16:32:21 <jlebon> #startmeeting fedora_coreos_meeting
16:32:21 <zodbot> Meeting started Wed Mar 17 16:32:21 2021 UTC.
16:32:21 <zodbot> This meeting is logged and archived in a public location.
16:32:21 <zodbot> The chair is jlebon. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:32:21 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:32:21 <zodbot> The meeting name has been set to 'fedora_coreos_meeting'
16:32:27 <lorbus> .hello2
16:32:27 <jlebon> #topic roll call
16:32:27 <zodbot> lorbus: lorbus 'Christian Glombek' <cglombek@redhat.com>
16:32:29 <lucab> .hello2
16:32:31 <travier> .hello siosùm
16:32:32 <dustymabe> .hello2
16:32:33 <travier> .hello siosm
16:32:33 <zodbot> lucab: lucab 'Luca Bruno' <lucab@redhat.com>
16:32:36 <zodbot> travier: Sorry, but you don't exist
16:32:39 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dusty@dustymabe.com>
16:32:41 <cyberpear> .hello2
16:32:42 <zodbot> travier: siosm 'Timothée Ravier' <travier@redhat.com>
16:32:45 <zodbot> cyberpear: cyberpear 'James Cassell' <fedoraproject@cyberpear.com>
16:32:47 <jlebon> #chair lorbus travier dustymabe lucab cyberpear
16:32:47 <zodbot> Current chairs: cyberpear dustymabe jlebon lorbus lucab travier
16:33:20 <PanGoat> .hello jaimelm
16:33:21 <zodbot> PanGoat: jaimelm 'Jaime Magiera' <jaimelm@umich.edu>
16:33:23 <vrutkovs> .hello2
16:33:24 <zodbot> vrutkovs: vrutkovs 'Vadim Rutkovsky' <vrutkovs@redhat.com>
16:33:53 <jlebon> #chair PanGoat vrutkovs
16:33:53 <zodbot> Current chairs: PanGoat cyberpear dustymabe jlebon lorbus lucab travier vrutkovs
16:34:45 <jlebon> got a decent sized group :)
16:34:54 <travier> jlebon: Can we go podman, countme, then the others?
16:34:55 <jlebon> just waiting one more minute
16:35:11 <bgilbert> .hello2
16:35:12 <zodbot> bgilbert: bgilbert 'Benjamin Gilbert' <bgilbert@backtick.net>
16:35:15 <jlebon> travier: sure can
16:35:19 <jlebon> #chair bgilbert
16:35:19 <zodbot> Current chairs: PanGoat bgilbert cyberpear dustymabe jlebon lorbus lucab travier vrutkovs
16:35:40 <jlebon> ok cool, let's start!
16:35:48 <jlebon> #topic Action items from last meeting
16:35:53 <jlebon> lucab to refresh the existing nm-cloud-setup PR, dropping the auto-enable part
16:36:00 <jlebon> * lucab to track the nm-cloud-setup kola testing in a ticket and followup on that
16:36:07 <jlebon> * cverna to coordinate adding sumantro to release executors
16:36:11 <jlebon> * walters to write up a ticket for https://hackmd.io/GPB-x3XHToiYHdEnYNA1kA
16:36:16 <jlebon> * walters to go have lunch
16:36:28 <jlebon> #info walters presumably had lunch that day
16:36:41 <lucab> https://github.com/coreos/fedora-coreos-config/pull/760#issuecomment-800193352 refreshed
16:36:56 <jlebon> #info walters wrote up https://github.com/coreos/fedora-coreos-tracker/issues/764
16:37:09 <jlebon> #info lucab refreshed https://github.com/coreos/fedora-coreos-config/pull/760
16:37:14 <lucab> (the testing part not yet)
16:37:31 <jlebon> didn't hear anything about cverna's action item, so i'm assuming that's still pending
16:37:37 <jlebon> lucab: should i re-action?
