18:00:07 #startmeeting Infrastructure (2016-01-21) 18:00:07 Meeting started Thu Jan 21 18:00:07 2016 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:07 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:07 The meeting name has been set to 'infrastructure_(2016-01-21)' 18:00:07 #meetingname infrastructure 18:00:07 #topic aloha 18:00:07 #chair smooge relrod nirik abadger1999 lmacken dgilmore mdomsch threebean pingou puiterwijk pbrobinson 18:00:07 The meeting name has been set to 'infrastructure' 18:00:08 Current chairs: abadger1999 dgilmore lmacken mdomsch nirik pbrobinson pingou puiterwijk relrod smooge threebean 18:00:08 #topic New folks introductions / Apprentice feedback 18:00:56 who all is around for a infra meeting? 18:00:58 .hello smdeep 18:01:00 smdeep: smdeep 'Sudeep Mukherjee' 18:01:02 Hi 18:01:05 Hi 18:01:06 hi 18:01:09 * dotEast2015 present 18:01:10 Hi 18:01:10 hi 18:02:37 .hello odin2016 18:02:38 odin2016: odin2016 'None' 18:02:52 also.. here. 18:02:53 Any new folks like to give a short introduction or apprentices with questions or comments? 18:03:24 I would like to ask someone to look at ticket 3294, if someone will have some time 18:03:29 nirik, I was having some issues recently so was not able to reply to emils 18:03:43 going forward I will be able to give more time 18:03:54 .ticket 3294 18:03:55 nirik: #3294 (Enable varnish caching for applications) – Fedora Infrastructure - https://fedorahosted.org/fedora-infrastructure/ticket/3294 18:03:58 smdeep: no problems. ;) 18:04:09 nirik, thanks 18:04:33 aldapetr: I think threebean or puiterwijk might be best to look at that if we can catch them with some free time sometime. :) 18:04:33 aldapetr: Will take a look at that after the meeting 18:04:40 well, there you go. ;) 18:04:57 thanks 18:05:22 it sounds great :) 18:05:37 cool. any other new folks or apprentices with questions? 18:06:13 ok, lets go on to status/info 18:06:32 #topic announcements and information 18:06:32 #info Applied updates to most machines (but no reboots) - kevin 18:06:32 #info added some space to ppc, s390, and arm koji storage - kevin 18:06:32 #info composer.stg is now Fedora 23 (to help with pungi4 work) - kevin 18:06:32 #info We now have a manual playbook to sync fas db from prod -> stg - kevin 18:06:34 #info Wrote up info about our datacenter visit last week: - kevin 18:06:34 #link http://www.scrye.com/wordpress/nirik/2016/01/17/a-vist-to-a-fedora-datacenter/ 18:06:42 anything there anyone wants to add to or discuss? 18:07:29 #info We now have a playbook to automatically update firmware on Dell boxes, and will upgrade all of them 18:07:41 very nice. ;) 18:07:49 puiterwijk: was the inventory group of dells up to date? 18:08:07 nirik: not sure. My playbook runs against "all" since I'm planning on adding the other brands as well 18:08:19 (except for IBM because I don't trust their downloads) 18:08:20 ok. 18:08:39 threebean: you around? 18:08:42 So the playbook has a dict mapping machine types to firmware blobs, and then applies them based on the ways the vendors have their updates 18:09:08 .fas sayanchowdhury 18:09:08 sayan|think: sayanchowdhury 'Sayan Chowdhury' 18:10:03 #topic FY16 items to finish up before the fiscal year ends - kevin 18:10:16 so, we had things we listed last year we wanted to try and get done... 18:10:27 we should see if there's anything we can finish up in the next month or so. 18:10:47 https://fedoraproject.org/wiki/Fedora_Engineering/FY16_Plan 18:11:30 off hand we have pending: the wiki upgrade... the arm/ppc koji hubs into ansible 18:12:04 fas3 we should figure out what we want to do and commit 18:12:59 nirik: yes. Could we schedule a sync-up on fas3 next week? 18:13:00 anyhow, hopefully the ppc/arm koji hub stuff can get done soon. I need to catch pbrobinson 18:13:17 SmootherFrOgZ: sure, pick a time/date. :) 18:13:33 * SmootherFrOgZ nods 18:14:06 SmootherFrOgZ: perhaps you could send to the list a proposed time/date and folks could hash it out there... 18:14:18 yup, wfm 18:15:32 cool. would be nice to move that forward some. 18:16:33 alright, so if everyone could take a look and move forward anything we would like to that would be great. 18:16:58 I also don't have anything lined up for learn about today. ;( Do people have items that they would _like_ to have someone teach about? 18:18:41 * threebean here now 18:18:57 threebean: want to do a quick retrospective of the debt week? 18:19:04 yeah, that's be great. 