13:02:24 <mcatanzaro> #startmeeting workstation
13:02:24 <zodbot> Meeting started Mon Jul  1 13:02:24 2019 UTC.
13:02:24 <zodbot> This meeting is logged and archived in a public location.
13:02:24 <zodbot> The chair is mcatanzaro. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:02:24 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
13:02:24 <zodbot> The meeting name has been set to 'workstation'
13:02:38 <mcatanzaro> #meetingname workstation
13:02:38 <zodbot> The meeting name has been set to 'workstation'
13:03:12 <cmurf> .hello chrismurphy
13:03:12 <zodbot> cmurf: chrismurphy 'Chris Murphy' <bugzilla@colorremedies.com>
13:03:42 <mcatanzaro> #chair cmurf juhp[m] langdon mclasen
13:03:42 <zodbot> Current chairs: cmurf juhp[m] langdon mcatanzaro mclasen
13:03:48 <mcatanzaro> #topic Roll call
13:03:50 <juhp[m]> .hello petersen
13:03:51 <zodbot> juhp[m]: petersen 'Jens Petersen' <petersen@redhat.com>
13:04:01 <mclasen> .hello mclasen
13:04:02 <zodbot> mclasen: mclasen 'Matthias Clasen' <mclasen@redhat.com>
13:04:42 <langdon> .hello2
13:04:44 <zodbot> langdon: langdon 'Langdon White' <langdon@redhat.com>
13:05:22 <mcatanzaro> Ryan is scheduled to chair today but Ryan has resigned, so I'll chair again... cmurf, you're next week's chair as a heads-up, we try to rotate every week and it's coincidence that I've been running every meeting for two months in a row...
13:06:31 <cmurf> roger
13:07:36 <mclasen> mcatanzaro: thanks for doing it
13:08:39 <mcatanzaro> .hello catanzaro
13:08:40 <zodbot> mcatanzaro: catanzaro 'Michael Catanzaro' <mcatanzaro@gnome.org>
13:09:03 <mcatanzaro> OK, we have five, that's quorum (barely)
13:09:19 <mcatanzaro> #topic Ryan resigned
13:09:54 <mcatanzaro> #undo
13:09:54 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x7f6d327a7d90>
13:10:07 <mcatanzaro> Anyone know if kalev or cschalle are coming? otaylor is excused of course
13:10:40 <mcatanzaro> We need to discuss WG membership again (one more open slot) and it's a shame to have people missing for that, but oh well
13:10:45 <mcatanzaro> #topic Ryan resigned WG membership
13:10:46 <juhp[m]> I believe cschalle is away
13:10:48 <mclasen> cschaller and kalev are both on pto
13:11:01 <mcatanzaro> OK so five is all we'll have
13:11:55 <mcatanzaro> Well basically I propose we add one new WG member to replace Ryan. Previously we had discussed Steven Snow and he said he was interested... I think that was managed by otaylor? I actually don't know Steven, anyone else familiar with him?
13:12:17 <juhp[m]> So how many members are we currently btw?
13:12:55 <mclasen> I don't know steven
13:13:04 <cmurf> I don't either.
13:13:19 <langdon> i know john snow (no joke) but not steven
13:13:22 <cmurf> There are nine in the list, eight current minus ryanlerch.
13:13:46 <juhp[m]> Okay, right
13:14:52 <cmurf> There's no actual problem with eight. A tie breaker is possible with 9 but not ensured as they could abstain.
13:15:24 <juhp[m]> Right
13:15:26 <cmurf> Also a 5-4 tie break isn't the most compelling decision.
13:15:44 <juhp[m]> Nod
13:15:48 <langdon> dumb meeting time q.. does this time follow utc? or some tz daylight savings?
13:16:02 <cmurf> 1300 UTC I believe
13:16:26 <langdon> cool.. ok.. i just like to set it in my calendar the correct way else i foget
13:16:31 <langdon> *forget even
13:16:45 <juhp[m]> I thought traditionally it was EST/EDT based?
13:16:53 <mcatanzaro> juhp[m] is right, it follows US Eastern time... Eastern Daylight Time -> Eastern Standard Time in November
13:17:09 <langdon> hahahaha glad i asked
13:17:28 <mcatanzaro> So 1200 UTC in the winter
13:17:28 <cmurf> right exactly
13:17:41 <cmurf> it's UTC on paper, in practice it's gonna get changed twice a year
13:17:44 <langdon> mcatanzaro: well.. the "right" winter ;)
13:17:52 <mcatanzaro> Anyway it sounds like none of us know Steven
13:17:52 <juhp[m]> I notice since I have been living in tz's without DST for a long time...