16:38:06 <travier> I think the nm-cloud PR is ready
16:38:06 <lucab> jlebon: no need to, it's already noted in GH
16:38:21 <jlebon> ack
16:38:35 <jlebon> won't reaction cverna's item. i'll follow up with him after meeting
16:38:46 <jlebon> ok, moving on!
16:39:11 <jlebon> #topic Joint community meeting with Podman folks
16:39:15 <jlebon> #link https://github.com/coreos/fedora-coreos-tracker/issues/768
16:39:23 <jlebon> travier: want to take this one?
16:39:26 <travier> Sure
16:39:33 <lucab> jlebon: if you are happy with my reply, just merge the PR, or let me know and I'll change it
16:40:05 <jbrooks> .hello jasonbrooks
16:40:05 <travier> This is a proposal to replace next week meeting by a joint meeting with podman folks around the topics listed in the issue
16:40:05 <zodbot> jbrooks: jasonbrooks 'Jason Brooks' <jbrooks@redhat.com>
16:40:24 <PanGoat> same time or later in the day?
16:40:25 <travier> #proposal Replace next week meeting with a joint podman/fcos meeting
16:40:27 <darkmuggle> .hello2
16:40:29 <zodbot> darkmuggle: darkmuggle 'None' <me@muggle.dev>
16:40:29 <travier> Same time
16:40:38 <jlebon> #chair darkmuggle
16:40:38 <zodbot> Current chairs: PanGoat bgilbert cyberpear darkmuggle dustymabe jlebon lorbus lucab travier vrutkovs
16:40:39 <PanGoat> yes
16:41:04 <jlebon> to be clear, this would be open to anyone to assist, right?
16:41:18 <travier> Yes, will be open to anyone and recorder
16:41:20 <travier> recorded
16:41:44 <jlebon> the list of topics seems like it's worth more than 1h
16:41:50 <cyberpear> similar to the monthly Podman Community meeting?
16:42:04 <jlebon> maybe we should pare it down to 2 or 3 instead
16:42:42 <jlebon> focusing on the most pressing ones
16:42:55 <travier> I'm not really familiar with how podman community meeting are run
16:43:26 <travier> Hopefully some are rather short like the first one which is now pretty much decided
16:43:51 <travier> but of course if one hour isn't enough we can schedule others
16:44:05 <jlebon> proposal itself SGTM :)
16:44:19 <jlebon> anyone opposed or with alternative suggestions?
16:44:20 <lorbus> +1
16:44:54 <travier> I'm +1 (as expected :))
16:45:18 <jlebon> travier: did you want to chat about the priority of each proposed topics here?
16:45:41 <travier> We can if folks want to but wasn't particularly planning on doing that
16:45:57 <bgilbert> +1 to meeting
16:46:08 <travier> If someone want some topics more than the others then feel free to raise that in the tracker?
16:46:33 <jlebon> travier: yeah, i think we should have a clearer agenda prepared before the meeting, but we can do that offline
16:46:48 <jlebon> ok cool, sounds like there's agreement :)
16:47:06 <jlebon> #agreed we will replace next week meeting with a joint podman/FCOS meeting
16:47:14 <dustymabe> I do have a question about one of the topics
16:47:39 <dustymabe> it's possibly a rat hole, though
16:48:15 <dustymabe> not sure I should bring it up now or later
16:48:31 <bgilbert> dustymabe: if it's the moby-engine one, let's not prefetch the conversation but I think it's worth mentioning briefly here
16:48:42 <dustymabe> bgilbert: yes, it's that one
16:49:11 <bgilbert> AFAIK this is coming from the Podman folks, and I don't know what they plan to say
16:49:25 <lorbus> kill the whale
16:49:29 <PanGoat> heh
16:49:30 <jlebon> maybe for now we can agree that we won't decide to remove it just from the meeting?
16:49:34 <travier> Feel free to add any worries in the tracker
16:50:01 <dustymabe> I think bgilbert filled in the context I needed
16:50:10 <travier> I wasn't planning on making decisions in the meeting so much as moving the discussion forward
16:50:25 <jlebon> travier: +1
16:50:33 <jlebon> ok, ready to move on?
