| *** TuanLA has joined #openstack-chef | 01:04 | |
| sc` | ono. local testing is failing | 02:29 |
|---|---|---|
| sc` | ooh. packaging | 02:31 |
| sc` | keystone is failing because credential_setup no longer exists | 02:31 |
| sc` | and now it's cascading fail, for whatever reason | 03:05 |
| sc` | frickler: that addendum you made broke the monolith build on my test machine. it needs to be xenial-updates to get the right packages, because reasons | 03:20 |
| sc` | i'm rolling 498976 into 495562 since i've proposed conflicting changes to do different things to the same recipe | 03:38 |
| sc` | since it's working with -updates, i'm also reverting that change to the repo. we can talk about it more when we are both online at the same time, but the code is kind of leading the direction here | 03:56 |
| openstackgerrit | Samuel Cassiba proposed openstack/cookbook-openstack-identity master: Initial identity Pike updates https://review.openstack.org/495565 | 04:05 |
| openstackgerrit | Samuel Cassiba proposed openstack/cookbook-openstack-common master: Initial common Pike updates https://review.openstack.org/495562 | 05:10 |
| frickler | sc`: sorry about that, I'd been wanting to write a comment about it once I see the test results | 07:01 |
| frickler | the issue with the nova packages was that they are too old in the -updates, still on beta something, and the requestlog module came in later | 07:02 |
| frickler | so I looked at http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/pike_versions.html and they had never versions for the non-updates branch | 07:03 |
| frickler | today they even seem to have the official release pkgs ready there | 07:04 |
| sc` | intriguing | 08:08 |
| sc` | well, then, i'll start respinning on that and trying to get back to convergence | 08:08 |
| *** openstackgerrit has quit IRC | 08:17 | |
| *** [d__d] has quit IRC | 08:18 | |
| *** [d__d] has joined #openstack-chef | 08:21 | |
| *** TuanLA has quit IRC | 11:02 | |
| *** TuanLA has joined #openstack-chef | 12:07 | |
| *** TuanLA has quit IRC | 12:47 | |
| sc` | frickler: so i was doing some nosing around this morning, and kolla bakes in a few repos: https://github.com/openstack/kolla/blob/master/docker/base/sources.list.ubuntu | 13:37 |
| sc` | they do xenial as well as xenial-updates | 13:37 |
| sc` | intriguing. E_CAFFEINE. they only pull in xenial-updates for the pike repo itself | 13:39 |
| sc` | even ansible only has xenial-updates that i can see | 13:40 |
| frickler | sc`: I'll ask over in #ubuntu-server for what is correct, feel free to join | 13:41 |
| sc` | will do | 13:41 |
| frickler | o.k., so using xenial-updates is correct most of the time, except right now when pike UCA hasn't been released yet | 13:48 |
| frickler | but then I'd stick to that in the cookbook and only change it locally for testing, even if that means that CI may not give positive results for now | 13:49 |
| sc` | i made it so rdo can be overridden in the environment file. perhaps similar can happen here for proposed vs release? | 13:49 |
| sc` | like, add updates by default, but put proposed behind is_development? | 13:50 |
| sc` | it's how i'm doing the dlrn deps | 13:50 |
| frickler | sc`: sounds reasonable, yes | 13:51 |
| sc` | i'm spinning a build with that theory | 13:51 |
| sc` | i just don't want to keep adding repos for only a short time like we did in the past :) | 13:52 |
| sc` | they come back again and again, so may as well make it so they can coexist | 13:52 |
| sc` | i can also remove the apt and yum cookbooks | 13:54 |
| sc` | their resources got folded back into chef-client after 12.9 | 13:55 |
| sc` | one change at a time though :) | 13:55 |
| sc` | the way it is, it's either on or off. if is_development is true, splat the repos out. if not, don't. hope that's sufficient | 14:01 |
| sc` | i don't see a way to make it more granular without additional node attributes | 14:02 |
| sc` | i could rename is_development to is_release, and toggle it to on per stable branches at the environment file, but leave default to false otherwise. just spitballing | 14:04 |
| sc` | i find if i don't mention it, i'll continue off in a quasi-vacuum | 14:05 |
| sc` | giggity. built in 19:25 on the ryzen box | 14:12 |
| sc` | the two repos seem to coexist together, and openstack builds. just trying to get that short-circuit testing with chefspec | 14:23 |
| sc` | i flipped it around from is_development to is_release and it seems to work better defaulting to false and using overrides to true | 14:32 |
| sc` | patches shall be flying today | 14:35 |
| sc` | positive *and* negative tests for the repos | 14:59 |
| sc` | rspec agrees and is happy | 14:59 |
| sc` | Touch Coverage: 100.0% | 15:01 |
| sc` | common is very happy | 15:01 |
| *** openstackgerrit has joined #openstack-chef | 15:03 | |
| openstackgerrit | Samuel Cassiba proposed openstack/cookbook-openstack-common master: Initial common Pike updates https://review.openstack.org/495562 | 15:03 |
| sc` | that means, by default proposed and rdo deps get dropped. in the environment file for pike in openstack-chef-repo, that is_release can be overridden to true to not emit the repos | 15:05 |
| openstackgerrit | Samuel Cassiba proposed openstack/openstack-chef-repo master: Initial repo Pike updates https://review.openstack.org/495567 | 15:32 |
| *** eroux has joined #openstack-chef | 16:17 | |
| sc` | looks like we got a green integration run: http://logs.openstack.org/67/495567/2/check/gate-openstack-chef-repo-chef-rake-integration-ubuntu-xenial-nv/978a71f/console.html | 19:41 |
| *** eroux has quit IRC | 21:37 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!