" target, and then have the dsvm-functional and dsvm-functional-sswan targets use that command? | 12:00 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: use optimized url for zuul status https://review.openstack.org/170935 | 12:00 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: add dbg function to hideci https://review.openstack.org/179573 | 12:00 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: add jquery visibility https://review.openstack.org/179362 | 12:00 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: create proxy for testing hideci https://review.openstack.org/179360 | 12:00 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: Revert "Revert zuul status in gerrit" https://review.openstack.org/179361 | 12:00 |
| pc_m | fungi: like... command = {|functional|command}? | 12:00 |
| fungi | pc_m: i don't think there's a way to inherit and then extend those variables from one env to the next, though you might check the documentation for tox. you'll probably just need to have that command appear explicitly in each env where it's needed | 12:03 |
| pc_m | fungi: ok. will do. thanks | 12:04 |
| fungi | pc_m: oh, though it is possible to add an env to the commands list of another env | 12:05 |
| fungi | for example nova (and a lot of other projects) have a pep8 env that calls the flake8 env | 12:05 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Convert node_set_provision_state to task https://review.openstack.org/177987 | 12:05 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Update recent Ironic exceptions https://review.openstack.org/177986 | 12:05 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Enhance error message in update_machine https://review.openstack.org/177985 | 12:05 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Enhance the OperatorCloud constructor https://review.openstack.org/179647 | 12:05 |
| AJaeger | fungi, we have an idle CI system with a IMHO very slow throughput. project-config changes need over an hour to get through, waiting very long for gate-infra-puppet-apply-precise - if I interpret status.openstack.org/zuul correctly | 12:05 |
| fungi | pc_m: wait, i just brained wrong. the pep8 env calls the flake8 command line utility. flake8 itself magically reads tox.ini too and looks for a "flake8" env containing its configuration | 12:06 |
| fungi | pc_m: so forget what i said about calling an env from another env | 12:06 |
| *** luqas has joined #openstack-infra | 12:07 | |
| fungi | AJaeger: yes, i think that jenkins-gearman bug clarkb and zaro were looking at last week is still an issue. i'll clear out more stale ready nodes for now and that should speed things back up | 12:07 |
| AJaeger | fungi, thanks! That jenkins-gearman bug is becoming famous ;/ | 12:08 |
| jhesketh | sdague: oh cool, nice stuff (re zuul status patches) | 12:08 |
| sdague | jhesketh: yeh, so the jquery visibility bit still needs that file to actually get deployed, I don't know where best to do that | 12:10 |
| sdague | jhesketh: do you? | 12:10 |
| *** mrmartin has quit IRC | 12:10 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/project-config: Don't short circuit when projects fails to clone https://review.openstack.org/179760 | 12:10 |
| sdague | also, the targettted status calls aren't working still | 12:11 |
| jhesketh | sdague: I just started reviewing at the top (the web server), so not sure which file that is sorry? | 12:11 |
| samueldmq | fungi, ^ thanks for your review | 12:11 |
| sdague | https://review.openstack.org/179362 | 12:11 |
| jhesketh | sdague: oh yeah, I know where that goes... gimme a sec | 12:12 |
| *** hdd has quit IRC | 12:12 | |
| openstackgerrit | Merged openstack-infra/project-config: Adds pypi-jobs to networking-hyperv https://review.openstack.org/178993 | 12:12 |
| jhesketh | sdague: it goes in system-config/modules/openstack_project/files/gerrit/GerritSiteHeader.html | 12:14 |
| sdague | jhesketh: ok, we also deploy it for zuul, so I wasn't sure if there was a common deploy point that we wanted | 12:14 |
| jhesketh | sdague: do it similar to the hideci.js in that file, the important part is the notify => Exec['reload_gerrit_header'] because of the way gerrit signs static files | 12:14 |
| sdague | right, my only concern was copying the file a second time into a tree | 12:15 |
| sdague | and if that's what people want or not | 12:15 |
| *** luqas has quit IRC | 12:15 | |
| jhesketh | sdague: oh right, I see.. Well we could link to the file on zuul.o.o but that's not great | 12:15 |
| fungi | sdague: sdague what do you mean by the targeted status calls aren't working still? http://zuul.openstack.org/status/change/179754,2 returns json for me right now | 12:15 |
| sdague | fungi: oh... but not over https | 12:16 |
| fungi | sdague: oh! right, we probably need to duplicate that block for https as well. i completely forgot :/ | 12:16 |
| sdague | ok, no prob, let me go look for that | 12:17 |
| jhesketh | sdague: but the actual file isn't copied into the tree, so something like this: http://git.openstack.org/cgit/openstack-infra/puppet-zuul/tree/manifests/init.pp#n249 | 12:17 |
| *** doug-fish has left #openstack-infra | 12:17 | |
| fungi | sdague: it's later in the same file you modified previously | 12:17 |
| fungi | sdague: one vhost template providing http and optionally https | 12:18 |
| jhesketh | sdague: but the notify part I mentioned is still important and the gerrit.pp is the place | 12:18 |
| sdague | jhesketh: any chance you want to write that puppet? I need to get my presentations for summit sorted early this week, so I'm going to loose this thread for a bit | 12:18 |
| *** shakamunyi has quit IRC | 12:18 | |
| jhesketh | sdague: also, looks like I had a paste failure... I meant to say it'd go into system-config/modules/openstack_project/manifests/gerrit.pp | 12:18 |
| jhesketh | sdague: heh, I was just about to offer | 12:18 |
| jhesketh | happy to | 12:18 |
| *** bswartz has quit IRC | 12:19 | |
| *** aysyd has joined #openstack-infra | 12:19 | |
| *** ashleighfarnham has joined #openstack-infra | 12:19 | |
| openstackgerrit | Sean Dague proposed openstack-infra/puppet-zuul: also add targeted status redirects to https https://review.openstack.org/179764 | 12:20 |
| sdague | jhesketh: thanks! | 12:20 |
| sdague | fungi: ok, that's hopefully the other fix | 12:20 |
| *** gordc has joined #openstack-infra | 12:22 | |
| mordred | jhesketh: ooh - you're still up | 12:24 |
| *** salv-orlando has joined #openstack-infra | 12:24 | |
| jhesketh | mordred: corright | 12:24 |
| *** ashleighfarnham has quit IRC | 12:24 | |
| mordred | jhesketh: I think I took care of your notes on https://review.openstack.org/177985 | 12:24 |
| jhesketh | mordred: looking | 12:24 |
| mordred | jhesketh: and good points, btw, thanks | 12:24 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/system-config: Serve up jquery-visibility.min.js for gerrit https://review.openstack.org/179765 | 12:26 |
| jhesketh | sdague: ^ | 12:27 |
| jhesketh | mordred: cool, lgtm | 12:27 |
| mordred | jhesketh: yay! | 12:27 |
| *** fhubik has quit IRC | 12:28 | |
| mordred | jhesketh: wow, we weren't jquerying already? | 12:28 |
| jhesketh | mordred: we weren't jquery-visibilitying | 12:28 |
| yolanda | mordred, i'm having issues with https://git.openstack.org/cgit/openstack-infra/nodepool/commit/?id=8b87890b978a54be10eff272f7bb6d9955e2d999 | 12:29 |
| *** fhubik has joined #openstack-infra | 12:29 | |
| yolanda | the change for nodepool and proxy | 12:29 |
| yolanda | you are not using that upstream, right? | 12:29 |
| *** luqas has joined #openstack-infra | 12:29 | |
| mordred | yolanda: we are not | 12:30 |
| yolanda | i'd say that this change doesn't work | 12:30 |
| yolanda | let me paste you the error | 12:30 |
| yolanda | mordred, http://paste.openstack.org/show/214843/ | 12:30 |
| mordred | yolanda: oh! we are | 12:30 |
| yolanda | that code is not raising a ProxyError, but a Timeout | 12:30 |
| yolanda | it happens when i do a manual image-update | 12:31 |
| mordred | yolanda: but we're not behind a proxy - so we've never seen that condition ourselves | 12:31 |
| yolanda | also, according to the logs, it even doesn't reach the provider_manager.py runTask method | 12:31 |
| mordred | hrm | 12:31 |
| *** hdd has joined #openstack-infra | 12:31 | |
| *** dizquierdo has quit IRC | 12:34 | |
| *** luqas has quit IRC | 12:34 | |
| *** alexpilotti has joined #openstack-infra | 12:35 | |
| *** woodster_ has joined #openstack-infra | 12:36 | |
| *** salv-orlando has quit IRC | 12:37 | |
| yolanda | mordred, so i guess the intention of that change is to override runtask, and add the proxy settings there, and call provider_manager.runTask instead of task_manage.runTask ? | 12:38 |
| *** _dguerri is now known as dguerri | 12:38 | |
| openstackgerrit | Merged openstack-infra/project-config: Don't short circuit when projects fails to clone https://review.openstack.org/179760 | 12:38 |
| *** dprince has joined #openstack-infra | 12:39 | |
| mordred | yolanda: yah - I think so | 12:40 |
| yolanda | looking at the code i'd say this runtask is not even reached, as i cannot see traces of errors there | 12:40 |
| yolanda | i even amended the code to accept Timeout exception, but same problem | 12:41 |
| mordred | yolanda: weird | 12:43 |
| yolanda | we have all of our images older than a day due to that error, and i'm unable to update manually, i'm testing more | 12:43 |
| *** radez_g0n3 is now known as radez | 12:44 | |
| *** dguerri is now known as _dguerri | 12:45 | |
| *** kgiusti has joined #openstack-infra | 12:46 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Convert node_set_provision_state to task https://review.openstack.org/177987 | 12:48 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Update recent Ironic exceptions https://review.openstack.org/177986 | 12:48 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Enhance error message in update_machine https://review.openstack.org/177985 | 12:48 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Enhance the OperatorCloud constructor https://review.openstack.org/179647 | 12:48 |
| *** dizquierdo has joined #openstack-infra | 12:48 | |
| *** dprince has quit IRC | 12:48 | |
| yolanda | mm, mordred, i think i got it | 12:51 |
| yolanda | command lines aren't using the thread interface , right? | 12:51 |
| *** mikeit has joined #openstack-infra | 12:53 | |
| mordred | yolanda: oh! you are right | 12:55 |
| *** links has joined #openstack-infra | 12:56 | |
| *** fhubik has quit IRC | 12:56 | |
| *** fhubik has joined #openstack-infra | 12:57 | |
| *** pc_m has quit IRC | 12:57 | |
| *** sdake has joined #openstack-infra | 12:57 | |
| anteaya | reed: I'm glad you are having a faster experience with git, I find it odd that ipv6 would be playing a role | 12:57 |
| *** macjack has left #openstack-infra | 12:57 | |
| yolanda | mordred, actually where it fails is in createservertask, and that's embedded inside a submittask method | 12:58 |
| *** jaypipes has joined #openstack-infra | 12:58 | |
| mordred | anteaya: that's a fascinating first sentence of the day | 12:58 |
| anteaya | I aim to facinate | 12:59 |
| anteaya | my job for the day is complete | 12:59 |
| anteaya | thank you | 12:59 |
| * mordred hands anteaya a japanese eggplant and a bowl of salsa | 12:59 | |
| anteaya | well thank you | 12:59 |
| fungi | mmmeggplantsalsa | 13:01 |
| *** bknudson has quit IRC | 13:01 | |
| *** BobH has joined #openstack-infra | 13:02 | |
| *** yfried is now known as yfried|afk | 13:02 | |
| anteaya | I never knew it could be good for breakfast | 13:03 |
| yolanda | mordred, ok, i think i got it, needed to add timeout exception as well to task_manager code. Will test it and send a patch upstream | 13:06 |
| *** _elmiko is now known as elmiko | 13:07 | |
| *** signed8bit has joined #openstack-infra | 13:09 | |
| *** signed8bit has quit IRC | 13:09 | |
| *** miqui has joined #openstack-infra | 13:09 | |
| jhesketh | sdague: I wonder with the live zuul status stuff if we should have a toggle to turn it on/off. If we merge this and find the optimisations aren't enough, turning it off again will require a number of reverts | 13:10 |
| *** jpeeler has joined #openstack-infra | 13:10 | |
| jhesketh | if we have a toggle switch it'll make it easier | 13:10 |
| *** mrmartin has joined #openstack-infra | 13:10 | |
| jhesketh | and also mean that if it works for the next 5 months but then zuul is under stupid amounts of load, we can change it back then too while we look at further optimisations | 13:11 |
| sdague | jhesketh: yes, I was thinking putting it behind an experimental flag would be good | 13:11 |
| sdague | which we could default off, then turn on inline, like the debug | 13:11 |
| sdague | so it would be easy to turn off | 13:11 |
| jhesketh | sdague: good idea | 13:11 |
| sdague | I think the only issue is we might need to change the caching model on review.openstack.org | 13:13 |
| *** hasharOut is now known as hashar | 13:13 | |
| sdague | so that it expires hideci more often | 13:13 |
| sdague | so that a change like that would deploy to people more quickly | 13:13 |
| jhesketh | sdague: what is it currently? | 13:13 |
| jhesketh | deployment is always going to be slow if we have to wait for jenkins+puppet etc | 13:14 |
| *** BobH has quit IRC | 13:14 | |
| sdague | yeh, that's not the concern, but js resources from the browser seem to have long lifetimes | 13:14 |
| *** dims has quit IRC | 13:14 | |
| sdague | so we might need an explicit expiration time to make browsers come back | 13:14 |
| *** dustins has joined #openstack-infra | 13:14 | |
| *** dims has joined #openstack-infra | 13:15 | |
| *** erlon_awaY has joined #openstack-infra | 13:15 | |
| jhesketh | sure, that's probably a good idea | 13:15 |
| mordred | jeblair_: when you return -I'm having an issue with gertty where it seems to have finished all of its syncing, and then I go to a change, and the change is not present, so it goes and figures out that it needs to sync new things | 13:15 |
| mordred | jeblair_: it's happening frequently - I would have expected all of the things to sync at some point, otherwise prep-for-being-offline becomes unhappy | 13:16 |
| fungi | mordred: is it a change for a subscribed project? | 13:16 |
| mordred | fungi: yes | 13:16 |
| fungi | odd | 13:16 |
| *** alexpilotti has quit IRC | 13:16 | |
| *** bcourt has joined #openstack-infra | 13:16 | |
| mordred | fungi, jeblair_: and in fact, like, every change I go to review in devstack-gate right now is doing this. | 13:17 |
| *** hdd has quit IRC | 13:17 | |
| *** yfried|afk is now known as yfried | 13:18 | |
| *** tiswanso has joined #openstack-infra | 13:18 | |
| *** hdd has joined #openstack-infra | 13:18 | |
| mrmartin | morning | 13:19 |
| fungi | mordred: hrm, running tip of master i am not having that problem on devstack-gate changes | 13:20 |
| *** jtriley has joined #openstack-infra | 13:20 | |
| * mordred is the SpamapS of gertty | 13:21 | |
| *** baoli has joined #openstack-infra | 13:21 | |
| mordred | wow. I JUST DID a ctrl-r on the devstack-gate screen, it finished, and tried to go to a review and I got the commits-not-present message | 13:22 |
| Shrews | mordred: give up. monday has won | 13:22 |
| fungi | problem with your local git cache? | 13:22 |
| mordred | fungi: my local git cache is my source tree | 13:22 |
| fungi | yeah, mine too. just didn't know if there might be something funky with your devtsack-gate clone in there | 13:23 |
| *** bswartz has joined #openstack-infra | 13:23 | |
| mordred | fungi: and I'm running tip of master too | 13:23 |
| mordred | fungi: well, it was doing it on other repos too | 13:23 |
| mordred | fungi: this is just the one I'm on in this morning's reviews | 13:23 |
| *** dizquierdo has quit IRC | 13:24 | |
| *** jcoufal_ has joined #openstack-infra | 13:24 | |
| * mordred gives Shrews a case of the mondays | 13:24 | |
| openstackgerrit | Merged openstack-infra/release-tools: Add a --stable-series argument to release_notes.py https://review.openstack.org/178194 | 13:24 |
| *** mestery_ is now known as mestery | 13:25 | |
| openstackgerrit | Merged openstack-infra/release-tools: Add option to format release notes for email https://review.openstack.org/178195 | 13:26 |
| *** amitgandhinz has joined #openstack-infra | 13:26 | |
| *** jcoufal has quit IRC | 13:27 | |
| *** erlon_awaY is now known as erlon | 13:27 | |
| mordred | fungi, sdague: 167412 - looks safe to merge now (this is temporarily enabling trove plugin stuff in d-g) - any gotchas you're aware of? | 13:28 |
| *** peristeri has joined #openstack-infra | 13:28 | |
| openstackgerrit | Merged openstack-infra/release-tools: Update rccut and rcdelivery to match Kilo process https://review.openstack.org/177202 | 13:28 |
| openstackgerrit | Jeremy Stanley proposed stackforge/gertty: Be more verbose on non-tagged versions https://review.openstack.org/179776 | 13:28 |
| sdague | mordred: yeh, it should be | 13:29 |
| fungi | lgtm, but didn't approve in case you were still reviewing | 13:29 |
| sdague | it's basically a hack to deal with the fact that project-config can't be pre gated | 13:29 |
| fungi | oh, though its dependency chain is still in need of work | 13:30 |
| fungi | but yeah, that one should be able to merge, it's just changes depending on it which need some updating | 13:30 |
| mordred | sdague: what does ;& mean in a base case statement? | 13:30 |
| *** sdake_ has joined #openstack-infra | 13:31 | |
| sdague | mordred: fall through | 13:31 |
| mordred | doesn't ;; also mean that? | 13:31 |
| sdague | no, ;; is break | 13:31 |
| mordred | well then | 13:31 |
| sdague | I thought at least | 13:31 |
| sdague | I had some sample code to ensure this did the right thing | 13:32 |
| *** dprince has joined #openstack-infra | 13:32 | |
| *** pabelanger has joined #openstack-infra | 13:32 | |
| *** pabelanger has joined #openstack-infra | 13:32 | |
| mordred | it's fine - I just didn't know that syntax | 13:32 |
| *** craigbr has joined #openstack-infra | 13:32 | |
| sdague | yeh, is a bash 4 thing apparently | 13:33 |
| sdague | because default shell case doesn't have fall through | 13:33 |
| sdague | which... is weird | 13:33 |
| sdague | because that's the point of case in my mind | 13:33 |
| *** bknudson has joined #openstack-infra | 13:33 | |
| *** sdake has quit IRC | 13:34 | |
| openstackgerrit | Sirushti Murugesan proposed openstack-infra/project-config: Add grenade jobs for Heat https://review.openstack.org/178352 | 13:35 |
| *** coolsvap is now known as coolsvap|afk | 13:37 | |
| fungi | dims: you're abandoning wip changes? | 13:38 |
| mordred | in which project(s)? | 13:39 |
| mordred | I would not like my WIP changes abandoned | 13:39 |
| fungi | oslo.* | 13:39 |
| mordred | k | 13:39 |
| mordred | as long as it's not infra | 13:39 |
| fungi | i'm restoring one that's still waiting on an ubuntu bug fix | 13:39 |
| dims | fungi: which one? (still tuning the script) | 13:40 |
| *** luqas has joined #openstack-infra | 13:40 | |
| fungi | dims: 120228 | 13:41 |
| fungi | unfortunately bugs in ubuntu lts take a while to get fixed | 13:42 |
| dims | thanks fungi! | 13:42 |
| openstackgerrit | Merged openstack-infra/shade: Enhance error message in update_machine https://review.openstack.org/177985 | 13:42 |
| openstackgerrit | Merged openstack-infra/shade: Update recent Ironic exceptions https://review.openstack.org/177986 | 13:42 |
| fungi | dims: you're welcome | 13:43 |
| mordred | sdague: could we do something clever to make a passthrough config to devstack? every time I approve a patch like 179662 it makes me think we're doing something wrong | 13:43 |
| sdague | mordred: 179662 ? | 13:43 |
| *** prad has joined #openstack-infra | 13:43 | |
| sdague | oh, yes, we already have that | 13:43 |
| openstackgerrit | Merged openstack-dev/pbr: Allow overwriting sphinx builder from command line https://review.openstack.org/117478 | 13:43 |
| *** jeblair_ is now known as jeblair | 13:44 | |
| *** superdan is now known as dansmith | 13:44 | |
| mordred | sdague: yeah? what is it? | 13:44 |
| sdague | grep for enable_plugin in project-config | 13:44 |
| sdague | DEVSTACK_LOCAL_CONFIG | 13:44 |
| sdague | that's it | 13:44 |
| mrmartin | fungi: the refactoring for puppet-askbot staging is almost ready, I just found some issues with askbot-devel. so it have issues with ubuntu precise cloudimg, so I made tests with trusty, and it works well | 13:45 |
| openstackgerrit | Yanis Guenane proposed openstack-infra/project-config: Add support for backport-potential commit flag https://review.openstack.org/175849 | 13:45 |
| mrmartin | so, what is the proper way of ask.o.o upgrade from precise to trusty? | 13:45 |
| mordred | sdague: col | 13:45 |
| mordred | cool | 13:45 |
| mordred | we shoudl start using that more | 13:45 |
| sdague | mordred: yes | 13:46 |
| *** mriedem has joined #openstack-infra | 13:46 | |
| *** dboik has joined #openstack-infra | 13:46 | |
| sdague | it gets a little janky with multi lines in the project config files, we probably want to come up with a better way to make that work | 13:46 |
| fungi | mrmartin: we build a new server running trusty and migrate from the current one to the new one (similar to how we migrated from the non-managed server to the current managed one) | 13:47 |
| mrmartin | ok great | 13:47 |
| sdague | but the intent was about 1/2 of d-g vars should be able to go away | 13:47 |
| sdague | just never got there | 13:47 |
| mordred | sdague: ++ | 13:47 |
| mordred | sdague: well, I'm going to start -1ing new variables | 13:47 |
| mrmartin | can we disable the puppet on the precise ask.o.o? | 13:47 |
| *** lascii is now known as alaski | 13:47 | |
| *** luqas has quit IRC | 13:48 | |
| *** julim has joined #openstack-infra | 13:48 | |
| mrmartin | This new patch contains some breaking changes, removes the slotX directory structure, and consume askbot-devel directly from github instead of pip | 13:48 |
| mordred | fungi: speaking of that ^^ https://review.openstack.org/#/c/179581/ | 13:48 |
| mordred | fungi: I'd argue we should stop spinning up new precise servers unless we really mean it | 13:49 |
| mrmartin | yeah, that's the case we like to upgrade the precise nodes to trusty | 13:50 |
| mordred | fungi: also, while I'm bothering you with things that do not matter (good morning): https://review.openstack.org/#/q/status:open+branch:master+topic:hound,n,z - and there is a server running at hound.openstack.org with those changes if you want to check it out | 13:50 |
| fungi | mordred: yep, saw that. +2 but we should get a little more consensus. worth mentioning briefly in the meeting tomorrow perhaps | 13:50 |
| fungi | mordred: on the default-to-trusty change i mean | 13:51 |
| *** luqas has joined #openstack-infra | 13:51 | |
| mordred | fungi: ++ | 13:51 |
| fungi | mrmartin: well, if we merge changes that won't work for the current production server, we're basically setting ourselves up to be unable to update that server further until we migrate... so we should be really sure we can arrange the migration maintenance in a short timeframe | 13:51 |
| mrmartin | ok. | 13:52 |
| samueldmq | mordred, ping - just saw your review on 'Pass the flag for Identity v3 API only jobs' | 13:52 |
| mordred | samueldmq: yah | 13:52 |
| mrmartin | but what you think about temporarily commenting out the askbot related manifests there? | 13:52 |
