15:01:24 <bowlofeggs> #startmeeting Bodhi stakeholders (2017-07-18)
15:01:24 <zodbot> Meeting started Tue Jul 18 15:01:24 2017 UTC.  The chair is bowlofeggs. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:24 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:24 <zodbot> The meeting name has been set to 'bodhi_stakeholders_(2017-07-18)'
15:01:25 <bowlofeggs> #meetingname bodhi_stakeholders
15:01:25 <zodbot> The meeting name has been set to 'bodhi_stakeholders'
15:01:27 <bowlofeggs> #topic salutations
15:01:28 <bowlofeggs> #chair acarter bowlofeggs caleigh dgilmore masta mboddu nirik pbrobinson puiterwijk trishnag Kellin
15:01:28 <zodbot> Current chairs: Kellin acarter bowlofeggs caleigh dgilmore masta mboddu nirik pbrobinson puiterwijk trishnag
15:01:29 <dustymabe> .hello dustymabe
15:01:30 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
15:01:47 <trishnag> o/
15:01:50 * nirik is here, but a bit groggy
15:01:50 <trishnag> .hello trishnag
15:01:51 <zodbot> trishnag: trishnag 'Trishna Guha' <trishnaguha17@gmail.com>
15:04:20 * masta_ lurks
15:04:36 <ryanlerch> .hello ryanlerch
15:04:37 <zodbot> ryanlerch: ryanlerch 'Ryan Lerch' <rlerch@redhat.com>
15:05:04 <bowlofeggs> there's not a ton on the agenda today
15:05:15 <bowlofeggs> #topic announcements and information
15:05:16 <bowlofeggs> #info A Bodhi 2.9.0 beta is deployed to stg: https://bodhi.stg.fedoraproject.org/
15:05:18 <bowlofeggs> #info Release notes available at https://bodhi.stg.fedoraproject.org/docs/release_notes.html
15:05:49 <bowlofeggs> 2.9 is mostly a bugfix release, but there's some features too
15:05:54 <bowlofeggs> lots of CLI stuff
15:05:56 <dustymabe> bowlofeggs: +1
15:06:01 <jcline> .hello jcline
15:06:02 <zodbot> jcline: jcline 'Jeremy Cline' <jeremy@jcline.org>
15:06:42 <bowlofeggs> it's got some preliminary patches from caleigh that will help bodhi support "batched updates"
15:06:59 <bowlofeggs> that's an exciting feature i think - it was discussed on devel@ 6ish months ago
15:07:35 <bowlofeggs> the idea is that bodhi will automatically batch "non-high severity" updates (meaning, non-security and non-urgent) into a weekly push instead of pushing all stable updates every day
15:07:58 <bowlofeggs> this is one of those rare bodhi features that will make a difference to end users, which is cool
15:08:07 <bowlofeggs> hopefully this will reduce the daily churn of fedora
15:08:09 <masta_> weekly updates?
15:08:19 <bowlofeggs> and end users should only see updates that are severe
15:08:30 <bowlofeggs> masta_: that would be a matter of releng policy, but i would suggest weekly
15:08:59 <masta_> so it's just a capability we have, right? we can batch things for later pushes?
15:09:00 <bowlofeggs> masta_: basically, there will be a cron script that will run on $schedule that will switch the batched updates to stable
15:09:21 <bowlofeggs> and schedule can be decided by releng, but i think weekly is probably a good recommendation
15:10:01 <masta_> interesting feature, cool
15:10:02 <bowlofeggs> masta_: all non-urgent updates will automatically go into request: batched instead of request: stable. when the cron job runs, it just switched the request: batched to request: stable
15:10:16 <bowlofeggs> nice and simple approach, and pacakgers will be able to override too
15:10:36 <bowlofeggs> security updates and high severity updates will still go straight to stable like they do today
15:10:51 <ryanlerch> bowlofeggs: is there any feedback in the UI for when an update is "staged"?
