18:00:23 <smooge> #startmeeting Infrastructure (2018-03-01)
18:00:23 <zodbot> Meeting started Thu Mar  1 18:00:23 2018 UTC.  The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:23 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:00:23 <zodbot> The meeting name has been set to 'infrastructure_(2018-03-01)'
18:00:23 <smooge> #meetingname infrastructure
18:00:23 <zodbot> The meeting name has been set to 'infrastructure'
18:00:23 <smooge> #topic aloha
18:00:23 <smooge> #chair smooge relrod nirik pingou puiterwijk tflink
18:00:23 <zodbot> Current chairs: nirik pingou puiterwijk relrod smooge tflink
18:00:37 <keitellf> .hello2
18:00:39 <zodbot> keitellf: keitellf 'Lee Keitel' <lee@onesimussystems.com>
18:01:34 <wesleyotugo> .hello2
18:01:35 <zodbot> wesleyotugo: wesleyotugo 'Dennis W. Otugo' <wesleyotugo@gmail.com>
18:01:38 <puiterwijk> Hello
18:01:55 * cverna waves at the crowd
18:02:01 <nirik> morning
18:02:05 * pingou here
18:02:08 <pingou> Morgen
18:02:09 * sayan is here
18:02:11 <smooge> hello
18:02:15 * tflink is also here
18:02:25 <smooge> #topic New folks introductions
18:02:26 <smooge> #info This is a place where people who are interested in Fedora Infrastructure can introduce themselves
18:04:41 <smooge> If there are no new people (and my internet connection can stay up).. I will move to the next section
18:04:56 <smooge> #info Infrastructure hackfest in april: https://fedoraproject.org/wiki/Infrastructure_Hackathon_2018
18:04:56 <smooge> #info The problem with 503's looks contained. Now they are 408s on the server to look at
18:04:56 <smooge> #info Rebuild mirrormanager container to be more resilient
18:04:56 <smooge> #info atop monitoring/diagnostic tool installed on all systems - mizdebsk
18:04:57 <smooge> #info Staging Koji synced with production - mizdebsk
18:04:58 <smooge> #info New s390x builder in staging Koji, moved from prod - mizdebsk
18:05:00 <smooge> #info There is a Bodhi-3.4.1 in staging - please test https://bodhi.stg.fedoraproject.org/docs/user/release_notes.html - bowlofeggs
18:05:37 <smooge> mizdebsk, what does atop do? We had htop on previously to help monitor
18:06:02 * relrod checks in
18:06:05 <nirik> it's a nicer top with more info on the one screen.
18:06:11 * nirik had not heard of it
18:06:28 <mizdebsk> atop is better than htop IHMO, it monitors many things and points out performance bottleneck
18:06:34 <keitellf> I've heard of it before but completely forgot about it
18:06:47 <mizdebsk> eg. you get red blinking text if your network is at it's max capacity, and so on
18:07:41 <bowlofeggs> .hello2
18:07:42 <zodbot> bowlofeggs: bowlofeggs 'Randy Barlow' <rbarlow@redhat.com>
18:08:47 <nirik> oh, one more info...
18:08:59 <nirik> #info Beta freeze starts next week (tuesday)
18:09:03 <smooge> woah.. that sa lot of info
18:09:38 <mizdebsk> next week? not in 2 weeks?
18:09:44 * nirik checks
18:10:01 <nirik> 2018-03-06
18:10:02 <cverna> #info fedora-docs CI in progress solution based on pagure's webhooks and centos-ci openshift/jenkins pipeline
18:10:08 <cverna> one more info :)
18:10:26 <nirik> cverna: yeah, thanks much for working on that. :)
18:10:38 <cverna> bstinson++ too
18:11:16 <smooge> oof well all good things have to end.
18:11:36 <mizdebsk> nirik, according to infra sop, freeze begins "2 weeks before each release", which is 2018-03-27
18:11:46 <pingou> mizdebsk: interesting
18:12:02 <mizdebsk> so freeze should start on 2018-03-13, no?
18:12:02 <nirik> mizdebsk: well, they adjusted the schedule this time to try more time...
18:12:13 <smooge> so we need to update the doc
18:12:15 <nirik> we could wait and do 2 weeks...
