16:00:09 <dcavalca> #startmeeting Hyperscale SIG
16:00:09 <centbot> Meeting started Wed Sep 29 16:00:09 2021 UTC.  The chair is dcavalca. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:09 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:00:17 <dcavalca> #chair dcavalca jvreeland
16:00:17 <centbot> Current chairs: dcavalca jvreeland
16:00:21 <dcavalca> morning folks
16:01:44 <Eighth_Doctor> hey all
16:02:02 <jvreeland> Hello
16:02:26 <dcavalca> let's get started
16:02:28 <dcavalca> #topic Followups
16:02:55 <dcavalca> I spent some more time looking at openssl3 in preparation for systemd switching over
16:03:18 <dcavalca> after discussion with the EPEL WG, there's agreement that EPEL is probably the right place for this, but the specfile is non-trivial
16:03:27 <dcavalca> and getting it to build there without some major refactoring won't be fun
16:03:53 <dcavalca> I don't have the bandwidth to tackle this right now, but when I do I'll probably try to reach out to the maintainers and see if they're open to reworking their specfile
16:04:10 <dcavalca> and if anybody else wants to take point on this, be my guest :)
16:05:12 <dcavalca> the only other followup I have is a reminder that Dojo is next week
16:05:15 <Eighth_Doctor> I've been doing some work to update spin packages so I can respin the Workstation live media this week
16:05:20 <dcavalca> which means I should probably start working on my slides :)
16:05:41 <Eighth_Doctor> refreshed Anaconda, branched PackageKit, and merged in updates of blivet
16:06:02 <dcavalca> nice
16:06:04 <Eighth_Doctor> I haven't yet rebuilt all the spin packages for AArch64, but I'm doing it as I go through updating things
16:06:24 <dcavalca> are we going to publish aarch64 live media too?
16:06:27 <dcavalca> (is that even a thing?)
16:06:30 <jvreeland> Eighth_Doctor: do you need an arch64 kernel for the live media?
16:06:47 <jvreeland> \
16:07:29 <Eighth_Doctor> yes
16:07:31 <Eighth_Doctor> and yes
16:07:41 <Eighth_Doctor> livecd-tools was adapted for aarch64 earlier in the year
16:07:52 <Eighth_Doctor> so we can produce aarch64 live media too :)
16:08:09 <Eighth_Doctor> I'd like for us to start building from the cs9 kernel though
16:08:42 <jvreeland> That sounds good to me
16:09:41 <dcavalca> yeah that works
16:09:58 <dcavalca> moving to on
16:10:04 <dcavalca> #topic Announcements
16:10:16 <dcavalca> we're due for reporting this month
16:10:27 <dcavalca> I've started drafting in https://hackmd.io/NyGzStIER6Gapkah0SayRA
16:10:47 <dcavalca> please add the stuff you've worked on there and fill in whatever details you deem worthy of dissemination
16:11:10 <dcavalca> this is due by oct 3, so not much time left
16:11:29 <dcavalca> oh that's a sunday
16:11:39 <dcavalca> ok, let's say this is due by this Friday then :)
16:12:06 <Eighth_Doctor> if jvreeland can do the kernel rebase today or tomorrow, I can start building ISOs and have an announce ready by Friday and have a section in the report
16:13:05 <jvreeland> That sounds reasonable
16:13:40 <jvreeland> I can give an update later today/tomorrow morning with progress if something comes up to but
16:13:51 <dcavalca> I'll try to flesh out the rest later today or tomorrow
16:14:17 <Eighth_Doctor> I'm going to start testing builds to see if any other userspace packages need updating
16:14:30 <Eighth_Doctor> it seems like the rpm stack was already updated recently, so I hope not
16:14:50 <dcavalca> same, I have a couple more things I need to build and test, I'll try to prioritize that so we can hopefully have it done before the deadline
16:14:55 <Eighth_Doctor> oh, I guess I should update btrfs-progs to 5.14.1, to go along with the 5.14 kernel :)
16:14:56 <dcavalca> yeah, malmond did that the other day
16:16:06 <dcavalca> on a somewhat related note, we now have pretty user-facing docs up at https://sigs.centos.org/hyperscale
16:16:18 <dcavalca> these are generated from https://pagure.io/centos-sig-hyperscale/sig using mkdocs
16:16:34 <jvreeland> Oh, that got setup quickly
16:16:37 <dcavalca> I've filled in some basic stuff, but it'd be nice to get these in ok shape in time for the report so we can point people to them
16:16:56 <Eighth_Doctor> yay!
16:16:59 <dcavalca> yeah, arrfab announced it today on centos-devel
16:17:03 <arrfab> yep :)
16:17:25 <Eighth_Doctor> when I make the next Workstation update, I'll fill in stuff on that page
16:17:26 <dcavalca> also Eighth_Doctor you should put up a PR against the material theme to add an icon for Pagure :)
16:17:43 <dcavalca> https://github.com/squidfunk/mkdocs-material/tree/master/material/.icons/fontawesome/brands
16:17:50 <Eighth_Doctor> lol sure, where?
