18:00:01 #startmeeting Infrastructure (2017-01-26) 18:00:01 Meeting started Thu Jan 26 18:00:01 2017 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:01 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:01 The meeting name has been set to 'infrastructure_(2017-01-26)' 18:00:01 #meetingname infrastructure 18:00:01 #topic aloha 18:00:01 #chair smooge relrod nirik abadger1999 lmacken dgilmore threebean pingou puiterwijk pbrobinson 18:00:01 The meeting name has been set to 'infrastructure' 18:00:01 Current chairs: abadger1999 dgilmore lmacken nirik pbrobinson pingou puiterwijk relrod smooge threebean 18:00:01 #topic New folks introductions 18:00:07 .hello bt0dotninja 18:00:08 morning everyone. 18:00:08 bt0: bt0dotninja 'Alberto Rodriguez Sanchez' 18:00:24 Hello everyone 18:00:30 o/ 18:00:43 We may have a lightly attended meeting today... many people are out at FOSDEM and Devconf 18:00:55 hey bt0. 18:01:07 My name is Alberto Rodriguez from Mexico, i want help at first with some SOPs, maybe taking one per week 18:01:30 I'm currently working in the https://pagure.io/fedora-infrastructure/issue/5159 and getting some info from the ansible playbooks :) 18:01:37 bt0: great! Welcome and do feel free to jump in with questions when you hit them... 18:01:47 many thanks :D 18:01:47 * doteast here 18:01:55 here 18:02:20 .hello jcline 18:02:21 jcline: jcline 'Jeremy Cline' 18:03:02 * nirik will wait a few more minutes for folks to arrive 18:04:14 ok, then lets move along... 18:04:19 #topic announcements and information 18:04:19 #info mass update/reboot cycle completed - kevin/smooge/patrick 18:04:19 #info Got hit by a pingback DDOS attack, weathered it ok - kevin 18:04:19 #info python-cffi update issues on: pagure01, pkgs02, batcave01 - kevin 18:04:20 #info FOSDEM and Devconf. People away 18:04:29 Hey, if someone give a talk in fosdem and not in this list 18:04:33 https://pagure.io/fedora-commops/issue/98 18:04:34 any other status/info items folks want to mention? or anything there anyone wants to ask about? 18:05:13 not from me 18:05:23 bt0: cool. Lots of talks (as usual) 18:05:25 and want to be listed please to comment the ticket :) () 18:06:18 #info if you are giving a FOSDEM talk, please make sure you are on the list in https://pagure.io/fedora-commops/issue/98 18:06:50 I was going to ask about fas3 status, but I don't think any of the people working on it are around today. 18:07:30 #topic fedorahosted migration status 18:07:40 I got fudcon-planning moved last week. 18:07:51 #info fudcon-planning moved. 18:07:53 I don't have much of anything other than the infra mailing list thread I started about app documentation. I figured email would be a better way to hold the initial discussion. 18:08:31 jcline: indeed. I think it's a great idea... we often just don't get around to documenting things well... 18:08:59 #info nirik will update and try and help migrate some more from the list at: https://fedoraproject.org/wiki/Infrastructure/Fedorahostedmigrations 18:09:34 I was wondering if we could do something similar for sysadmin work 18:10:00 doteast: sure. we have talked about compiling up the SOPs and putting them in a pagure docs... 18:10:40 yes, that and something that goes inline with getting started 18:10:43 perhaps it would make sense to fold it all into one fedora infrastructure thing 18:10:55 be all dev opsy 18:11:09 so, one of the items on the getting started mentions that you might need to create a setup locally 18:11:10 .hello edvm 18:11:11 x-ip: edvm 'Emiliano Dalla Verde Marcozzi' <6564766d@gmail.com> 18:11:23 #topic Updated documentation 18:12:23 so, I was thinking of creating some docs on how to create some minimal setups to work on 18:12:36 for ansible testing or ? 18:13:04 .hello sayanchowdhury 18:13:05 sayan: sayanchowdhury 'Sayan Chowdhury' 18:13:28 any.... for example, say one want to work on mirrormanager, how can one create a local setup to start with 18:13:44 may be even an ansible playbook to create the setup 18:13:54 yeah, that gets back somewhat to the apps docs jcline was talking about... 18:13:55 this might be for any of the fi apps 18:14:08 some of our apps have vagrant or docker setups 18:14:12 some don't. 18:14:19 something like ansible + vagrand??? 18:14:40 Yeah, so I wanted to (as part of that docs project) document the standard Vagrant/docker/ansible setup we use/should add to all projects 18:15:01 yeah, it would be nice to have a consistent way to locally setup an app for testing. 18:15:02 nirik, no for dev, but also to follow issues with the app 18:15:28 should that be defined on each app README file? For example, 'To start contributing on this app, you can use ...' instead of having a central system where we write app docs. 18:15:53 another example to setup torrent 18:15:57 x-ip: well, thats where it is now, but it's all over the place... different apps are setup completely differently 18:16:04 I've got a pretty standard Vagrantfile + Ansible setup I have stashed in a templates folder, I was planning on dropping these into the docs project 18:16:43 As long as we all agree on that stuff, we can add it to missing apps, fix up ones that are doing something unique and different, etc. 18:17:11 doteast: yeah, that brings up also what one of 'our apps' is. I think we have different levels of things... we have apps "we" maintain and build, but also some apps we just run that some other upstreams maintain. 