14:00:27 #startmeeting Infrastructure (2018-10-11) 14:00:27 Meeting started Thu Oct 11 14:00:27 2018 UTC. 14:00:27 This meeting is logged and archived in a public location. 14:00:27 The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:27 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:00:27 The meeting name has been set to 'infrastructure_(2018-10-11)' 14:00:27 #meetingname infrastructure 14:00:27 The meeting name has been set to 'infrastructure' 14:00:27 #topic aloha 14:00:27 #chair nirik pingou puiterwijk relrod smooge tflink threebean cverna mkonecny 14:00:27 Current chairs: cverna mkonecny nirik pingou puiterwijk relrod smooge tflink threebean 14:00:48 * cyberworm54 here 14:00:49 .hello2 14:00:50 puiterwijk: puiterwijk 'Patrick "マルタインアンドレアス" Uiterwijk' 14:00:56 .hello2 14:00:57 odin2016: odin2016 'David Shier' 14:01:00 hi 14:01:01 morning 14:01:08 morning 14:01:11 .hello2 14:01:12 Kal3ssiN: Sorry, but you don't exist 14:01:15 hey 14:01:16 =( 14:01:18 .hello2 14:01:19 aikidouke: aikidouke 'Zach Villers' 14:01:28 .hello linkdupont 14:01:29 sub_pop: linkdupont 'Link Dupont' 14:01:29 hi 14:01:30 o/ 14:01:40 Kal3ssiN: it works if your FAS username is identical to your IRC nick 14:02:04 would '.hello2 kalessin'work? 14:02:10 .hello kalessin 14:02:11 puiterwijk: kalessin 'None' 14:02:12 That does 14:02:22 ack, thanks puiterwijk 14:03:06 .hello2 14:03:08 bowlofeggs: bowlofeggs 'Randy Barlow' 14:03:29 hello everyone 14:04:31 hello 14:05:01 hi guys! 14:05:02 #topic New folks introductions 14:05:02 #info This is a place where people who are interested in Fedora Infrastructure can introduce themselves 14:06:02 hello everyone 14:06:14 I've been poking at fpdc a lot with cverna, so I thought I'd say hello. 14:06:25 I intro myself a couple of weeks ago, but didnt interact too much ¬¬ 14:06:38 hello to you both! welcome. 14:07:02 welcome! 14:07:16 welcome 14:08:26 is there any onboard guide for infra? 14:08:59 https://fedoraproject.org/wiki/Infrastructure/GettingStarted 14:09:21 thanks chris787 14:10:52 ok, who would I ask to joing fi-apprentice group? 14:11:11 s/joing/join 14:11:53 After you have sent an introduction, you can send a message to the oncall person who can add you to the group 14:12:26 If you have already sent the introduction, just let the oncall person know and they will do so when they have a chance 14:13:18 Kal3ssiN: in #fedora-admin channel ;) 14:14:17 #topic announcements and information 14:14:17 #info tflink is on extended PTO 14:14:17 #info Final Freeze has started 2018-10-09 14:14:17 #info version 0.13.1 of release-monitoring.org deployed in production - mkonecny 14:15:11 #info Changes to the Ansible repo are inbound shortly, anyone who has access to the ansible repo should follow infra-list 14:15:22 FYI, openshift 3.11 is out... I might update stg soon... 14:16:03 puiterwijk, are these changes like 'get email, read email, rm -rf ~/ansible; git clone new place; go back to your work' 14:16:27 smooge: yes, pretty much 14:17:21 puiterwijk: will it do AI and blockchains now 14:17:56 bowlofeggs: artifical blockchain is the word on the street. 14:18:00 bowlofeggs: no 14:18:12 bowlofeggs: hbut it will be moving over to Pagure! 14:18:19 So it will do PRs 14:18:56 hahaha 14:19:25 artificial blockchain written in a functional contract language 14:19:58 hey, no digging at functional languages. :( 14:20:02 dig at contracts all you want. :P 14:20:40 relrod, well it it was erlang based 14:20:54 * smooge moves into a different standup routine 14:20:57 #topic Oncall 14:20:58 #info bowlofeggs is on call from 2018-10-04 -> 2018-10-11 14:20:58 #info smooge is on call from 2018-10-11 -> 2018-10-18 14:20:58 #info relrod is on call from 2018-10-18 -> 2018-10-25 14:20:58 #info smooge is on call from 2018-10-25 -> 2018-11-01 14:20:58 #info ??? is on call from 2018-11-01 -> 2018-11-?? 