| samueldmq | mordred, so I just need to put what I need in DEVSTACK_LOCAL_CONFIG in my project-config change | 13:52 |
| samueldmq | mordred, right ? | 13:52 |
| mordred | samueldmq: yup! | 13:52 |
| samueldmq | mordred, that's happening in a lot of places | 13:53 |
| mordred | samueldmq: no need to teach d-g about it at all | 13:53 |
| *** emagana has quit IRC | 13:53 | |
| sdague | mordred: +1 | 13:53 |
| samueldmq | mordred, do we need to refactor all the other places we are already doing this ? | 13:53 |
| fungi | mrmartin: if it's not too hard to adjust it so that it works on trusty for the ask-dev server but continues working on precise for the production ask server, that would make things a little easier to test and coordinate, but if that's a lot of extra work then i think just updating the module and migrating the server within a maintenance is probably okay | 13:53 |
| samueldmq | mordred, I can do this refactor if you want me too | 13:53 |
| mordred | samueldmq: it would be nice - although as sdague mentions, if there are jobs taht need to set more than one thing, it's a little odd syntax-wise, so we might want to sort that | 13:54 |
| mrmartin | ok. | 13:54 |
| mordred | samueldmq: but probably all of the jobs that just set one thing that's just a passthrough and does not trigger additional logic in d-g are good candidates | 13:54 |
| mrmartin | we have some open issues with askbot-devel which must be closed first, so when it works well, I'll come back with some plan | 13:54 |
| samueldmq | mordred, great I will do it later today | 13:54 |
| samueldmq | mordred, will ping you once I submit the patch, thanks | 13:54 |
| openstackgerrit | Claudiu Belu proposed openstack-infra/project-config: Adds pypi-jobs to compute-hyperv https://review.openstack.org/179782 | 13:55 |
| samueldmq | mordred, yeah I will get those good candidates first, it should already be a nice cleanup | 13:55 |
| fungi | mrmartin: great. i'll try to make sure i keep reviewing those a priority, but ping me when you have related changes i haven't noticed/reviewed just in case | 13:55 |
| *** luqas has quit IRC | 13:55 | |
| mrmartin | fungi, ok nothing urgent here | 13:55 |
| mrmartin | we have only one large issue with google-auth but it still requires some work from askbot devs. | 13:56 |
| *** chlong has joined #openstack-infra | 13:58 | |
| fungi | okay, thanks for the update! | 14:01 |
| *** btully has joined #openstack-infra | 14:02 | |
| *** yfried is now known as yfried|afk | 14:02 | |
| *** btully has quit IRC | 14:03 | |
| *** fhubik has quit IRC | 14:03 | |
| *** zz_naotok has quit IRC | 14:04 | |
| *** zz_naotok has joined #openstack-infra | 14:06 | |
| *** eharney has joined #openstack-infra | 14:08 | |
| *** ashleighfarnham has joined #openstack-infra | 14:08 | |
| *** otter768 has joined #openstack-infra | 14:08 | |
| *** chlong has quit IRC | 14:08 | |
| *** ayoung has joined #openstack-infra | 14:09 | |
| *** dguitarbite has quit IRC | 14:09 | |
| *** _dguerri is now known as dguerri | 14:12 | |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Add query for nova unit test bug 1451389 https://review.openstack.org/179788 | 14:13 |
| openstack | bug 1451389 in OpenStack Compute (nova) "Nova gate broke due to failed unit test" [Critical,In progress] https://launchpad.net/bugs/1451389 - Assigned to Sean Dague (sdague) | 14:13 |
| *** otter768 has quit IRC | 14:13 | |
| *** dguerri is now known as _dguerri | 14:13 | |
| *** _dguerri is now known as dguerri | 14:13 | |
| *** ashleighfarnham has quit IRC | 14:13 | |
| *** e0ne has quit IRC | 14:14 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:14 | |
| *** salv-orlando has joined #openstack-infra | 14:16 | |
| *** rossella_s has quit IRC | 14:16 | |
| *** rossella_s has joined #openstack-infra | 14:16 | |
| *** mrunge has quit IRC | 14:17 | |
| *** pc_m has joined #openstack-infra | 14:17 | |
| *** signed8bit has joined #openstack-infra | 14:19 | |
| *** yfried|afk is now known as yfried | 14:22 | |
| *** BharatK has quit IRC | 14:23 | |
| *** stevemar has joined #openstack-infra | 14:23 | |
| *** gsagie has joined #openstack-infra | 14:24 | |
| *** chlong has joined #openstack-infra | 14:26 | |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Add query for nova unit test bug 1451389 https://review.openstack.org/179788 | 14:27 |
| openstack | bug 1451389 in OpenStack Compute (nova) "Nova gate broke due to failed unit test" [Critical,In progress] https://launchpad.net/bugs/1451389 - Assigned to Sean Dague (sdague) | 14:27 |
| *** dboik_ has joined #openstack-infra | 14:30 | |
| *** salv-orlando has quit IRC | 14:32 | |
| *** armax has joined #openstack-infra | 14:33 | |
| *** gsagie has left #openstack-infra | 14:33 | |
| *** gokrokve_ has joined #openstack-infra | 14:34 | |
| *** dboik has quit IRC | 14:34 | |
| openstackgerrit | Merged openstack-infra/devstack-gate: Temporarily add trove plugin configuration https://review.openstack.org/167412 | 14:34 |
| openstackgerrit | Paul Belanger proposed openstack-infra/system-config: WIP: Initial commit for grafana https://review.openstack.org/179208 | 14:34 |
| *** zz_jgrimm is now known as jgrimm | 14:35 | |
| *** ajmiller has joined #openstack-infra | 14:35 | |
| openstackgerrit | Gal Sagie proposed openstack-infra/project-config: Enable a check job for stackforge/dragonflow https://review.openstack.org/175152 | 14:36 |
| *** gokrokve has quit IRC | 14:37 | |
| *** jgrimm is now known as zz_jgrimm | 14:38 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/project-config: Run shade functional tests on neutron and nova-net https://review.openstack.org/179797 | 14:40 |
| mordred | sdague: is that correct? ^^ | 14:40 |
| *** wenlock has joined #openstack-infra | 14:40 | |
| *** mtanino has joined #openstack-infra | 14:40 | |
| *** unicell1 has joined #openstack-infra | 14:40 | |
| *** wenlock1 has joined #openstack-infra | 14:40 | |
| sdague | I don't think you need L54 in - https://review.openstack.org/#/c/179797/1/jenkins/jobs/shade.yaml,cm | 14:41 |
| *** packet has joined #openstack-infra | 14:41 | |
| sdague | otherwise, I think this is right | 14:42 |
| *** unicell has quit IRC | 14:42 | |
| *** emagana has joined #openstack-infra | 14:42 | |
| mordred | sdague: what if we change the default in the future? | 14:44 |
| *** claudiub_ has quit IRC | 14:44 | |
| mtreinish | mordred: heh, you really think that'll happen? :) | 14:46 |
| *** Somay has joined #openstack-infra | 14:46 | |
| *** mase_x200 has quit IRC | 14:46 | |
| mordred | mtreinish: *sad trombone* | 14:47 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/system-config: Create rubygems mirror from rubygems.org https://review.openstack.org/178026 | 14:47 |
| *** tonytan4ever has joined #openstack-infra | 14:52 | |
| sdague | mordred: so... if the default flips we're going to have to update it a bunch of places in project config | 14:52 |
| mordred | sdague: k. | 14:53 |
| sdague | I'm +2 either way | 14:53 |
| sdague | just saying it's not strictly needed | 14:53 |
| *** alexpilotti has joined #openstack-infra | 14:53 | |
| *** vhoward has quit IRC | 14:54 | |
| *** vhoward has joined #openstack-infra | 14:55 | |
| mordred | nod | 14:55 |
| *** gokrokve_ has quit IRC | 14:55 | |
| AJaeger | mordred, just gave a +2 - feel free to self-approve once you finished the discussion here ;) | 14:55 |
| *** gokrokve has joined #openstack-infra | 14:55 | |
| *** links has quit IRC | 14:56 | |
| *** gokrokve has quit IRC | 14:57 | |
| openstackgerrit | Peter Hamilton proposed openstack/requirements: Update PyKMIP version in requirements https://review.openstack.org/164449 | 14:58 |
| *** gokrokve has joined #openstack-infra | 14:58 | |
| *** sushilkm has joined #openstack-infra | 14:58 | |
| *** sushilkm has left #openstack-infra | 14:58 | |
| openstackgerrit | Victor Stinner proposed openstack/requirements: Add mysqlclient dependency https://review.openstack.org/179745 | 14:58 |
| *** craigbr has quit IRC | 15:00 | |
| *** sdake has joined #openstack-infra | 15:01 | |
| *** sarob has joined #openstack-infra | 15:02 | |
| *** sdake__ has joined #openstack-infra | 15:03 | |
| *** sdake_ has quit IRC | 15:03 | |
| openstackgerrit | Merged openstack-infra/puppet-jeepyb: Modernize template syntax https://review.openstack.org/168147 | 15:03 |
| openstackgerrit | Merged openstack-infra/puppet-jeepyb: Parameterize git revision of jeepyb https://review.openstack.org/170960 | 15:03 |
| openstackgerrit | Merged openstack-infra/puppet-jeepyb: Add helpers to execute manage_projects https://review.openstack.org/171538 | 15:03 |
| *** sarob has quit IRC | 15:03 | |
| *** sarob has joined #openstack-infra | 15:03 | |
| *** sdake has quit IRC | 15:07 | |
| mordred | jeblair: so - basically every time I go to a new change now, I get the no-revs error, and then gertty adds about 100 things to the queue, processes them, then I have the change | 15:07 |
| mordred | jeblair: the 100-ish number is consistent | 15:07 |
| *** asselin has joined #openstack-infra | 15:07 | |
| mordred | jeblair: also, I have nothing in my logs to indicate that gertty thinks anything is broken | 15:07 |
| *** ociuhandu_ has joined #openstack-infra | 15:07 | |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: Disable voting for puppet-syntax-future on havana/icehouse https://review.openstack.org/179809 | 15:09 |
| *** arxcruz has quit IRC | 15:10 | |
| pabelanger | Looking for some feedback for provisioning grafana. Right now, there is a manual process for connecting data sources. If you read here: http://docs-draft.openstack.org/08/179208/5/check/gate-system-config-docs/e8333d4//doc/build/html/grafana.html | 15:11 |
| pabelanger | however, need to setup default admin credentials and looking for which openstack-infra group to be using | 15:12 |
| pabelanger | and even a default location for password | 15:12 |
| openstackgerrit | Merged openstack-infra/puppet-jenkins: Allow for plugin pinning https://review.openstack.org/160549 | 15:12 |
| pabelanger | Basically, once I am done with that info, it would be ready for review | 15:12 |
| openstackgerrit | Merged openstack-infra/tripleo-ci: Bump up delorean pinning https://review.openstack.org/177176 | 15:14 |
| *** claudiub has joined #openstack-infra | 15:14 | |
| *** spzala has joined #openstack-infra | 15:14 | |
| mordred | pabelanger: I don't suppose it's possible to set those things directly in the grafana db is it? | 15:15 |
| pabelanger | mordred, checking | 15:16 |
| mordred | because, I mean, I know that we don't want to create graphs through the web | 15:16 |
| jeblair | pabelanger: i was talking with timrc recently about grafana and he suggested that it would be possible to write a short python script to translate a yaml file and write it into the db | 15:16 |
| openstackgerrit | Merged openstack-infra/puppet-jenkins: Use jenkins group instead of nogroup https://review.openstack.org/175368 | 15:17 |
| jeblair | pabelanger: iirc, he said the db format was basically just a json blob describing each dashboard | 15:17 |
| pabelanger | jeblair, Ya, that was mentioned last week. Seems like the right approach | 15:17 |
| jeblair | timrc meet pabelanger | 15:17 |
| jeblair | pabelanger meet timrc | 15:17 |
| pabelanger | timrc, ohai | 15:18 |
| *** jcoufal_ has quit IRC | 15:18 | |
| jeblair | pabelanger: so maybe check with timrc and see if he's started anything, and if not, fight it out between the two of you? :) | 15:18 |
| pabelanger | jeblair, timrc Sure. I'll dive into DB right now to see what it looks like | 15:20 |
| openstackgerrit | Merged openstack-infra/project-config: Run shade functional tests on neutron and nova-net https://review.openstack.org/179797 | 15:20 |
| *** craigbr has joined #openstack-infra | 15:22 | |
| jeblair | mordred: i'm getting asymtoptically closer to post-travel usefulness, i'll get with you on gertty soon -- though in the mean time, if you have time to capture a 'refresh, missing git refs' cycle with debug logging enabled ("-d" on command line); and probably quit shortly after to keep log size small, that will probably help. | 15:22 |
| mordred | jeblair: on it | 15:23 |
| *** dannywilson has joined #openstack-infra | 15:23 | |
| openstackgerrit | Merged openstack-infra/puppet-log_processor: Setting dependencies in metadata.json https://review.openstack.org/153821 | 15:24 |
| mordred | jeblair: are gertty logs safe to pastebin? | 15:24 |
| *** craigbr has quit IRC | 15:25 | |
| *** sarob_ has joined #openstack-infra | 15:26 | |
| jeblair | mordred: i believe so | 15:26 |
| *** hdd has quit IRC | 15:27 | |
| *** dannywilson has quit IRC | 15:27 | |
| mordred | jeblair: http://paste.openstack.org/show/214861 | 15:28 |
| *** sarob has quit IRC | 15:28 | |
| *** e0ne has joined #openstack-infra | 15:28 | |
| *** sks has quit IRC | 15:28 | |
| *** sabeen has joined #openstack-infra | 15:29 | |
| *** dannywilson has joined #openstack-infra | 15:29 | |
| *** ashleighfarnham has joined #openstack-infra | 15:29 | |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder: Added parallelization options https://review.openstack.org/75514 | 15:30 |
| *** Somay has quit IRC | 15:30 | |
| *** ociuhandu_ has quit IRC | 15:31 | |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder: Added gerrit trigger missing options https://review.openstack.org/167175 | 15:32 |
| *** maishsk has joined #openstack-infra | 15:32 | |
| *** mtanino has quit IRC | 15:33 | |
| *** harlowja_at_home has joined #openstack-infra | 15:35 | |
| *** sarob has joined #openstack-infra | 15:35 | |
| openstackgerrit | Merged openstack-infra/askbot-theme: updated theme to match version 0.7.53 https://review.openstack.org/177485 | 15:35 |
| *** yamahata has joined #openstack-infra | 15:36 | |
| fungi | jeblair: also when you're settled back in, i've noticed that gertty doesn't deal well with reviewing multi-parent changes like a merge commit (it dies in flames, actually). i haven't had time to dig into possible fixes for it, but will write a story with the trace in a bit | 15:36 |
| openstackgerrit | Merged openstack-infra/askbot-theme: deleted unused file https://review.openstack.org/177486 | 15:36 |
| jeblair | fungi: yeah, there's a "TODO: support multiple parents" in the database schema :) | 15:36 |
| openstackgerrit | Merged openstack-infra/askbot-theme: css changes for language nav and dropdown menu https://review.openstack.org/179472 | 15:36 |
| jeblair | fungi: so it's a substantial project | 15:37 |
| fungi | jeblair: i figured i wasn't the only one to spot that. good to know | 15:37 |
| jeblair | fungi: though, i mean, hopefully "not crashing" might be a less substantial bugfix | 15:37 |
| fungi | i thought a stopgap might be to just treat a multi-parent change as a non-parented change and move on | 15:38 |
| clarkb | jeblair: question about gearman, if you send a RESET_ABILITIES does that clear out a pending GET JOB? | 15:38 |
| clarkb | zaro: I think ^ may be the problem and it is much less complicated than I thought | 15:38 |
| jeblair | clarkb: hehe, i do not immediately know the answer to your pop quiz! i'll look | 15:38 |
| *** sputnik13 has quit IRC | 15:38 | |
| clarkb | jeblair: we are seeing nodepool nodes remain "ready" for long periods of time and reading jenkins logs inidicates they go in that state if they send a GET JOB then immediately register functions | 15:39 |
| *** sarob_ has quit IRC | 15:39 | |
| clarkb | jeblair: I had thought the run loop was getting live locked but in this case I think it may be the server state is changing around the GET JOB because we RESET ABILITIES | 15:39 |
| jeblair | clarkb: we send 'get job' before registering functions? that sounds buggy, right? | 15:40 |
| clarkb | jeblair: yes, and yes I think that is a bug | 15:40 |
| clarkb | jeblair: however we do receive the JOB_UNIQ response we just never seem to handle it | 15:40 |
| *** mfisch has quit IRC | 15:40 | |
| jeblair | clarkb: JOB_ASSIGN_UNIQ? | 15:40 |
| *** Longgeek_ has joined #openstack-infra | 15:40 | |
| *** ildikov has quit IRC | 15:41 | |
| clarkb | jeblair: I have a change up with logs inthe ocmmit message https://review.openstack.org/#/c/178929/ | 15:41 |
| *** salv-orlando has joined #openstack-infra | 15:41 | |
| clarkb | jeblair: you should probably ignore the reset of the commit message though as zaro indicates my change does not work beyond the mvn run tests (which I Had run locally) | 15:41 |
| *** Guest65197 has joined #openstack-infra | 15:41 | |
| jeblair | clarkb: that makes it sound like geard is doing what we want in this situation (which is behavior i would expect -- i would expect geard to react to adding functions and then wake the worker as appropriate) | 15:42 |
| *** Longgeek has quit IRC | 15:42 | |
| jeblair | clarkb: oh, but maybe your question is -- does geard think the job it assigned is no longer running? | 15:42 |
| clarkb | jeblair: correct | 15:42 |
| jeblair | clarkb: got it | 15:42 |
| clarkb | jeblair: because the worker ends up not running any job at that point and does not go offline or emit events saying it did anything | 15:43 |
| clarkb | jeblair: so we essentially have a worker that can run a job but never does | 15:43 |
| clarkb | (and this disrupts nodepool's deficit calculations) | 15:43 |
| jeblair | clarkb: i do not think geard would take any action there; i would expect it to think the worker is running the job but slowly. however, that would suggest that zuul would see stuck jobs | 15:43 |
| timrc | pabelanger, Hi, reading up. | 15:45 |
| timrc | jeblair, pabelanger I actually think it would be best if we used the grafana rest API to update / overwrite dashboards. | 15:45 |
| *** Guest84883 has quit IRC | 15:46 | |
| timrc | I have something somewhere on my laptop which does that. The import script would then be a simple matter of converting yaml to json and then using the rest API to post new dashboards / update existing ones. | 15:46 |
| jeblair | timrc: ooh there's one of those? that sounds even better | 15:46 |
| timrc | jeblair, Yep. | 15:47 |
| clarkb | jeblair: I am not sure that we have seen that, I think I have seen jobs earlier in the pipeline that are queued but not running even though jobs that run on the same node type are running later in the pipeline | 15:47 |
| clarkb | I will keep an eye out for ^ today | 15:47 |
| *** mwagner_lap has quit IRC | 15:47 | |
| timrc | jeblair, At the time we talked I didn't realize there was a rest API but discovered there was on the long flight back from Ireland. | 15:47 |
| pabelanger | timrc, Oh, nice. Didn't know there was restapi | 15:47 |
| clarkb | fungi: fyi https://github.com/openstack-infra/git-review/pull/8 | 15:49 |
| jeblair | clarkb: i'm guessing we don't have any info on the job that was assigned (and apparently ignored)? | 15:49 |
| clarkb | jeblair: not that I could fine | 15:49 |
| clarkb | s/fine/find/ | 15:49 |
| fungi | clarkb: conveniently, git-review does not enforce a signed icla, so i'll be happy to git-am that pr into gerrit on the author's behalf, unless you wanted dibs | 15:51 |
| clarkb | fungi: well I am wondering if we should close won't fix because it runs on windows natively last I checked | 15:52 |
| *** mattfarina has joined #openstack-infra | 15:52 | |
| clarkb | but I don't know enough about windows to know why one would try to use a cygwin utility without cygwin paths | 15:52 |
| openstackgerrit | Darragh Bailey proposed openstack-infra/jenkins-job-builder: Ensure tests run with ConfigParser object https://review.openstack.org/179822 | 15:53 |
| *** sigmavirus24 is now known as Apsu_ | 15:53 | |
| pabelanger | timrc, jeblair: okay, sweet. API is there, simple json blob returned. Going to now hack up a dashboard now | 15:53 |
| pabelanger | timrc, do you have any code started? | 15:53 |
| *** Apsu_ is now known as sigmavirus24 | 15:53 | |
| *** mtanino has joined #openstack-infra | 15:54 | |
| timrc | pabelanger, Not really, I just validated I could create and overwrite JSON dashboards | 15:54 |
| timrc | via the rest api | 15:54 |
| pabelanger | roger | 15:54 |
| pabelanger | going to go get some food, but will start work on something after lunch | 15:55 |
| timrc | pabelanger, Do you think it would be best if the import code ran on the grafana host itself and posted against the localhost:3000? I didn't know how that would actually work. | 15:55 |
| *** hdd has joined #openstack-infra | 15:55 | |
| pabelanger | timrc, My thoughts are to have it work like JJB does. At least thats my preference | 15:56 |
| timrc | That way we could use http basic auth and store those creds in hiera. Maybe a post pipeline job to ssh and curl? | 15:56 |
| pabelanger | I've done both JJB local and remote | 15:56 |
| timrc | pabelanger, Ah okay. | 15:56 |
| fungi | clarkb: unless maybe people trapped on 'doze prefer a cygwin-provided shell/environment (i haven't tried it in nearly two decades at this point, honestly) | 15:56 |
| *** devvesa has quit IRC | 15:56 | |
| *** armax has quit IRC | 15:56 | |
| clarkb | jeblair: I wonder if the sequence of events is, worker send GET JOB, server receives GET JOB and sends JOB ASSIGN UNIQ, worker sends RESET ABILITIES, worker gets JOB ASSIGN UNIQ, server gets RESET ABILITIES, "kills" previously assigned UNIQ | 15:56 |
| *** hashar is now known as hasharAway | 15:57 | |
| clarkb | fungi: ya in any case its probably worht a follow up with the author to understand the usecase then they/you/we can push the change to gerrit if needed | 15:57 |
| *** craigbr has joined #openstack-infra | 15:57 | |
| *** maishsk has quit IRC | 15:58 | |
| *** sabeen2 has joined #openstack-infra | 15:58 | |
| jeblair | clarkb: i don't believe geard has the assumptiong that a worker can only have one job, so i don't think it would alter the previously assigned job | 15:58 |
| fungi | clarkb: i am asking it on the pr now | 15:58 |
| *** whayutin_ has joined #openstack-infra | 15:59 | |
| jeblair | clarkb: my current understanding is: zuul submits job to geard, geard assigns it to worker, worker receives assignment and ignores job, geard assumes job is running as long as that worker is connected, and so does zuul | 15:59 |
| *** ashleighfarnham has quit IRC | 15:59 | |
| clarkb | jeblair: I am almost sure that isn't what happens because we will have a lot of leaked nodes but no leaked jobs | 15:59 |
| jeblair | clarkb: so i would expect to see stuck jobs in zuul unless we were "lucky" enough for them to be canceled gate jobs or something. | 15:59 |
| jeblair | clarkb: i agree, but i can't falsify that yet :/ | 15:59 |
| clarkb | ok | 15:59 |
| *** asselin has quit IRC | 16:00 | |
| jeblair | clarkb: even zuul's "look for missing jobs" thing, which is mostly a holdover from jenkins days, just checks that gearman knows about the job | 16:00 |
| jeblair | geard that is | 16:00 |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder: (hipchat) Accept list of hipchat rooms. https://review.openstack.org/153831 | 16:00 |
| *** maishsk has joined #openstack-infra | 16:01 | |
| *** gokrokve_ has joined #openstack-infra | 16:01 | |
| *** gokrokve_ has quit IRC | 16:01 | |
| *** sabeen has quit IRC | 16:01 | |
| fungi | clarkb: actually, the pr mentions http://babun.github.io/ so maybe that's the key | 16:01 |
| *** whayutin_ has quit IRC | 16:02 | |
| *** SumitNaiksatam has quit IRC | 16:02 | |
| *** whayutin_ has joined #openstack-infra | 16:02 | |
| openstackgerrit | Ken Dreyer proposed stackforge/python-jenkins: support six 1.3 https://review.openstack.org/179826 | 16:02 |