13:18:45 <mcatanzaro> We could increase by one, or not at all, or even by two. otaylor had proposed increasing the WG size up to 11
13:19:12 <mcatanzaro> But anyway, I guess we should defer discussion until we have someone who knows Steven? I think otaylor does.
13:19:23 <juhp[m]> I was thinking like ~10
13:19:25 <cmurf> i think that's unworkable without some stronger promises of attendance; it makes meeting quorum harder
13:19:35 <mcatanzaro> Well I suppose we can define quorum to be four instead of five in the meantime?
13:19:45 <cmurf> that isn't what quorum meens
13:19:49 <cmurf> or even means
13:20:05 <juhp[m]> True
13:20:06 <mclasen> mcatanzaro: turning dst off would be the right way to fix it
13:20:12 <mcatanzaro> I think our quorum is one half of our membership?
13:20:22 <cmurf> ahh yes that's fair
13:20:51 <juhp[m]> Yeah
13:20:52 <cmurf> the membership is 8, a 50% quorum is reasonable, so 4 makes quorum
13:20:58 <juhp[m]> Rounded up
13:21:33 <mcatanzaro> Ideally we should try to improve attendance of course, but can't fault people for taking holidays and parental leave
13:21:48 <mcatanzaro> We didn't have quorum again last week, btw, that was a shame
13:22:09 <juhp[m]> In that sense even membership increases the chance of "quorum" all things being equal
13:22:12 <mcatanzaro> Oh and welcome langdon! I believe this is your first meeting as a WG member, right?
13:22:45 <langdon> ha.. i think so.. because the same no-quorum meeting is the one i started to show up for and the my brain fell out of my head :/
13:22:47 <juhp[m]> Hey Langdon!
13:23:23 <cmurf> on second thought, i'm not sure 50% can be quorum, because 50% isn't a majority.
13:23:32 <mcatanzaro> Don't make it harder for us :P
13:23:42 <cmurf> not intentionally
13:23:46 <juhp[m]> Right but seems good enough
13:23:58 <mcatanzaro> Well there are no other issues on the agenda this week... do we want to discuss membership more or move on to open floor?
13:25:39 <juhp[m]> Anything for F31?
13:26:21 <mcatanzaro> #topic Open floor
13:27:03 <cmurf> there's a good blog write up on Fedora 31 floating about, I dont have a URL handy at this second, quite a lot of things are on the burners.
13:27:33 <juhp[m]> Yes
13:27:34 <langdon> the phoronix one?
13:27:41 <cmurf> cschaller
13:27:58 <cmurf> https://blogs.gnome.org/uraeus/2019/06/24/on-the-road-to-fedora-workstation-31/
13:28:20 <mcatanzaro> #info cschaller continues to blog about Fedora Workstation
13:28:31 * langdon again raises fist at ctrl-L = "clear" &  "jump to bar in FF"
13:30:51 <cmurf> So the Fedora Media Writer not launching on macOS, is macOS version specific, and puiterwijk is looking into a work around.
13:31:33 <langdon> is that high in his queue? or should we try to find someone who might have more bandwidth?
13:31:55 <cmurf> It's a good question but it takes esoteric knowledge about Apple's app signing process.
13:32:16 <langdon> ohh.. yeah.. ok..
13:32:18 <cmurf> I don't know where it is in his queue.
13:32:23 <puiterwijk> Well, do note that: 1. it only impacts macOS 10.13 (i.e. not people who run the latest version), and 2. there's a client-side local workaround
13:32:25 <langdon> and you probably have to have the keys
13:32:47 <langdon> do we have a note on the get.fp.o page?
13:32:57 <puiterwijk> https://pagure.io/fedora-infrastructure/issue/7902#comment-577150
13:33:24 <langdon> hi puiterwijk! :)
13:33:35 <puiterwijk> hi langdon
13:33:47 <mcatanzaro> #topic Fedora Media Writer not launching on macOS
13:34:02 <mcatanzaro> #info Fedora Media Writer doesn't work on macOS 10.13
13:34:17 <mcatanzaro> This 10.13 is still supported by Apple though, right?
13:34:18 <puiterwijk> *for some systems*
13:34:35 <langdon> puiterwijk: where is the fun in consistency!?!?
13:34:37 <mcatanzaro> #undo
13:34:37 <zodbot> Removing item from minutes: INFO by mcatanzaro at 13:34:02 : Fedora Media Writer doesn't work on macOS 10.13
13:34:56 <mcatanzaro> #info Fedora Media Writer doesn't work on certain macOS 10.13 systems
13:35:03 <cmurf> It's super goofy, I'm not happy with Apple. I had a clean install of 10.13, fully updated, and ran into this which is why I thought everyone was affected.