18:12:24 <nirik> or we could freeze at the same time packages/releng does
18:12:29 <mizdebsk> ok, i thought infra freeze was not aligned with devel freeze
18:12:32 * pingou would be in favor of 2 weeks
18:12:47 <nirik> it's alined with releases.
18:12:53 <pingou> if the freeze last longer we end up with work piling up :(
18:13:00 <smooge> so when we have not frozen at the same time with the releng/packages we have broken them
18:13:02 <nirik> the entire purpose is so that we don't make some change that prevents the release from going out
18:13:08 <pingou> should we do a light freeze the first week?
18:13:22 <pingou> then regular freeze until the release?
18:13:24 <nirik> well, we can still do things in freeze, just need acks
18:13:57 <smooge> pingou, what is your team needing to get done next week?
18:14:27 <pingou> I guess bowlofeggs will want the new bodhi to be out
18:14:30 * puiterwijk notes that that "2 weeks before each release" is unlikely to be true anyway... :) It'd be "2 weeks before scheduled release" at best (unrelated to further arguments, but just a note)
18:14:45 <pingou> pagure 4.0 will likely not be ready by then
18:14:51 <pingou> hubs is not impacted
18:15:06 <pingou> mirrormanager needs to get out and only got the data it needed to be tested today
18:15:19 <smooge> ok I think bodhi will need to go through the ack system because it really affects development
18:15:51 * nirik has no problem reviewing things and acking them... although once we have PR's it would be even nicer.
18:15:55 <pingou> if I recall correctly Randy's schedule, it may slip in before the freeze
18:16:12 <nirik> I think we should stay alined with the release freezes... prevents confusion.
18:16:22 <pingou> ok
18:16:24 <smooge> mirrormanager would be also needing acks but it also needs to get done before too far before release date
18:16:31 <nirik> but it will be more stress on us being longer... but the idea is that 3 weeks allows us to avoid slipping
18:16:45 <pingou> :)
18:16:55 <smooge> it is an interesting hypothesis you have there
18:17:00 * pingou will not make any snarky remarks
18:17:02 * pingou will not make any snarky remarks
18:17:03 <nirik> well, it wasn't my proposal. ;)
18:17:04 * pingou will not make any snarky remarks
18:17:39 <bowlofeggs> i can just ask for acks on bodhi
18:17:41 <bowlofeggs> no big deal
18:17:53 <bowlofeggs> i could also try to deploy it on monday, before the freeze
18:18:08 <nirik> bowlofeggs: also note that tuesday is bodhi enablement for f28
18:18:25 <bowlofeggs> right, so i guess the freeze should start after that
18:18:40 <bowlofeggs> the bodhi release is a small .z release, just bug fixes, and not very many patches
18:18:41 <pingou> git commit && git push && ansible ...
18:18:51 <pingou> && email -t "Freeze is in effect"
18:18:58 <bowlofeggs> pingou: you forgot the -f on the git push :)
18:18:58 <nirik> indeed. ;)
18:19:07 <pingou> bowlofeggs: permission denied :)
18:19:12 <bowlofeggs> haha
18:19:25 * nirik can update the SOP (or put in a PR for others to review) That SOP needs other work I have been meaning to do
18:19:59 <pingou> 3 weeks of freeze should give us some time :D
18:20:01 * pingou ducks
18:20:07 <nirik> indeed
18:20:10 <smooge> I am doing a bunch of work on the SOPs this weekend. I was trying to figure out what the process was for that
18:20:31 <nirik> smooge: file PR's against the pagure.io/infra-docs project
18:21:13 <smooge> so fork a copy, make changes, then file PR's against it from m copy
18:21:31 <smooge> aaah sorry I am runnng my meeting off the rails
18:21:32 <nirik> pingou: on another topic... what was the feedback on darkserver? are we ready to do some actions there?