| *** weshay has quit IRC | 16:02 | |
| *** Guest65197 has quit IRC | 16:02 | |
| clarkb | fungi: maybe, I am just confused becuase I would expect the cygwin path to work in cygwin and the non cygwin path to work in normal windows, we shouldn't be crossing path streams | 16:03 |
| *** emagana has quit IRC | 16:03 | |
| fungi | also, my wife and i just had an offer accepted on a house, so i'm likely not around much later today as we're scrambling to get mortgage paperwork finalized | 16:03 |
| clarkb | but that likelyreflects my poor understanding of windows | 16:03 |
| clarkb | fungi: gl | 16:03 |
| *** SumitNaiksatam has joined #openstack-infra | 16:03 | |
| fungi | clarkb: i consider that i have basically _no_ understanding of windows at this point, except maybe still how to spell it | 16:04 |
| *** gokrokve has quit IRC | 16:04 | |
| *** emagana has joined #openstack-infra | 16:04 | |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder: Add support for 'raw' to inject direct xml to job configs. https://review.openstack.org/160158 | 16:05 |
| *** dguerri is now known as _dguerri | 16:06 | |
| jeblair | fungi: congrats! | 16:06 |
| *** mikeit has quit IRC | 16:06 | |
| *** unicell1 has quit IRC | 16:06 | |
| *** gokrokve has joined #openstack-infra | 16:06 | |
| *** gokrokve has quit IRC | 16:06 | |
| fungi | thanks--hopefully no expensive surprises on the inspection | 16:07 |
| zaro | morning | 16:08 |
| *** mpavone has quit IRC | 16:08 | |
| *** yamahata has quit IRC | 16:09 | |
| *** otter768 has joined #openstack-infra | 16:09 | |
| clarkb | zaro: good morning, jeblair and I were discussing possibility that RESET ABILITIES being sent is potentially related to the gearman plugin trouble. jeblair was looking at the geard side. Might be worth reading some of the recent scrollback for that and let us know what you think | 16:09 |
| *** sergsh_ has quit IRC | 16:10 | |
| *** sergsh has quit IRC | 16:10 | |
| *** markus_z has quit IRC | 16:11 | |
| *** salv-orlando has quit IRC | 16:12 | |
| *** maishsk has quit IRC | 16:13 | |
| *** otter768 has quit IRC | 16:14 | |
| *** yfried has quit IRC | 16:14 | |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: Disable voting for puppet-syntax-future on havana/icehouse https://review.openstack.org/179809 | 16:14 |
| EmilienM | AJaeger: thx for your review ^ | 16:14 |
| *** ashleighfarnham has joined #openstack-infra | 16:14 | |
| *** scheuran has quit IRC | 16:15 | |
| *** hdd has quit IRC | 16:17 | |
| *** alexsyip has joined #openstack-infra | 16:17 | |
| *** zz_jgrimm is now known as jgrimm | 16:18 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config: Disable voting for puppet-syntax-future on havana/icehouse https://review.openstack.org/179809 | 16:20 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config: Disable voting for puppet-syntax-future on havana/icehouse https://review.openstack.org/179809 | 16:20 |
| *** _dguerri is now known as dguerri | 16:21 | |
| mordred | fungi: woot! | 16:24 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: use optimized url for zuul status https://review.openstack.org/170935 | 16:24 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: add dbg function to hideci https://review.openstack.org/179573 | 16:24 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: add jquery visibility https://review.openstack.org/179362 | 16:24 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: create proxy for testing hideci https://review.openstack.org/179360 | 16:24 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: Revert "Revert zuul status in gerrit" https://review.openstack.org/179361 | 16:24 |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder: Support the use of lists in trigger-build builder. https://review.openstack.org/150597 | 16:25 |
| *** maishsk has joined #openstack-infra | 16:27 | |
| *** whoops has joined #openstack-infra | 16:33 | |
| *** yamahata has joined #openstack-infra | 16:34 | |
| *** Guest19502 has joined #openstack-infra | 16:34 | |
| *** unicell has joined #openstack-infra | 16:37 | |
| *** baoli has quit IRC | 16:38 | |
| *** maishsk has quit IRC | 16:40 | |
| *** flashgordon is now known as jogo | 16:40 | |
| *** MarkAtwood has joined #openstack-infra | 16:40 | |
| *** emagana has quit IRC | 16:41 | |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: use optimized url for zuul status https://review.openstack.org/170935 | 16:42 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: add dbg function to hideci https://review.openstack.org/179573 | 16:42 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: add jquery visibility https://review.openstack.org/179362 | 16:42 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: create proxy for testing hideci https://review.openstack.org/179360 | 16:42 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: Revert "Revert zuul status in gerrit" https://review.openstack.org/179361 | 16:42 |
| *** emagana has joined #openstack-infra | 16:42 | |
| sdague | greghaynes: doh, there was a pep8 at the bottom of the last patch | 16:42 |
| greghaynes | sdague: fun | 16:43 |
| *** harlowja_at_home has quit IRC | 16:43 | |
| sdague | greghaynes: so the issue with nginx is that it's not just a proxy, it's also rewriting urls inside of content on the fly to get sub resources to go back to the proxy | 16:44 |
| greghaynes | sdague: yea, you can do all that fun stuff in nginx, but it would probably end up being about as much config as you had code | 16:44 |
| sdague | and if anyone figures out how to do that for the openid request from gerrit so that you can actually log in, that would be super awesome | 16:44 |
| sdague | greghaynes: ok, well that would be beyond my level of experience :) | 16:44 |
| *** hdd has joined #openstack-infra | 16:44 | |
| sdague | and this requires nothing beyond python stdlib except requests | 16:45 |
| *** cody-somerville has joined #openstack-infra | 16:45 | |
| *** sarob_ has joined #openstack-infra | 16:46 | |
| greghaynes | yep, and its readable, which is a big win over $webserver_config | 16:47 |
| sdague | I wrote a version to test ui in elastic recheck, then thought "I wonder if this would work in gerrit" | 16:48 |
| sdague | which led to this tweet - https://twitter.com/sdague/status/594102249875451904 | 16:48 |
| *** sarob_ has quit IRC | 16:48 | |
| sdague | because, gerrit's post back channel as a single page app, is kind of a pain | 16:48 |
| mordred | reed: ping - question about sending out a survey to people in #openstack-horizon if yo uhave a minute | 16:49 |
| reed | mordred, sure | 16:49 |
| *** gokrokve has joined #openstack-infra | 16:50 | |
| fungi | clarkb: revisiting that git-review patch briefly, it looks like the use case is git running in cygwin but python running outside cygwin, so git-review is getting conflicting file paths between what its callouts to git return and what it tries to access via python | 16:50 |
| * fungi has no idea if this is a sane/common configuration | 16:50 | |
| *** bcornec has joined #openstack-infra | 16:52 | |
| bcornec | Hello everybody. I was directed to this IRC chat room by Devananda while reporting an issue I had on the ironic channel. Hope that's fine. | 16:53 |
| *** hdd has quit IRC | 16:53 | |
| mordred | bcornec: hey! welcome to the madhouse | 16:53 |
| bcornec | The issue I had is with i18n with the git review command. I put info on the wiki so that others can work around it, but I think it should behave differently (meaning without the need to abandon my LANGUAGE setup). Cf: https://wiki.openstack.org/wiki/Documentation/HowTo/FirstTimers#i18n | 16:54 |
| bcornec | Thx mordred ! | 16:54 |
| *** maishsk has joined #openstack-infra | 16:54 | |
| openstackgerrit | Terry Wilson proposed openstack/requirements: global-requirements: Add ovs https://review.openstack.org/177891 | 16:54 |
| *** dizquierdo has joined #openstack-infra | 16:55 | |
| clarkb | bcornec: what version of git-review do you have? I believe that the latest version should override the envs LANGUAGE and set it to C for each git command | 16:56 |
| clarkb | bcornec: that way you don't have to set it yourself | 16:56 |
| openstackgerrit | Claudiu Belu proposed openstack-infra/project-config: Adds pypi-jobs to compute-hyperv https://review.openstack.org/179782 | 16:56 |
| bcornec | $ rpm -q git-review | 16:57 |
| bcornec | git-review-1.23-3.mga4 | 16:57 |
| sdague | or and mordred while you are active - https://review.openstack.org/#/c/179764/ - we need the redirect on https in zuul as well | 16:57 |
| *** blahdeblah has joined #openstack-infra | 16:57 | |
| mordred | sdague: on it | 16:57 |
| clarkb | bcornec: can you test 1.24 and see if the problem persists? You can install it to a virtualenv and run it from there | 16:58 |
| *** sdake__ is now known as sdake | 16:58 | |
| mordred | bcornec: virtualenv foo ; foo/bin/pip install git-review ; foo/bin/git-review ... in case "install to a virtualenv" is new for you | 16:58 |
| clarkb | bcornec: a quick check of the git history shows we added the fix in 1.24 | 16:59 |
| clarkb | but if it isn't fixed we should know that too | 16:59 |
| *** bswartz has quit IRC | 17:00 | |
| bcornec | Thanks testing now... | 17:01 |
| fungi | okay, heading out to run some errands, bbiaw | 17:01 |
| *** ivar-lazzaro has joined #openstack-infra | 17:01 | |
| bcornec | mordred: everything pythong is new for me ;-) | 17:01 |
| bcornec | s/pythong/python/ | 17:01 |
| *** abregman has quit IRC | 17:02 | |
| clarkb | though I note it sets LANG=C but does not set LANGUAGE, possible it needs an update | 17:02 |
| bcornec | LANGUAGE was needed for sure on my side. | 17:03 |
| *** baoli has joined #openstack-infra | 17:03 | |
| mordred | cool. good to know | 17:04 |
| openstackgerrit | Clark Boylan proposed openstack-infra/nodepool: Document running Nodepool for testing https://review.openstack.org/179840 | 17:04 |
| clarkb | nibalizer: ^ | 17:04 |
| *** dboik_ has quit IRC | 17:04 | |
| *** dboik has joined #openstack-infra | 17:05 | |
| mordred | clarkb: I only see us doing it on the git branch command | 17:05 |
| mordred | clarkb: perhaps we should do it on every command? | 17:05 |
| *** dguerri is now known as _dguerri | 17:06 | |
| *** baoli has quit IRC | 17:06 | |
| clarkb | oh I thought it was in the subprocess wrapper | 17:06 |
| *** zul has quit IRC | 17:06 | |
| mordred | clarkb: it's not | 17:06 |
| clarkb | if not, then ya it should be pushed down to there | 17:06 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/release-tools: Add script to cleanup of translations https://review.openstack.org/172827 | 17:06 |
| *** baoli has joined #openstack-infra | 17:06 | |
| openstackgerrit | Merged openstack-infra/puppet-zuul: also add targeted status redirects to https https://review.openstack.org/179764 | 17:07 |
| *** zul has joined #openstack-infra | 17:07 | |
| pabelanger | jeblair, timrc: suggestions on naming for grafana yaml app? | 17:08 |
| openstackgerrit | Monty Taylor proposed openstack-infra/git-review: Push language override down into the cmd wrapper https://review.openstack.org/179841 | 17:08 |
| timrc | pabelanger, grafyaml jk... | 17:08 |
| pabelanger | grafana dashboard builder was first choice, but gdb seems a bad naming ;) | 17:08 |
| openstackgerrit | Monty Taylor proposed openstack-infra/git-review: Push language override down into the cmd wrapper https://review.openstack.org/179841 | 17:08 |
| pabelanger | timrc, Oooh | 17:08 |
| mordred | clarkb, bcornec: ^^ I think that should be the thing we're talking about? | 17:08 |
| pabelanger | timrc, I'm not even mad | 17:08 |
| clarkb | mordred: ya that looks right | 17:09 |
| *** Longgeek_ has quit IRC | 17:10 | |
| mordred | uhm | 17:11 |
| *** achanda has joined #openstack-infra | 17:11 | |
| bcornec | mordred: for what I can understand yes it seems that is at least needed. | 17:11 |
| clarkb | mordred: yes? | 17:11 |
| mordred | nm | 17:12 |
| mordred | I'm stoopid | 17:12 |
| anteaya | never | 17:12 |
| *** _nadya_ has joined #openstack-infra | 17:13 | |
| *** jistr has quit IRC | 17:15 | |
| openstackgerrit | Claudiu Belu proposed openstack-infra/project-config: Adds publish-to-pypi to networking-hyperv https://review.openstack.org/179844 | 17:15 |
| *** Guest19502 has quit IRC | 17:15 | |
| *** Guest19502 has joined #openstack-infra | 17:15 | |
| *** tnovacik has quit IRC | 17:15 | |
| *** bswartz has joined #openstack-infra | 17:16 | |
| *** sputnik13 has joined #openstack-infra | 17:16 | |
| *** Krinkle|detached is now known as Krinkle | 17:16 | |
| *** dizquierdo has quit IRC | 17:17 | |
| pabelanger | timrc, https://github.com/pabelanger/grafyaml now lives :) | 17:17 |
| mordred | this is exciting | 17:17 |
| *** dizquierdo has joined #openstack-infra | 17:17 | |
| *** patrickeast has joined #openstack-infra | 17:18 | |
| timrc | pabelanger, Sweet. | 17:19 |
| cody-somerville | w00t | 17:21 |
| anteaya | pabelanger: congratulations | 17:21 |
| *** Sukhdev has joined #openstack-infra | 17:22 | |
| *** luqas has joined #openstack-infra | 17:22 | |
| *** armax has joined #openstack-infra | 17:23 | |
| *** ociuhandu has quit IRC | 17:24 | |
| mordred | sdague, dhellmann, clarkb: where did we ever get with markerlib support for things in requirements/ | 17:24 |
| mordred | ? | 17:24 |
| mordred | is that real? or just a vision of the future still? | 17:24 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/system-config: Create rubygems mirror from rubygems.org https://review.openstack.org/178026 | 17:24 |
| *** mrmartin has quit IRC | 17:25 | |
| *** _nadya_ has quit IRC | 17:25 | |
| *** Guest19502 is now known as mfisch | 17:26 | |
| *** gokrokve has quit IRC | 17:27 | |
| clarkb | mordred: vision of the future iirc, lifeless had a ml response explaining why | 17:27 |
| mordred | k | 17:27 |
| mordred | so a requirements change to add a requirement that is really only intended for python3 is not happyface | 17:27 |
| *** mrmartin has joined #openstack-infra | 17:28 | |
| timrc | pabelanger, Luckily converting yaml to json relatively simple. I don't think there's a python grafana cli that implements the rest interface. Are you just going to put a thing wrapper around requests to post dashboards? | 17:28 |
| *** mrmartin has quit IRC | 17:28 | |
| timrc | is relatively* | 17:28 |
| *** mrmartin has joined #openstack-infra | 17:29 | |
| timrc | Man thing wrapper? _thin_ wrapper. | 17:29 |
| timrc | It's going to be one of those Mondays. | 17:29 |
| *** dboik_ has joined #openstack-infra | 17:29 | |
| pabelanger | timrc, Ya, I don't see any python libs for grafana / or CLI. So, ya for now just build out some simple CRUD for it. It likely can be moved out at some future point. | 17:30 |
| *** erikmwilson is now known as Guest14807 | 17:30 | |
| *** erikmwil_ has joined #openstack-infra | 17:30 | |
| *** Guest14807 has quit IRC | 17:30 | |
| *** erikmwil_ is now known as erikmwilson | 17:30 | |
| *** salv-orlando has joined #openstack-infra | 17:30 | |
| *** erikmwilson_ has joined #openstack-infra | 17:30 | |
| *** armax has quit IRC | 17:31 | |
| *** ociuhandu has joined #openstack-infra | 17:31 | |
| *** tiswanso has quit IRC | 17:32 | |
| *** tiswanso_ has joined #openstack-infra | 17:33 | |
| *** luqas has quit IRC | 17:33 | |
| *** dboik has quit IRC | 17:34 | |
| *** harlowja has joined #openstack-infra | 17:34 | |
| anteaya | timrc: I was hoping for thing wrapper | 17:34 |
| *** baoli has quit IRC | 17:36 | |
| openstackgerrit | Merged openstack-infra/shade: Rename get_endpoint() to get_session_endpoint() https://review.openstack.org/178859 | 17:36 |
| *** baoli has joined #openstack-infra | 17:37 | |
| clarkb | nibalizer: so looking up the puppet beaker stuff. Have questions about https://review.openstack.org/#/c/179249/8/spec/acceptance/basic_swift_spec.rb comments says run puppet twice but it appears to run puppet 3 times, what is going on there? | 17:37 |
| clarkb | nibalizer: and why would you ever want t catch failures? shouldn't a failure alwys be a failure and raised as such? | 17:37 |
| openstackgerrit | Merged openstack-infra/shade: Convert node_set_provision_state to task https://review.openstack.org/177987 | 17:38 |
| clarkb | EmilienM: ^ I guess same questions for you | 17:38 |
| *** baoli has quit IRC | 17:38 | |
| *** salv-orlando has quit IRC | 17:39 | |
| clarkb | also not a fan of the inlined heredoc, the test file should be a fixture so that it is reuseable as a "known working" | 17:39 |
| *** baoli has joined #openstack-infra | 17:39 | |
| *** gokrokve has joined #openstack-infra | 17:41 | |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Add query for libvirt socket hangup bug 1451506 https://review.openstack.org/179851 | 17:42 |
| openstack | bug 1451506 in OpenStack Compute (nova) "spawn failed with "libvirtError: internal error: received hangup / error event on socket" in the gate" [Undecided,New] https://launchpad.net/bugs/1451506 | 17:42 |
| crinkle | clarkb: i think that's an error | 17:42 |
| nibalizer | clarkb: probably thats because that particular case isn't idempotent | 17:42 |
| *** russellb has quit IRC | 17:42 | |
| crinkle | clarkb: convention with beaker-rspec is to use the heredoc | 17:42 |
| *** jogo has quit IRC | 17:42 | |
| timrc | anteaya, :) | 17:42 |
| nibalizer | and doesn't swift use exported resources? so it'd have to run twice ya? | 17:42 |
| nibalizer | crinkle: ^? | 17:42 |
| crinkle | maybe | 17:43 |
| crinkle | if so the comment should be updated | 17:43 |
| clarkb | nibalizer: crinkle I see so its "run twice, then run a third time" not "run twice" | 17:43 |
| *** russellb has joined #openstack-infra | 17:43 | |
| crinkle | yeah the comment may have been cargoculted | 17:43 |
| *** jogo has joined #openstack-infra | 17:43 | |
| nibalizer | clarkb: so this is more the pattern: http://git.openstack.org/cgit/stackforge/puppet-keystone/tree/spec/acceptance/basic_keystone_spec.rb#n61 | 17:43 |
| clarkb | for the heredoc is there an advantage to inlining that file? | 17:44 |
| clarkb | its clearly a test fixture and should be treated as one... | 17:44 |
| mordred | clarkb, nibalizer, crinkle: the more I look at things, the more I think we should do beaker-rspec so that we can do things the same way as the puppet-openstack stuff and enhance each other's happy | 17:44 |
| clarkb | mordred: thats funny, the more I look at it the less I like it :) | 17:44 |
| mordred | I was able to read and understand and reason about writing new tests looking at the stuff in puppet-openstack | 17:44 |
| clarkb | mordred: it comes off as very hacky | 17:44 |
| crinkle | clarkb: the advantage is we can see it immediately without looking at a fixture | 17:45 |
| crinkle | and it's conventional | 17:45 |
| *** gokrokve has quit IRC | 17:45 | |
| *** gokrokve has joined #openstack-infra | 17:45 | |
| mordred | clarkb: that may be - but if it's the recommended puppet way, and it's what EmilienM and crinkle and folks are doing in the other puppet things, I think we'd be better off working together than doing it all a bit differently ourselves | 17:45 |
| *** mwagner_lap has joined #openstack-infra | 17:45 | |
| * EmilienM reads backlog | 17:46 | |
| mordred | basically, I'm not commenting on whether it's a good tool - more whether we have another option that is better _enough_ to warrant divergence | 17:46 |
| clarkb | mordred: you have to do a fair bit of setup t make it work in nodepool nodes for example, and it completely bypasses the zuul-cloner goodness we have (though shouldn't be required, just means an extra step) | 17:46 |
| mordred | clarkb: well, we should figure out how to fix it to not bypass zuul cloner so that EmilienM is also not unhappy | 17:46 |
| mordred | which might mean making beaker itself less bong | 17:46 |
| mordred | or - we should convince EmilienM to do something different with us | 17:47 |
| crinkle | I think we're pretty set on using beaker-rspec for those modules | 17:47 |
| crinkle | but i have no opinion on whether infra decides to follow suit | 17:47 |
| nibalizer | clarkb: i really really want to get zuul cloner and beaker playing nice with each other | 17:47 |
| mordred | I spend enough time yelling at openstack devs about divergence, that I believe diverging myself would be quite hypocritical | 17:47 |
| mordred | nibalizer: ++ | 17:47 |
| nibalizer | because zuulcloner + a gate for the puppet openstack modules would be sooo cool | 17:48 |
| * nibalizer gets chills | 17:48 | |
| clarkb | mordred: ya its definitely nice to converge, I am just pointing out the first impression problems I see with it | 17:48 |
| mordred | clarkb: sure | 17:48 |
| nibalizer | clarkb: also beaker is opensource, we can take feedback upstream and patches too if we want | 17:48 |
| mordred | clarkb: also- I mean, I'm sstill not really sure if I should write thigns in rspec or serverspec or both or somethign else | 17:48 |
| mordred | I will say I think that serverspec is very nicely readable | 17:49 |
| clarkb | nibalizer: well these issues appaer to be conventions | 17:49 |
| clarkb | nibalizer: for example rspec expects to install the module for you | 17:49 |
| mordred | but beaker-rspec seems to be different? | 17:49 |
| clarkb | er sorry beaker | 17:49 |
| nibalizer | clarkb: conventions evolve | 17:49 |
| *** luqas has joined #openstack-infra | 17:49 | |
| clarkb | nibalizer: and heredocing fixtures inline in code | 17:49 |
| nibalizer | beaker-rspec sits above beaker because largely of conventions disagreements (as i understand it) | 17:49 |
| clarkb | mordred: beaker-rspec is just run rspec tests after beakering | 17:49 |
| clarkb | mordred: the openstack puppet modules seem to use beaker-rspec not beaker alone | 17:50 |
| mordred | yah | 17:50 |
| mordred | but something in the beaker-rspec docs mentions serverspec | 17:50 |
| mordred | although it does not seem like that's actually used/true | 17:51 |
| *** luqas has quit IRC | 17:51 | |
| mordred | https://github.com/puppetlabs/beaker-rspec/blob/master/lib/beaker-rspec/helpers/serverspec.rb | 17:51 |
| clarkb | fwiw https://review.openstack.org/#/c/179249/8/spec/spec_helper_acceptance.rb is the thing that would need to be made to grok zuul-cloner aiui | 17:51 |
| clarkb | right now, it appears to want to install all the things for you | 17:51 |
| *** gulic has joined #openstack-infra | 17:52 | |
| clarkb | except it doesn't really understand the gate and testing things together, which is fine until you want to gate things together (unsure if that is necessary) | 17:52 |
| *** gulic has left #openstack-infra | 17:52 | |
| mordred | yah - we'd need to be able to have that find zuul cloner repos to install from | 17:52 |
| mordred | clarkb: I think we'll want to gate things together for sure | 17:52 |
| crinkle | clarkb: yeah, actually gating things together would be nice, right now it just git clones from master | 17:52 |
| mordred | clarkb: we ahve enough thigns with depends-on relationships | 17:52 |
| *** ZZelle has quit IRC | 17:53 | |
| *** ZZelle has joined #openstack-infra | 17:53 | |
| nibalizer | ya so my plan to hack that was to shell out to zuul cloner from inside spec_helper_acceptance, but thats kinda gross | 17:53 |
| mordred | so maybe we'll need to write a ruby library called beaker-rspec-zuul or something, and have that have a function that knows how to clone things, and will ook for zuul-cloner prepared repos and clone from them, otherwise clone from source | 17:53 |