18:17:26 like mediawiki, etc 18:18:02 jcline: I have no horse in that race... if there's something that will work well for all of them, uses free software, then it's ok for me. ;) 18:18:09 nirik, well the more we have of these guides, the more on-boarding we get I guess 18:19:07 another thing that comes out of this is it might encourage ppl to talk about issues with existing apps/systems 18:19:45 doteast: sure. Also, we talked long ago about making some kind of 'public' private repo. Since our ansible repo is open this public private repo could be a template of the variables you need to set to use our ansible setup somwhere else. 18:20:05 I mean, its not necessary to have one setup that fits all, but we think about them as we move along 18:20:06 or just docs on how to set things. 18:20:28 Do you think Docker could be a good tool to help in this task? Creating docker images? 18:21:33 we may not want to get too big a plans here... start with docs and work out rather than try and do everything at once. 18:21:53 sounds like a plan 18:22:00 For example, the Python Argentinian web source code is on github, and we encourage users who want to contribute, to use Docker and docker-compose to get quickly an instance running locally with latest code: https://github.com/PyAr/pyarweb 18:22:22 +1 to the start with docs 18:22:28 +1 18:22:36 x-ip: sure... I leave it to jcline and other app folks if docker is better than a vagrant setup... it's all similar I think in the end. 18:23:05 jcline: by docs you mean the in-code documentation 18:23:14 ? 18:23:28 I much prefer Vagrant because I all use it for is to provision a libvirt-based VM and everything else is ansible, but I know others have different opinions (and that's totally fine) so we can document both methods 18:24:04 nirik, I think it might be intimating for new people to sif through ansible to get a setup and work on a ticker 18:24:10 the end goal should be the same: an easy way to run the app locally so you can make changes and see them there before proposing them upstream. 18:24:16 sayan, I was thinking about a docs project about Fedora apps in general. One thing I wanted to document is how to get a sphinx project going with mostly autogenerated docs 18:24:41 sayan, I sent an email to the infra list with a less brief summary 18:24:45 doteast: true 18:25:28 hey, this might also change our weekly meeting theme of "learn about something".... 18:25:36 jcline: I was planning to write blog on sphinx, autodoc after the learning on setting up for ircb 18:26:10 doteast: 'read this new doc about something' ? ;) 18:26:54 nirik, yeah ;) 18:26:57 jcline, not sure to understand you. Do you mean to document on how to get running a sphinx doc project so other devs can use sphinx to create the documentation for their apps? 18:26:58 one thing, kinda bold, but interesting, so to change the theme to "how to break app X" 18:27:17 x-ip, yeah, it's pretty meta 18:27:32 docs on making docs 18:27:39 But I've noticed a lot of projects don't have sphinx projects, or they aren't using the autodoc feature, or they aren't generating REST API documentation 18:27:40 I know nirik has been looking into knowing single point of failures 18:28:25 jcline: also, this might be a potential FAD later in the year... get a bunch of folks together and hack on adding docs to apps... 18:28:29 nirik, we dwell about how to break app X in the meeting 18:28:31 ;) 18:28:33 And in that same theme, a lot of projects lay out their tests different, or lay out their flask app differently. 18:28:48 doteast: well I sadly know most of them I think, but sure, improvements are welcome. 18:28:59 They all work, but it's jolting when jumping from one project to another as we all do 18:29:16 jcline: so perhaps a 'best practices for fedora infrastructure apps' document at first? 18:29:16 nirik, the idea is to get people help you ;) 18:29:35 It's mostly depends on the style of the dev 18:29:36 nirik, yeah 18:29:50 * nirik nods 18:29:53 so a style guide kind of thing for fedora infra apps 18:30:11 sayan, yeah, I understand that, and I think it's okay, but if we can all sit down and agree on things we'll all do the same, it'll make everyone's experience easier 18:31:06 jcline: then, should we have a vFAD soon then? 18:31:08 I've already knocked up some docs for personal use and one thing I've noted is where to expect differing styles and what is common. For example, rst-style docblocks vs numpy vs google. 18:32:43 sayan, we could 18:33:24 I thought a good way to get consensus on best practices is for us to a have a docs repo, and then someone who's interested in a practice makes a PR and we talk about whether it's something we all want to do or not. 18:34:03 sounds reasonable to me. 18:34:11 Based on that discussion we can choose to document the practice, or document that it'll vary in ways X, Y, Z, etc. 18:34:55 It the SOPs are part of this we can also get PR reviews on those which I think would be nice. I started to write one for Anitya, but I was uncertain about various bits of it and generally feedback is helpful for me. 18:35:46 we wanted the SOP's to be available even if most things were down... but we can figure out ways to address that. 18:36:11 Yeah, we can mirror the docs places. 