13:35:30 <puiterwijk> mcatanzaro: basically, there's a bug in the local gatekeeper cache on 10.13, and if you get hit by it, you need to clear your cache before you can open stapled software
13:35:43 <mcatanzaro> So not our fault
13:35:48 <puiterwijk> No
13:36:03 <mcatanzaro> "No, it is our fault" or "no, not our fault"?
13:36:06 <puiterwijk> The reason it showed up with this latest build is because I started stapling the latest build
13:36:13 <puiterwijk> "not our fault"
13:36:17 <puiterwijk> It's a bug in macOS 10.13
13:36:19 <cmurf> I did test the local work around and it does fix the problem. But then Apple has no tech note on this for regular users. They do have the incantation buried in a note for developers.
13:36:35 <mcatanzaro> And could affect anything anybody tries to download if that software is stapled? Whatever that means...?
13:36:40 <puiterwijk> mcatanzaro: so, what I have on my todo list is to make a non-stapled build, and get that tested and republished
13:37:00 <mcatanzaro> okidokie
13:37:01 <puiterwijk> mcatanzaro: stapling means I submitted the build to Apple to check for malicious stuff, and I attached their signature that no, nothing malicious was found
13:37:33 <juhp[m]> I see
13:37:48 <puiterwijk> macOS 10.14 (I really want to type OSX...) will also go out to the internet to get the notarization results if it's not stapled, but if you staple them it has an easier time at doing so.
13:38:11 <puiterwijk> Unfortunately, it looks like while 10.14 is perfectly fine with it all, 10.13 has a bug at verifying this.
13:38:56 <mclasen> do we have any sense how important this is ? how many users do we have for the media writer on os x ?
13:39:00 <mclasen> 1, 10, 100 ?
13:39:01 <cmurf> The top solution Apple recommends is upgrade to 10.14
13:39:34 <puiterwijk> cmurf: to be fair, I would suggest that too.
13:40:21 <puiterwijk> Since macOS Sierra is no longer supported, I *guess* Apple will stop supporting High Sierra by fall, with the release of.... 10.15 (can't remember the codename)
13:40:37 <langdon> and 10.14 released 9+ months ago
13:40:56 <cmurf> Yeah, typically Apple "supports" the current two major versions. But the support starts to slack really badly except for security updates, with the n-1 version.
13:41:01 <puiterwijk> And 10.14 is a free upgrade for all 10.13 users, iirc
13:41:02 <cmurf> And 10.13 is that n-1 version.
13:41:20 <juhp[m]> This isn't really particular to Workstation, right?
13:41:24 <puiterwijk> So... honestly, while I will try to get a non-stapled build, it's not high on my priority queue.
13:41:38 <puiterwijk> juhp[m]: no. This is particular to the system you run media writer on to get your workstation media generated
13:41:51 <cmurf> Yeah but every major version drops hardware support. So there are people stuck with 10.13, but that would be hardware older than 2011.
13:42:29 <langdon> omg.. OSX came out in 2001?!??! like 18 years ago?!?!
13:42:44 <cmurf> Another factor for puiterwijk to balance out is that there's a decent chance sometime at or soon after 10.15 release ships, that notarization+stapling will become required for apps.
13:42:44 * langdon notes he is old
13:42:47 <puiterwijk> langdon: OSX no longer exists, it's now macOS :P
13:42:50 <cmurf> So anyway...
13:43:36 <langdon> puiterwijk: ha... that must make it better!
13:43:54 <puiterwijk> langdon: it makes my brain worse though... I still want to type OSX when I'm typing quick :)
13:44:00 <langdon> ha
13:44:44 <cmurf> Well they changed the name to macOS as opposed to iOS, and now they have iPadOS
13:45:02 <langdon> cmurf: but it is all the same OS, iirc?
13:45:05 <puiterwijk> Anyway, it's on my todo list, but this are the reasons I didn't put it very high in my list.