16:18:13 <dcavalca> at least I think that's the repo
16:18:37 <dcavalca> also feel free to play with the theme/options, I tried picking sane settings but I've never used this before
16:18:48 <Eighth_Doctor> yeah I'll look into a symbolic icon for pagure then
16:18:59 <Eighth_Doctor> looks like it's a vendored fontawesome tree though
16:19:05 <Eighth_Doctor> so I'll talk to them
16:19:15 <dcavalca> cool, thanks
16:19:21 <dcavalca> oh, this should probably go in the minutes
16:19:31 <Eighth_Doctor> hehe
16:19:38 <dcavalca> #info user-facing docs up at https://sigs.centos.org/hyperscale
16:19:58 <dcavalca> I have a couple more quick things
16:20:12 <dcavalca> I've tagged a build of grep 3.7
16:20:24 <dcavalca> mostly because the 3.6 one we had before was still signed with SHA1
16:20:37 <dcavalca> #info grep 3.7 build is now available
16:20:51 <dcavalca> also, one of our engineers found a _nasty_ bug in jq
16:20:56 <dcavalca> https://bugzilla.redhat.com/show_bug.cgi?id=2008717
16:21:05 <dcavalca> this impacts all distros afaict
16:21:16 <dcavalca> it's fixed upstream, but upstream hasn't had a release since 2018
16:21:51 <dcavalca> I'm going to see if I can get this sorted out in a SIG build for the time being, but we'll see how that works out as the PR isn't pretty
16:23:13 <dcavalca> anything else for announcements?
16:23:23 <Eighth_Doctor> can it be backported in RHEL 9 and RHEL 8?
16:23:34 <dcavalca> Eighth_Doctor: I hope so, we've filed tickets for all of them
16:24:49 <dcavalca> let's move on to
16:24:53 <dcavalca> #topic Tickets
16:25:06 <Eighth_Doctor> are you going to try to send patches/MRs/PRs?
16:25:27 <dcavalca> Eighth_Doctor: yes, after this meeting, assuming the patch is actually backportable
16:26:15 <dcavalca> bkircher: any update on #67 for the qemu package?
16:26:33 <Eighth_Doctor> 👍️
16:27:00 <bkircher> no progress, sorry
16:27:35 <dcavalca> all good
16:27:51 <Eighth_Doctor> I branched and released PackageKit (#70)
16:27:52 <dcavalca> I think that was the only ticket we had pending? we already covered the sig report and the docs
16:28:16 <Eighth_Doctor> it carries a backported fix from upstream so that the CentOS repos actually work
16:28:31 <Eighth_Doctor> yup
16:28:53 <dcavalca> alright, we're doing good time today
16:28:56 <Eighth_Doctor> there's the powertools stuff
16:29:12 <dcavalca> yeah, I don't really have an update there, and michel is still on leave
16:29:35 <Eighth_Doctor> it seems like bstinson doesn't want to fix this for CentOS Stream 9 in a way that allows us to avoid a scriptlet to switch on the repo
16:30:11 <dcavalca> I remember we had a long discussion about this in one of the office hours but tbh I kinda forgot what the conclusion was
16:30:49 <Eighth_Doctor> that meeting conclusion was Brian was going to try to have stuff moved from CRB to AppStream
16:30:59 <Eighth_Doctor> I'm not very optimistic of them ever doing that, though
16:31:21 <Eighth_Doctor> they put it there so that RH Support can tell customers they won't support it
16:31:37 <dcavalca> if that's the goal, we should just file tickets every time we find a CRB dependency then to have it moved
16:31:45 <dcavalca> both for EPEL and for Hyperscale
16:32:14 <Eighth_Doctor> our two options are to either 1) use the scriptlet and force it on since bstinson rejected a repos subpackage to install/enable it or 2) shipping a copy of the powertools/crb defintions ourselves
16:32:29 <Eighth_Doctor> yeah, but that doesn't scale
16:32:38 <Eighth_Doctor> and it won't stop people from depending on them anyway
16:32:43 <dcavalca> between the two I'd rather take the scriptlet, as horrible as that is
16:32:54 <dcavalca> Eighth_Doctor: no, but at least it raises visibility
16:33:02 <dcavalca> I agree we're still going to need CRB for the forseeable future
16:33:02 <Eighth_Doctor> the problem is that CRB isn't actually just development packages
16:33:11 <Eighth_Doctor> it's a dumping ground for everything they can't get away with not shipping
16:34:10 <Eighth_Doctor> I've also considered shipping our own version of the centos-repos package
16:34:47 <dcavalca> uh I'd rather not do that
16:35:06 <Eighth_Doctor> I don't want to either
16:35:51 <Eighth_Doctor> but I also don't want to waste my time if I know that nobody is going to fix anything even if we file tickets
16:36:14 <Eighth_Doctor> the thread on -devel about CRB reinforced my view on this
16:36:22 <Eighth_Doctor> err epel-devel
16:36:38 <dcavalca> I'd say let's assume good intent for the time being and follow the process that was outlined
16:36:45 <Eighth_Doctor> https://lists.fedoraproject.org/archives/list/epel-devel@lists.fedoraproject.org/thread/6H3RVJFXZLIQVXGQLVBFAJ66QIMGIMTW/
16:36:52 <dcavalca> if that doesn't work out we can escalate as needed, but I want to at least make an effort to try it out
16:37:21 <Eighth_Doctor> 🤷‍♂️
16:37:23 <Eighth_Doctor> okay then
16:38:30 <Eighth_Doctor> okay
16:38:38 <Eighth_Doctor> does someone want to start that process?