18:21:33 <smooge> ok back
18:21:38 <nirik> smooge: yep
18:22:25 <pingou> nirik: my outcome was: there is interest but it would likely make sense to merge it into abrt but there isn't anyone working on this
18:22:34 <pingou> (nor is there in our side someone working on fixing it)
18:23:05 <pingou> so I'm kinda tempted to say: backup the DB and let's announce and nuke it
18:23:20 <nirik> ok. +1
18:23:23 <cverna> +1
18:23:29 <smooge> sounds like a monday task
18:23:40 <cverna> what about tagger ?
18:23:44 <nirik> you want to do the announcement, and pick a sunset time and I can fold up the chairs on that date?
18:23:53 <pingou> let me reply to the thread first
18:23:59 <cverna> I don't think I saw the thread on the mail list
18:24:02 <pingou> but sure, sounds like a plan
18:24:07 <pingou> cverna: for darkserver
18:24:15 <pingou> for tagger, basically gnome-software doesn't use it
18:24:16 <cverna> pingou: tagger
18:24:24 <pingou> I've meant to ping Ralph first about it
18:24:30 <nirik> It was great source of badges. ;)
18:24:37 <pingou> then I think I'll reach out to commops and devel
18:24:47 <pingou> because the badges angle does make it interesting in some ways
18:24:58 <nirik> https://apps.fedoraproject.org/tagger/
18:24:58 <cverna> fedora-packages uses tagger, but I would glad not too
18:25:01 <cverna> :)
18:25:11 <nirik> tagger also has a issue about new packages not showing up.
18:25:24 <nirik> but perhaps that got fixed?
18:26:09 <nirik> https://pagure.io/fedora-infrastructure/issue/6323
18:26:16 <pingou> it's not hard, it's just changing the fedmsg consumer listening for new packages being added
18:26:27 <pingou> but does need someone to fix it (if we want to keep it)
18:26:46 <nirik> right
18:27:03 <nirik> probibly should file that upstream and close the infra ticket
18:27:29 <pingou> +1
18:28:26 <nirik> anyhow, now we derailed things. ;)
18:28:46 <smooge> ok putting the train back on the tracks
18:29:13 <smooge> actually this does count as "Ticket cleanup"
18:29:23 <smooge> so it wasn't too far off the rails.
18:29:49 <smooge> ok next up
18:29:52 <nirik> tagger is a flask app at least
18:29:57 <smooge> #topic Apprentice Open office minutes
18:29:57 <smooge> #info A time where apprentices may ask for help or look at problems.
18:31:22 <cverna> wesleyotugo: any progress on https://pagure.io/fedora-infrastructure/issue/6665
18:31:33 <cverna> wesleyotugo: do you need help ?
18:33:25 <smooge> mizdebsk added an easy fix today for monitoring mailman if anyone is interested in taking it
18:37:02 <smooge> ok any questions or items?
18:37:16 * pingou has one thing for the open-floor
18:37:19 <smooge> #topic Open Floor
18:37:28 <smooge> pingou, you are up
18:37:36 * nirik opens the floor, looks in it, closes.
18:37:49 <keitellf> nirik: good move
18:37:59 <pingou> I just wanted to run https://pagure.io/fedora-infrastructure/issue/6717 by folks
18:38:17 <pingou> it's basically a small service that flags commits in pagure when they successfully build in koji
18:38:34 <pingou> looks like this https://src.fedoraproject.org/rpms/perl-IO-HTML/c/69a460858b8c1f47bb6320faa6f310dab6808184
18:39:04 <bowlofeggs> sounds cool to me
18:39:08 <mizdebsk> can be useful, but why pagure-specific thing and not plain git tags?
18:39:55 <mizdebsk> can we have both at the same time?
18:39:56 <nirik> pingou: does this solve that ticket also about github 'badges' (which confused me at first)
18:40:05 <cverna> could it be part of simple-koji-ci ?