13:45:07 <mcatanzaro> No
13:45:12 <mcatanzaro> iPads used to run iOS
13:45:17 <puiterwijk> langdon: iOS/iPadOS and macOS are two entirely different codebases :)
13:45:17 <mcatanzaro> Now iPadOS
13:45:23 <mcatanzaro> iOS, macOS, iPadOS, tvOS, and watchOS, all different
13:45:25 <langdon> lol
13:45:29 <puiterwijk> mcatanzaro: iPadOS is... basically a fork/branch of iOS :)
13:45:35 <cmurf> i'm not sure they're all that different
13:45:42 <mcatanzaro> Probably not all THAT different
13:45:49 <mcatanzaro> Maybe like Fedora/Debian different? Who knows
13:45:53 <puiterwijk> I think they just named it differently to indicate the behaviour is different between them, but the code is the same
13:46:03 <puiterwijk> mcatanzaro: I think more like "#if define IS_IPAD" different
13:46:13 <puiterwijk> "#ifdef IS_IPAD"*... I can write C
13:46:31 <cmurf> it's probably just subsetting
13:46:57 <juhp[m]> I think we are definitely OT now
13:46:58 <puiterwijk> Anyway. So, mediawriter. If people say it's highly critical, please mark that in the infra ticket and bring it up to my mgr. Otherwise I'll get to it when I get to it.
13:47:02 <puiterwijk> juhp[m]: agreed.
13:47:23 <cmurf> And I've got a brief update on the zram business.
13:49:30 <cmurf> I've posted to anaconda list and have figured out where the change needs to go to get zram enabled for LiveOS install media.
13:49:33 <cmurf> https://www.redhat.com/archives/anaconda-devel-list/2019-June/msg00011.html
13:50:17 <cmurf> My plan is to come up with PR to do exactly that, stomp out the old swap stuff in that fedora-live-base.ks
13:50:23 <cmurf> And use anaconda's zram for now.
13:52:00 <cmurf> pbrobinson has a zram package in Fedora that includes zram-swap.service I think mostly used for Fedora IoT. And I'm discussing with systemd folks the idea of having one ship with systemd, and then standardizing on that.
13:52:03 <cmurf> That's the update.
13:52:24 <pbrobinson> cmurf: it's used on all arm images
13:52:33 <cmurf> ok super
13:52:48 <pbrobinson> cmurf: and IoT as well
13:53:30 <pbrobinson> cmurf: there'
13:53:51 <pbrobinson> cmurf: there is work from some people to drop the anaconda one and use mine BTW
13:53:58 * mclasen has no opinion on this - as soon as my system hits swap, its basically dead
13:54:15 <pbrobinson> weird that it's been silent basically for ever and 2 different queries in a week or so
13:54:17 <mcatanzaro> #topic zram
13:54:17 <cmurf> mclasen: If it's non-live, that's a separate issue that I want to look at
13:54:55 <cmurf> pbrobinson: might have something to do with my queries on anaconda-devel@ and systemd-devel@ because I mentioned your package specifically
13:55:02 <mclasen> I also don't care about live isos much - I think we should minimize the differences and get people to and through the installer as quickly as possible
13:55:20 <cmurf> basically checking out the differences in approach with the various zram setups out there and if systemd folks just want to ship one
13:56:02 <pbrobinson> cmurf: the systemd have previously refused, but they did do some improvements for us
13:56:02 <cmurf> mclasen: right and in the low memory VM situation, Live soaks up quite a lot more RAM, and swap on zram helps with that
13:56:34 <cmurf> *shrug*
13:56:38 <pbrobinson> we've had pretty decent success with in on Arm, since it went live I don't remember a single complaint or issue with ti
13:56:39 <pbrobinson> it
13:57:35 <cmurf> And also FWIW, netinstalls always use it when RAM is less than 2GiB
13:58:41 <cmurf> mclasen and mcatanzaro both have previously reported production workstation systems that implode with swap thrashing, so I want to look into that soon
13:59:05 <pbrobinson> cmurf: feel to reach out if you've got queries
13:59:37 <cmurf> That to me sounds like an application that might be able to take advantage of cgroup2 to constrain what might be unreasonable memory hogging
13:59:55 <cmurf> But to test I'm gonna need a reproducer from someone
14:00:01 * cmurf looks at mclasen and mcatanzaro
14:00:21 <cmurf> pbrobinson: ok thanks
14:00:39 <cmurf> pbrobinson: btw cmurf = chrismurphy :D
14:00:42 <mcatanzaro> cmurf: I have too much RAM now to reproduce that anymore, but yes, on my old computer it was really easy to hang the entire desktop
14:01:01 <mcatanzaro> Anyway we're out of time, I think we can close here
14:01:02 <mclasen> yeah, this is why systemic behaviors like this will never get fixed
14:01:59 <mcatanzaro> I have a feeling Red Hat could fix swap if Red Hat decides it wants to fix it. Seems to be able to accomplish pretty much anything it decides to accomplish recently. Otherwise it will indeed never get fixed. ;)
14:02:01 <mcatanzaro> Anyway
14:02:05 <mcatanzaro> #endmeeting