| *** os-chef-bot has quit IRC | 02:04 | |
| *** os4uinfo has joined #openstack-chef | 02:36 | |
| *** os4uinfo has quit IRC | 03:00 | |
| *** chlong_ has quit IRC | 03:08 | |
| *** chlong_ has joined #openstack-chef | 03:09 | |
| *** chlong_ has quit IRC | 03:17 | |
| *** chlong_ has joined #openstack-chef | 03:19 | |
| *** ch-mhass has joined #openstack-chef | 05:01 | |
| *** ch-mhass has quit IRC | 06:01 | |
| *** ch-mhass has joined #openstack-chef | 06:01 | |
| *** ch-mhass has quit IRC | 06:10 | |
| *** ch-mhass has joined #openstack-chef | 06:27 | |
| *** ch-mhass has quit IRC | 06:33 | |
| *** mrodden has quit IRC | 06:37 | |
| *** bodgix has quit IRC | 06:38 | |
| *** mrodden has joined #openstack-chef | 06:39 | |
| *** ch-mhass has joined #openstack-chef | 06:41 | |
| *** bodgix has joined #openstack-chef | 06:43 | |
| *** chlong_ has quit IRC | 12:43 | |
| *** Dinesh_Bhor has quit IRC | 13:32 | |
| *** chlong_ has joined #openstack-chef | 14:09 | |
| sc` | jklare: i think i fixed chef-provisioning-vagrant. i have to test further, but i may dispense with the whole vagrantfile and displacing chef-provisioning in favor of a pr upstream to get it fixed in chefdk | 14:26 |
|---|---|---|
| sc` | it's an addition of four characters if i'm right | 14:26 |
| *** ch-mhass has quit IRC | 14:34 | |
| bodgix | I've been reading on zuul and trying to wrap my head around testing in OpenStack. It seems to me that it wouldn't be possible to spin-up the whole OpenStack cluster on the machine where the zuul job is running even using very thin encapsulation like Docker. | 15:01 |
| bodgix | I know very little so far. But it looks to me that for a full integration tests, the zuul job would need to have access to a cloud and get some resources, VM(s) from there. | 15:02 |
| *** ch-mhass has joined #openstack-chef | 15:02 | |
| bodgix | doesn't have to be a dedicated undercloud though. I think it could be shared. just a project with some resources is enough | 15:03 |
| sc` | bodgix: this is what is offered - https://docs.openstack.org/infra/manual/testing.html | 15:21 |
| sc` | right now, we bootstrap based off of a shell script in openstack-common, with some bits in the rakefile from openstack-chef-repo | 15:22 |
| sc` | https://github.com/openstack-infra/project-config is where the ci related plumbing goes for actually defining what runs | 15:23 |
| sc` | https://github.com/openstack-infra/project-config/blob/master/zuul/layout.yaml#L736-L744 and https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/chef-jobs.yaml are a couple of relevant locations | 15:25 |
| bodgix | I was trying to find openstack-chef's stuff there. I found chef-rake-jobs in layout.yaml but couldn't find what they actually do in that file | 15:26 |
| sc` | that's the jjb template | 15:26 |
| bodgix | ah. ok I never got as far as jenkins template. Wasn't sure where Jenkins fits in the puzzle | 15:27 |
| sc` | one pain point is that we have little to no visibility aside from pulling logs, so it's been hunt-n-peck to find the right logs to archive | 15:27 |
| bodgix | yeah… I guess one thing is to stand up a cluster with chef. another problem is to run some tests. Wouldn't tempest be a good fit for an integration tests framework? | 15:28 |
| sc` | yeah, if chef converges :D | 15:28 |
| sc` | that's how it's supposed to go | 15:28 |
| sc` | converging in openstack ci is error prone partly due to things outside our control. this has been alleviated somewhat by the external ci, but that's wholly managed by jklare's team | 15:30 |
| bodgix | is there a document somewhere? | 15:31 |
| bodgix | documenting the current setup | 15:31 |
| sc` | i'd say most of the issues are outside our control. the biggest is rubygems.org being terrible at being reliable | 15:31 |
| sc` | unless jklare has some, i don't know of any | 15:31 |
| bodgix | :( didn't realise that about rubygems.org | 15:32 |
| sc` | the workflow we follow is the standard openstack development workflow. it's nothing different than any other project https://docs.openstack.org/infra/manual/developers.html | 15:32 |
| sc` | the current local development workflow is mirrored to https://github.com/openstack/openstack-chef-repo | 15:34 |
| sc` | it works, but heed caution running it on a laptop, even our developer spec macbooks | 15:34 |
| sc` | last time i ran it on mine, the poor little guy wanted to fly away | 15:36 |
| sc` | the main caveat is that in order to use the provisioning cookbook, you need no later than chefdk 1.1 | 15:37 |
| sc` | 1.3+ or 2.0 won't work | 15:38 |
| bodgix | ok. so I'm looking at bootstrap.sh and the rakefile. I don't see where the VM is actually converged. Looks like only unit tests, lint and style tests are ran | 15:42 |
| sc` | the rakefile is where that magic happens on the vm | 15:43 |
| sc` | https://github.com/openstack/openstack-chef-repo/blob/master/Rakefile#L194-L233 | 15:43 |
| bodgix | Ah… I was looking at the Rakefile in openstack-common | 15:44 |
| bodgix | Thanks | 15:44 |
| *** chlong_ has quit IRC | 18:40 | |
| *** _rai_ has quit IRC | 21:31 | |
| *** openstackstatus has quit IRC | 21:56 | |
| *** openstack has joined #openstack-chef | 21:59 | |
| *** ChanServ sets mode: +o openstack | 21:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!