16:50:39 <travier> +1
16:50:51 <jlebon> #topic Enabling DNF Count Me support in Fedora CoreOS
16:50:54 <jlebon> #link https://github.com/coreos/fedora-coreos-tracker/issues/717
16:51:08 <jlebon> travier: :)
16:51:17 <travier> Count Me support is now part of the latest rpm-ostree release.
16:51:20 <travier> Fedora Silverblue and IoT will have that enabled for F34.
16:51:32 <travier> FCOS can have that enabled at almost any time
16:52:04 <travier> I will be working on the messaging before we enable that
16:52:23 <jlebon> hmm, we discussed this in a previous meeeting, but don't see a recorded outcome in the ticket
16:52:25 <travier> This will require manual action from people that do not want to be counted
16:52:59 <travier> I remember that there was mostly agreement on the way forward.
16:53:40 <travier> It's been a while and we are crossing another step so though it would be best to bring that up again just in case
16:53:44 <walters> .hello2
16:53:45 <zodbot> walters: walters 'Colin Walters' <walters@redhat.com>
16:53:55 <dustymabe> yep. thanks for bringing it up again.
16:54:04 <jlebon> #chair walters
16:54:04 <zodbot> Current chairs: PanGoat bgilbert cyberpear darkmuggle dustymabe jlebon lorbus lucab travier vrutkovs walters
16:54:07 <travier> If nothing has changed and everyone is still ok with the plan then we'll move forward
16:54:13 <dustymabe> travier: if you can find what we agreed upon last time I don't see any reason to modify that (should be able to find it in meeting logs)
16:54:33 <travier> We just need to figure out how and where to tell people that this is happening
16:54:48 <smooge> hello
16:55:16 <lorbus> I think it'd make sense to activate it around the same time Silverblue and IoT do it, too
16:55:37 <lorbus> wouldn't need FCOS-specific messaging then
16:56:01 <jlebon> lorbus: FCOS doesn't require manual intervention to upgrade though
16:56:02 <travier> #link previous discussion: https://meetbot-raw.fedoraproject.org/teams/fedora_coreos_meeting/fedora_coreos_meeting.2021-01-20-16.30.html
16:56:24 <lorbus> jlebon: good point
16:56:25 <jlebon> GA is in a few weeks, so we should send a heads up e.g. to the list really soon
16:56:33 <jlebon> with actions to take if they want it disabled
16:56:45 <jlebon> but agreed otherwise on just matching with the f34 rebase
16:56:47 <lorbus> don't activate for upgrades, only for new installs maybe?
16:57:26 <travier> lorbus: this is more complex as we need a barrier release but can be done
16:57:46 <jlebon> personally, i think it's ok to activate on upgrades, we just need to be loud about it
16:57:52 <dustymabe> agree ^
16:58:02 <walters> my main worry about that is we're not extensively testing old upgrades and accumulating state like that hits our testing matrix
16:58:16 <walters> by far the biggest one of these will be cgroupsv2
16:58:53 <walters> but it probably wouldn't be really hard to have a kola upgrade test variant
16:59:35 <jlebon> yeah, everytime we divert between upgrade and new install, it makes maintenance harder
16:59:44 <bgilbert> I guess I'm confused.  we already have a mechanism for opting out of statistics.
17:00:03 <bgilbert> couldn't we import the user's choice there into the new system?
17:00:38 <travier> that was discussed last time but no conclusion was reached
17:00:40 <jlebon> bgilbert: what's the mechanism? pinger enablement?
17:00:45 <bgilbert> yup
17:00:49 <bgilbert> or rather, pinger opt-out
17:01:32 <travier> https://meetbot-raw.fedoraproject.org/teams/fedora_coreos_meeting/fedora_coreos_meeting.2021-01-20-16.30.log.html#l-149
17:01:34 <bgilbert> .......which we only documented in the CL migration docs.  never mind
17:02:16 <jlebon> is the process for disabling countme on traditional to just do `countme=0` in the yum repo files?