| clarkb | mordred: or maybe beaker should gets its hands out of everything... | 17:54 |
| mordred | that way anyone doing beaker-rspec things in a zuul context can just use beaker-rspec-zuul | 17:54 |
| clarkb | mordred: it doesn't need to install the modules, and it doesn't need to setup test nodes | 17:54 |
| jeblair | mordred, nibalizer: joining late, sorry, but can you say in 20 words why this should happen inside of a tool and not as a preparatory step before running it? | 17:54 |
| clarkb | it just needs to assert some state | 17:54 |
| mordred | clarkb: sure - but let's one step at a time - I think we are the exceptional case | 17:54 |
| jeblair | (i think clarkb might also be asking my question) | 17:54 |
| clarkb | jeblair: I think so :) | 17:55 |
| mordred | jeblair: because not everyone in the world has an infra-quality repo prep system at their disposal, so "do the thigns for me" is a feature fora lot of people | 17:55 |
| crinkle | jeblair: it's meant to be able to run from a user's laptop, so having it self contained is preferred | 17:55 |
| *** tnovacik has joined #openstack-infra | 17:55 | |
| mordred | jeblair: what we need is a way to easily say "please don't do that, kthxbai" in a way that doesn't break people trying to do local testing with, say, vagrant | 17:55 |
| claudiub | Hello, anyone can give me a helping hand regarding stackforge pushed versions through gerrit? It seems I have a little bit of a problem and I did not find a proper solution yet and I was wondering if you folks would know. | 17:55 |
| nibalizer | jeblair: what crinkle and mordred are saying is valid, I'm not 100% satisfied that what we have now (not includin zuul cloner) is enabling local development to match ci testing | 17:56 |
| nibalizer | so adding zuul cloner functionality as a hidden side effect, to me, sounds like a step in the wrong direction | 17:56 |
| jeblair | mordred: yeah; i feel like that's the approach to start with; as there are many testing tools out there that do not manipulate git repos, so i don't actually think we're weird. i appreciate the extra helpfulness, but it should be extra. if every computing tool had to understand an scm, the world would be a bad place. | 17:56 |
| jeblair | nibalizer: by 'what we have now' do you mean the proposal in 179249? | 17:57 |
| nibalizer | I worry that we're pushing vagrant files out to local developers and no one is validating that they work | 17:57 |
| jeblair | nibalizer: or something else? | 17:57 |
| EmilienM | clarkb: swift was a bad example | 17:57 |
| EmilienM | all other modules needs 1 puppet run | 17:57 |
| dhellmann | mordred, jd__: regarding lockfile, it's in murano's requirements list. murano is not in the requirements/projects.txt list though so that won't break anything, but now that murano is an official project maybe it would want to be in the projects list? | 17:57 |
| clarkb | EmilienM: thanks, I will have to look at the others too, I choose swift somewhat randomly | 17:58 |
| mordred | jeblair: yah - and it soudns like nibalizer is going to write us some upstream ruby to help with that | 17:58 |
| nibalizer | jeblair: 179249 is just the swift version of over a dozen patches that EmilienM has spun up, some are landed | 17:58 |
| *** sabeen2 has quit IRC | 17:58 | |
| jeblair | nibalizer: i agree -- i believe there is no warranty of functionality or fitness of purpose for anything in infra repos if we don't gate-test it. :) | 17:58 |
| mordred | dhellmann: yes. in fact, I think it should be in the projects list | 17:58 |
| jeblair | (and maybe not even then ;) | 17:58 |
| nibalizer | jeblair: this is the file im worried about: https://review.openstack.org/#/c/179249/8/spec/spec_helper_acceptance.rb | 17:59 |
| *** harlowja has quit IRC | 17:59 | |
| *** harlowja_ has joined #openstack-infra | 17:59 | |
| clarkb | claudiub: what is the problem? | 17:59 |
| *** tonytan4ever has quit IRC | 17:59 | |
| claudiub | clarkb: hello. So, I've pushed 1.0.3 to https://github.com/stackforge/networking-hyperv and if I download the zip / tar.gz, unzip it and python setup.py install it, this occurs: Exception: Versioning for this project requires either an sdist tarball, or access to an upstream git repository. Are you sure that git is installed? | 17:59 |
| nibalizer | in that I think there are assumptions about running in infra, that wouldn't be there if a developer was using vagrant | 17:59 |
| mordred | claudiub: yes, that's totally not supported | 17:59 |
| *** luqas has joined #openstack-infra | 17:59 | |
| clarkb | claudiub: ya, you either need a git repository or an sdist | 17:59 |
| mordred | claudiub: github zip/tarball files are not and never will be a format that you can use | 17:59 |
| dhellmann | mordred: so do we remove lockfile and force murano to update before they can be added, or do we do it in the other order? | 17:59 |
| mordred | claudiub: what you want is http://tarballs.openstack.org/networking-hyperv/ | 18:00 |
| nibalizer | so I think we're in a grey area now, and we need to be careful about how we add zuul-cloner so that we enable both local development and CI | 18:00 |
| mordred | claudiub: http://tarballs.openstack.org/networking-hyperv/networking-hyperv-master.tar.gz if you want a thing that tracks master | 18:00 |
| clarkb | nibalizer: to some degree, but say I am making a change to one module that requires an update to a dep too, I need to be able to express that even on my laptop | 18:00 |
| *** erikmwil_ has joined #openstack-infra | 18:00 | |
| *** erikmwilson is now known as Guest38131 | 18:00 | |
| *** erikmwil_ is now known as erikmwilson | 18:00 | |
| mordred | clarkb: ++ | 18:00 |
| *** erikmwilson has quit IRC | 18:00 | |
| *** armax has joined #openstack-infra | 18:00 | |
| clarkb | nibalizer: I think vagrant areas are likely where we need to more directly work around the laptop use case. "Install the correct stuff" is something that should be the same whereever | 18:00 |
| jeblair | EmilienM, nibalizer: i see -- i think since puppet modules are joining the big tent we should probably work together on a solution that works for puppet-openstack and infra | 18:01 |
| mordred | in fact, I know the tchaypo frequently talks about wanting to locally test complex interaction chains | 18:01 |
| clarkb | nibalizer: and many tools solve that by making you do the install | 18:01 |
| *** erikmwilson_ is now known as erikmwilson | 18:01 | |
| *** luqas has quit IRC | 18:01 | |
| claudiub | ah, I see. cool. So, another question.. in order to have certain versions for the tarballs, there has to be a specific job in project-config, right? | 18:01 |
| mordred | jeblair: ++ | 18:01 |
| clarkb | nibalizer: which I think is a very valid way to do it | 18:01 |
| dhellmann | mordred: actually, it looks like they may not be using that dependency, let me try a patch | 18:01 |
| claudiub | also, if the pypi job is defined on the project-config, pip install should not have any problems, right? | 18:01 |
| mordred | claudiub: that's right - although the pypi jobs will only release something to pypi when you push a signed tag | 18:02 |
| *** emagana has quit IRC | 18:02 | |
| *** chlong has quit IRC | 18:02 | |
| nibalizer | clarkb: so you think we should set up the right repos in the right places before the test runs, and not do those things in spec_helper_acceptance ? | 18:03 |
| claudiub | mordred: I assume - name: openstack-server-publish-jobs should be added to layout/zuul.yaml, right? | 18:03 |
| jeblair | nibalizer, mordred, clarkb: is this similar to devstack vagrant testing? where you want to use vagrant to run devstack in a local vm based on git repos on your host? | 18:03 |
| clarkb | nibalizer: correct, because I think its better to have the user think about what they should do and do it correctly instead of doing it poorly and most likely wrong for them | 18:04 |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Add query for libvirt socket hangup bug 1451506 https://review.openstack.org/179851 | 18:04 |
| claudiub | mordred: and - openstack-publish-jobs to project.yaml | 18:04 |
| openstack | bug 1451506 in OpenStack Compute (nova) "spawn failed with "libvirtError: internal error: received hangup / error event on socket" in the gate" [Undecided,Confirmed] https://launchpad.net/bugs/1451506 | 18:04 |
| openstackgerrit | Doug Hellmann proposed openstack/requirements: Remove lockfile from requirements https://review.openstack.org/151224 | 18:04 |
| *** SumitNaiksatam has quit IRC | 18:04 | |
| jeblair | clarkb, nibalizer: ++ | 18:04 |
| nibalizer | jeblair: i have no information about that workflow, but sounds related | 18:04 |
| *** Guest38131 has quit IRC | 18:04 | |
| dhellmann | mordred, jd__ : The global requirements patch now depends on the patch to remove it from murano. | 18:05 |
| *** EmilienM is now known as EmilienM|mtg | 18:05 | |
| *** sandywalsh_ has joined #openstack-infra | 18:05 | |
| *** SumitNaiksatam has joined #openstack-infra | 18:05 | |
| dmsimard | Would love a core review on https://review.openstack.org/#/c/179809/ when anyone gets a chance, thanks :) | 18:05 |
| *** sandywalsh has quit IRC | 18:06 | |
| nibalizer | clarkb: okay, and we could write a little 'grab the tip of master for everything' script for the people who aren't doing anything fancy? | 18:06 |
| clarkb | nibalizer: ya | 18:06 |
| clarkb | nibalizer: similar to how devstack works | 18:06 |
| clarkb | devstack will use what you give it or if lazy give you $BRANCH | 18:07 |
| nibalizer | clarkb: okay i think we can do this, note that we'll have to put /some/ logic in spec_helper_acceptance since thats our entry point into running stuff on the vagrant vm before the actuall tests start | 18:07 |
| greghaynes | sdague: Commented on https://review.openstack.org/#/c/179573/4/modules/openstack_project/files/gerrit/hideci.js | 18:07 |
| *** jcoufal has joined #openstack-infra | 18:08 | |
| *** maishsk has quit IRC | 18:08 | |
| *** packet has quit IRC | 18:09 | |
| *** Sukhdev has quit IRC | 18:09 | |
| sdague | greghaynes: *shrug*? that was how I figured out not to crush all the important object info :) | 18:09 |
| sdague | if you have a better way, I'm all for it | 18:09 |
| * sdague doesn't even pretty to play a js expert on tv | 18:09 | |
| *** SumitNaiksatam has quit IRC | 18:09 | |
| sdague | s/pretty/pretend/ | 18:10 |
| *** otter768 has joined #openstack-infra | 18:10 | |
| greghaynes | heh, I probably shouldnt pretent | 18:10 |
| greghaynes | pretend | 18:10 |
| claudiub | clarkb, mordred: anyways, thanks a lot for the help! :D | 18:10 |
| anteaya | greghaynes: but can you pretty | 18:10 |
| *** maishsk has joined #openstack-infra | 18:10 | |
| greghaynes | :) | 18:11 |
| greghaynes | sdague: The thing I pasted should work I think, ill test it out using your fancy proxy patch though | 18:11 |
| *** emagana has joined #openstack-infra | 18:11 | |
| sdague | ok, I did it that way so that the objects came through natively in the chrome console, which lets you click through them | 18:12 |
| greghaynes | yes, IIRC that is a matter of the type of the first argument you pass in | 18:13 |
| greghaynes | if your first arg is not a string then itll output objects | 18:13 |
| *** rbradfor has quit IRC | 18:14 | |
| nibalizer | clarkb: so I think we have s_p_a do some inspection to see if it is in a zuul environment (maybe check $ZUUL_REF or something), then if not do the git clone from tip of master, probably need s_p_a to clean /etc/puppet/modules and copy the zuul-cloner'd repos into /etc/puppet/modules though. Also need it to do the pure dependency installs such as stahnma-epel etc | 18:14 |
| *** otter768 has quit IRC | 18:15 | |
| *** chlong has joined #openstack-infra | 18:15 | |
| clarkb | nibalizer: can we check for the existence of the modules? or is it possible these nodes are reused even with vagrant? | 18:15 |
| *** Krinkle is now known as Krinkle|detached | 18:16 | |
| nibalizer | clarkb: beaker works by starting up a vm and then doing things to it, then turning it off | 18:17 |
| nibalizer | so making that work with side effects is gonna be a bit tricky | 18:17 |
| *** sabeen has joined #openstack-infra | 18:18 | |
| nibalizer | and yes sometimes people run beaker with BEAKER_destroy=no or BEAKER_provision=no which allows for reusing the vm | 18:18 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Enhance the OperatorCloud constructor https://review.openstack.org/179647 | 18:18 |
| clarkb | nibalizer: hrm, right because something in the new VM has to figure it out | 18:18 |
| nibalizer | mostly when debugging the testing harness in my experience | 18:18 |
| nibalizer | but I can't speak for all developers | 18:18 |
| nibalizer | it does make it a ton faster if you know what you're doing isn't corrupting the tests you are doing | 18:19 |
| mordred | so we'd want normal beaker to do the thing people expect from a normal beaker | 18:19 |
| mordred | and we'd want to be able to run it in a way that says not only BEAKER_provision=no - but also "hi, we've done some prep for you" | 18:19 |
| clarkb | mordred: sort of, normal beaker is wrong from a install the list of correct things perspective | 18:19 |
| mordred | what do you mean? | 18:20 |
| clarkb | mordred: normal operation appears to be install master or latest release | 18:20 |
| clarkb | mordred: it does not grok "install this development commit so I can iterate on my dev workflow" | 18:20 |
| mordred | yah | 18:20 |
| openstackgerrit | Merged openstack-infra/project-config: Disable voting for puppet-syntax-future on havana/icehouse https://review.openstack.org/179809 | 18:20 |
| mordred | well, that CLEARLY needs to be fixed | 18:20 |
| nibalizer | mordred: so we don't set BEAKER_provision=no, we tell it to use a very dumb hypervisor called 'just ssh into localhost, itll be ok' | 18:20 |
| clarkb | nibalizer: its super close to metal | 18:21 |
| clarkb | nibalizer: not even a container in the way | 18:21 |
| nibalizer | clarkb: right well so... beacuse the beaker consumers are also using the github workflow, they get around that :P | 18:21 |
| nibalizer | so they have a stack of commits checked out locally | 18:21 |
| *** EmilienM|mtg is now known as EmilienM | 18:21 | |
| nibalizer | on a brnach when they(we) run beaker locally | 18:21 |
| mordred | clarkb: we need to make beaker work for the "clone from master" case like normal puppet people expect - we need to improve their experience by adding a thign that knows how to run beaker using the git commits they have locally and not pushed, and we need to support zuul-cloner envs | 18:22 |
| mordred | clarkb: once that's done, it'll be good :) | 18:22 |
| clarkb | nibalizer: how does that help if beaker is installing via puppet module? | 18:22 |
| nibalizer | as for depending on an unlanded commit in another module, idunno how pepole do that other than derping with spec_helper_acceptance | 18:22 |
| clarkb | nibalizer: (I want to know because that is essentially how zuul-cloner works) | 18:22 |
| nibalizer | clarkb: only applies to the repo you're working in | 18:22 |
| clarkb | nibalizer: ah, ok | 18:22 |
| *** zaneb has joined #openstack-infra | 18:22 | |
| nibalizer | so for some extra diversity, here is another s_p_a https://github.com/elastic/puppet-elasticsearch/blob/master/spec/spec_helper_acceptance.rb | 18:23 |
| nibalizer | in my experience, these things vary kinda wildly | 18:23 |
| jeblair | nibalizer: s_p_a means what? | 18:24 |
| nibalizer | also https://github.com/puppetlabs/puppetlabs-postgresql/blob/master/spec/spec_helper_acceptance.rb | 18:24 |
| nibalizer | jeblair: spec_helper_acceptance.rb, it is the script that runs before the testing in beaker | 18:24 |
| jeblair | where's the p come from? | 18:24 |
| nibalizer | so this is where you install puppet dependencies, and even puppet itsefl sometimes | 18:25 |
| jeblair | to avoid collision with sha? (ha!) | 18:25 |
| nibalizer | jeblair: i dont know, i think i meant s_h_a but ended up saying s_p_a (my bad!) | 18:25 |
| nibalizer | hehe | 18:25 |
| clarkb | right its basically our run install_puppet.sh script step | 18:25 |
| jeblair | nibalizer: ok, np. maybe you just invented a thing. :) | 18:25 |
| anteaya | nibalizer: I like s_p_a as acronyms go | 18:26 |
| clarkb | except our install_puppet.sh understands these things | 18:26 |
| *** dizquierdo has quit IRC | 18:27 | |
| *** mattfarina has quit IRC | 18:28 | |
| *** Sukhdev has joined #openstack-infra | 18:29 | |
| *** nikita09 has joined #openstack-infra | 18:30 | |
| *** sdake_ has joined #openstack-infra | 18:30 | |
| EmilienM | nibalizer: I think fbo found out why we can't use beaker5 on OS infra | 18:31 |
| EmilienM | https://github.com/puppetlabs/beaker/blob/master/lib/beaker/host_prebuilt_steps.rb#L528 | 18:31 |
| EmilienM | and http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/puppet-module-jobs.yaml#n39 | 18:31 |
| *** dboik_ has quit IRC | 18:31 | |
| asselin__ | nibalizer, can you reevaluate your -1 on this? https://review.openstack.org/#/c/167425/ | 18:32 |
| *** tonytan4ever has joined #openstack-infra | 18:33 | |
| clarkb | mordred: hahaha see EmilienM's links. We very recently had a discussion about this | 18:33 |
| nibalizer | asselin__: sure | 18:33 |
| openstackgerrit | Ken Dreyer proposed stackforge/python-jenkins: support six 1.3 https://review.openstack.org/179826 | 18:33 |
| *** tiswanso_ has quit IRC | 18:33 | |
| EmilienM | clarkb: it's time to collaborate seriously, isn't ? :) | 18:33 |
| nibalizer | EmilienM: i dont understand? | 18:33 |
| EmilienM | nibalizer: when I started to play with beaker | 18:34 |
| EmilienM | I tried beaker 5 | 18:34 |
| clarkb | EmilienM: it was a week or two ago that mordred and I were arguing about how ansible should know how to pass envs through to remote end | 18:34 |
| EmilienM | and I failed and did not investigate because I wanted to move forward with your work | 18:34 |
| clarkb | EmilienM: and it appears that that is one thing beaker depends on | 18:34 |
| *** sdake has quit IRC | 18:34 | |
| EmilienM | nibalizer: fbo found our the same issue and it seems it comes from the links I shared here | 18:34 |
| *** sdake_ is now known as sdake | 18:34 | |
| nibalizer | EmilienM: yea so the issue is that infra and beaker fight over the ssh config file? | 18:34 |
| clarkb | (I am still not sure I know the answer around that, but having the option to with a tool like ansible seems useful) | 18:34 |
| EmilienM | nibalizer++ | 18:35 |
| mordred | clarkb: for the record, I've not been arguing that passing env vars is not something it shoudl do - I've been arguing that the framing of the way that you want to do it is not the way they do it | 18:35 |
| waynr | does anyone know what a gerrit patchSet "number" is? | 18:36 |
| *** samueldmq_ has joined #openstack-infra | 18:37 | |
| *** boris-42 has joined #openstack-infra | 18:37 | |
| clarkb | waynr: yes, its the number assigned to the patchset. It is an incrementing counter for each patchset starting at 1 | 18:37 |
| waynr | oh right, that makes sense | 18:38 |
| clarkb | waynr: in gerrit a change has a series of patchsets. Unlike github you don't continuously append to your pull request, instead you push a new version of a chnage. Each version of a change is a patchset and is assigned a unique number | 18:38 |
| nibalizer | asselin__: you have feedback :) | 18:38 |
| waynr | yeah i was just confused, the number makes sense to me now | 18:38 |
| nikita09 | hi, does Jenkins job builder support passing list parameters to macros ??? | 18:40 |
| nikita09 | i want to pass list of branches to my git macro, that my job will be building, im not sure if its allowed to pass a list of items as parameter to a macro.. | 18:41 |
| waynr | I find the implementation of JJB macros confusing and inconsistent (can't reference otherwise globally-defined JJB variables, can't pass yaml "objects") so I tend to stay away from them | 18:41 |
| clarkb | nikita09: I think so, you have to annotate the type otherwise you get the macro expanded n time where n = length of list | 18:41 |
| waynr | oh maybe you cant pass yaml objects | 18:41 |
| clarkb | waynr: you should be able to pass yaml objects | 18:41 |
| waynr | s/cant/can/ | 18:41 |
| openstackgerrit | Merged openstack/requirements: also install a new pbr into the venvs https://review.openstack.org/155410 | 18:42 |
| clarkb | http://ci.openstack.org/jenkins-job-builder/definition.html#variable-references | 18:42 |
| clarkb | waynr: nikita09 see ^ | 18:42 |
| waynr | clarkb: i seem to recall not being able to simply use '{obj:key}' format for passing objects like that, but haven't tried recently enough--maybe it's fixed now? | 18:43 |
| waynr | i also definitely didn't try with the type annotation | 18:43 |
| clarkb | I have no idea, we don't use the feature anywhere, I have yet to find a case where it is necessary | 18:43 |
| clarkb | but you can in theory use it | 18:44 |
| clarkb | and it is in theory tested by JJBs tests | 18:44 |
| *** gokrokve has quit IRC | 18:44 | |
| waynr | yeah i started using it when I first tried JJB but quickly stopped when i realized I couldn't reference JJB variables defined elsewhere inside a macro | 18:44 |
| clarkb | waynr: you have to pass variables into macros same as any job template | 18:44 |
| clarkb | so you can use them they just have to be passed through | 18:45 |
| nikita09 | clarkb: thanks...) how do i annotate the type ? im not sure what that means ..:/ | 18:45 |
| waynr | clarkb: yeah but if i define a variable in a project, then list job-group in that project's "jobs" the variable defined at the project-level will be available at the job-template level for job templates defined in the job-group | 18:45 |
| nikita09 | like '{obj:list"'??? | 18:45 |
| clarkb | nikita09: the documentation there should show you, !!python/tuple for a tuple looks like | 18:45 |
| waynr | but for macros if I remember correctly i always have to explicitly pass variables when calling the macro, otherwise variables defined at other levels in the jjb object hierarchy aren't available | 18:46 |
| clarkb | waynr: you can think of putting the job template in the job list of a project as an explicit pass | 18:46 |
| clarkb | waynr: there is nothing global about that var, you are explicitly passing it through to every entry in the jobs list | 18:47 |
| nikita09 | clarkb: got it / thanks alot :) | 18:47 |
| waynr | but i'm saying that the variables from the project get passed to the job-template even if that job-template is only transitively referenced in the list of jobs for that project which is very different from the macro behavior | 18:47 |
| waynr | when I say "global" what I really mean is "gets passed in from outer scopes" | 18:47 |
| *** gokrokve has joined #openstack-infra | 18:48 | |
| *** subscope_ has joined #openstack-infra | 18:49 | |
| *** shivharis has joined #openstack-infra | 18:50 | |
| *** tsg has joined #openstack-infra | 18:50 | |
| *** SumitNaiksatam has joined #openstack-infra | 18:51 | |
| anteaya | mtreinish: thanks for the post to that email, I was wondering what the issue was | 18:51 |
| mtreinish | anteaya: oh, the depends-on thing? Sure, np | 18:52 |
| *** ildikov has joined #openstack-infra | 18:52 | |
| mtreinish | yeah it took me a sec to figure out what he was asking | 18:52 |
| anteaya | mtreinish: I looked at it too, had no idea, and marked it as something to track for the solution | 18:52 |
