16:30:27 <lucab> #startmeeting fedora_coreos_meeting
16:30:27 <zodbot> Meeting started Wed Jul  1 16:30:27 2020 UTC.
16:30:27 <zodbot> This meeting is logged and archived in a public location.
16:30:27 <zodbot> The chair is lucab. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:30:27 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:30:27 <zodbot> The meeting name has been set to 'fedora_coreos_meeting'
16:30:37 <darkmuggle> .hello02
16:30:38 <lucab> #topic roll call
16:30:45 <darkmuggle> .hello2
16:30:46 <zodbot> darkmuggle: darkmuggle 'None' <me@muggle.dev>
16:31:05 <lucab> .hello2
16:31:06 <zodbot> lucab: lucab 'Luca Bruno' <lucab@redhat.com>
16:31:40 <lucab> #chair darkmuggle
16:31:40 <zodbot> Current chairs: darkmuggle lucab
16:32:11 <cyberpear> .hello2
16:32:12 <zodbot> cyberpear: cyberpear 'James Cassell' <fedoraproject@cyberpear.com>
16:33:41 <lucab> #chair cyberpear
16:33:41 <zodbot> Current chairs: cyberpear darkmuggle lucab
16:34:11 <lucab> I don't see a lot of movement today...
16:34:53 <lucab> I'll wait a few more minutes for people to show up, if any
16:35:14 <bgilbert> .hello2
16:35:14 <zodbot> bgilbert: bgilbert 'Benjamin Gilbert' <bgilbert@backtick.net>
16:35:25 <lucab> #chair bgilbert
16:35:25 <zodbot> Current chairs: bgilbert cyberpear darkmuggle lucab
16:36:46 <lucab> I'd say let's go
16:36:56 <lorbus> ,hello2
16:37:00 <lucab> #topic Action items from last meeting
16:37:03 <lorbus> .hello2
16:37:04 <zodbot> lorbus: lorbus 'Christian Glombek' <cglombek@redhat.com>
16:37:11 <lucab> #chair lorbus
16:37:11 <zodbot> Current chairs: bgilbert cyberpear darkmuggle lorbus lucab
16:37:42 <lucab> we didn't have any action items from last meeting
16:38:13 <lucab> so let's see the meeting tickets
16:38:42 <lucab> there's one from Dusty, https://github.com/coreos/fedora-coreos-tracker/issues/553
16:39:24 <lucab> but I think his IRC bouncer droppedf of the network
16:39:55 <lucab> anyone has more context on this? Or should we keep it for the next meeting?
16:40:12 <darkmuggle> probably best to punt till next meeting
16:40:32 <lucab> ack
16:41:02 <cyberpear> looks like a generic question of "what things should be configurable via kargs versus only ignition"
16:41:04 <lucab> we don't have other items for today. Perhaps we can go for a round of misc updates before open floor?
16:41:38 <lucab> cyberpear: happy to take that too
16:42:06 <lucab> #topic Customization via kargs vs Ignition
16:42:32 <lucab> cyberpear: any specific case to start from?
16:43:10 <cyberpear> the most common cases I come across are network-type configs and fips config
16:43:37 <cyberpear> fips=1 on the cmdline during install on Fedora/RHEL type hosts puts the installed system into FIPS mode
16:43:40 <lucab> I see
16:43:49 <lucab> so a bit of context
16:43:51 <cyberpear> net.ifnames=0 turns off the "predictable naming"
16:44:16 <lucab> we use Ignition to configure pretty much everything that runs in the real root filesystem
16:44:46 <lucab> however that leaves out a few cases
16:45:09 <lucab> those examples are cases from two categories:
16:45:38 <lucab> 1) kernel configuration which is applied once, on boot
16:46:03 <lucab> 2) runtime configuration for things that runs in initramfs, possibly before Ignition runs
16:46:49 <lucab> fips is in the first bucket, ifnames is in the second bucket
16:47:38 <lucab> for both, there isn't much that can be done via Ignition, as when it runs it is too late
16:48:04 <cyberpear> right, plus both of those things generally require at least a reboot to configure afterwards, IIUC, even if you had a way for ign to do it
16:49:09 <cyberpear> so I guess that pretty much answers it generally, I suppose
16:49:29 <cyberpear> the specific ticket referenced asks about persisting ifnames that have been explicitly requested by the kargs
16:49:44 <lucab> yes, I think we have a few ticket tracking this topic
16:50:07 <cyberpear> which sounds like something that perhaps should be added to persistent kargs, assuming the user wouldn't want different IFNAMES just for first boot
16:50:09 <lucab> I know of https://github.com/coreos/fcct/issues/57 and https://github.com/coreos/fedora-coreos-tracker/issues/292
16:52:01 <lucab> cyberpear: I think dusty's question was slightly different but I think that "just persistent the kargs" could be an option, yes
16:52:19 <cyberpear> for the cgroupsv2, I think there's been some discussion podman/runc/crun-side to enable cgroupv1 for containers that don't work w/ v2
16:52:52 <cyberpear> i.e., have the system use cgroupv2 and have it pass v2 to a container that requires the older version
16:53:03 <cyberpear> *pass v1 to a container requiring the older versino
16:54:11 <lucab> I don't have enough knowledge in that area
16:55:10 <lucab> cyberpear: any other question on this topic?
16:55:24 <cyberpear> nope, we can move on
16:55:40 <lucab> ack
16:55:50 <lucab> #topic Misc updates
16:56:43 <lucab> #link https://discussion.fedoraproject.org/t/change-to-network-interface-naming-on-new-installs/21307
16:57:27 <lucab> we are currently going through this, in order to fix a mistake we made and didn't realize for a bit
16:58:18 <lucab> we have already put in place the legacy-pinning script on all the streams
16:58:48 <lucab> and the switch is happening starting from testing & next
17:00:09 <lucab> #link https://github.com/coreos/zincati/blob/master/docs/usage/updates-strategy.md#periodic-strategy
17:00:31 <lucab> this also reached stable now (I need to add a bit more doc)
17:01:33 <lucab> #link https://github.com/coreos/fedora-coreos-docs/pull/94
17:01:47 <lucab> and this one too just hit stable
17:02:05 <lucab> and that's pretty much what I had on my plate recently
17:02:37 <lucab> not sure if other people have other things they want to highlight
17:04:19 <lucab> (I'll take it as a no)
17:04:26 <lucab> #topic Open Floor