18:40:21 <wesleyotugo> cverna: Not really, I still haven't got bowlofeggs  input yet, so it is pending
18:40:44 <nirik> https://pagure.io/fedora-infrastructure/issue/6572
18:41:35 <cverna> wesleyotugo: no input means OK :)
18:41:51 <cverna> bowlofeggs: can you check https://pagure.io/fedora-infrastructure/issue/6665
18:43:15 <bowlofeggs> wesleyotugo: i didn't realize you were waiting on me - let's sync up in #fedora-apps
18:44:20 <bowlofeggs> ah, i don't get notifications on infra tickets since there isn't a way for me to subscribe to individual tickets since i'm a member
18:44:32 <bowlofeggs> so i can either get all of them (firehose) or none :(
18:45:17 <pingou> mizdebsk: git tags can be removed
18:45:42 <pingou> nirik: I hope it can be used for it yes
18:45:47 <mizdebsk_> you could have server-side hook preventing removal
18:46:09 <pingou> mizdebsk_: we used to have git tags from koji directly
18:46:24 <pingou> and this got removed, I believe because tags can be removed, but maybe worth bringing this back with releng
18:46:54 <nirik> I don't think we ever did tags... we talked about it many times, but never did it.
18:46:54 <pingou> cverna: while there is some overlap they are quite different in what they try to achieve, so I think it makes more sense to keep them separate
18:47:32 <nirik> ie, tagging after a successfull build the hash used
18:47:39 <pingou> nirik: https://src.fedoraproject.org/rpms/guake/releases
18:48:01 <nirik> that was manual I am pretty sure.
18:48:23 <nirik> the old "Makefile" thing from cvs had a 'make tag'
18:48:57 <pingou> when did we move to git (from cvs) ?
18:49:35 <mizdebsk_> 2010
18:49:49 <nirik> I was going to say 2011.. but yeah, sometime around there
18:50:06 <pingou> so maybe it was make tag
18:50:13 <pingou> I really thought it was koji doing it
18:50:31 <nirik> Date:   Mon Nov 7 22:58:36 2011 +0000
18:50:31 <nirik> Initial setup of the repo
18:50:38 <smooge> ok I am going to need to wrap this up in 5 minutes for another meeting.
18:50:53 <mizdebsk_> nirik, Date:   Thu Jul 29 17:07:15 2010 +0000
18:50:53 <mizdebsk_> dist-git conversion
18:51:17 <smooge> warring timestamps
18:51:18 <nirik> oh, I was looking at fedpkg... you could be right
18:51:19 <wesleyotugo> cverna: I had no idea, thanks 🙂
18:51:38 <wesleyotugo> bowlofeggs: Thanks that would be fantastic
18:51:46 <pingou> all the tags in guake are from before that dist-git conversion
18:52:03 <pingou> mizdebsk_: should we make a ticket to track that feature?
18:52:15 <mizdebsk_> sure, i can open it
18:52:27 <nirik> what feature? and where? tagging builds in git?
18:52:32 <mizdebsk> yes
18:52:33 <pingou> yes
18:52:44 <pingou> for fbom, yay/nay/meh?
18:52:45 <nirik> so possibly releng?
18:52:51 <pingou> nirik: sounds good
18:53:11 <nirik> we could also discuss more on list... I am not fully sure what use this has...
18:53:22 <mizdebsk> i was thinking flag-build-o-matic upstream
18:53:34 <pingou> it may become handy with stream branching nirik
18:53:45 <nirik> possibly
18:53:46 <pingou> mizdebsk: not entirely sure we should have fbom doing the git tag
18:54:08 <pingou> I see fbom as a 3rd party utility, living in our cloud, low-to-no maintenance
18:54:21 <pingou> I'm not sure I want it to commit to all git repos
18:54:31 <pingou> (even if it's just tags)
18:54:33 <mizdebsk> it's not commit, just tak
18:54:36 <mizdebsk> tag*
18:54:52 <mizdebsk> but i see your point
18:55:00 <nirik> lets continue in #fedora-apps since we are out of time
18:55:01 <nirik> ?
18:55:06 <pingou> mizdebsk: it could still be one of the options to do it
18:55:08 <pingou> sure
18:55:38 <smooge> last thing
18:55:49 <smooge> US Daylight Savings starts on the 11th
18:56:01 <smooge> but the time for the meeting will still be 1800 UTC
18:56:06 <pingou> and 2 weeks later for EU
18:56:19 <nirik> daylightsavingstime--
18:56:33 <smooge> #info The time of clock stupidity is upon us. Call your government to stop this from happening again
18:56:35 <pingou> nirik: isn't it +1 this time? :)
18:56:46 <smooge> #endmeeting