| anteaya | mtreinish: I didn't know tempest-lib check jobs installed from last release not source | 18:53 |
| lifeless | mordred: what requirements change? | 18:53 |
| mtreinish | anteaya: all the openstack libs and clients install from releases (unless it's a source job) | 18:53 |
| anteaya | mtreinish: well shows how much I know | 18:54 |
| mordred | lifeless: ECONTEXT | 18:54 |
| anteaya | mtreinish: thank you | 18:54 |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/requirements: Adding os-brick for managing initiators https://review.openstack.org/177372 | 18:54 |
| lifeless | 05:27 < clarkb> mordred: vision of the future iirc, lifeless had a ml response explaining why | 18:54 |
| lifeless | 05:27 < mordred> k | 18:54 |
| lifeless | 05:27 < mordred> so a requirements change to add a requirement that is really only intended for python3 is not | 18:54 |
| lifeless | happyface | 18:54 |
| mordred | lifeless: gotcha | 18:55 |
| mtreinish | anteaya: heh, a few people have been tripped on it before, especially because we used to install everything from git not that long ago | 18:55 |
| anteaya | mtreinish: not that long ago -- last week maybe? | 18:55 |
| mordred | lifeless: was about a proposed patch to requirements adding new mysqlclient library - except the intent was to only use it for python 3 (at least that's what the mailing list came down to) | 18:55 |
| mordred | lifeless: so I was poking at whether I shold -1 broadly - or -1 saying "add markerlib thing" | 18:56 |
| *** shivharis has quit IRC | 18:56 | |
| lifeless | yeah | 18:56 |
| lifeless | so today | 18:56 |
| mtreinish | anteaya: I think it was an early kilo thing, but it might have been earlier, my sense of time gets skewed | 18:56 |
| lifeless | add to requirements-py3.txt | 18:56 |
| lifeless | but | 18:56 |
| *** tsg has quit IRC | 18:56 | |
| anteaya | mtreinish: ah | 18:56 |
| lifeless | point out reallllly clearly that that means that we can't build wheels | 18:56 |
| lifeless | for that project | 18:56 |
| anteaya | mtreinish: yes time moves in strange ways | 18:56 |
| lifeless | [that get published to PyPI] | 18:56 |
| lifeless | without infra changes - and it would be better to do the environment marker work than change infra | 18:57 |
| *** sslypush1nko is now known as sslypushenko | 18:57 | |
| zaro | clarkb: hey sorry for late reply. read your conversation about gearman plugin. | 18:58 |
| zaro | i'm looking at it and trying to debug some more to see if i can validate your theories. | 18:59 |
| anteaya | zaro: I've lost track, what needs to happen before Saturday? | 18:59 |
| *** rockyg has joined #openstack-infra | 18:59 | |
| anteaya | zaro: is there an etherpad started yet? | 18:59 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/system-config: Create rubygems mirror from rubygems.org https://review.openstack.org/178026 | 19:00 |
| zaro | anteaya: https://etherpad.openstack.org/p/gerrit-2.10-upgrade | 19:00 |
| *** radez is now known as radez_g0n3 | 19:01 | |
| *** julim has quit IRC | 19:02 | |
| anteaya | zaro: great thank you | 19:02 |
| *** whayutin_ is now known as weshay | 19:03 | |
| anteaya | zaro: what a stunning lavender | 19:03 |
| pleia2 | hehe | 19:03 |
| *** julim has joined #openstack-infra | 19:04 | |
| *** craigbr has quit IRC | 19:04 | |
| *** rbradfor has joined #openstack-infra | 19:04 | |
| anteaya | isn't it a great colour? | 19:04 |
| *** hdd has joined #openstack-infra | 19:05 | |
| pleia2 | it is | 19:05 |
| *** yfried has joined #openstack-infra | 19:07 | |
| *** chlong has quit IRC | 19:07 | |
| nikita09 | clarkb: well...didnt go as expected :( im getting error about 'obj:' ... full paste is here : http://fpaste.org/218313/43076639/ | 19:07 |
| *** dboik has joined #openstack-infra | 19:09 | |
| openstackgerrit | Matthew Treinish proposed openstack/requirements: Remove tempest from projects.txt on stable https://review.openstack.org/179887 | 19:09 |
| *** sdake_ has joined #openstack-infra | 19:09 | |
| lifeless | mtreinish: wat? | 19:09 |
| clarkb | nikita09: I don't think youare passing those vars across | 19:09 |
| *** sandywalsh_ has quit IRC | 19:09 | |
| clarkb | nikita09: I don't know why that is the case but looking at the error that seems to be what is happening? | 19:10 |
| openstackgerrit | Matthew Treinish proposed openstack/requirements: Remove tempest from projects.txt on stable https://review.openstack.org/179888 | 19:10 |
| clarkb | nikita09: its saying it doesn't know what obj is but it should realize thats special before that and do the replacement | 19:10 |
| openstackgerrit | Matthew Treinish proposed openstack/requirements: Remove tempest from projects.txt on stable https://review.openstack.org/179889 | 19:10 |
| nikita09 | clarkb: just to be clear...variable refereces can be made to macros...not just job-templates as the docs say.. | 19:10 |
| nikita09 | rite? | 19:10 |
| clarkb | nikita09: orrect | 19:11 |
| mtreinish | lifeless: grenade fun, it is g-r syncing stable on tempest master | 19:11 |
| clarkb | nikita09: what version of jjb do you have? | 19:11 |
| lifeless | mtreinish: ah, so - | 19:11 |
| mtreinish | which is causing the wrong version of tempest-lib from being installed for the grenade job | 19:11 |
| nikita09 | clarkb: 1.1.0 | 19:11 |
| lifeless | mtreinish: I'm worried about uninstallability happening | 19:11 |
| lifeless | mtreinish: can you you think about someway to prevent that? | 19:11 |
| openstackgerrit | Merged openstack-infra/shade: Enhance the OperatorCloud constructor https://review.openstack.org/179647 | 19:12 |
| mtreinish | lifeless: so tempest should be venv isolated in the gate cases I'm thinking of because it should always be a clone followed by tox | 19:12 |
| clarkb | nikita09: ok that version should be new enough according to git logs | 19:12 |
| mtreinish | lifeless: I'm not sure I understand what you mean by uninstallability | 19:13 |
| mordred | clarkb, jeblair, Shrews: https://review.openstack.org/#/c/178886/ - I'd like to land that shade patch and then cut a release | 19:13 |
| *** sdake has quit IRC | 19:13 | |
| *** dprince has quit IRC | 19:13 | |
| nikita09 | clarkb: yes..i have almost latest version... i just want to pass a list to my git scm macro.... i tried !!python/tuple and it also gave same "keyerror obj" error..:/ | 19:14 |
| *** tiswanso has joined #openstack-infra | 19:14 | |
| *** samueldmq_ has quit IRC | 19:14 | |
| mtreinish | lifeless: we wouldn't be able to use branchless tempest on a stable devstack if it wasn't venv isolated | 19:14 |
| clarkb | nikita09: zaro may know more | 19:15 |
| *** sandywalsh has joined #openstack-infra | 19:15 | |
| mordred | the isolation makes sense to me - tempest is not part of an openstack install - it's a thing taht sits outside of the install and judges it | 19:16 |
| fungi | okay, sort of back now | 19:16 |
| mordred | fungi: https://review.openstack.org/#/c/178886/ if you have a sec - I'd like to get that landed and cut a release | 19:16 |
| jeblair | mordred: ack; i hope to be reviewing after lunch | 19:17 |
| mordred | jeblair: thanks | 19:17 |
| mtreinish | mordred: yeah, although I don't think we want the isolation on master (we do use it for gate runs though) But, having tempest installed systemwide on a devstack install is a better dev story for most | 19:18 |
| *** mjturek1 has joined #openstack-infra | 19:18 | |
| zaro | nikita09: are you doing multiple scm or just one? | 19:18 |
| mordred | jeblair, clarkb, fungi, Shrews: also https://review.openstack.org/#/c/175891/ ... those are the last two that need to go in this release | 19:19 |
| lifeless | mtreinish: co-installability, not uninstallability | 19:19 |
| lifeless | mtreinish: with master | 19:19 |
| nikita09 | zaro: i have multiple scm macros...but im in my im using just one scm | 19:19 |
| *** ildikov has quit IRC | 19:19 | |
| lifeless | mtreinish: I have a chain of assumptions; let me spell it out | 19:19 |
| *** mjturek1 has quit IRC | 19:19 | |
| lifeless | mtreinish: ops value us having everything co-installable; I assume that that includes tempest | 19:20 |
| zaro | clarkb: are we using jjb from master or the latest release (v 0.5.5)? | 19:20 |
| *** mjturek1 has joined #openstack-infra | 19:20 | |
| lifeless | mtreinish: global-requirements inclusion in projects.txt is how we ensure that co-installability is maintained for a given branch, including master | 19:20 |
| clarkb | zaro: we use master, latest release is 1.20 | 19:20 |
| zaro | clarkb: jjb/geard | 19:20 |
| clarkb | er 1.2.0 | 19:20 |
| *** yfried is now known as yfried|afk | 19:20 | |
| clarkb | zaro: do you mean gearman-plugin? | 19:20 |
| zaro | nikita09: multiple scm requires a special plugin | 19:21 |
| lifeless | mtreinish: ... and I just noticed your patch is to s/kilo. :) | 19:21 |
| zaro | nikita09: do you have that? | 19:21 |
| zaro | clarkb: no python gear | 19:21 |
| clarkb | zaro: nikita09's error appears to be in jjb not jenkins | 19:21 |
| zaro | clarkb: the server | 19:21 |
| mtreinish | lifeless: oh, heh. Yeah it;s just for stable :) | 19:21 |
| clarkb | zaro: python gear we install from latest release | 19:21 |
| *** emagana has quit IRC | 19:21 | |
| nikita09 | zaro: WHICH plugin?? jenkins multiple SCM plugin?? | 19:21 |
| nikita09 | zaro: i have that jenkins plugin... | 19:22 |
| zaro | clarkb: ok. i was testing with v 5.5 | 19:22 |
| *** ZZelle_ has joined #openstack-infra | 19:22 | |
| *** salv-orlando has joined #openstack-infra | 19:22 | |
| nikita09 | zaro: i also tested passing multiple branches to git scm macro...got same 'keyerror obj' ... :/ i -passed like this >> branch: !!python/list ['branch1',branch2','branch3'] and it git scm >>> branches : '{obj:branch}' | 19:23 |
| openstackgerrit | Merged openstack/requirements: Bump requests version support https://review.openstack.org/173290 | 19:24 |
| openstackgerrit | Merged openstack/requirements: update openstackclient requirement to 1.0.3 https://review.openstack.org/177337 | 19:25 |
| mtreinish | jeblair: if you get a sec: https://review.openstack.org/175008 should be super simple :) | 19:26 |
| *** e0ne has quit IRC | 19:26 | |
| *** claudiub has quit IRC | 19:26 | |
| clarkb | mordred: it seems really weird that the flag to use tasks for image uploads is a cloud flag | 19:26 |
| clarkb | mordred: but thats because rax glance v2 is not glance v2 | 19:27 |
| clarkb | mordred: right? | 19:27 |
| Shrews | mordred: oh, this can merge now https://review.openstack.org/168673 | 19:27 |
| Shrews | err | 19:27 |
| Shrews | https://review.openstack.org/168672, rather | 19:28 |
| nikita09 | zaro: the thing is, JJB git SCM doesnt understand what 'obj' is ...i see thats a feature described for project templates, but not sure if its also supposed to work for git scm too.... | 19:28 |
| Shrews | mordred: though, hrm... guess we need swiftclient to make another release, so maybe not | 19:29 |
| clarkb | nikita09: I am pretty sure that we use the same substitution method on all the things, but it is possible there is a big | 19:29 |
| lifeless | clarkb: no, its because each v2 cloud can choose to run the tasks stuff or not | 19:29 |
| clarkb | er s/big/bug/ | 19:29 |
| *** _nadya_ has joined #openstack-infra | 19:29 | |
| clarkb | lifeless: then wtf are api versions for? | 19:29 |
| *** alexpilotti has quit IRC | 19:29 | |
| lifeless | clarkb: [also rax isn't quite standard, but thats actually a distraction here] | 19:29 |
| mordred | clarkb: well, yes and no | 19:29 |
| lifeless | clarkb: confusing users. | 19:29 |
| clarkb | lifeless: awesome | 19:29 |
| zaro | nikita09: git001.yaml test here: http://git.openstack.org/cgit/openstack-infra/jenkins-job-builder/tree/tests/scm/fixtures has config for multiple branches. | 19:29 |
| mordred | clarkb: it's a feature that takes a priori knowledge that cannot be inferred from the keystone catalog | 19:29 |
| zaro | nikita09: does the corresponding xml look correct to you? | 19:30 |
| clarkb | mordred: because it is not reflected in either the catalog of glance api version? | 19:30 |
| mordred | clarkb: so, the API for tasks that rackspace uses is the v2 tasks api | 19:30 |
| mordred | clarkb: that's right | 19:30 |
| clarkb | mordred: can we go back and -2 this glance stuff and make them delete v2 and tasks? | 19:30 |
| mordred | clarkb: so - there are two axes | 19:30 |
| mordred | clarkb: well, v2 isn't a problem | 19:30 |
| *** e0ne has joined #openstack-infra | 19:30 | |
| mordred | clarkb: PUT works in v2 | 19:30 |
| clarkb | mordred: then have them make a v3 that works? | 19:30 |
| mordred | there is "is this v1 or v2" | 19:30 |
| clarkb | mordred: sure except that v2 doesn't have a way of telling you to use tasks or not | 19:30 |
| mordred | and there is "must you use tasks" | 19:30 |
| lifeless | clarkb: I feel your pain, but thats just going to be worse still | 19:30 |
| clarkb | mordred: and if the are optional but sometimes required the api should tell you | 19:30 |
| mordred | clarkb: well, the "must use tasks" is a deployment policy decision at rax | 19:30 |
| zaro | nikita09: does that work in a macro? | 19:30 |
| clarkb | mordred: either via a capabilities requests or via making it required | 19:31 |
| mordred | clarkb: AIUI, what they did is reduce the available quota for PUT calls to 0 | 19:31 |
| mordred | or something | 19:31 |
| * mordred waves hands | 19:31 | |
| nikita09 | clarkb: can you please link me to an example job config where you guys use same list parameters passed ? using obj:key?? | 19:31 |
| clarkb | mordred: sure all I am suggesting is that it is glance's job to reflect this to the user | 19:31 |
| mordred | clarkb: 100% agree | 19:31 |
| clarkb | nikita09: we have none, we never use them | 19:31 |
| mordred | clarkb: I jjust think it's an even bigger problem | 19:31 |
| mordred | clarkb: because the same thign is potentially the case with every single thing that an operator can set in policy.json | 19:31 |
| clarkb | mordred: fun | 19:32 |
| mordred | clarkb: this just happens to be one that one of our deployers did set and we hit it and learned about it | 19:32 |
| mordred | clarkb: so, I'd argue that policy.json itself should be shot in the face | 19:32 |
| clarkb | so policy.json can affect the functionality of the api wihtout having any way to reflect such things to the user before they use the api | 19:32 |
| *** ildikov has joined #openstack-infra | 19:32 | |
| mordred | yes | 19:32 |
| clarkb | or at least before they use the api to interrogate functionality | 19:32 |
| clarkb | thats insane | 19:32 |
| mordred | and that exists as an idea across all of openstack | 19:32 |
| mordred | "deployer freedom" man | 19:32 |
| zaro | nikita09: http://git.openstack.org/cgit/openstack-infra/jenkins-job-builder/tree/tests/yamlparser/fixtures/custom_distri.yaml | 19:32 |
| *** salv-orlando has quit IRC | 19:33 | |
| zaro | nikita09: and http://git.openstack.org/cgit/openstack-infra/jenkins-job-builder/tree/tests/yamlparser/fixtures/expand-yaml-for-template-job/dimensionality-test001.yaml | 19:33 |
| nikita09 | zaro: thanks..yes i see same example in docs..but thats referring to project template...i was wondering if there is similar for git scm.. | 19:33 |
| *** mjturek2 has joined #openstack-infra | 19:34 | |
| clarkb | mordred: my only suggestion for shade then is to somehow distinguish between glance tasks and shade tasks | 19:34 |
| *** e0ne has quit IRC | 19:35 | |
| zaro | nikita09: no, unfortunately no test for that. so probably not supported in scm | 19:35 |
| clarkb | mordred: the docstring is ambiguous as to which task is in use here | 19:35 |
| nikita09 | zaro: i have tried passing list of branches in both ways... using [] and using @branch | 19:35 |
| nikita09 | - branch 1 - branch 2 - branch3 | 19:35 |
| clarkb | mordred: left a comment | 19:36 |
| nikita09 | zaro: both times it says 'keyerror: obj' so it means this feature is not enabled for git scm ..right? | 19:36 |
| zaro | nikita09: doesn't the git001.yaml do that? | 19:36 |
| *** mjturek1 has quit IRC | 19:36 | |
| zaro | nikita09: i see a list of branches getting passed to git scm | 19:37 |
| *** e0ne has joined #openstack-infra | 19:37 | |
| *** jgrimm is now known as zz_jgrimm | 19:37 | |
| nikita09 | zaro: git001.yaml is not a macro... | 19:38 |
| *** mjturek2 has quit IRC | 19:38 | |
| *** e0ne has quit IRC | 19:39 | |
| *** e0ne has joined #openstack-infra | 19:39 | |
| nikita09 | zaro: i have macro of git scm just like git001.yaml but in that, under `branches` , i passed a list of branches from my jobs... so branches : '{obj:mybranchlist}' and in my job , mybranchlist: !!python/list [a,b,c ,] etc etc | 19:39 |
| anteaya | mtreinish: tempest is being invoked here: https://etherpad.openstack.org/p/YVR-ops-user-committee | 19:40 |
| *** luqas has joined #openstack-infra | 19:40 | |
| anteaya | mtreinish: you may want to climb in | 19:40 |
| nikita09 | but it doesnt get what obj is ..... | 19:40 |
| *** yfried|afk is now known as yfried | 19:40 | |
| mtreinish | anteaya: I support both of those bullet points | 19:40 |
| anteaya | mtreinish: tim bell may appreciate your support | 19:41 |
| *** e0ne has quit IRC | 19:41 | |
| mtreinish | those are things I've been pushing for some time, but never got volunteers before | 19:41 |
| zaro | nikita09: ahh, sorry i misunderstood your use case. | 19:41 |
| anteaya | next time he reads that etherpad | 19:41 |
| anteaya | mtreinish: good to know, I didn't know that | 19:41 |
| nikita09 | zaro: np :) is it fine if i report a bug on storyboard for it ?? | 19:42 |
| anteaya | mtreinish: I wonder how many user committee folks know your thoughts | 19:42 |
| mtreinish | anteaya: should I just +1 it in the etherpad? Sorry I don't know the contxt of it | 19:42 |
| *** e0ne has joined #openstack-infra | 19:42 | |
| *** mjturek1 has joined #openstack-infra | 19:42 | |
| anteaya | this is tim bells etherpad | 19:42 |
| anteaya | sorry let me get context for you | 19:42 |
| zaro | nikita09: yes, please do. | 19:42 |
| *** e0ne has quit IRC | 19:42 | |
| pabelanger | Anybody know of a yaml schema validator? | 19:43 |
| pabelanger | ... for python | 19:43 |
| anteaya | mtreinish: context: http://lists.openstack.org/pipermail/openstack-dev/2015-May/063102.html | 19:43 |
| *** e0ne has joined #openstack-infra | 19:43 | |
| clarkb | pabelanger: voluptuous, zuul uses it | 19:43 |
| *** e0ne has quit IRC | 19:43 | |
| anteaya | mtreinish: this is the who can vote for tc thread, and this is tim saying he is trying to organize the user committee better via this etherpad | 19:43 |
| pabelanger | clarkb, Ah, yes. I knew that | 19:44 |
| pabelanger | thanks | 19:44 |
| mtreinish | anteaya: oh heh, no wonder I missed it, it was part of that thread I was keeping away from | 19:44 |
| anteaya | mtreinish: via, using in support of | 19:44 |
| anteaya | mtreinish: ah sorry to drag you in | 19:44 |
| anteaya | mtreinish: but tempest | 19:44 |
| anteaya | mtreinish: and things you want to support | 19:44 |
| mtreinish | hehe, no worries, I don't need to chime in on the thread, just comment on the etherpad :) | 19:44 |
| *** e0ne has joined #openstack-infra | 19:44 | |
| *** luqas has quit IRC | 19:45 | |
| *** sushilkm has joined #openstack-infra | 19:45 | |
| *** e0ne has quit IRC | 19:45 | |
| *** sushilkm has left #openstack-infra | 19:45 | |
| anteaya | mtreinish: yes I concur | 19:47 |
| mtreinish | anteaya: ok I left some comments | 19:47 |
| nikita09 | well...this also seems like a valid bug in jjb.. using awk statements that have {} , jjb thinks this this mapping to something.. liike awk '{print $1}' , jjb thinks 'print $1' is key in our macro :) lol https://storyboard.openstack.org/#!/story/2000180 | 19:48 |
| *** emagana has joined #openstack-infra | 19:48 | |
| *** tiswanso has quit IRC | 19:48 | |
| mtreinish | if the blinding red didn't give that away :) | 19:48 |
| anteaya | mtreinish: thank you | 19:48 |
| anteaya | mtreinish: ha ha ha | 19:48 |
| *** unicell has quit IRC | 19:49 | |
| anteaya | I always have to look for my sunglasses when you speak up | 19:49 |
| anteaya | :) | 19:49 |
| anteaya | not sure what they would want in terms of a ci here | 19:49 |
| jeblair | nikita09: should work if you double them "{{foo}}" | 19:50 |
| *** yfried is now known as yfried|afk | 19:50 | |
| mtreinish | anteaya: honestly we probably could leverage infra to just have a job which installs tempest somewhere and configures it to talk to public facing cloud X | 19:50 |
| mtreinish | I know that's something that came up in a summit session at some point in the past | 19:51 |
| *** emagana_ has joined #openstack-infra | 19:51 | |
| mtreinish | I think fungi or clarkb proposed just running a periodic job to have tempest run against infra's rax account | 19:51 |
| mtreinish | (which it turns out would fail miserably :) ) | 19:51 |
| anteaya | awesome | 19:52 |
| *** emagana has quit IRC | 19:52 | |
| anteaya | well I'm more at the let's keep discussing phase than let's decide on the solution, but that might be an interesting solution | 19:52 |
| anteaya | mtreinish: any thoughts on where I might dig up the etherpad that references said idea? | 19:53 |
| mtreinish | I'm not sure it would be in etherpad notes, iirc it was just a random comment at a session in HK (but super hazy) | 19:54 |
| fungi | mtreinish: yeah, you'd want a dedicated tenant for that, and make sure the job was nonconcurrent with other jobs using the same tenant probably | 19:54 |
| anteaya | mtreinish: can you be a bit more vague please | 19:54 |
| mtreinish | it might have been portland too | 19:54 |
| anteaya | mtreinish: way too much detail there | 19:54 |
| mtreinish | anteaya: ^^^ :) | 19:54 |
| anteaya | mtreinish: there we go, thanks | 19:54 |
| anteaya | spread it a little thinner | 19:54 |
| openstackgerrit | Joshua Harlow proposed openstack-infra/release-tools: Include 'source_url' in notes and use readme parser function https://review.openstack.org/176528 | 19:54 |
| anteaya | fungi: have we the ability to dedicate a tenant now, if we should want to? | 19:55 |
| mtreinish | fungi: sure, we probably would actually need multiple separate tenants to do it right (so we could run in parallel) | 19:55 |
| fungi | anteaya: we'd have to request one | 19:56 |
| * mtreinish remembers he needs to kill his rax node running his blog and close his account as it's no longer free | 19:57 | |
| anteaya | fungi: okay thanks, not jumping the gun, just trying to understand what is required should this route be selected | 19:57 |
| fungi | anteaya: not looking at the etherpad, i would infer this is for the benefit of confirming that tempest testing some real cloud returns sane results (to validate tempest's operation? to validate the cloud it's hitting? both?) | 19:57 |
| mtreinish | mikal: ^^^ unless you want to fix that for me :) | 19:57 |
| anteaya | fungi: both | 19:57 |
| *** e0ne has joined #openstack-infra | 19:57 | |
| mtreinish | fungi: yeah, both | 19:57 |
| *** _nadya_ has quit IRC | 19:57 | |
| anteaya | fungi: the way it is worded to verify the cloud code does as expected | 19:57 |
| fungi | anteaya: also once we have an infra-cloud, this might be waaaaay easier to organize | 19:58 |
| anteaya | but tempest code would be evaluated too as a result | 19:58 |
| anteaya | fungi: ohhhh | 19:58 |
| clarkb | mtreinish: oh ya I need to do that too | 19:58 |
| anteaya | fungi: might we be open to trying to recruite some ops folks to help with infra-cloud? | 19:58 |