14:21:00 #info Summary of last week: (from bowlofeggs) 14:21:13 bowlofeggs, the spot light is on you and the hook is ready 14:22:19 * smooge waits for bowlofeggs to get the clown makeup on and into the tiny car 14:23:42 it was pretty smooth i'd say 14:23:57 cool 14:24:02 nothing too crazy 14:24:02 you got a lot of things dealt with 14:24:08 didn't need sysadmin-main privs 14:24:16 yeah i did have a few bigs things, like the registry dying 14:24:23 but that would probably have been my responsibility anyway 14:24:36 mostly i just triaged tickets ☺ 14:25:00 yep. thanks 14:25:01 there was one time i had to summon the ops ☺ 14:25:14 other than that i'd say pretty quiet 14:25:20 and nirik took the weekend for me 14:26:59 #topic RFR revamp - kevin 14:27:07 nirik, you have the floor 14:28:51 so, I wanted to toss out an idea I have had for a while and see what people think... 14:29:17 we currently have a long and complex request for resources process. This has protected us from being left holding the bag 14:29:24 when people make an app and disappear. 14:29:46 I was thinking we may want to revamp the process with some new factors in mind: 14:30:04 openshift makes it easier to spin things up and work on them 14:30:14 more puiterwijk clones? 14:30:17 cloud allows people to work on things in isolation 14:30:49 so perhaps we want to add some rfr paths more suited to those... 14:31:08 and be less afraid to delete stuff if it doesn't work out. 14:31:14 yeah that sounds cool to me, but i really think we also need the clones so long as they dont' attack 14:31:14 Also, maybe make some changes based on supportability (SLEs)? 14:31:26 yeah! 14:31:35 yeah tying SLE to how complex the RFR is makes a lot of sense 14:31:39 bowlofeggs: no promises are made as to the future behaviour of clones 14:31:46 adding SLE info as part of the process... perhaps request apps display that somehow 14:31:59 (sle is Service level expectations) 14:32:07 if you want a strong SLE, you need to do more work, if you are ok with a weak one you are on your own 14:32:10 hahaha 14:32:35 Also, I have some plans on how to make the security audit bit for supported apps more measurable and easy to perform, which should be landing soon 14:32:36 right... 14:32:51 +1 14:33:12 also for openshift some things may be much easier than traditional... like we can make monitoring just built in... 14:33:38 would these apps be tagged in some way that makes it obvious to infra and end-users? 14:33:41 in any case I can write up a proposal, I just wanted to see if everyone thoght it was a good idea 14:33:57 aikidouke: yeah, that's the "apps display that somehow" (SLE) 14:34:10 domain name? 14:34:12 at one point we were going to try and use domains, but I think we need something on apps or in front of them or something 14:34:31 unless we add more domains to our openshift that won't help there 14:34:33 so if I make smooge-on-a-shelf.fpo or something, passersby wont expect smooge-on-a-shelf to have five 9s of uptime 14:34:41 lol 14:35:31 yeah. I am not sure how best to do that. 14:35:46 we could have a link, but people don't click on all links 14:36:13 Maybe on status.fp.o make categories? 14:36:14 maybe just a banner or something, require/ask creator to use a template for anything public 14:36:19 we talking just web apss, or? 14:36:23 we could have fedorastatus.org divide up things... yeah 14:36:23 I was going to say banner as well. 14:36:25 so that when people go look at whether it's known down, they can see it there 14:37:30 yeah, mostly web apps. 14:38:18 also, I was thinking a new RFR process would strongly encourage openshift 14:38:20 I mean, for webapps, there's gotta be someway to impose a banner on random things like that. 14:38:48 sure, but in a way thats not intrusive... people wouldn't like a big SLE: LOW over their app 14:38:56 I'd agree with that, we have it, it makes life easy for people (in theory) 14:39:33 doesn't have to be big, was thinking maybe something thin right that the top that could be be dismissed... 14:39:47 I saw something like what I am thinking earlier today. 