18:19:15 #link https://lists.fedoraproject.org/archives/list/infrastructure@lists.fedoraproject.org/message/IGBX7YAKI2SP6CYD5NSJLADIXSW4TX3M/ 18:19:25 #topic Tech Debt Week Retrospective 18:19:47 I liked pingou's idea of doing a short meetup each day... I think that would help people focus and get more done. 18:19:48 ok, I wrote that note to the infra list.. but just wanted to open up some more room for discussion. 18:19:54 * threebean agrees 18:20:15 one thing that makes that slightly more complicated is the timezone difference. 18:20:20 yeah. 18:20:33 it would be morning PST, mid-day EST, and evening in EMEA. 18:20:41 but, that's not a blocker. still a good idea. 18:20:41 there were some folks that showed up but just got pointed to the wiki page, which was good, but a bit vuage... 18:20:48 * threebean nods 18:21:01 focusing on a few apps might be better too... not sure. 18:21:12 yeah, just to help narrow it down. 18:21:34 the stuff we did on dev side this time was almost entirely focused on python-fedora, which was good. maybe next time we focus on some apps instead. 18:22:25 yep. 18:22:27 I was thinking for next time maybe we should call for people to sign up on the wiki ahead of time too. I think we had some miscommunication about who would and wouldn't be around for it. 18:22:43 yeah, thats a good idea. 18:22:45 have a list of participants, like we do for a FAD. 18:23:24 yep. 18:23:35 * threebean hmms. 18:23:51 Maybe have people also write a bit of a summary at the end of the day? 18:24:12 And/or at the beginning of what they hope to do? 18:24:38 well, we had a wiki page part for things that were done. 18:24:41 Although that might make contributing a bit of a higher bar 18:25:14 threebean: I don't think timezone is necessarily bad 18:25:17 Right, that's true. 18:25:21 yeah, but if you are going to work on something like that during a week, it's probably not a bad thing. 18:25:32 threebean: your start of the day is my mid-day but then I just plan for afternoon+ morning d+1 18:26:10 maybe going out on a limb here, but there's one aspect of the whole thing that makes it a PITA, irrespective of how we organize it. working on tech debt kind of necessarily means putting aside the other things we would otherwise be working on.. many of which have some kind of deadline or expectation of urgency. so - tech debt fighting is competing with those things. 18:26:27 yeah. 18:26:46 maybe instead of a whole week, dedicate a day or two a month ? 18:26:54 pingou: yeah - i wasn't saying it was bad. we can't just all start at the same time and divy up tasks for the day like we do at a FAD. 18:27:19 true there 18:27:42 pcreech: we could certainly try that :) 18:27:48 one thing we could try instead of a set meeting at a time, is several times each 24h period just say "lets sync up" and do a 5min meeting to tell what people are working on or point people to things to do? 18:28:00 might help break it up into more focused, managable chunks 18:28:18 nirik: +1 18:28:45 nirik, +1 18:29:06 hm - well. do we want to try it again? 18:29:07 might be distracting, but could work... dunno 18:29:15 and if so, how soon? 18:29:50 I'm up for tryng it again, but we also I think talked about a similar docs thing... ie, time to just update/fix docs... would that be a thing to try next? 18:30:25 sure - we could choose that as a focus. 18:30:26 what if we declare the debt week well in advance to plan out deadlines accordingly? 18:30:44 sayan|think: sometimes things come up not much in advance tho. ;( 18:31:10 nirik: how wide are you thinking to cast that net? include docs for infra SOP and infra apps? more stuff? 18:31:45 mostly those... infra docs and apps... I guess perhaps wiki pages? 18:31:58 sure. infra docs and apps sounds good to me ;) 18:32:12 I know there's lots to do there... 18:32:18 yeah, we'll max out. 18:32:28 some apps don't have SOPs... lots of the infra docs are out of date, etc. 18:32:53 * threebean may spend the whole time making diagrams. 18:33:08 ha, yeah. 18:33:15 time to show off some mad inkscape skills? 18:33:17 lol 18:33:19 so when? I think it would be good before the f24 cycle really hits. 18:33:21 tflink: ditaa! 18:33:21 threebean: you can't sped more than 90% of the week on diagrams! :) 18:33:34 spend* 18:33:39 puiterwijk: add that to the formal rules. 18:33:41 feb 16th - 19th? 18:33:50 absolutely! 