| fungi | but yeah, validating our infra-cloud, while something we probably should do, doesn't necessarily help rackspace. that's where stuff like refstack is aiming to fill the gap i think | 19:58 |
| anteaya | should I seed the etherpad with that idea? | 19:58 |
| *** wenlock has quit IRC | 19:59 | |
| anteaya | fungi: funny so far none of the ops folks have mentioned refstack | 19:59 |
| *** dboik has quit IRC | 19:59 | |
| fungi | i am unsure they're even aware | 19:59 |
| *** mriedem has quit IRC | 19:59 | |
| fungi | that's probably one of those places where we're getting too siloed and people are working in parallel toward similar goals without even knowing about each other | 19:59 |
| *** dboik has joined #openstack-infra | 19:59 | |
| * mtreinish probably should try to be nice and not say anything | 19:59 | |
| openstackgerrit | Merged openstack-dev/pbr: Stop re-invoking pip https://review.openstack.org/175369 | 20:00 |
| *** jcoufal has quit IRC | 20:00 | |
| anteaya | mtreinish: good strategy | 20:00 |
| anteaya | fungi: possibly | 20:00 |
| fungi | mtreinish: all i'll say is that at least one if not both groups might benefit from some additional help from outside their silo | 20:00 |
| jeblair | fungi: ++ | 20:00 |
| anteaya | fungi: that is a diplomatic thing to say | 20:00 |
| anteaya | and I can agree with it | 20:01 |
| mtreinish | fungi: heh, I wasn't taking issue with that. That's a good statement | 20:01 |
| *** dboik_ has joined #openstack-infra | 20:02 | |
| *** claudiub has joined #openstack-infra | 20:03 | |
| *** wenlock1 has quit IRC | 20:04 | |
| *** dboik has quit IRC | 20:05 | |
| *** SumitNaiksatam has quit IRC | 20:07 | |
| * tchaypo joins the conversation hours late | 20:07 | |
| tchaypo | nibalizer: you mentioned “adding zuul-cloner” but I couldn’t figure out what you were talking about adding it to | 20:08 |
| *** subscope_ has quit IRC | 20:08 | |
| clarkb | tchaypo: the openstack puppet module beaker tests | 20:08 |
| clarkb | tchaypo: right now they only test the current change, there is no integration testing or cogating going on | 20:08 |
| tchaypo | and the concern is that adding it to the tests makes it hard to run outside infra? | 20:09 |
| *** Somay has joined #openstack-infra | 20:10 | |
| nibalizer | tchaypo: basically | 20:10 |
| nibalizer | tchaypo: tests on a local workstation should do the right thing | 20:10 |
| *** doude has quit IRC | 20:10 | |
| *** doude has joined #openstack-infra | 20:10 | |
| tchaypo | where “right thing” means “same tests as CI will do”? | 20:10 |
| greghaynes | fungi: Hey, any idea where the change to add https://zuul.openstack.org/status/change/ is? | 20:11 |
| *** otter768 has joined #openstack-infra | 20:11 | |
| clarkb | tchaypo: yes, but using vagrant | 20:11 |
| fungi | greghaynes: i +2'd it, i think it needs a second set of eyes | 20:11 |
| clarkb | tchaypo: since locally the expectation from beaker is to spin up VMs | 20:11 |
| *** SumitNaiksatam has joined #openstack-infra | 20:12 | |
| clarkb | mordred: reading https://review.openstack.org/#/c/175891/24/shade/cmd/inventory.py shade inventory is an ini file, why would we output yaml and json but not ini? | 20:12 |
| clarkb | Shrews: ^ maybe you know? | 20:12 |
| greghaynes | clarkb: inventory plugin | 20:12 |
| mordred | clarkb: 2 things | 20:12 |
| mordred | clarkb: that's not an ansible inventory plugin | 20:12 |
| mordred | clarkb: that's a script that a user can use to get a printout of what's in their clouds without necessarily seeing the extra ansible things like group info | 20:13 |
| fungi | greghaynes: oh, it's https://review.openstack.org/179764 and merged several hours ago | 20:13 |
| mordred | clarkb: I wind up looking at printouts of my ansible inventory a lot of times to just see somethinigs about the data | 20:13 |
| greghaynes | fungi: ah! that would explain why my digging through open reviews isnt finding it | 20:13 |
| mordred | clarkb: second - a dynamic ansible inventory is not ini - it is a program that outputs json | 20:13 |
| *** radez_g0n3 is now known as radez | 20:14 | |
| mordred | clarkb: so a program that outputs inventory in an ini format is not a thing that anything would consume | 20:14 |
| clarkb | mordred: can I take that json and pass it to ansible -i ? | 20:14 |
| mordred | clarkb: that's not how the mechanism works, no | 20:14 |
| fungi | greghaynes: working example https://zuul.openstack.org/status/change/179706,4 | 20:14 |
| mordred | clarkb: what you do is you pass the location of teh script to ansible -i | 20:14 |
| clarkb | mordred: then I think you need to also output ini | 20:14 |
| mordred | clarkb: no | 20:15 |
| clarkb | mordred: hold on | 20:15 |
| *** cloudnull has quit IRC | 20:15 | |
| fungi | greghaynes: that's a requirements change in the gate right now | 20:15 |
| mordred | clarkb: you tell ansible what the script is, and ansible runs it for you | 20:15 |
| greghaynes | fungi: ahoy | 20:15 |
| mordred | clarkb: and the script outputs json | 20:15 |
| clarkb | mordred: yes I get that | 20:15 |
| clarkb | mordred: BUT | 20:15 |
| clarkb | mordred: what if I want to edit the file first? I don't want to write a new script, just take output, make some changes then pass it back in | 20:15 |
| clarkb | mordred: its just really weird to me that the interface here is inconsistent | 20:15 |
| mordred | clarkb: this script is not an ansible inventory | 20:15 |
| *** otter768 has quit IRC | 20:16 | |
| Shrews | this is *dynamic* inventory | 20:16 |
| clarkb | mordred: except you just said I could pass it to ansible for invetory purposes | 20:16 |
| mordred | this is not a tool for generating static inventories | 20:16 |
| mordred | clarkb: sorry - what I meant was | 20:16 |
| mordred | you can pass an ansible dynamic inventory script to ansible | 20:16 |
| mordred | this is not an ansible inventory script | 20:16 |
| clarkb | ok, then what is it? | 20:16 |
| mordred | this is a program to produce some information about your clouds for whatever purpose you might find useful | 20:16 |
| *** rockyg has quit IRC | 20:17 | |
| mordred | I personally use it to debug looking at all of my cloud info | 20:17 |
| clarkb | and one of those useful things is not giving the data back to ansible? | 20:17 |
| Shrews | clarkb: http://docs.ansible.com/intro_dynamic_inventory.html | 20:17 |
| mordred | no | 20:17 |
| clarkb | ok | 20:17 |
| mordred | beause I havea n ansible dynamic inventory plugin that uses the same underlying code as this, but does othe rthings too | 20:17 |
| mordred | like add group information and ansible_ssh_hostname varaibles | 20:17 |
| rcarrill` | i use the ansible inventory to target plays against servers that have a certain nova metadata | 20:17 |
| tchaypo | I think http://docs.ansible.com/developing_inventory.html is the better link for explaining what a dynamic inventory script looks like | 20:17 |
| rcarrill` | it's pretty cool | 20:17 |
| *** rcarrill` is now known as rcarrillocruz | 20:17 | |
| mordred | yup | 20:17 |
| clarkb | so I understand whay a dynamic inventory script is | 20:18 |
| *** cloudkiller has joined #openstack-infra | 20:18 | |
| clarkb | I understand it outputs jsob | 20:18 |
| clarkb | I also understand that ansible -i can take an ini input file | 20:18 |
| clarkb | so an inventory script should support that too imo | 20:18 |
| clarkb | but if this isn't an inventory script then whatever | 20:18 |
| lifeless | jsob? | 20:18 |
| clarkb | sorry json | 20:18 |
| lifeless | do you men json or bson... | 20:18 |
| lifeless | thanks | 20:18 |
| mordred | lifeless: json | 20:19 |
| greghaynes | jsob is what you do when you pass json via command line arguments | 20:19 |
| fungi | boson | 20:19 |
| fungi | awesome | 20:19 |
| clarkb | greghaynes: ++ | 20:19 |
| mordred | clarkb: I could see the thing you're saying - but it's not really a thing with ansible inventories so far | 20:19 |
| mordred | clarkb: I'd be happy though to add that as a feature to the openstack one | 20:19 |
| mordred | clarkb: if you thnk it's a thing you'd find useful | 20:19 |
| clarkb | mordred: sure it is, I run command get ini file. Now I can do what I want with that and pass it to -i | 20:19 |
| mordred | clarkb: since it would not be ahrd | 20:19 |
| mordred | clarkb: no, what I'm saing is that that's not how people in ansible world do this | 20:19 |
| mordred | and none of the current ansible tools do it | 20:19 |
| mordred | however, it's an interesting idea | 20:20 |
| greghaynes | clarkb: im curious why you would want to do that? | 20:20 |
| * fungi is going to create a bunch of json-like serialization languages named after subatomic particles | 20:20 | |
| mordred | and easy to add | 20:20 |
| clarkb | mordred: no, instead they reinvent the world and write a NEW script to do dynamic inventory | 20:20 |
| *** doug-fish has joined #openstack-infra | 20:20 | |
| mordred | well, no - there's no need to do that | 20:20 |
| mordred | you can have multiple inventoies | 20:20 |
| sdague | fungi: hmmm... so when I try to urlopen from python https://zuul.openstack.org/status/change/179706,4 | 20:20 |
| rcarrillocruz | yeah | 20:20 |
| clarkb | greghaynes: first thing that comes to mind is to be extra sure that any restricted roll outs are restricted | 20:20 |
| mordred | so if you want to suppliment the openstack dynamic inventory | 20:20 |
| rcarrillocruz | you point to a folder containing several inventories | 20:20 |
| clarkb | greghaynes: ansible can't touch a host if it doesn't know about it | 20:20 |
| mordred | you just put some more static inventory next to it | 20:20 |
| sdague | I get an SSL error | 20:20 |
| sdague | URLError: <urlopen error [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed (_ssl.c:590)> | 20:20 |
| clarkb | greghaynes: then I don't have to worry about bugs in my inventory | 20:20 |
| mordred | clarkb: see what I just said ^^ | 20:20 |
| mordred | clarkb: you make your inventory point to a dir - one of the things in the dir is a script, the other things are static files | 20:21 |
| clarkb | mordred: thats additive not restricting nodes | 20:21 |
| greghaynes | clarkb: you can also do -l to limit your inventory ontop | 20:21 |
| greghaynes | er, limit ontop your inventory | 20:21 |
| clarkb | mordred: so I am not sure how I would say "make sure this node never shows up when I run the new scary thing" | 20:21 |
| greghaynes | -l !host or somesuch | 20:21 |
| rcarrillocruz | you can exclude in the 'hosts' section | 20:21 |
| mordred | clarkb: so - I understand your general usecase, but I don't have a concrete example of it in mind, so it's hard to conjecture about how it might be solved for real... that said, I think adding a thing that outputs some ini would be pretty easy to do | 20:22 |
| clarkb | greghaynes: rcarrillocruz ok did not know bout -l | 20:22 |
| mordred | and I'm happy to do it | 20:23 |
| greghaynes | :) its nifty | 20:23 |
| openstackgerrit | Doug Hellmann proposed openstack/requirements: Remove lockfile from requirements https://review.openstack.org/151224 | 20:23 |
| mordred | if anything - it could be used for debugging purposes | 20:23 |
| mordred | "give me a static ini-style output I can futz with" | 20:23 |
| fungi | sdague: i'll take a look at why my browser is cool with it but other tools might not be | 20:23 |
| fungi | sdague: it's possible the chain cert there is wrongish or something | 20:23 |
| *** cloudkiller has quit IRC | 20:23 | |
| *** sigmavirus24 has quit IRC | 20:23 | |
| *** d34dh0r53 has quit IRC | 20:23 | |
| mordred | clarkb: also - fwiw, you can do static file ansible inventories in json and yaml too | 20:24 |
| openstackgerrit | Michael Krotscheck proposed openstack/requirements: Updated oslo_middleware to 1.2.0 https://review.openstack.org/179926 | 20:24 |
| mordred | clarkb: they do not have to be ini | 20:24 |
| *** hockeynut has quit IRC | 20:24 | |
| clarkb | mordred: so I have to say this is where the docs break down | 20:24 |
| clarkb | mordred: everything I have read is ini | 20:24 |
| mordred | I believe they make all the docs ini because it's the easiest thing for people to do | 20:24 |
| *** d34dh0r53 has joined #openstack-infra | 20:24 | |
| *** hockeynut has joined #openstack-infra | 20:24 | |
| clarkb | and honestly yaml is probably easier to read and I would likely just use that if I knew I could | 20:24 |
| mordred | I'm 99% sure you can do that | 20:24 |
| mordred | I _know_ you can for variables files | 20:24 |
| *** cloudkiller has joined #openstack-infra | 20:25 | |
| mordred | and for local facts | 20:25 |
| tchaypo | nibalizer: do you have specific concerns about zuul_cloner, or is it just the complexity? | 20:25 |
| mordred | clarkb: I'll verify and get back to you | 20:25 |
| sdague | fungi: so I can force this back to http for my testing proxy, I'm just surprised | 20:25 |
| clarkb | tchaypo: I think its "how do we make beaker and zuul_cloner play nice" | 20:25 |
| *** sigmavirus24 has joined #openstack-infra | 20:25 | |
| mordred | clarkb: and yes- much docs hacking is no my TDL | 20:25 |
| clarkb | tchaypo: importantly beaker should not run zuul-cloner imo | 20:25 |
| clarkb | tchaypo: because that adds a weird dependency, Instead beaker should say "give me the thngs to use else use sane default" basically what devstack does | 20:26 |
| *** tiswanso has joined #openstack-infra | 20:26 | |
| mordred | ++ | 20:26 |
| fungi | sdague: yeah, i believe crd doesn't kick in when you're supplying local javascript in the debugger console or something | 20:26 |
| mordred | I totally agree with clarkb there ... | 20:26 |
| tchaypo | maybe I should step back and ask what problem zuul_cloner could be used to solve | 20:26 |
| mordred | "if I made one for you, use it, otherwise, do this" | 20:26 |
| clarkb | tchaypo: it solves the problem of what should be tested | 20:26 |
| *** e0ne has quit IRC | 20:26 | |
| clarkb | tchaypo: currently beaker doesn't really solve that at all and also doesn't give you a good interface to communicate your decisions to it | 20:27 |
| mordred | tchaypo: I think clarkb is agreeing with you - that we shoudl not teach beaker about zuul cloner | 20:27 |
| sdague | fungi: crd? | 20:27 |
| mordred | but instead, should teach beaker how to deal with repos that have already been managed by zuul cloner (or anything else) | 20:27 |
| mordred | yeah? | 20:27 |
| clarkb | mordred: ya | 20:27 |
| mordred | and I agree with that | 20:27 |
| clarkb | and from what I am reading that is basically going to look like a hook to beaker the s_h_a thing that would do the correct thing for us and not beaker itself | 20:27 |
| mordred | that way the scripts run by beaker can look like things that want to clone and work with local things like people expect them to be | 20:27 |
| mordred | when they're run by $person on $laptop | 20:28 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: Do not automatically target bugs for stable releases https://review.openstack.org/179447 | 20:28 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: Add list_oslo_unreleased_changes.sh https://review.openstack.org/179396 | 20:28 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: Print release notes as part of releasing the library https://review.openstack.org/179445 | 20:28 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: Move get_last_tag to functions https://review.openstack.org/179444 | 20:28 |
| *** dustins has quit IRC | 20:29 | |
| tchaypo | as in, use these three local repos for these modules, but get HEAD for everything else? | 20:29 |
| *** ociuhandu has quit IRC | 20:29 | |
| *** dustins has joined #openstack-infra | 20:29 | |
| *** Sukhdev has quit IRC | 20:29 | |
| sdague | fungi: is there anything special about zuul's cert? | 20:30 |
| *** e0ne has joined #openstack-infra | 20:30 | |
| clarkb | tchaypo: that would be one way, I think its probably easier to do all local repos or use HEAD for everything though | 20:30 |
| clarkb | tchaypo: that way you get clear failure modes rather than seemingly random workingness | 20:30 |
| *** jraim has quit IRC | 20:30 | |
| *** boris-42 has quit IRC | 20:30 | |
| *** kun_huang has quit IRC | 20:30 | |
| clarkb | "it works for module X but not Y, what is going on?: | 20:30 |
| *** cyeoh has quit IRC | 20:31 | |
| *** apmelton has quit IRC | 20:31 | |
| *** Trozz has quit IRC | 20:31 | |
| *** dboik_ has quit IRC | 20:31 | |
| *** gondoi has quit IRC | 20:31 | |
| openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: update make_library_stable_branch.sh to set defaultbranch https://review.openstack.org/173919 | 20:31 |
| *** tiswanso has quit IRC | 20:31 | |
| *** Trozz has joined #openstack-infra | 20:31 | |
| *** Nakato has quit IRC | 20:32 | |
| *** tiswanso has joined #openstack-infra | 20:32 | |
| clarkb | mordred: last inventory question, shouldn't the --list and --host options imply no print(output_format_dict)? | 20:32 |
| *** persia has quit IRC | 20:32 | |
| *** cloudkiller is now known as cloudnull | 20:32 | |
| clarkb | oh wait I see, thoseare passed to output_format_dict and it renders them and inventory depending, nvm | 20:32 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Switch tasks vs put on a boolean config flag https://review.openstack.org/178886 | 20:32 |
| *** Nakato has joined #openstack-infra | 20:33 | |
| mordred | clarkb: ^^ fixed your comment - good call on the confusing | 20:33 |
| clarkb | mordred: I just read it and its "use task manager, use tasks for uploads" | 20:33 |
| *** hockeynut has quit IRC | 20:33 | |
| *** jraim has joined #openstack-infra | 20:33 | |
| clarkb | and went wait a minute | 20:33 |
| *** fungi has quit IRC | 20:33 | |
| *** persia has joined #openstack-infra | 20:34 | |
| *** persia has quit IRC | 20:34 | |
| *** persia has joined #openstack-infra | 20:34 | |
| *** jvrbanac has quit IRC | 20:34 | |
| *** apmelton has joined #openstack-infra | 20:34 | |
| *** hockeynut has joined #openstack-infra | 20:34 | |
| *** unicell has joined #openstack-infra | 20:34 | |
| *** julim has quit IRC | 20:34 | |
| *** tchaypo has quit IRC | 20:34 | |
| *** cyeoh has joined #openstack-infra | 20:34 | |
| *** d34dh0r53 has quit IRC | 20:34 | |
| *** sigmavirus24 has quit IRC | 20:34 | |
| *** gondoi has joined #openstack-infra | 20:34 | |
| *** dboik has joined #openstack-infra | 20:34 | |
| *** kun_huang has joined #openstack-infra | 20:34 | |
| *** jvrbanac has joined #openstack-infra | 20:35 | |
| *** d34dh0r53 has joined #openstack-infra | 20:35 | |
| clarkb | mordred: actually one more question :) why are you using flamel? yaml.dump(thing, default_flow_style=False) should work too right? | 20:35 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: Add script to update .gitreview files https://review.openstack.org/174594 | 20:35 |
| *** fungi has joined #openstack-infra | 20:35 | |
| *** sigmavirus24 has joined #openstack-infra | 20:35 | |
| *** boris-42 has joined #openstack-infra | 20:36 | |
| mordred | clarkb: flamel does a few other things that are important that I did not want to cargo cult in - most importantly it manages unicode representaiton and also knows how to deal with ordereddicts | 20:36 |
| *** tchaypo has joined #openstack-infra | 20:37 | |
| clarkb | ok | 20:37 |
| mordred | clarkb: granted, it's only 14 lines of python - but it's in a library, so why not :) | 20:37 |
| tchaypo | woo, that was fun | 20:37 |
| clarkb | out of curiousity what unicode things need to be done? it should all be utf8 | 20:37 |
| mordred | def unicode_representer(dumper, uni): | 20:37 |
| clarkb | and pyyaml will handle ^ for you | 20:37 |
| mordred | node = yaml.ScalarNode(tag=u'tag:yaml.org,2002:str', value=uni) | 20:37 |
| mordred | return node | 20:37 |
| tchaypo | I love it when I randomly miss 10 minutes or irc in the middle of conversation | 20:37 |
| clarkb | mordred: I am not sure that is necessary | 20:38 |
| clarkb | mordred: but ok | 20:38 |
| mordred | clarkb: I believe I hit it - it's possible it isnt' - I'll go back and figure it out | 20:38 |
| clarkb | yaml strings are always utf8 | 20:38 |
| clarkb | unless you do other crazness which we arne't | 20:38 |
| mtreinish | lifeless: fwiw, I use arch linux on most of my machines and pbr works fine :) | 20:39 |
| mordred | clarkb: these strings don't necessarily always come from a sane place | 20:39 |
| fungi | sdague: yep, wrong/incorrect chain... fixing now | 20:39 |
| clarkb | mordred: hrm how do you handle ordered dict? either you serialize to regular dict and lose ordering or you serialize a python object which yaml should already know how to do | 20:40 |
| *** fawadkhaliq has joined #openstack-infra | 20:40 | |
| *** dizquierdo has joined #openstack-infra | 20:41 | |
| clarkb | (or I suppose some third representation that only a special yaml loader can load) | 20:41 |
| openstackgerrit | Ramy Asselin proposed openstack-infra/puppet-openstackci: Create a log server like http://logs.openstack.org https://review.openstack.org/167425 | 20:41 |
| mordred | clarkb: the first - serialize to a regular dict and lose ordering | 20:41 |
| mordred | clarkb: if you serilize to a representaiton, then you could only load back into python with pyyaml | 20:41 |
| clarkb | mordred: correct, but the alternative is to lose ordering which is supposedly important | 20:42 |
| mordred | well, you don't actually lose ordering it turns out | 20:42 |
| clarkb | both seem bad | 20:42 |
| mordred | in the flamel case | 20:42 |
| clarkb | mordred: you should, python hashes starting with python3 (and any use of tox) have random seeds | 20:42 |
| mordred | right. so - this is mainly to keep things human readable - not to keep actual important data an ordered fashion | 20:43 |
| *** zz_jgrimm has quit IRC | 20:43 | |
| mordred | so far it's working to output the stuff in a mostly sane order - i"m not 100% how that's working | 20:43 |
| openstackgerrit | Ramy Asselin proposed openstack-infra/puppet-openstackci: Simple script to setup and test log server https://review.openstack.org/169117 | 20:44 |
| mordred | but in flamel's case, it means when you read the json in and then write it back out as yaml, the keys are in pretty much teh same arragement | 20:44 |
| mordred | and vice versa | 20:44 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: use optimized url for zuul status https://review.openstack.org/170935 | 20:44 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: add dbg function to hideci https://review.openstack.org/179573 | 20:44 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: add jquery visibility https://review.openstack.org/179362 | 20:44 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: create proxy for testing hideci https://review.openstack.org/179360 | 20:44 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: Revert "Revert zuul status in gerrit" https://review.openstack.org/179361 | 20:44 |
| *** maurosr has quit IRC | 20:44 | |
| sdague | fungi: cool, thank you | 20:44 |
| sdague | greghaynes: so... I think there isn't a sufficiently better interface for dbg atm given the contraints of js | 20:44 |
| sdague | so how do you feel about just living with the slightly wonky interface for now, which is at least useful | 20:45 |
| clarkb | mordred: it looks like flamel also ships a loader that preserves order, but it does so by assuming all dicts are ordered from what I can tell | 20:45 |
| clarkb | mordred: either way we lose some data | 20:45 |
| *** e0ne has quit IRC | 20:45 | |