15:11:01 <bowlofeggs> anyways, this won't actually be in 2.9.0 - 2.9.0 just has some of the database work needed to make it happen - it'll be in a later release
15:11:10 <bowlofeggs> just wanted to highlight that progress has been made ☺
15:11:31 <masta_> bowlofeggs, does this open the way for highly urgent fixes having a fast path ?
15:11:55 <bowlofeggs> ryanlerch: the plan is to have the UI say "batched" instead of "stable" on the request. hopefully with a tooltip to explain what it means
15:12:14 <ryanlerch> bowlofeggs: coolio!
15:12:35 <bowlofeggs> masta_: i would think probably so in the sense that i think the daily mashes should go faster since there should be fewer updates in them. i'm not certain though
15:12:49 <masta_> rel-eng has an old issue/ticket for that, and we never conceived a great way to move forward urgent CVE's outside the normal push work flow
15:12:56 <bowlofeggs> i'm not confident how much the number of updates affects the time to mash
15:13:03 <dustymabe> bowlofeggs: does this mean we could not create drpms daily
15:13:13 <bowlofeggs> i'd expect it to help some, but i'm not sure it's linear (i'd guess it probably isn't linear)
15:13:20 <dustymabe> but only when a new rpm makes it into batched?
15:13:51 <bowlofeggs> dustymabe: the drpms would be created when the updates gets mashed, so most updates would only be done on the schedule (weekly, for example)
15:14:49 <dustymabe> k
15:14:49 <nirik> I suspect it makes very little speed difference if it's 1 update or 10.
15:14:53 <bowlofeggs> masta_: it is currently possible to push a specific set of builds, which i think could be used to push an urgent cve
15:15:07 <bowlofeggs> masta_: bodhi-push --help talks about a --builds flag
15:15:53 <bowlofeggs> nirik: yeah it might not - the mashing time is probably just due to making the repo data and the repo isn't getting smaller
15:16:05 <bowlofeggs> nirik: but the drpm generation bit should go faster at least
15:16:11 <bowlofeggs> so there should be some difference
15:16:22 <bowlofeggs> i'm not sure how much difference it would be
15:16:28 <bowlofeggs> the feature is more about end users than it is about mash time
15:16:48 <ryanlerch> bowlofeggs: nirik: just out of interest, what is the mash time at the moment?
15:16:49 <bowlofeggs> the goal is to reduce how often your computer tells you it wants to install 79 updates ☺
15:16:51 <ryanlerch> ballpark
15:17:01 <bowlofeggs> ryanlerch: i'd throw out 8-12 hours?
15:17:10 <dustymabe> bowlofeggs: so part of what my thoughts are on the weekly bathed updates are that there is a batched update but also another stream people can follow
15:17:13 <bowlofeggs> that's to mash everything
15:17:27 <dustymabe> that represents what is pending the next batch
15:17:31 <nirik> I think it's less than that, but I haven't looked recently
15:17:44 <bowlofeggs> yeah i haven't actually timed it either
15:17:46 <bowlofeggs> just a guess
15:18:14 <bowlofeggs> dustymabe: that could be cool, though that would be yet another repo to mash and mirror (unless we didn't think it needs to be mirrored)
15:18:19 <bowlofeggs> dustymabe: we already have updates-testing
15:18:23 <bowlofeggs> whcih isn't exactly the same
15:18:25 <bowlofeggs> but kinda similar
15:18:33 <dustymabe> bowlofeggs: what happens with rawhide today?
15:18:45 <dustymabe> don't those rpms get put into the repos faster ?
15:18:55 <bowlofeggs> dustymabe: i don't know much about how rawhide works
15:18:59 <nirik> once a day...
15:19:02 <bowlofeggs> yeah the rawhide repo seems to be daily
15:19:11 <bowlofeggs> bodhi doesn't make rawhide
15:19:21 <dustymabe> nirik: what creates the repo for rawhide? is that the nightly pungi run?
