15:02:58 #startmeeting Infrastructure (2018-09-13) 15:02:58 Meeting started Thu Sep 13 15:02:58 2018 UTC. 15:02:58 This meeting is logged and archived in a public location. 15:02:58 The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:02:58 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:02:58 The meeting name has been set to 'infrastructure_(2018-09-13)' 15:02:59 #meetingname infrastructure 15:02:59 The meeting name has been set to 'infrastructure' 15:02:59 #topic aloha 15:02:59 #chair nirik pingou puiterwijk relrod smooge tflink threebean 15:02:59 Current chairs: nirik pingou puiterwijk relrod smooge tflink threebean 15:03:13 .hello2 15:03:14 * relrod waves 15:03:14 bcotton: bcotton 'Ben Cotton' 15:03:15 .hello2 15:03:16 bowlofeggs: bowlofeggs 'Randy Barlow' 15:03:21 ./hello2 15:03:34 crud lol 15:03:38 .hello2 15:03:39 aikidouke: aikidouke 'None' 15:03:41 * cverna is around 15:03:47 .hello2 15:03:47 queue-tip: Sorry, but you don't exist 15:03:58 NOTE: I will be stopping this meeting at :55 because the next meeting starts at 00 15:03:59 * mkonecny here 15:04:06 I always wondered what that command does 15:04:23 .hello2 15:04:24 mkonecny: Sorry, but you don't exist 15:04:41 #topic New folks introductions 15:04:41 #info This is a place where people who are interested in Fedora Infrastructure can introduce themselves 15:04:42 morning 15:05:29 hello any new people here this week? 15:06:34 #topic announcements and information 15:06:35 #info tflink is on extended PTO 15:06:35 #info Beta Freeze is in effect. All changes to frozen systems in Infrastructure will require +1 15:06:35 #info Bodhi 3.10.0 beta in staging https://bodhi.stg.fedoraproject.org/docs/user/release_notes.html 15:06:35 #info odd network problem on arm network worked around. 15:06:36 #info odd install problem on aarch64 systems is ongoing. 15:06:38 #info Anitya 0.13.0 deployed on staging https://stg.release-monitoring.org - mkonecny 15:06:40 #info Fedora Product Center repo created https://github.com/fedora-infra/fdc - cverna 15:06:52 been a busy week so far 15:06:55 any other announcements 15:07:16 #info Hurricane Florence may affect some servers like fedorapeople 15:07:46 queue-tip, mkonecny: the .hello2 command looks you up in Fedora's Account System (FAS) by your IRC nick and then says your name to everyone 15:07:59 so it requires you to have a FAS account, and for your IRC nick to be registered with it 15:08:26 #info fedora-messaging-1.0.0b1 is available in infra-stg repos as far back as epel7 and in rawhide, f29 Fedora repos 15:08:43 cool 15:08:58 jcline, that updates/replaces fedmsg? 15:09:09 or something completely different 15:09:10 bowlofeggs: Strange, I have it there 15:09:33 smooge, basically 15:10:02 Replaces, that is 15:11:35 mkonecny: yeah it does look like you have it set up correctly to me 15:11:37 cool 15:11:37 weird 15:11:42 #topic Oncall 15:11:42 #info Relrod is on call from 2018-09-06->2018-09-13 15:11:42 #info Smooge is on call from 2018-09-13->2018-09-20 15:11:42 #info Nirik is on call from 2018-09-20->2018-09-27 15:11:42 #info ??? is on call from 2018-09-27->2018-10-04 15:11:42 jcline: I wanted to look at recording from the workshop about fedora-messaging that was on the Fedora Flock, but it isn't on youtube 15:11:43 #info Summary of last week: (from Relrod) 15:12:00 It's been pretty quiet, probably because of freeze 15:12:03 cool 15:12:14 mkonecny, I think there wasn't any recording equipment in the room we were in 15:12:25 I will take over after this meeting. I would like to have an alternative because I may be without power 15:12:38 I can be backup 15:12:50 jcline: :-( 15:13:24 thanks relrod 15:13:31 ok next up 15:13:42 #topic recent openshift additions - kevin 15:13:47 nirik, you are up 15:14:05 oh yeah, wanted to mention my recent stuff with openshift stg... 15:14:22 I enabled metrics and logs... so now we can see cpu/load/etc. 15:14:35 and look at logs, but it's kinda complex... kibana 15:14:57 Also multitenant is enabled in stg (meaning all projects are network isolated from each other) 15:15:11 I'm going to roll these out to prod after freeze. 