| *** sdake has joined #openstack-infra | 20:45 | |
| sdague | fungi: also, in this go around of all these patches I put all the zuul embedded logic behind a kill switch - https://review.openstack.org/#/c/179361/6/modules/openstack_project/files/gerrit/hideci.js,cm | 20:46 |
| *** maurosr has joined #openstack-infra | 20:46 | |
| sdague | so flipping it off is a single global | 20:46 |
| greghaynes | sdague: oh, my specific suggestion was to use console.log(arguments[0]); and start the for loop at 1 and do console.dir(arguments[i]); | 20:46 |
| sdague | if it's wrong | 20:46 |
| *** markvan__ is now known as markvan | 20:46 | |
| sdague | greghaynes: ok, what does that buy us? | 20:46 |
| *** zz_jgrimm has joined #openstack-infra | 20:46 | |
| mordred | clarkb: yah. in this case I mainly wanted to make sure that we weren't outputting unreadable pyyaml unicode tags - I don't think the ordereing really comes in to play | 20:46 |
| mordred | at least not for the inventory output thing | 20:46 |
| clarkb | mordred: ok | 20:46 |
| nibalizer | tchaypo: I think we should use zuul_cloner | 20:46 |
| greghaynes | sdague: youll always get a nagivable object in the case youre using - using console.log doesnt necessarially give you that | 20:46 |
| sdague | because console.log(obj) does a useful thing, at least in chrome | 20:47 |
| nibalizer | but I want to advocate early for ensuring a happy path forward for local development | 20:47 |
| sdague | oh, it's always doing it here | 20:47 |
| mordred | clarkb: but you're right - I'll go verify a few of those things because I'm curious now | 20:47 |
| *** hasharAway is now known as hashar | 20:47 | |
| greghaynes | sdague: yes, apparently that depends on argument type and browser | 20:47 |
| sdague | well, all of the interface does :) | 20:47 |
| sdague | console is kind of a ghetto interface that no one standardized | 20:47 |
| greghaynes | sdague: yes, supposedly .dir is consistent about giving you an object though. Im not hugely bent on it I just know ive run into a lot of cases of "huh, why am I all of a sudden not getting an object" | 20:48 |
| *** samueldmq_ has joined #openstack-infra | 20:48 | |
| *** amitgandhinz has quit IRC | 20:48 | |
| sdague | ok | 20:49 |
| greghaynes | sdague: not sure if you saw https://review.openstack.org/#/c/170935/ ye | 20:49 |
| *** pabelanger has quit IRC | 20:49 | |
| *** sdake_ has quit IRC | 20:49 | |
| greghaynes | sdague: my brain hurt reasoning that one out, so let me know if it makes sense | 20:49 |
| *** mrmartin has quit IRC | 20:50 | |
| jogo | greghaynes: FYI one of your wheel patches needs a rebase https://review.openstack.org/#/c/164927/ | 20:50 |
| sdague | so... yeh, the challenge is the following.... | 20:50 |
| tchaypo | nibalizer: why do you think we should use it? | 20:51 |
| greghaynes | jogo: oy | 20:51 |
| greghaynes | jogo: again :( | 20:51 |
| sdague | we trigger a timer loop for zuul polling based on the current url | 20:51 |
| jogo | greghaynes: yeah :( | 20:51 |
| sdague | which self propogates, because javascript doesn't have threads | 20:51 |
| sdague | if you navigate to a different url | 20:51 |
| *** sdake_ has joined #openstack-infra | 20:52 | |
| sdague | when the current request cycles back around it needs to realize "oh, we're not there any more, stop" | 20:52 |
| sdague | if you have a better suggestion on how to structure that code, please pastebin or post a solution | 20:52 |
| sdague | because... thems as good as I got | 20:52 |
| greghaynes | sdague: ah. So that part is desired. What about the visibility interaction though? | 20:52 |
| greghaynes | I didnt realize that was a feature though ;) | 20:52 |
| sdague | you are right, we want to restart that on a return to visible | 20:53 |
| sdague | oh, we do that | 20:53 |
| sdague | #516 | 20:53 |
| sdague | we just start the top loop again | 20:54 |
| asselin__ | nibalizer, I made the change to ::jenkins but I think it will fail b/c it's not merged yet. Let's do these one at a time | 20:54 |
| zaro | clarkb: i'm seeing something odd with gearman. | 20:54 |
| asselin__ | nibalizer, so I plan to revert back...let me know your opinion though... | 20:54 |
| nibalizer | asselin__: okay let me review | 20:54 |
| greghaynes | sdague: ah! ok. I think that has some race issues then | 20:54 |
| sdague | yeh, for visibility that was a jeblair requested feature that if the tab goes hidden in the browser that we stop polling | 20:54 |
| zaro | clarkb: i send a job to the server and keep the connection until job ends and that works fine. | 20:54 |
| sdague | so we aren't adding load for hidden tabs | 20:54 |
| sdague | greghaynes: probably :) | 20:54 |
| greghaynes | sdague: I think if you flip between tabs before the timeout occurs you end up with two loops | 20:55 |
| *** fawadkhaliq has quit IRC | 20:55 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/puppet-zuul: Correct the variable for the chain cert https://review.openstack.org/179935 | 20:55 |
| fungi | sdague: ^ i finally tracked the typo back through all my puppeghetti | 20:55 |
| sdague | greghaynes: yeh | 20:55 |
| *** miqui has quit IRC | 20:55 | |
| zaro | clarkb: when i send a job to the server without keeping a connection then the server reponds with NO_JOB | 20:55 |
| *** salv-orlando has joined #openstack-infra | 20:55 | |
| sdague | it's not the end of the world though | 20:55 |
| nibalizer | asselin__: you mean in line 30 of https://review.openstack.org/#/c/167425/10/manifests/logserver.pp ? | 20:55 |
| clarkb | zaro: ya, if you disconnect from the server then the server assumes you are no longer valid | 20:55 |
| zaro | clarkb: but clearly the job is registered. | 20:55 |
| *** sdake has quit IRC | 20:56 | |
| asselin__ | nibalizer, yes | 20:56 |
| clarkb | zaro: and es it doesn't unregister the job | 20:56 |
| fungi | sdague: i'm _shockec_ (okay not really) that my browser didn't just asplode all over that error | 20:56 |
| zaro | clarkb: the job doesn't get executed | 20:56 |
| sdague | fungi: go go SSL? | 20:56 |
| zaro | clarkb: in the past the job would get executed. | 20:56 |
| nibalizer | asselin__: why would that fail? there is a jenkins class in the jenkins module now right? | 20:56 |
| sdague | yeh, that's not making me feel super excited about security of browsers :) | 20:56 |
| clarkb | zaro: I think the job should stop running if it started | 20:57 |
| waynr | clarkb greghaynes (or anyone) interested in pairing on github/zuul integration work? I am at the point where zuul is sending jobs to gearman based on updates to PRs but i'm not sure why the zuul log doesn't seem to reference job names...maybe it's not supposed to | 20:57 |
| sdague | greghaynes: so, do you have a suggestion on how to block that double update? | 20:57 |
| zaro | clarkb: huh? i expect the job to execute to completion. | 20:57 |
| zaro | clarkb: does zuul keep a connection open for every job it sends to the gearman server? | 20:58 |
| asselin__ | nibalizer, there's a jenkinsuser.pp | 20:58 |
| asselin__ | which is what I had before...right? | 20:58 |
| greghaynes | sdague: let me poke at it a bit, it takes me a little while to reason about that bit of code ;) | 20:59 |
| clarkb | zaro: no, zuul has one connection to the server iirc then the server has one connection per job | 20:59 |
| *** bswartz has quit IRC | 20:59 | |
| clarkb | zaro: if a worker connection goes away then server tells client that job failed | 20:59 |
| *** spzala has quit IRC | 20:59 | |
| zaro | clarkb: ohh yeah. that makes sense. I'm referring to connection from client to server. | 21:00 |
| *** kgiusti has left #openstack-infra | 21:00 | |
| waynr | clarkb: do you have any advice for monitoring/inspecting the gearman queue? | 21:00 |
| mordred | waynr: neat! | 21:00 |
| asselin__ | nibalizer, back in a bit.... | 21:00 |
| *** Sukhdev has joined #openstack-infra | 21:00 | |
| * mordred isn't going to be helpful - it purely just waiting for the day that work magically makes his life better | 21:00 | |
| zaro | clarkb: client sends job to zuul. worker sends GRAB_JOB server reponds with NO_JOB | 21:01 |
| clarkb | waynr: connect to gearman via telent/nc and run status against it to get a full listing | 21:01 |
| mordred | and with that - I'm going to the airport - I'll be back online in a bit | 21:01 |
| clarkb | waynr: you can also configure it to emit graphite events and track things with graphite | 21:01 |
| waynr | okay i'll give that a shot | 21:01 |
| zaro | clarkb: here's my simple client: https://github.com/zaro0508/gearman-plugin-client/blob/master/gear_client.py | 21:01 |
| *** luqas has joined #openstack-infra | 21:02 | |
| *** hdd has quit IRC | 21:02 | |
| openstackgerrit | Merged openstack/requirements: Follow project renames https://review.openstack.org/179706 | 21:02 |
| waynr | clarkb: just FYI, looking at my 'lsof' output it seems like gearman keeps one connection open per worker (rather than one per job) | 21:02 |
| zaro | clarkb: here's the log http://paste.openstack.org/show/214898/ | 21:03 |
| zaro | clarkb: this definately used to work. so i'm guessing there is some regression going on. | 21:03 |
| clarkb | waynr: sorry, ya that is correct | 21:03 |
| zaro | clarkb: i'm not sure how any jobs run because i'm pretty sure zuul sends jobs to gearman server just like this. | 21:04 |
| fungi | sdague: looks like the version of firefox i'm running has the ca that signed our zuul cert in its trust ring, and since apache was serving two copies of the server cert (rather than the server cert plus the chain cert) the browser probably just ignored one of them | 21:04 |
| clarkb | zaro: zuul alwys has a persistent connection | 21:05 |
| *** tonytan4ever has quit IRC | 21:05 | |
| sdague | fungi: gotcha | 21:05 |
| sdague | hey, slightly off topic, but does anyone have a link to where the ops tracks are listed with time slots? | 21:05 |
| lifeless | sdague: https://libertydesignsummit.sched.org/ ? | 21:06 |
| lifeless | sdague: https://docs.google.com/spreadsheets/d/1VsFdRYGbX5eCde81XDV7TrPBfEC7cgtOFikruYmqbPY/edit#gid=569963128 has rooms but not topics | 21:06 |
| sdague | lifeless: yep, thank you | 21:06 |
| *** luqas has quit IRC | 21:07 | |
| clarkb | ok comments on https://review.openstack.org/#/c/175891/ not sure if any of them are worth a -1 | 21:07 |
| sdague | the sched link seems useful enough | 21:07 |
| waynr | hmm it's not obvious to me how projects specified in layout.yaml get mapped to gearman functions and how gearman knows how many of its workers are capable of running each job | 21:07 |
| clarkb | Shrews: mordred SpamapS ^ feedback appreciated | 21:07 |
| clarkb | waynr: the names must match | 21:07 |
| clarkb | waynr: basically the name specific in layout.yaml is the gearman function name | 21:07 |
| zaro | clarkb: hmm, ok. more debugging.. | 21:07 |
| lifeless | sdague: http://libertydesignsummit.sched.org/type/design+summit/ops#.VUffnN_va00 may be better | 21:07 |
| *** ldnunes has quit IRC | 21:08 | |
| SpamapS | clarkb: I was just reviewing that right now as well. | 21:08 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/system-config: Create rubygems mirror from rubygems.org https://review.openstack.org/178026 | 21:09 |
| waynr | clarkb: i see functions that match the names i have specified in my layout.yaml, but I also see functions that exist nowhere in my project.yaml such as "merger:update", and I'm not sure which part of that job name actually needs to exist on my jenkins masters | 21:09 |
| openstackgerrit | Darragh Bailey proposed openstack-infra/git-review: Use plumbing rev-parse to get the branch name https://review.openstack.org/179937 | 21:09 |
| mordred | clarkb: I think they're worth fixing. new patch coming soon | 21:09 |
| EmilienM | clarkb, nibalizer: does it make sense if I file a bug for latest beaker not working? I would like to investigate more | 21:10 |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/project-config: Experimental Devstack job for Identity v3 API only https://review.openstack.org/179661 | 21:10 |
| clarkb | waynr: merger:update is a merger function registered by the merger worker | 21:10 |
| sdague | lifeless: they say Ops: so I was able to get what I wanted out of it | 21:10 |
| SpamapS | mordred: I'm also looking at why it takes 10s... ;) | 21:10 |
| sdague | thanks again | 21:10 |
| EmilienM | not working (in OS infra) | 21:10 |
| clarkb | waynr: you get that regardless of your layout.yaml | 21:10 |
| SpamapS | mordred: should take like 1s at the worst. | 21:10 |
| lifeless | sdague: cool | 21:10 |
| waynr | clarkb: do i need to configure a specific worker as the "merger" worker? | 21:10 |
| clarkb | waynr: layout.yaml does not register any functions, layout.yaml describes to the gearman client what functions to run on which events | 21:10 |
| clarkb | waynr: your workers all register the functions for you. In this case likel jenkins and the merger worker | 21:11 |
| waynr | oh...so the workers register the functions? | 21:11 |
| clarkb | waynr: its a separate service you run to merge things. Zuul may start one automagically for you like it starts the geard server though, check ps | 21:11 |
| waynr | okay | 21:11 |
| clarkb | waynr: yes workers advertise what work can be done, clients select from those lists | 21:11 |
| waynr | so there is a gearman function showing up in my gearmand status that doesn't appear on any of my workers | 21:12 |
| greghaynes | sdague: How about just a todo comment about theres a known race here :) | 21:12 |
| waynr | but it is a name that exists as a result of the interpolation of variables into my zuul project template | 21:12 |
| greghaynes | sdague: then I can poke at it in a followup | 21:13 |
| sdague | greghaynes: I'm good with that | 21:13 |
| waynr | oh wait i'm wrong, there is no way that function comes from any of my layout.yaml sorry | 21:13 |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/project-config: Experimental Devstack job for Identity v3 API only https://review.openstack.org/179661 | 21:14 |
| samueldmq | mordred, ^ project config patch is now updated to use DEVSTACK_LOCAL_CONFIG, thanks | 21:14 |
| *** dustins has quit IRC | 21:15 | |
| *** amitgandhinz has joined #openstack-infra | 21:15 | |
| fungi | clarkb: waynr: technicality, but zuul (the scheduled daemon) doesn't start the merger daemon, that's generally a second initscript | 21:15 |
| waynr | clarkb: so zuul will register gearman workers for stuff like "merger:update" | 21:15 |
| waynr | oh | 21:15 |
| fungi | s/scheduled/scheduler/ | 21:15 |
| mordred | samueldmq: woot! | 21:16 |
| waynr | i'm confused right now about where some of these gearman functions are coming from and why zuul doesn't seem to try to run jobs named in my layout.yaml for the particular project I am updating | 21:16 |
| fungi | waynr: those merger:.* functions are registered by the separate zuul-merger daemon which is running | 21:16 |
| waynr | ooohhh | 21:17 |
| waynr | well i didn't start any zuul-merger daemon just FYI | 21:17 |
| clarkb | right which is why I think zuul may fork one off for you | 21:18 |
| *** mmedvede has quit IRC | 21:18 | |
| clarkb | or run it internally | 21:18 |
| fungi | if it does, that's a new change | 21:18 |
| fungi | we still start ours from a separate zuul-merger initscript | 21:18 |
| openstackgerrit | Sean Dague proposed openstack-infra/system-config: use optimized url for zuul status https://review.openstack.org/170935 | 21:18 |
| sdague | there you go greghaynes | 21:19 |
| sdague | I hope you appreciate the comment :) | 21:19 |
| fungi | last i looked the only separate daemon process forked by the zuul scheduler daemon was the internalized geard | 21:19 |
| clarkb | quick grep says we don't fork a different process | 21:19 |
| *** hdd has joined #openstack-infra | 21:19 | |
| greghaynes | haha | 21:19 |
| *** hdd has quit IRC | 21:20 | |
| waynr | clarkb fungi that makes sense, because the merger:update function has 2 queued jobs that don't seem to be leaving the queue | 21:20 |
| waynr | so there is no worker actually accepting those jobs | 21:21 |
| fungi | waynr: and no workers? you probably need to start the zuul-merger daemon | 21:21 |
| waynr | okay, but i'm not sure i want anything merged by zuul...just want to run tests on jenkins based on PR updates | 21:21 |
| *** eharney has quit IRC | 21:22 | |
| fungi | waynr: you do want things merged by zuul. the zuul-merger is what serves the merged testing refs to your workers | 21:22 |
| fungi | so they have something to test | 21:22 |
| fungi | the zuul-merger creates your zuul refs | 21:22 |
| clarkb | ya github doesnt solve this problem for you | 21:22 |
| waynr | oh, i think i see what you're saying | 21:22 |
| waynr | because there may be multiple open PRs that could be merged | 21:23 |
| fungi | for example you proposed patch x to branch y, the zuul merger attempts to merge x to a local clone of y and then that gets served as ref z | 21:23 |
| *** blahdeblah has left #openstack-infra | 21:23 | |
| fungi | and the jenkins workers check out that ref to run jobs | 21:24 |
| nibalizer | github does expose those refs for you, per PR | 21:24 |
| nibalizer | its right there in the api "merge commit ref" or something | 21:24 |
| *** wenlock has joined #openstack-infra | 21:24 | |
| clarkb | nibalizer and those update as masyer moves? | 21:24 |
| nibalizer | doesn't help at gate time though, I think | 21:24 |
| nibalizer | clarkb: i think so, but :shrug: | 21:24 |
| nibalizer | i know i've never pressed the green button and then github has said 'oops sorry I mean it can't merge' | 21:25 |
| waynr | yeah github has refs of the form "refs/pulls/<PR#>" | 21:25 |
| nibalizer | but none of the repos I merge on are very high velocity | 21:25 |
| greghaynes | nibalizer: but theres also the case of it can merge but tests fail because master moved | 21:25 |
| *** radez is now known as radez_g0n3 | 21:25 | |
| clarkb | that doesnt address deps though which is what you mean by gate I think | 21:25 |
| greghaynes | and you just werent testing that rebase | 21:25 |
| waynr | nibalizer: github usually tells you before you try to merge whether or not a change can be merged | 21:25 |
| waynr | since the merge target is usually a known ref, whenever that ref is updated github will check to see if open PRs can be merged and provide an indication of whether or not it they can on the PR page | 21:26 |
| waynr | s/usually/always/ | 21:27 |
| fungi | nibalizer: are you saying there's a way to have github serve you the result which would be obtained from merging pr x to branch y as ref z without actually also merging x to y in the public repo and serving that to everyone as the new tip of y? | 21:27 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/shade: Split security group list operations https://review.openstack.org/179945 | 21:27 |
| clarkb | fungi ya but it wont do x depends on a because gate | 21:27 |
| waynr | fungi: I think we're saying that github does expose the result of merging x to y in the form of refs/pulls/<PR#> | 21:27 |
| fungi | well, it might if github prs can have dependencies specified in github, but as i avoid it for the most part i really don't know much about github's more esoteric workflow options | 21:28 |
| waynr | i don't think github has the concept of inter-project dependencies | 21:28 |
| greghaynes | I dont think theres a way | 21:28 |
| fungi | oh, i see, a way to request an arbitrary series of prs merged in order even when there's no dependency specified in github | 21:29 |
| *** Krinkle|detached is now known as Krinkle | 21:29 | |
| greghaynes | every time thats come up in a project ive just seen it done manually | 21:29 |
| clarkb | you cpuld do it with zuul | 21:29 |
| nibalizer | fungi: I use this: http://paste.ubuntu.com/10986825/ | 21:29 |
| clarkb | and thats why you need mergers so gate works | 21:30 |
| waynr | also I think I have been talking about a simplified single-PR case...in the case of ordering multiple PRs and checked that they are merge/test compatible, that is a different story | 21:30 |
| fungi | yeah, so anyway, those are the things the zuul-merger process does, and i don't think we can replace any of that when using teh github trigger | 21:30 |
| nibalizer | so ill cd into puppet-postgresql, then pr 284 and it will checkout the code of that pull request | 21:30 |
| waynr | so yeah i agree, zuul would need to handle detection of dependencies between PRs | 21:30 |
| waynr | githubs's notion of PR allows one PR to be a subset of another PR since there can be multiple patches in a given PR, i'm not sure how the commit tree changes once the subset PR is merged into master | 21:31 |
| fungi | waynr: well, and ordering multiple prs is basically necessary in the gate pipeline because you may approve several at once and want to know that they don't break each other | 21:31 |
| fungi | waynr: so your simplified solution would be more like the check pipeline and then a no-op gate | 21:32 |
| fungi | i.e. independently test the pull requests, hope they don't have negative interactions when you approve them | 21:32 |
| waynr | so it seems like the reason my intended build jobs aren't running then is that the zuul-merger hasn't done its thing...time to play with that now i guess | 21:34 |
| fungi | waynr: and i doubt that's sufficient, since zuul lets you have more than one trigger for a pipeline, so you might be comingling prs from github with changes in gerrit on a single project | 21:34 |
| fungi | and would need to have them both support a common set of jobs in the same pipelines together | 21:35 |
| *** jtriley has quit IRC | 21:35 | |
| *** mwagner_lap has quit IRC | 21:35 | |
| waynr | i think my goal for now is to implement the simplified "either or" case where a project either has "github" trigger or "gerrit" trigger | 21:36 |
| waynr | i'm sure it should be possible to specify that kind of mutual exclusion in the layout schema | 21:36 |
| waynr | (to provide users immediate feedback if they try to use both styles of trigger on given pipeline) | 21:36 |
| waynr | unless i am misunderstanding zuul configuration | 21:37 |
| waynr | which i probably am | 21:37 |
| waynr | dang it's not even 3 yet | 21:38 |
| clarkb | waynr: a single project will only ever come from one most likely, but it may be nice to mix projects on github and projects on gerrit together in the same pipeline for cogating. That said I don't think you need to support that off the bat if its more complicated | 21:38 |
| fungi | i think your options there are to document the exclusivity of those triggers in a pipeline, have zuul log an error if someone mistakenly adds both, and optionally refuse to load the config | 21:39 |
| *** dkranz has quit IRC | 21:39 | |
| waynr | yeah | 21:39 |
| fungi | enforcing that exclusivity in the voluptuous validation would probably be possible | 21:39 |
| waynr | that's what i'm hoping | 21:40 |
| fungi | which also makes it easier for someone to run a gate job validating their zuul layout to catch it | 21:40 |
| *** samueldmq has quit IRC | 21:40 | |
| *** samueldmq_ is now known as samueldmq | 21:40 | |
| *** peristeri has quit IRC | 21:40 | |
| *** pc_m has quit IRC | 21:41 | |
| waynr | i really prefer gerrit's notion of changes and patchsets over github pr's | 21:42 |
| clarkb | ++ | 21:43 |
| pleia2 | ++ | 21:44 |
| *** dizquierdo has quit IRC | 21:45 | |
