| *** saneax is now known as saneax-_-|AFK | 01:07 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: WIP: Organize playbooks folder https://review.openstack.org/441547 | 03:55 |
|---|---|---|
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Organize playbooks folder https://review.openstack.org/441547 | 04:00 |
| SpamapS | jeblair: that's great news.. our lhf pickers are just getting started. :) | 04:25 |
| *** Cibo has joined #zuul | 11:33 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Provide file locations of config syntax errors https://review.openstack.org/441606 | 15:42 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Create tox-tarball job https://review.openstack.org/441609 | 16:26 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Create tox-tarball job https://review.openstack.org/441609 | 16:30 |
| pabelanger | jlk: how often do you use tags in ansible? Follow up question, do you like them? | 16:34 |
| SpamapS | pabelanger: I'll answer for myself: sometimes, and not really. ;) | 16:40 |
| SpamapS | jlk: they might be well suited to testing matrices though. | 16:41 |
| SpamapS | I usually see them as a band-aid to prevent refactoring roles. | 16:41 |
| pabelanger | agree, I am not a fan of them myself. However, I could see somebody wanting to use them in zuulv3 | 16:42 |
| SpamapS | err | 16:42 |
| SpamapS | yeah | 16:42 |
| SpamapS | pabelanger: I'd say the stdlib of playbooks and roles should never use them. | 16:43 |
| pabelanger | ya | 16:43 |
| SpamapS | Would rather see when: xxx | 16:44 |
| pabelanger | As an example, if we create a job today, and want to have a pre-run, run, and post-run. We need 3 playbooks to achieve that. However, an optimization could be, 1 single playbook, that used tags for pre-run, run and post-run things. In fact, they could actually map to pre_tasks, tasks, post_tasks in your playbook. | 16:45 |
| pabelanger | Not a big deal now, but something I've been thinking about the more playbooks I start to add | 16:46 |
| SpamapS | wha? | 16:46 |
| SpamapS | OH | 16:46 |
| SpamapS | I see | 16:46 |
| SpamapS | so same playbook | 16:46 |
| SpamapS | but run in different contexts | 16:47 |
| pabelanger | right | 16:47 |
| SpamapS | yeah no, that's what roles are for IMO | 16:47 |
| pabelanger | I do have roles today | 16:47 |
| pabelanger | let me get an example | 16:47 |
| SpamapS | when: zuul.context=pre would be better IMO | 16:47 |
| SpamapS | But a disciplined approach to some zuul tags might be the same. | 16:49 |
| SpamapS | They only really get ugly when you start mixing -t with --skip-tags | 16:49 |
| pabelanger | Ya, I am not a fan of the skip-tag approach myself. But, I don't want to count them out either, if people value them. | 16:50 |
| SpamapS | well AFAIK, we're not going to allow arbitrary invocation of ansible-playbook | 16:51 |
| pabelanger | but ya, nothing major now, just brain dumping on some optimizations in the future (maybe) at the cost of more ansible magic in playbooks | 16:51 |
| SpamapS | so if we control the -t argument, I could see it working out | 16:51 |
| SpamapS | Another problem with tags is there's no "default tag", so once you adopt them, they have to go everywhere. | 16:52 |
| SpamapS | (the default tag is no tag ;) | 16:52 |
| pabelanger | agree | 16:52 |
| mordred | so - yah - I think for me I'd rather have multiple playbooks | 16:53 |
| mordred | than tags/sections | 16:53 |
| SpamapS | same | 16:53 |
| mordred | because that allows for more mix/match | 16:53 |
| SpamapS | and anything shared between them is a role | 16:53 |
| mordred | (as a for instance, I'm looking forward to running the devstack playbook as a _pre_ playbook for shade | 16:53 |
| mordred | since devstack failing is a setup fail, not a test fail | 16:54 |
| mordred | in the case of shade's tests | 16:54 |
| mordred | but I do just want the standard results of running devstack-gate :) | 16:54 |
| SpamapS | mordred: so.. Rust people like Zuul | 16:54 |
| mordred | SpamapS: \o/ | 16:54 |
| SpamapS | mordred: there's a slight chance redox OS may start BonnyCI'ing | 16:54 |
| mordred | SpamapS: and Zuul people like Rust | 16:54 |
| mordred | SpamapS: AWESOME | 16:54 |
| SpamapS | yeah it's a symbiotic echochamber | 16:54 |