14:40:14 or we could spin up a microservice that has a endpoint for all apps sles! bowlofeggs can write it... :) 14:40:22 smooge-on-a-shelf has 1 5 of uptime 14:40:29 beucase, honestly, the average end user / tester is not going to care what the app sle status is. 14:40:39 unless/until it's down. 14:40:43 fair. 14:40:49 nirik: haven't you heard? nanoservices are the new hotness 14:40:52 so, how about this: 14:40:52 smooge: lol 1 5 14:40:57 picoservices 14:41:22 bowlofeggs: the-new-hotness? https://github.com/fedora-infra/the-new-hotness 14:41:41 we modify our proxies httpd 'app is down' page to talk about sles and point to status where we have them in neat rows 14:41:57 smooge: unless they are yocto services, I think we should wait. 14:42:09 nirik: I like that. 14:42:30 nirik, that sounds good 14:42:37 and/or the openshift one 14:42:40 mkonecny: haha yep 14:42:45 pingou, did you get an answer to your question on domain name? 14:43:08 smooge: I was referring to our old idea 14:43:15 which seems to no longer be the consensus 14:43:17 I think domain names won't do it anymore... we have too much mix. ;( but we could add another openshift domain and try it I suppose 14:43:21 we should just make one giant gigaservice that does everything fedora needs 14:43:23 so I got my answer yes :) 14:43:28 then there's only 1 SLE 14:44:04 it's all our web apps combined, and it's our DB, operating system, framework, chat system, and so forth 14:44:05 I like it. we can retire all our hardware in favor of a single node too! 14:44:10 serverless 14:44:20 ok enough 14:44:25 run it on shields!!! 14:44:29 right, we are drifting. ;) 14:44:33 * smooge puts everyone back in the clown car 14:44:38 blockchain? 14:44:41 * bowlofeggs goes into the glovebox 14:44:45 #topic US daylight savings time ends 2018-11-04. Meeting time? 14:44:48 bowlofeggs: again? 14:44:59 pingou: hah, yeah, it's where i belong 14:45:18 that ends already? dang. 14:45:19 poll: how many people actually keep gloves in their gloveboxes? 14:45:22 So the US does its daylights savings time weird 14:45:26 stupid DST 14:45:39 smooge: we ignore DST, and just keep to the same UTC time? 14:45:40 i propose that the meeting time stay in UTC 14:45:48 since we are a global audience 14:46:02 I'll plus ne that all day... 14:46:03 EU wants to get rid of this daylight saving time, finally 14:46:14 puiterwijk, so I am confused by how this works.. does that put the meeting at 0600 for nirik or 0800 for nirik 14:46:27 I think 0600 is way too early 14:46:44 * nirik agrees. :) 14:46:49 i think it does go to 0600 for nirik :( 14:46:51 I believe our side of the planet is going forward, so that'd be 0600 yes. 14:46:58 UTC++ 14:47:02 but it will be called '7am' 14:47:04 because what we currently call 10:00 will change to 09:00 when we "fall back" 14:47:09 smooge: so, I'm fine adjusting the time to one or two hours in the future, but then we stay there... 14:47:13 (in my time zone) 14:47:29 Basically, stop the constant shifting with DST, and just stick to one time. 14:47:29 yeah... dst in fall becomes earlier.. 14:47:35 puiterwijk, I agree 14:47:45 i'd be in favor of changing the time to help nirik, but let's not change it again when we go back to DST 14:47:54 I agree 14:48:04 I think 1500 would be where we should move to 14:48:08 pingou: it's latest for you, what do you think? 14:48:24 right, which means we just shift it a bit forward and never speak of DST ahain. 14:48:24 I like. 14:48:25 nirik: it's same for me. It'd be 18:00 for us 14:48:43 6pm will likely not work for me 14:48:46 ah, I thought pingou was one zone futher... not sure why. 14:48:55 nirik: nope, same tz :) 14:48:58 where are you pingou? 14:49:08 pingou: 7pm? That'd be 8am for nirik I think 14:49:10 odin2016: France 14:49:22 puiterwijk: past 5:30 pm it's kids time for me 14:49:29 pingou: that i what I thought... I'M in germany. 