15:15:15 nirik++ 15:15:17 * nirik digs up some urls. 15:15:48 oh, and puiterwijk fixed logs and such from the web interface, but non firefox as firefox does something bad with websockets. 15:16:32 https://kibana.app.os.stg.fedoraproject.org/app/kibana 15:16:44 nirik: could the multitenant stuff be used to move bodhi staging into prod openshift, while continuing to keep it isolated from production stuff (like db) 15:16:47 the monitoring appears next to each app (small graphs) 15:17:03 bowlofeggs: well, partly, but it still needs more... 15:17:16 cool 15:17:27 if that is achievable, it could be nice ☺ 15:17:28 it needs egress stuff to make sure it goes out on the stg network too... but that does exist. 15:17:59 you can basically set that per node... 15:18:06 and then label that node as say 'staging' 15:18:15 and then tell the staging apps to use that label 15:18:26 ah interesting 15:18:27 so we will get there, just more testing and working 15:18:38 so the pods can't be on the same host as prod, but can be on the same openshift deployment 15:18:59 cool, i look forward to it (no rush, it's a nice to have ☺) 15:19:00 but pods get moved around at times right? 15:19:21 yeah, but based on node selector/label. 15:19:34 or you can tell it only to put those on nodes labeled etc. 15:19:36 we would need more than 2 nodes then? 15:19:40 yep. 15:19:56 I added 2 nodes in stg also... it has 4 now. 15:20:01 ok cool.. 15:20:06 oh and https://registry-console-default.app.os.stg.fedoraproject.org/registry should now work too 15:20:11 so you can see images/etc 15:20:43 Anyhow, I'd love for us to start moving more things to openshift. 15:20:51 perhaps we could do a VFAD after f29 is out... 15:21:33 cool 15:21:34 ok 15:21:38 next up? 15:21:46 #topic python3 porting help needed for supybot-plugins - kevin 15:21:58 hey nirik.. can you come back? 15:22:18 sure! 15:22:33 so, we have several supybot plugins we use in zodbot. 15:22:44 they are currently python2 and there's bugs to retire them in f30 15:22:54 it would be nice if we could port them to python3 and prevent that 15:23:25 https://bugzilla.redhat.com/show_bug.cgi?id=1627142 15:23:33 https://bugzilla.redhat.com/show_bug.cgi?id=1627228 15:23:45 https://bugzilla.redhat.com/show_bug.cgi?id=1627225 15:23:55 so, any help welcome. it should be pretty easy I hope 15:24:21 that could be some good material for the office hours if an apprentice want to look at it 15:24:31 indeed 15:25:19 would like to do take one or more - looks like there are some comments from Miro on the first one? 15:25:35 yeah, miro offered to help too 15:26:20 ok - I'll follow the bugs and see if I can do anything to help :) 15:26:30 thanks! 15:27:05 Are there any tests ready for python3? 15:27:56 tests for these? or apps in general ? 15:27:57 .hello2 15:27:57 x3mboy: x3mboy 'Eduard Lucena' 15:28:33 mkonecny, ^^ 15:29:07 aikidouke: let me know if you need help :) 15:29:12 nirik: in general 15:29:24 cverna++ ty! 15:29:54 normally tests for apps have been the province of the developer of the app 15:30:20 yeah... any normal tests should work I would think... nothing specific I know of to test python3 ability 15:30:52 nirik: thanks, this is what I wanted to know 15:31:07 #topic Tickets discussion 15:31:08 #info https://pagure.io/fedora-infrastructure/report/Meetings%20ticket 15:31:22 .ticket 7233 15:31:24 smooge: Issue #7233: RFE: Machine-readable freeze flag? - fedora-infrastructure - Pagure - https://pagure.io/fedora-infrastructure/issue/7233 15:32:25 so this is a discussion on how to know when we are in freeze or not 15:32:25 threebean asked for this... 15:32:29 yep. 15:33:19 i can see that if we put it in PDC replacement then both ops, dev, qa and releng can all query one place for what they are doing 15:33:31 but so would be having it in the ansible tree 15:34:14 it could be a nice feature for bodhi too, because bodhi could ask "are you sure?" if someone tries to push during a freeze 15:34:16 I think we can do the ansible tree now and the PDC thing in how many months? 