16:38:54 <Eighth_Doctor> I don't have the time to go through and file tickets for all the KDE deps that exist only in PowerTools/CRB
16:39:05 <dcavalca> do we have a list of packages somewhere?
16:39:26 <dcavalca> this feels like something we can bring up in the next EPEL WG meeting and split the work between us
16:40:08 <dcavalca> conveniently, that's this afternoon
16:40:27 <Eighth_Doctor> I think Mike Rochefort made a list before
16:40:36 <Eighth_Doctor> omenos is I think his handle?
16:40:38 <mroche> It should be on centos-devel
16:40:56 <dcavalca> oh I remember that, yeah we should dig it up again
16:41:05 <dcavalca> that's definitely a good starting point
16:41:18 <Eighth_Doctor> semi-related, CS9 doesn't have libiscsi-devel in CRB, which means we can't build QEMU
16:41:21 <Eighth_Doctor> https://bugzilla.redhat.com/show_bug.cgi?id=2004762
16:41:37 <Eighth_Doctor> (I filed that bug for something else, but it'll be needed for qemu too)
16:41:47 <mroche> dcavalca: It's the "Dependent on PowerTools" thread
16:41:55 <dcavalca> thanks mroche
16:42:07 <Eighth_Doctor> ah so mroche :)
16:42:15 <mroche> https://lists.centos.org/pipermail/centos-devel/2021-July/077149.html
16:43:13 <mroche> Michel filed a bugzilla for this: https://bugzilla.redhat.com/show_bug.cgi?id=1983216
16:44:08 <Eighth_Doctor> that wasn't done: https://gitlab.com/redhat/centos-stream/rpms/centos-release/-/commit/0f4dbe586240f620206a9b3620455111d9da1f29
16:47:56 <dcavalca> probably also worth to bring this up again during the next office hours and/or at Dojo
16:48:22 <Eighth_Doctor> yup
16:49:06 <dcavalca> let's move to
16:49:09 <dcavalca> #topic Membership
16:49:18 <dcavalca> which I don't think we have anything for this week
16:50:05 <Eighth_Doctor> nope, don't think we do
16:50:16 <dcavalca> #topic Misc
16:50:28 <dcavalca> anything else folks wanna talk about in the last 10 min or so?
16:51:18 <mroche> Quick one on #7
16:51:36 <dcavalca> mroche: sure thing
16:52:17 <mroche> It seems like the suggestion is to rebase on the next RHEL kernel when that Stream launches. Is there still the idea to provide the BTRFS work directly into the stream kernel or has that sailed?
16:53:40 <mroche> (thinking of people who would want to rebuild the Stream kernel to flip on the BTRFS support but not move off that releases version)
16:53:48 <dcavalca> mroche: RHEL 9 / c9s isn't going to ship with btrfs if that's what you're asking
16:54:08 <dcavalca> but yes, the idea would be to try and maintain that support in their tree to make it easier to work with
16:54:21 <mroche> Yes, that's what I was referring to
16:54:22 <Eighth_Doctor> dcavalca: I believe he's asking about how we will ship btrfs enablement based on the c9s kernel
16:55:15 <mroche> Thinking of how those two concepts would interact with this: https://pagure.io/centos-sig-hyperscale/sig/issue/7#comment-745520
16:56:06 <mroche> As it seems that work would stagnate/stop on the previous RHEL kernel as Hyperscale moves forward
16:56:31 <Eighth_Doctor> the idea is we'll ship the c9s kernel on c8s too
16:56:33 <dcavalca> mroche: I think tbh the specifics of how this is going to work are still kinda up in the air
16:57:00 <dcavalca> but the general idea is that we'd keep tracking the CentOS Stream kernel as it gets developed, not necessarily the "fixed in time" rhel9 kernel
16:57:00 <Eighth_Doctor> we still need to build our first kernel based on the c9s kernel tree
16:57:41 <Eighth_Doctor> the advantage is that folks like the kmods sig people can take that tree and carve out stuff for kmod packages
16:58:28 <mroche> :thumbsup:
16:59:40 <mroche> Sorry if that question was a bit confusing, it'd be much easier to convey in person.
16:59:52 <Eighth_Doctor> this is where we're tracking backports we want to push into RHEL's kernel tree: https://pagure.io/centos-sig-hyperscale/sig/issue/72
17:00:03 <dcavalca> mroche: all good
17:00:17 <dcavalca> looks like we're out of time for this week
17:00:23 <dcavalca> thanks everyone
17:00:31 <dcavalca> #endmeeting