| mordred | SpamapS: definitely let them know we all think their stuff is the absolute best | 16:55 |
| SpamapS | I also had a nice chat with some Travis users who said the main reason they don't gate on it is it explodes for !(their code) reasons a lot. | 16:55 |
| SpamapS | mordred: also Bonny got our first beta user yesterday.. Emily Dunham (the person who spoke after you about the volcano) has hooked up the rust book repo "too-many-lists" to BonnyCI. :) | 16:56 |
| mordred | SpamapS: AWESOME | 16:56 |
| SpamapS | yeah | 16:56 |
| SpamapS | it exploded | 16:56 |
| SpamapS | a lot | 16:56 |
| SpamapS | and docs | 16:56 |
| SpamapS | zomg | 16:56 |
| SpamapS | but beta \o/ | 16:56 |
| mordred | yay! and yeah, she's a sympathetic human who will work with you on it | 16:56 |
| SpamapS | and I love how everyone who sees it goes through the "it should [insert half of zuul v3's features]" | 16:57 |
| mordred | SpamapS: also yah re: travis. I see the same complaint from the ansible folks about shippable | 16:57 |
| SpamapS | and I go "yes! ... soon | 16:57 |
| mordred | SpamapS: +100 | 16:57 |
| SpamapS | mordred: the lesson there is that in addition to using zuul to make the model awesome.. there's an opportunity to use diverse clouds and excellent ops to make the quality higher. | 16:57 |
| * SpamapS afk's suddenly | 16:58 | |
| SpamapS | enjoy your saturday | 16:58 |
| mordred | SpamapS: yup. in fact, there are some key infra features about test stability that are not related to zuul | 16:58 |
| mordred | like the mirror subsystem | 16:58 |
| mordred | SpamapS: you too! | 16:58 |
| mordred | pabelanger: so - back to your tag thing ... | 16:58 |
| mordred | pabelanger: for my money, I'd rather have separate things - but I think it may be important to expose the ability for a user to express a tag in a job definition | 16:59 |
| mordred | pabelanger: same as how we allow them to express variables to pass in | 16:59 |
| mordred | pabelanger: because just because _I_ don't want to use tags to organize playbooks and whatnot - doesn't mean someone else might want to, yeah? | 17:00 |
| pabelanger | mordred: agree! Was just getting the pulse of people and tags | 17:00 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Create tox-tarball job https://review.openstack.org/441609 | 17:00 |
| pabelanger | cause, I can see somebody wanting a single playbook, multiples | 17:01 |
| SpamapS | Yeah, good mirrors, tight image control, cloud standards, cloud diversity, open source control of most components... | 17:01 |
| mordred | SpamapS: ++ | 17:01 |
| SpamapS | "never be a victim of the internet" | 17:01 |
| mordred | pabelanger: yah - or even somebody who already has a playbook that they just want to have zuul run that already has tags | 17:02 |
| pabelanger | indeed | 17:02 |
| pabelanger | mordred: when you have some down time: topic:zuulv3-ansible looking for feeback or +2s :) | 17:03 |
| mordred | pabelanger: awesome | 17:05 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Create tox-tarball job https://review.openstack.org/441609 | 17:09 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Create tox-tarball job https://review.openstack.org/441609 | 17:17 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Add net-info to bootstrap role https://review.openstack.org/441617 | 17:51 |
| *** saneax-_-|AFK is now known as saneax | 18:11 | |
| *** Cibo has quit IRC | 19:00 | |
| *** saneax is now known as saneax-_-|AFK | 19:09 | |
| Shrews | We support tags in the ansible tests in shade, but that's only for selecting individual tests. That's the only useful thing I've ever used them for. | 20:32 |
| Shrews | mordred: btw, omg duke v. unc #2 | 20:32 |
| Shrews | i hear it might be entertaining | 20:33 |
| *** EmilienM_ has joined #zuul | 21:13 | |
| *** rbergeron has quit IRC | 21:13 | |
| *** fungi has quit IRC | 21:14 | |
| *** persia has quit IRC | 21:14 | |
| *** EmilienM has quit IRC | 21:14 | |
| *** EmilienM_ is now known as EmilienM | 21:15 | |
| *** EmilienM has quit IRC | 21:15 | |
| *** EmilienM has joined #zuul | 21:15 | |
| *** fungi has joined #zuul | 21:18 | |
| *** persia has joined #zuul | 21:18 | |
| mordred | Shrews: yah man | 21:57 |
| mordred | gonna be good | 21:57 |
| *** rbergeron has joined #zuul | 22:02 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!