15:34:42 honestly the ansible tree is probably good enough for a permanent solution 15:34:57 pdc replacement is officially FDC (Fedora Product Center) :) 15:35:02 it's just a URL for some json/yaml/toml/data_type_of_the_month 15:36:34 yep 15:36:59 anything more here? 15:37:45 .ticket 7226 15:37:47 smooge: Issue #7226: DNS for fedoraproject.org email alias forwarding - fedora-infrastructure - Pagure - https://pagure.io/fedora-infrastructure/issue/7226 15:37:54 I was thinking just an ansible var... but if people have better ideas I am all ears. 15:37:54 if no ideas we can move on. 15:37:54 I'll probibly wait and do this after freeze 15:38:25 I left this on from last week... 15:38:35 no change, waiting on reporter I think 15:38:47 okie dokie 15:38:59 #topic Apprentice Open office minutes 15:39:00 #info A time where apprentices may ask for help or look at problems. 15:39:10 ok anything for this part? 15:39:52 should I resend my clear 2fa request? 15:40:21 aikidouke: yeah, I don't think I saw the updated one? 15:40:21 * relrod still hasn't done any onboard videos yet, will try to do some soonish. 15:40:34 ok thanks nirik will do 15:40:40 (hopefully correctly) 15:40:58 aikidouke: is your key up on the keyservers? 15:41:04 it is sir 15:41:18 cool. 15:41:21 relrod, thanks 15:41:27 relrod: liked your talk :) 15:41:36 aikidouke: thank you! 15:41:44 I am hopefully going to watch it tomorrow. 15:42:25 I've been trying to think about ways that an apprentice could clone the infra repo locally and then cut out some variables and run just a mini version of the app that they wanted to work on, if that makes sense 15:42:42 that would be useful for many things 15:43:01 obv wouldnt work well for builders, etc 15:43:02 or use openshift as a playground :-) 15:43:36 what level of os ninja would one need for that mkonecny? 15:45:02 in a playground, local OR openshift, what things would not be needed? fedmsg? stg? just trying to think about things that would cut down on complexity 15:45:44 I'm just thinking about using one instance of openshift for apprentice training, maybe prepared with some test application 15:45:50 but thats probably a bigger discussion for an apprentice workday? 15:46:06 i would agree on that 15:46:08 yeah, as more of our apps use openshift that should be doable... 15:46:23 and people could just use oc cluster up locally. 15:46:33 * aikidouke nods 15:46:56 is after freeze a good time for a workday? after F29 beta/final release? 15:47:37 yeah, after final might be... 15:47:44 Octoberish 15:48:04 I am going to move to Open Floor 15:48:07 👌🏼 15:48:08 #topic Open Floor 15:48:38 any open projects? 15:49:16 if not I can close out and give you all coffee time 15:49:29 coffee! 15:49:34 #info Release Readiness and Go/No-Go meetings are later today! 15:50:00 day of meetings. what fun! 15:50:07 #info Stay Safe US Southeast :) 15:50:14 aikidouke++ 15:50:14 bcotton: Karma for aikidouke changed to 3 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:50:14 ah yes. these are important meetings that need concentration... so please just ping the oncall person and have them contact whoever you needed to see when they are available 15:50:50 yeah, hopefully everyone has their bread and milk stockpiled. 15:50:59 don't forget your cinnamon 15:51:08 i've got a month's worth of soylent on hand! 15:51:08 because you will need it for the french toast 15:51:10 milk will not keep well :-) 15:51:22 vanilla extract is important, too 15:51:27 oh yes 15:51:48 what I love is that canned and powdered milk is rarely bought 15:51:50 ok anyway 15:51:57 have a good week everyone 15:52:00 say good night bowlofeggs 15:52:26 #endmeeting