14:49:37 odin2016: cool :) 14:50:20 so he's on my time... the meeting is presently at 1600 here... or 4 PM 14:50:30 in the period during US time change and EU one, I can try to adjust but long term it won't work for me 14:50:43 ok understood 14:50:47 4 PM for me currently 14:50:52 but if that only blocks me, I don't think we should block on this 14:50:55 Yep. So if we move forward 1 hour for DST, it'd be 17:00, until EU switches too, then 18:00 14:51:03 is eu changing this year? I know they are thinking of dropping it 14:51:17 nirik: we are this year, from 2019 it's per-country I believe 14:51:18 I've not heard we aren't yet. 14:51:24 nirik: the EU gave an opinion which all states have to implement 14:51:25 nirik: yes 14:51:30 (hopefully in the same way) 14:51:34 so they are currently on DST and are looking to end it. It is a matter if DST starts up in March or not 14:51:51 nirik: it will take some time till they get rid of DST 14:51:52 AI believe the EU at least realizes that deciding to skip DST within a month of it happening is a terrible idea.... 14:51:53 so they will swap their clocks.. and may not swap them ever again 14:51:59 mkonecny: yeah. ;( 14:52:02 how much faster than the speed of light would nirik or pingou need to go so that they can gain/lose an hour? and what would their ping time be like from the future/past 14:52:19 aikidouke: 42 14:52:26 ok I will take this to the list. 14:52:32 :) this is the correct answer 14:52:32 it isn't going to get solved here 14:52:35 smooge: so we end up with correct timezone in the end, nice 14:52:37 pingou: beat me to it. 14:52:51 #topic Tickets discussion 14:52:51 #info https://pagure.io/fedora-infrastructure/report/Meetings%20ticket 14:52:52 64 bytes from localhost (::1): icmp_seq=1 ttl=64 time=0.136 ms 14:52:52 ping localhost -c 1 14:53:02 lol 14:53:06 again, I think nirik has more added-value than I in this meeting and should be prioritized 14:53:17 I didn't put in any tickets for meeting... 14:53:23 ok I have no network to pagure 14:53:24 but if there's any people want to talk about 14:54:19 so my network may be going out 14:54:27 or we are low on time too... 14:54:33 smooge: yeah i've been having lots of timeouts to pagure this week too 14:54:43 #topic Open Floor 14:54:57 smooge: you and i live close - i wonder if it's something to do with the routes/proxies we use? 14:55:59 Well, hopefully at some point we'll be able to move it to PHX2, which might make that significantly better. 14:56:02 looking at the firefox tool its actually timing out on libavatar 14:56:03 so I've been fooling around with zod/supy/limnoria and updating plugins to py3 and i noticed we have a Fedora Wiki search function, but nothing for quick docs 14:56:29 i asked the docs team and right now Antora has no search feature (or our deploy doesnt) 14:57:12 does Zod need to be able to search the quick docs? 14:57:19 no 14:57:36 * aikidouke nods 14:57:42 well that's easy then :) 14:57:43 ok that is my unvarnished opinion 14:57:54 search is hard. 14:58:02 yup 14:58:08 but yeah, thats way beyond the scope of a bot plugin. 14:58:21 I think zodbot has a lot of things added onto it which were 'fun' and 'cool' and then the person promoting it left and we have a poopton of broken half-features 14:58:25 do people still use Zod to search the wiki? 14:58:44 aikidouke: I.... did not even know that was a thing 14:58:45 aikidouke, maybe. 14:58:51 I would say it needs to go too 14:58:53 ahh...this could get to be a really big thing 14:58:55 I bet some people will use it. But it can go 14:59:08 maybe once I have something to share, I can do a list mail 14:59:19 aikidouke: if they want to search wiki, there's a searchbar on the wiki, and it's indexes in better search engines 14:59:43 yep...agreed...okey dokey 14:59:50 +1 14:59:56 ty! 14:59:59 indexed* 15:00:03 no problem. thanks for the work 15:00:13 anything else for open floor? 15:00:19 aikidouke: if you need acks on getting rid of old cruft, just ping any of us, and you'll probably get the +1! :D 15:00:33 thanks puiterwijk! 15:01:10 umm..not a topic but python 3.7 has some bugs with pipenv or something? I think nb filed one or two..just a heads up 15:01:42 okie dokie 15:01:47 (not specifically 3.7, but pep something that changes a behavior...yadda) 15:01:56 I am going to end this in 30 seconds 15:02:03 but thats an entirely different story all together 15:02:30 you can take off the facepaint in 3...2....1 15:02:32 #endmeeting