18:34:01 oo, that's sooner than I was thinking, nirik. but, ok. 18:34:09 Oh, sorry, the "absolutely" was about the formal rules 18:34:22 well, alpha freeze is in mar 18:34:32 for docs, sounds ok to me 18:34:33 there's devconf inbetween now and then, which cuts into the time available to get other stuff done. 18:34:38 16 - 19 sounds reasonable to me, although I think it's a bit close to the FY end and close to a lot of conferences 18:34:48 can we wait until the alpha freeze week? 18:35:02 oooh... yeah, docs shouldn't affect freeze 18:35:26 thats mar 8th 18:35:43 +1 18:35:49 Sounds good 18:36:22 so mar 8th to 11th... or even 8th to 14th (so there's a weekend?) 18:36:57 +1 to the weekend idea 18:37:03 (so I might actually be helpful) 18:37:08 cool :) 18:37:14 yeah, there might be people who only have time in weekends. 18:37:27 wfm 18:37:36 Right, but then we should probably note that during the weekends less people might be available 18:37:42 (to help*) 18:37:56 And/or at irregular times 18:37:59 #info Docs clean up week (infra docs, applications documentation, wiki) on 2016-03-08 to 2016-03-14 18:38:02 sure 18:38:09 cool. anything else on this? 18:38:14 otoh, those of us who are fulltime could try to organize ourselves so that at least one of us is always around on the weekend. 18:38:36 question: do we allow ipsilon in the list of apps to document? :D 18:38:37 while making sure that burden doesn't fall unfairly on a single one of us. 18:38:46 pingou: absolutely. 18:38:54 sure. :) 18:39:04 puiterwijk: you're not allowed to answer that one :D 18:39:20 pingou: hah. Why not? I can't tell that I plan to document my own stuff? :) 18:39:28 puiterwijk: you're going to have us help writing that do you don't want to do :-p 18:39:42 that doc* 18:40:07 #topic Open Floor 18:40:14 Anything for open floor? 18:40:31 pagure 1.0 is around the corner 18:40:35 cool. 18:40:36 We are almost ready to move darkserver into el7 18:40:41 (Have I been saying this for too long already?) 18:40:42 also cool. ;) 18:40:58 pingou: I think you have :) 18:41:08 dang, but it is :) 18:41:31 Our initial test run worked very well. 18:41:33 pingou: I need to look sometime and see what issues in pagure needs before it could replace trac more. ;) 18:41:50 i'm having way more trouble than expected getting PDC deployed. still making progress, but its slow. 18:41:51 nirik: cool, looking forward 18:42:18 threebean: just to make sure: the SAML issues are resolved, right? Or do you need help with that? 18:42:34 (since you didn't come back to me about it, I guessed SAML wasn't the issue anymore) 18:42:58 puiterwijk: resolved now. although it did take some more work. 18:43:22 stupid idea for the doc week: provide a kickstart and instruction on how to set up a vm w/ it 18:43:39 so that we can (if needed) start documenting how to deploy an app from scratch 18:43:43 Ah, okay. Would you have time to talk that through so I can document it all, threebean? 18:43:45 pingou, provide a Vagrantfile 18:43:47 there were reverseproxy issues where mod_auth_mellon wouldn't use the x forwarded stuff.. so I had to fake it with ServerName and UseCanonical. 18:44:03 pingou: we could make our ks repo available with some work. 18:44:28 threebean: ah, that's most likely because we're using non-defacto-standard headers. I've been meaning to add the defacto standard to the templates as well. 18:44:30 nirik: I was looking at one the other day, we know it's "public" to all apprentices anyway right? 18:44:57 pingou: yeah, but we could clean up thigns and make a new repo with no history 18:45:19 nirik: I'd like that :) 18:45:27 we could. I don't see any reason why not. 18:46:28 anyhow, anything else? or shall we give everyone 15min of their day back and call it a meeting? 18:46:41 I was wondering if we could/should expand the ks w/ specific packages (ie: install upon image creation rather than yum/dnf'ing it later) 18:46:43 +1 to close 18:47:22 pingou: well, I think doing that with ansible would be preferable, since the kickstart would crash if a package didn't exist or anything, and it would give us lots and lots of kickstarts 18:47:27 pingou: that would expand the amount of ks'es a lot... I prefer a base then ansible installs the app 18:47:33 and it's already tricky enough finding the correct kickstart sometimes 18:48:40 Also: a kickstart only does something upon install, and we don't want to reinstall VMs everytime we need to add a package, and we don't want to keep ansible and kickstart in sync with package installs 18:49:05 I was approaching the ks as a template but it's not something we do really 18:50:00 yeah... lets discuss this more out of meeting? 18:50:05 Thanks for coming everyone! 18:50:08 #endmeeting