15:19:22 <nirik> and it takes about 8 hours... because it also makes all the images and such
15:19:31 <bowlofeggs> (though i like the suggestion of making a "bikeshed" distro with rawhide + bodhi)
15:19:32 <nirik> yes, cron job that runs pungi
15:20:28 <bowlofeggs> anyways, i'm excited about the feature and caleigh is doing good work ☺
15:20:42 <bowlofeggs> hard to say if it'll be in 2.10 or not, but it could be!
15:20:47 <bowlofeggs> i do think it'll be in 2.11 at least
15:20:56 <bowlofeggs> so ~next 1-2 months
15:21:25 <bowlofeggs> we are going to give developers the ability to override and go straight to stable, so i'm also hoping to be able to gather statistics on how often they choose to do that
15:21:30 <bowlofeggs> because we'd like them to generally not do that
15:21:46 <bowlofeggs> but i also don't want to force a policy on them
15:22:05 <bowlofeggs> so the plan is to monitor and see how much people do it and react if it seems like it's done too much
15:22:14 <bowlofeggs> "react" might mean just writing a plea e-mail ☺
15:22:46 <bowlofeggs> or possibly making the UI ask nicely not to do that unless it's truly importnat, and if it's important why isn't it marked severe? ☺
15:23:10 <nirik> hum?
15:23:19 <nirik> direct to stable with no karma?
15:23:21 <dustymabe> what's the point in enabling them to go straight to stable?
15:23:38 <nirik> thats against updates policy. ;)
15:24:58 <bowlofeggs> nirik: sorry - what i mean is skipping the "batched" state
15:25:08 <bowlofeggs> users can't go to batched until they meet the stable requirements
15:25:25 <nirik> ah, I see...
15:25:29 <bowlofeggs> but if they are in batched, we will allow them to click a button to go to stable to skip the batching
15:25:42 <nirik> yeah, that makes sense. ;)
15:25:44 <bowlofeggs> and i want to monitor it because i'm worried that too many people will do that
15:25:47 <dustymabe> i see - so basically the ability to skip batched
15:25:50 <dustymabe> cool +1
15:25:50 <bowlofeggs> yes
15:25:57 <dustymabe> you should provide a 'comment box'
15:26:01 <dustymabe> for them to say why they are doing it
15:26:09 <dustymabe> and we can look at those reasons over time
15:26:24 <bowlofeggs> i want to give the freedom to do that, but i want to also see if it's happening so much that the feature isn't helping the users
15:26:31 <bowlofeggs> yeah
15:26:35 <dustymabe> btw I have to run - just wanted to say if possible let's get stg in order so we can run bodhi there and test the pungi repo creation with bodhi
15:26:40 <bowlofeggs> i think at first i'll just watch for how much it happens
15:26:44 <bowlofeggs> and go from there
15:26:46 <dustymabe> (not including mash, which I know doesn't work in stg)
15:26:52 <bowlofeggs> anyways, let's move on
15:26:54 <dustymabe> ok I have to run
15:26:56 <dustymabe> bye all
15:27:31 <bowlofeggs> dustymabe: i still can't do taht today - i need to fix a bug for kevin still regarding alternate arches and mash configs
15:27:48 <bowlofeggs> dustymabe: i spent way too much time trying to get staging working yesterday and haven't gotten to that unfortunately
15:28:05 <bowlofeggs> #topic Looking forward
15:28:06 <bowlofeggs> #info Bodhi's high priority issue list https://github.com/fedora-infra/bodhi/issues?q=is%3Aopen+is%3Aissue+label%3A%22High+priority%22
15:28:08 <bowlofeggs> #info High priority means it's important, but not a show stopper
15:28:09 <bowlofeggs> Any filed issues that aren't on these lists that should be?
15:30:37 <bowlofeggs> <crickets>
15:31:23 * nirik doesn't think so
15:31:56 <bowlofeggs> cool
15:32:07 <bowlofeggs> #topic Open floor
15:39:53 <bowlofeggs> ok, let's call it
15:39:57 <bowlofeggs> thanks for coming everyone!
15:40:04 <bowlofeggs> #endmeeting