18:36:15 sorry... what means SOP's ? 18:36:49 standard operating procedure 18:36:50 x-ip: standard operating procedure 18:36:57 one good thing that might come out of this is to bring forth things that need attention and could help people to contribute better 18:37:22 thx 18:37:31 So for example, releng has all these SOPs: https://docs.pagure.org/releng/sop.html 18:38:26 Yeah, we just don't want to be in the case where say pagure was down, and someone wanted to look at the pagure sop to see what to try to do to bring it back... and that sop was on pagure. ;) 18:39:02 heh 18:39:22 nirik, yup. We can mirror builds to, say, readthedocs, and have local clones, and have git hooks to push to other places. 18:39:53 or I was thinking we could sync it to status also (since thats a place you might be looking already) 18:40:10 but yeah, there are ways around that problem 18:40:18 Yeah 18:41:00 anyhow, any other thoughts on docs for now? 18:42:06 nope 18:42:12 ok. 18:42:15 #topic Apprentice Open office hours 18:42:16 Okay, well how about as an action item, if the discussion on the mailing list is also in favour how about I start an infra-docs project on Pagure and start making PRs for what I have so far? 18:42:25 any apprentices with questions or comments or needing things to do? 18:42:39 jcline: sounds ok to me. 18:42:50 nope 18:43:02 i'm still diving in the sops and playbooks 18:43:24 i would like to contribute in python code... but have no idea where to start. The last time i contributed was two years ago (i think) on pypkgdb which now i think is obsolete. 18:43:47 nirik, yes, I could spend some cycles 18:44:31 s/contribute in pythoncode/contribute with python code/gy 18:44:36 x-ip: we have pkgdb2 now. ;) https://fedoraproject.org/easyfix/ has some easy fix tickets 18:45:02 doteast: what was the status on ansible-review? were we going to clean up the stuff it complains about and add it to a hook? 18:45:17 x-ip, if you want something fairly easy, I've been meaning to do https://github.com/fedora-infra/anitya/issues/383 for a while, but just haven't had time 18:46:01 nirik, I go the feeling that you''r not interested any more since it was lacking some feature you we looking for, I can remember which 18:46:17 s/go/got 18:46:18 doteast: huh... not that I recall. 18:46:37 I thought it had a bunch of stuff we needed to fix first... but it was during a freeze 18:46:52 or we needed to decide policy 18:46:58 x-ip, Anitya has a Vagrant environment so it should be straightforward to get a development setup going: https://anitya.readthedocs.io/en/latest/contributing.html 18:47:19 as I remember you're expecting it to do lint by default or something, but it didn't 18:47:36 nirik, anyway I revive that memory 18:47:37 nirik, thanks. jcline, great, do you mind if i ping you on IRC to get more in depth with the project? or ask doubts? :) 18:48:03 there's a seperate ansible-lint, but thats ok... I wanted the review thing so it could stop people commiting bad things. ;) 18:48:09 x-ip, ping away. I'd also like to hear if the contribution documentation is lacking/confusing. 18:48:34 x-ip: join #fedora-apps channel 18:48:39 doteast: for example it should yell at people doing lineinfile a lot or non idempotent things. 18:49:04 nirik, ok, will reserect again... 18:49:27 doteast: that would be great... let me know if you need input. :) 18:50:17 doteast: I think the Red Hat IT folks have some default policy... I can't recall if it's public, but can look. 18:51:20 ok, any other apprentices with questions? 18:51:22 nirik, sure.... that it would make it richer since they already dug it up 18:51:36 doteast: FWIW, they wrote ansible-review too. :) 18:51:36 jcline: do you have a initial template for vagrant/ansible setup somewhere? 18:51:48 I can then create one for the tahrir-* packages 18:51:58 nirik, I know , I reached-out and talked to the guy who wrote it 18:52:06 ah cool. :) 18:52:13 #topic Open Floor 18:52:38 anyone have anything for open floor? Questions, comments, 🍺s of beverages? 18:52:51 I have a quick question 18:53:45 sayan, don't ask to ask, just ask ;) 18:54:19 * doteast revising his irc atiquate 18:54:28 I created a wrong badge, and now want to update the value 18:54:56 update the value in the database, should I go ahead an create an infrastructure ticket? 18:55:29 sayan: not sure. I guess so... there might be a way to edit them... a playbook? I cannot recall 18:56:21 https://infrastructure.fedoraproject.org/infra/docs/badges.rst 18:56:30 by value you mean the way it's awarded? 18:56:32 or / 18:56:58 nirik: nope, there ain't one to update the value. there is one to delete the badge but it's throwing an error 18:57:21 ok then, yeah, if it needs direct db changes file a ticket for it 18:58:07 nirik: yeah, direct db changes. 18:58:25 Okay, I'll create one 18:58:49 cool. 18:59:05 ok, we are almost out of time... so will close out in a minute or less here. 18:59:15 nirik, please include me if you find more about RH default policy for review 18:59:21 Thanks for coming everyone, was a good meeting with some good discussion. Just what I like to see. :) 18:59:26 doteast: can do. 18:59:38 #endmeeting