| mtreinish | clarkb, fungi: do you have any thoughts on: https://review.openstack.org/176564 I'm debating doing something like that vs spinning up a separate liberty db | 21:49 |
| *** rfolco has quit IRC | 21:51 | |
| clarkb | mtreinish: is that going to take locks out on the db and prevent us from writing to it for long periods of time? | 21:51 |
| *** Somay has quit IRC | 21:52 | |
| mtreinish | clarkb: honestly dunno, maybe, the sqlalchemy calls are here: http://git.openstack.org/cgit/openstack-infra/subunit2sql/tree/subunit2sql/db/api.py#n644 | 21:53 |
| mtreinish | given our past experience everything gets to be kinda slow with the db at this size, I forget how long that last migration took | 21:54 |
| *** _dguerri has quit IRC | 21:54 | |
| fungi | hours. i too forget exactly | 21:54 |
| *** zz_jgrimm has quit IRC | 21:54 | |
| *** zz_zz_ja has quit IRC | 21:55 | |
| clarkb | mtreinish: http://docs.sqlalchemy.org/en/improve_toc/orm/query.html check out the warning for snychronize_session there | 21:56 |
| *** zz_jgrimm has joined #openstack-infra | 21:56 | |
| clarkb | mtreinish: I think its bascally gonna do a select for all the things that match the filter, then delete them | 21:57 |
| clarkb | so at the very least it won't be super efficient | 21:57 |
| *** samueldmq has quit IRC | 21:57 | |
| *** zz_ja has joined #openstack-infra | 21:57 | |
| clarkb | but I am noexpert here | 21:58 |
| clarkb | just worried that it will be expensive. Worth noting that we drop an entire index per day with ES because thats cheap | 21:58 |
| mtreinish | clarkb: yeah I remember reading that when I wrote the api call. I figured evaluate was the best tradeoff but I really didn't have enough background to weigh things well | 21:59 |
| openstackgerrit | Darragh Bailey proposed openstack-infra/git-review: Isolate tests from user/system git configuration https://review.openstack.org/179954 | 21:59 |
| mtreinish | clarkb: yeah, I'm no expert either. I was just trying to come up with a solution so we don't run out of space on the trove node and this was my first thought | 22:00 |
| *** dguerri has joined #openstack-infra | 22:00 | |
| *** dboik has quit IRC | 22:02 | |
| *** rbradfor has quit IRC | 22:03 | |
| *** rbradfor has joined #openstack-infra | 22:06 | |
| fungi | mtreinish: i would say before going forward with something like 176564 we should probably do a couple manual runs, spaced as far apart as we're considering setting the cron job, and then time the second one | 22:06 |
| *** harlowja_ has quit IRC | 22:06 | |
| *** harlowja has joined #openstack-infra | 22:07 | |
| fungi | and also try to pick a trough time in the historical activity graphs to minimize impact | 22:07 |
| *** zz_ja has quit IRC | 22:07 | |
| *** gordc has quit IRC | 22:07 | |
| mtreinish | fungi: sure, that sounds like a good idea | 22:08 |
| mtreinish | if I've learned anything around this db it's that the automation never works the first time anyway :) | 22:09 |
| *** zz_jgrimm has quit IRC | 22:09 | |
| *** mmedvede has joined #openstack-infra | 22:09 | |
| mtreinish | fungi: but you think the db pruning approach is the way to go over spinning up a second db or something else? | 22:10 |
| fungi | (and often not the second or third either) | 22:10 |
| mtreinish | fungi: heh, yep | 22:10 |
| fungi | i hadn't previously considered adding more databases, but i suppose that could be engineered | 22:10 |
| *** maurosr has quit IRC | 22:11 | |
| waynr | wooo with zuul-merger running jobs are correctly triggered | 22:11 |
| clarkb | waynr: woot | 22:11 |
| *** otter768 has joined #openstack-infra | 22:12 | |
| *** zz_jgrimm has joined #openstack-infra | 22:12 | |
| waynr | so how does zuul get the merged branch to the jenkins jobs? | 22:13 |
| fungi | mtreinish: unfortunately any automation around creating and removing the trove instances themselves, if it were only exercised every 6 months, would probably bitrot and break between every invocation | 22:13 |
| fungi | so per-cycle databases would likely be a manual process to add and remove | 22:13 |
| clarkb | waynr: the zuul merger has created a zuul ref for you in the repo with that change something like refs/zuul/Zsomesha1, you then have to serve this to your workers in a manner that makes sense for you | 22:14 |
| *** zz_ja has joined #openstack-infra | 22:14 | |
| *** maurosr has joined #openstack-infra | 22:14 | |
| fungi | mtreinish: how about renaming tables in the existing database? start writing new data to empty tables and then prune down the old tables which shouldn't block on other writes at that point? | 22:15 |
| waynr | okay that part might not be functional at the moment | 22:15 |
| clarkb | waynr: zuul also drops info on the location of where to get the refs in the zuul's job params, and its configurable as far as what base url (including protocol) to use | 22:15 |
| clarkb | waynr: we serve via http with apache sitting in front of our zuul mergers but ssh:// should work as well as git:// iirc | 22:15 |
| fungi | mtreinish: i have no idea if that's any more or less optimized than trying to prune the tables to which you're appending | 22:15 |
| mtreinish | fungi: heh, yeah that's a fair point I don't want to introduce another manual process if I can help it. I mean it was just an alternate idea I was floating around | 22:16 |
| *** otter768 has quit IRC | 22:16 | |
| mtreinish | fungi: oh, that's an interesting idea. Yeah I don't know if that would be more efficient or not | 22:16 |
| *** zz_jgrimm has quit IRC | 22:16 | |
| fungi | mtreinish: also splitting your data across multiple databases means you stop being able to use one sql query to access all your data and have to perform potentially heavy (depending on your use case) post-query correlation work | 22:17 |
| mtreinish | it would also make dumping the delete stuff into a subunit2sql command a bit more complicated since it wouldn't know about the renamed table | 22:17 |
| *** ayoung is now known as ayoung-dadmode | 22:17 | |
| mtreinish | unless I added a migration to create that table | 22:17 |
| fungi | at least in the trove server-per-database model | 22:17 |
| mtreinish | fungi: yeah, that's what I was thinking. I mean the only reason to do the split would be because of resource limitations on a server. (I can live with some slow queries because of the table size) | 22:19 |
| mtreinish | my assumption was that it would make correlation difficult, but each db would be an isolated data set for a release | 22:19 |
| *** maurosr has quit IRC | 22:19 | |
| *** zz_jgrimm has joined #openstack-infra | 22:19 | |
| mtreinish | so you probably wouldn't need to do direct mapping between dbs | 22:20 |
| jeblair | fungi: do you know if trove supports the federated engine? | 22:20 |
| *** amitgandhinz has quit IRC | 22:20 | |
| *** ildikov has quit IRC | 22:20 | |
| *** maurosr has joined #openstack-infra | 22:20 | |
| fungi | jeblair: absolutely no idea whatsoever | 22:20 |
| mtreinish | just thinking out loud about this, not actually sure about anything | 22:21 |
| jeblair | mtreinish, fungi: subunit2sql uses innodb, right? | 22:21 |
| fungi | jeblair: i didn't even realize that was a thing until you just mentioned it | 22:21 |
| mtreinish | jeblair: yeah | 22:21 |
| fungi | "storage engine that accesses data in tables of remote databases rather than in local tables" | 22:21 |
| fungi | neat idea | 22:21 |
| jeblair | if we used myisam, we could use the merge engine. that might be okay with this dataset. but if we want to stick with innodb, i think merge is not available, but federated would work | 22:22 |
| *** mwagner_lap has joined #openstack-infra | 22:22 | |
| mtreinish | jeblair: well iirc I have a migration test which makes sure all the created tables get set to innodb, which is only there because I copy and pasted most of the test from nova | 22:22 |
| mtreinish | I remember mordred telling me myisam might be a better fit for some of the tables | 22:22 |
| jeblair | and can be used to set up a sharding system where we put, say, one month worth of data in a database at a time, but can query the whole set transparently, and then drop those databases as we want | 22:22 |
| jeblair | mordred: ^ may have more input | 22:23 |
| waynr | clarkb: so when i cd to the directory where zuul-merger cloned my project, I don't see any reffs named 'refs/zuul/*' | 22:29 |
| mordred | aroo | 22:30 |
| mordred | reading | 22:30 |
| mordred | OK. yes, it's likely a great choice for subunit2sql, don't use federated | 22:31 |
| mordred | myisam and merge, likely great choice | 22:31 |
| *** spzala has joined #openstack-infra | 22:31 | |
| clarkb | waynr: nothing in .git/refs/zuul? | 22:32 |
| mordred | on phone, will write more in a bit | 22:32 |
| mtreinish | mordred: ok, I can push up a patch to add a migration to convert the tables from innodb to myisam | 22:32 |
| * mtreinish can't wait to see how long that'll take | 22:32 | |
| *** mmedvede has quit IRC | 22:32 | |
| mordred | rbradford: around? wanna chat with mtreinish here? | 22:32 |
| fungi | mtreinish: assuming there's even sufficient space on the instance? | 22:33 |
| mordred | mtreinish: we'll get fancy | 22:33 |
| *** aysyd has quit IRC | 22:33 | |
| mordred | these are all append only, right? | 22:33 |
| mordred | and you want to roll old rows off eventually? | 22:33 |
| *** ildikov has joined #openstack-infra | 22:33 | |
| mtreinish | mordred: a few tables are append only, one isn't | 22:33 |
| mordred | kk. let me look more deeply on laptop in 20 | 22:34 |
| waynr | clarkb: nada | 22:34 |
| clarkb | waynr: hrm, you should check your zuul-merger logs then | 22:34 |
| clarkb | waynr: it should log everything it did for that project when the change came in | 22:34 |
| mtreinish | mordred: yeah, that was the idea, I had a wip patch up now to do a cron delete on old rows on the append only tables | 22:35 |
| waynr | i'm watching them, running in non-daemonized mode | 22:35 |
| mtreinish | fungi: heh, well last time there was enough space, but trove didn't configure mysql to use all of it correctly | 22:35 |
| mtreinish | but yeah, we'd likely hit that again | 22:35 |
| clarkb | waynr: so you should have a log that says "created ref blahblahblah | 22:36 |
| mordred | kk | 22:36 |
| clarkb | waynr: do you see anything like that? | 22:36 |
| waynr | nope | 22:36 |
| clarkb | "2015-05-04 22:35:41,006 DEBUG zuul.Repo: CreateZuulRef master/Z755990147e834176a684021c571cd261 at 80c53f437fab2a5d0974b18dce720ef3a654d652" is an example from one of our mergers | 22:37 |
| *** baoli has quit IRC | 22:37 | |
| SlickNik | mtreinish: That sounds like a trove bug. Any more information on it? | 22:38 |
| waynr | yeah i see where that log message gets emitted in the code, there must be something about my setup that prevents it from trying? | 22:38 |
| SlickNik | (sorry jumping in, and don't have all of the context — reading scrollback now) | 22:39 |
| clarkb | waynr: ya it may be bailing before that for some reason? do you get lines for resetting repo, updating repo, checking out $REF? | 22:39 |
| clarkb | waynr: those are the steps it does before it does a merge, OH! it may be trying to talk to gerrit there? | 22:39 |
| clarkb | maybe? | 22:39 |
| waynr | clarkb: oh it looks like it is only trying to run the "merge:update" job, not "merge:merge" | 22:40 |
| waynr | so it never reaches the merge codepath | 22:40 |
| clarkb | waynr: ah ya, updates are there if you don't need a zuul ref and just need the latest code in place | 22:40 |
| waynr | what determines that? | 22:41 |
| clarkb | waynr: merge is called when you have a thing that needs merging. Basic difference between the two, is you do an update for things like "test code that has already merged" eg ref updates or tags being made and you do merger:merge when you have unmerged code that is being proposed and needing to be tested. Eg new patchset or worflow +1 | 22:41 |
| waynr | interesting | 22:42 |
| fungi | waynr: no errors logged by the zuul-merger about being unable to clone from review.openstack.org? (or whatever gerrit it's pointed at? or is this with some github mojo already?) | 22:42 |
| waynr | fungi: the zuul-merger is definitely cloneing correctly from github, it seems like that url got passed from the EventTrigger object I created during webhook handling | 22:43 |
| fungi | waynr: okay, so you can see where it has managed to clone the repos on disk at least | 22:43 |
| waynr | yeah | 22:43 |
| waynr | and it sorta makes sense that only merge:update is being run | 22:43 |
| *** bknudson has quit IRC | 22:44 | |
| mtreinish | SlickNik: oh, yeah IIRC it was the temp storage for mysql wasn't on the volume used for the general storage, so when I did an alter table to add a column it ran out of space | 22:44 |
| *** cody-somerville has quit IRC | 22:44 | |
| mtreinish | SlickNik: so we had to do this: https://review.openstack.org/#/c/164838/ | 22:45 |
| mtreinish | SlickNik: I can dig up the irc logs from when we tried doing this | 22:45 |
| mtreinish | SlickNik: the details are in http://eavesdrop.openstack.org/irclogs/%23openstack-infra/%23openstack-infra.2015-03-16.log somewhere | 22:46 |
| *** shashankhegde has joined #openstack-infra | 22:46 | |
| mtreinish | I think we pinged you at one point, so maybe grep for SlickNik :) | 22:46 |
| SlickNik | mtreinish: Thanks much — that helps! | 22:48 |
| jhesketh | Morning | 22:48 |
| * clarkb remembers the pbr thing | 22:50 | |
| lifeless | clarkb: pbr thing? | 22:50 |
| clarkb | dims: dhellmann mordred is there a good reason for having all of oslo-core be able to tag pbr releases? | 22:50 |
| lifeless | oh yeah | 22:51 |
| clarkb | lifeless: ya, remembering the release trouble we had and that the acl was funny because ^ | 22:51 |
| lifeless | I spotted that and WTF'd | 22:51 |
| *** mattfarina has joined #openstack-infra | 22:51 | |
| mordred | only because of Oslo association | 22:51 |
| lifeless | clarkb: I'm in pbr-release now, thanks to dims but we should sort the other thing out | 22:51 |
| lifeless | all of oslo can't release all of oslo | 22:51 |
| *** mattfarina has quit IRC | 22:51 | |
| lifeless | but all of oslo can release pbr | 22:51 |
| openstackgerrit | Matthew Treinish proposed openstack-infra/subunit2sql: Add migration to convert append only tables to MyISAM https://review.openstack.org/179964 | 22:52 |
| mtreinish | mordred: ^^^ I doubt that will work as is... | 22:52 |
| mtreinish | oh and it'll definitely fail the mysql migration tests (because they check for innodb :) ) | 22:53 |
| clarkb | right having oslo-release be able to release pbr is fine, my question is why oslo-core too | 22:53 |
| lifeless | -> the list, better for tz's since dhellmann is already off for the night | 22:54 |
| *** ZZelle_ has quit IRC | 22:55 | |
| clarkb | jeblair: btw, I don't see anything in gear that would make RESET_ABILITIES nuke a running job | 22:55 |
| lifeless | I'll send a mail | 22:55 |
| fungi | vague recollection is that originally oslo-core was a very small group handling a handful of libraries. fast forward a couple release cycles and some of the acls didn't get updated to match that early state | 22:55 |
| clarkb | jeblair: so I think that is placing the blame back on the geraman plugin implementation. I am likely going to take a stab at rewriting the work() loop so that we can have state transitions all in one place | 22:56 |
| fungi | er, s/early/later/ | 22:56 |
| clarkb | jeblair: at least then this should be easier to reason about | 22:56 |
| *** Krinkle is now known as Krinkle|detached | 22:56 | |
| *** armax has quit IRC | 22:57 | |
| clarkb | unless | 22:57 |
| clarkb | zaro: is there any chance that the gearman java lib is maintaining some state too that we don't have control over? | 22:57 |
| clarkb | zaro: or is it purely just for input and output and dispatching the handleevent method? | 22:57 |
| waynr | \ | 22:58 |
| *** _nadya_ has joined #openstack-infra | 22:58 | |
| clarkb | mordred: have you been able to look at my comments on https://review.openstack.org/#/c/175891/24 ? my only real concern there with fixing it later is you intended to make a release | 22:59 |
| *** prad has quit IRC | 23:00 | |
| *** Sukhdev has quit IRC | 23:00 | |
| *** _nadya_ has quit IRC | 23:02 | |
| *** zz_jgrimm has quit IRC | 23:03 | |
| *** zz_ja has quit IRC | 23:03 | |
| *** sdake has joined #openstack-infra | 23:04 | |
| *** zz_ja has joined #openstack-infra | 23:06 | |
| *** zz_jgrimm has joined #openstack-infra | 23:06 | |
| *** camunoz has quit IRC | 23:07 | |
| *** rlandy has quit IRC | 23:07 | |
| *** sdake_ has quit IRC | 23:08 | |
| mordred | I have, I agree with them and will fix when I take off | 23:08 |
| openstackgerrit | Merged openstack-infra/shade: Switch tasks vs put on a boolean config flag https://review.openstack.org/178886 | 23:09 |
| openstackgerrit | Ramy Asselin proposed openstack-infra/puppet-openstackci: Create a log server like http://logs.openstack.org https://review.openstack.org/167425 | 23:10 |
| asselin__ | nibalizer, you around? I'm still confused | 23:11 |
| asselin__ | ::jenkins class doesn't have an ssh_key...? | 23:11 |
| asselin__ | did you mean ::jenkinsuser? | 23:12 |
| *** shashankhegde has quit IRC | 23:12 | |
| *** maishsk has quit IRC | 23:13 | |
| *** ajo has quit IRC | 23:13 | |
| *** cody-somerville has joined #openstack-infra | 23:14 | |
| asselin__ | anyone see this error on icehouse? 2015-05-04 23:12:39.096 | 'keystonemiddleware' is not a global requirement but it should be,something went wrong | 23:14 |
| fungi | asselin__: i do know that there was no keystonemiddleware in icehouse | 23:15 |
| fungi | asselin__: it was split from keystoneclient during the juno development cycle | 23:15 |
| *** maishsk has joined #openstack-infra | 23:15 | |
| asselin__ | fungi, ok....nevermind...my error.....thanks | 23:16 |
| *** whoops has quit IRC | 23:16 | |
| openstackgerrit | Matthew Treinish proposed openstack-infra/subunit2sql: Add migration to convert append only tables to MyISAM https://review.openstack.org/179964 | 23:17 |
| lifeless | Daviey: oh hai :) | 23:19 |
| nibalizer | asselin__: did it fail? | 23:21 |
| nibalizer | let me look | 23:21 |
| nibalizer | sorry I'm donig like 100 things today | 23:21 |
| asselin__ | nibalizer, ok....np...me2. it's running now, just trying to make sence of it | 23:22 |
| *** shashankhegde has joined #openstack-infra | 23:22 | |
| nibalizer | so using what I meant was to use if ! defined(Class['::jenkins::jenkinsuser']) { instead of if ! defined(Class['jenkins::jenkinsuser']) { | 23:23 |
| nibalizer | thats all | 23:23 |
| openstackgerrit | Joshua Harlow proposed openstack/requirements: Bump futures up to 3.0 or greater https://review.openstack.org/179967 | 23:24 |
| asselin__ | nibalizer, ok! that makes more sense | 23:24 |
| nibalizer | ya should be a super easy/nonbreaking change | 23:24 |
| openstackgerrit | Davide Guerri proposed openstack-infra/shade: Add port resource methods https://review.openstack.org/179968 | 23:26 |
| openstackgerrit | Joshua Harlow proposed openstack/requirements: Bump futures up to 3.0 or greater https://review.openstack.org/179967 | 23:27 |
| *** zz_jgrimm has quit IRC | 23:27 | |
| openstackgerrit | Joshua Harlow proposed openstack/requirements: Bump futures up to 3.0 or greater https://review.openstack.org/179967 | 23:27 |
| *** samueldmq has joined #openstack-infra | 23:27 | |
| *** maurosr has quit IRC | 23:28 | |
| *** dangers is now known as dangers_away | 23:28 | |
| *** zz_ja has quit IRC | 23:28 | |
| *** maurosr has joined #openstack-infra | 23:29 | |
| *** zz_jgrimm has joined #openstack-infra | 23:29 | |
| *** zz_ja has joined #openstack-infra | 23:29 | |
| *** chlong has joined #openstack-infra | 23:30 | |
| openstackgerrit | Ramy Asselin proposed openstack-infra/puppet-openstackci: Create a log server like http://logs.openstack.org https://review.openstack.org/167425 | 23:30 |
| asselin__ | nibalizer, ^^ compare patch 9 & 12 | 23:31 |
| nibalizer | asselin__: kk | 23:32 |
| nibalizer | asselin__: precisely! | 23:32 |
| *** dannywilson has quit IRC | 23:32 | |
| asselin__ | nibalizer, I also understand you comment now. so LGTM | 23:33 |
| clarkb | zaro: jeblair ok, I think I may have finally figured this out! gearman plugin calls registerFunctions() which clears the function map in the instance, handleEvent's addNewJob() has a check for the function being in the functions map, if it doesn't come back with a valid function it sends a work fail then never sends a new send grab job from what I can tell | 23:34 |
| clarkb | zaro: jeblair I think this is the problem, its also completely unlogged. I will try taking a stab at fixing this | 23:34 |
| clarkb | zaro: jeblair: I think we can sanely fix this lock free if we put the events on a thread safe queue then have work() pull them off one by one and process them | 23:35 |
| clarkb | but that way we won't race the functionMap updates and won't need to handle them separately | 23:35 |
| *** stevemar has quit IRC | 23:37 | |
| *** hemna is now known as hemnafk | 23:37 | |
| *** claudiub has quit IRC | 23:37 | |
| *** viglesias has quit IRC | 23:37 | |
| *** bswartz has joined #openstack-infra | 23:38 | |
| *** viglesias has joined #openstack-infra | 23:38 | |
| *** gokrokve has quit IRC | 23:39 | |
| clarkb | or I suppose we can just clear out the grab job sent flag and have it try again later | 23:39 |
| clarkb | but I would prefer to not send the work fail at all | 23:39 |
| *** ajmiller has quit IRC | 23:41 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Add inventory command to shade https://review.openstack.org/175891 | 23:41 |
| mordred | clarkb: ok- I think I got your things there ^^ | 23:42 |
| *** dims_ has joined #openstack-infra | 23:42 | |
| mordred | and redid one thing - put a parameter around the exclusion in the get_hosts call ... the logic there is carry over from ansible, but non-ansible may want different things | 23:42 |
| *** dims has quit IRC | 23:44 | |
| clarkb | mordred: I like that better, it more clearly indicates if you have the host or not | 23:49 |
| *** cody-somerville has quit IRC | 23:49 | |
| clarkb | mordred: rather than just giving an ambiguous {} which may mean unconnected or does not exist | 23:49 |
| mordred | \o/ | 23:49 |
| *** jtriley has joined #openstack-infra | 23:49 | |
| mordred | clarkb: look - code review leading to better code | 23:49 |
| *** signed8bit is now known as signed8bit_ZZZzz | 23:50 | |
| jeblair | clarkb: ah good -- that explains the lack of stuck zuul jobs | 23:50 |
| *** freerunner has joined #openstack-infra | 23:51 | |
| *** ashleighfarnham has quit IRC | 23:51 | |
| *** freerunner has quit IRC | 23:52 | |
| *** freerunner has joined #openstack-infra | 23:52 | |
| clarkb | jeblair: zaro: I think my plan is to collapse the work() method's loop a bit (I am pretty sure that we can do a single driveSessionIO which should make that easier to reason about, then move handleSessionEvent into the worker thread so that functionMap updates are serialized, and finally use the "add noop event to kick off get job requests" everywhere since we already do that in one place and this | 23:54 |
| clarkb | will make that state change consistent | 23:54 |
| lifeless | new testtools release just cut | 23:56 |
| lifeless | its the first using pbr | 23:56 |
| lifeless | so - it might be a disaster. | 23:56 |
| *** SumitNaiksatam has quit IRC | 23:56 | |
| *** sdake_ has joined #openstack-infra | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!