17:02:45 <jlebon> would be nice to match that to be uniform, though it'd mark the files as untracked
17:02:57 <travier> yes unfortunately and as dustymabe noted last time this is problematic for us
17:03:06 <travier> https://coreos.github.io/rpm-ostree/countme/
17:03:35 <jlebon> travier: do you want to try a proposal or should i?
17:04:10 <travier> https://github.com/rpm-software-management/libdnf/issues/1068 > discusses adding something in libdnf to override that
17:04:17 <travier> jlebon: go ahead
17:05:28 <jlebon> #proposed we will enable countme support by default in upgrades and new installs as part of the f34 rebase. we will send an email to coreos a few weeks before with details about how to opt-out ahead of the rebase.
17:05:39 <travier> Maybe the disable option isn't really an option for us as we want to keep the repo configs updated thus we need something else in libdnf
17:05:53 <PanGoat> can we also post to the forum?
17:05:58 <PanGoat> lets hit everywhere
17:06:00 <travier> PanGoat: +1
17:06:06 <PanGoat> and twitter
17:06:39 <jlebon> agreed re. forum.  not sure about twitter
17:06:59 <dustymabe> last time we were talking about disabling a timer unit to effectively disalbe counting support
17:07:10 <PanGoat> Yeah, I understand most folks view that platform as news more than technical changees
17:07:23 <PanGoat> either way, defintely forum
17:07:31 <dustymabe> not sure how viable that option is
17:07:59 <travier> dustymabe: This will disable the new rpm-ostree logic but not the embedded libdnf one that is triggered when packages are layered
17:08:15 <dustymabe> ahh ok
17:08:17 <travier> Maybe I need to make the docs clearer
17:08:27 <bgilbert> I think I'd be in favor of coreos-status also
17:09:06 <jlebon> bgilbert: ughh, totally meant coreos-status actually.  do you think both makes sense?
17:09:20 <bgilbert> I agree with PanGoat, this should go everywhere
17:09:38 <dustymabe> everywhere++
17:10:26 <jlebon> ok, let me amend
17:10:59 <jlebon> #proposed we will enable countme support by default in upgrades and new installs as part of the f34 rebase. we will announce this on the mailing lists, twitter account, and forum a few weeks before with details about how to opt-out ahead of the rebase.
17:12:59 <travier> +1
17:13:12 <dustymabe> ack, fedmag article would be nice too if we can
17:13:20 <bgilbert> dustymabe: +1
17:13:23 <travier> I'll work on that
17:13:41 <PanGoat> yes
17:13:46 <jlebon> sgtm
17:13:47 <bgilbert> and we should get something into the F34 messaging
17:13:58 <bgilbert> jlebon: +1 on everything except upgrades, +0 on upgrades
17:14:18 <bgilbert> we're doing this a bit fast
17:15:07 <dustymabe> the actual f34 coming to stable won't happen for some time, right?
17:15:17 <dustymabe> can you elaborate on "a bit fast" ?
17:15:51 <jlebon> though actually...  `next` may move to f34 in two weeks (that's one of the topics after this one :) )
17:15:52 <travier> If it's too fast we can hold and do that in a more controlled fashion during the F34 cycle
17:16:03 <bgilbert> so I'm torn.  I think countme is pretty harmless from a privacy perspective
17:16:25 <travier> but it's potentially bad press if not explained
17:16:28 <bgilbert> but since it's a sensitive topic with no prior docs on opting out, I'd rather we left a looooong time for advance notice
17:16:36 <bgilbert> F35 would not be too long
17:16:37 <jlebon> right, we don't have to do it at rebase time. let's agree on... 6 weeks between announcement and turning it on?
17:16:46 <travier> agree
17:16:55 <travier> 2 months at least?
17:17:16 <jlebon> f35 feels excessive IMO
17:17:32 <dustymabe> yeah, that's a long time - I forget a lot in that amount of time :)
17:17:33 <bgilbert> it is excessive.  excessive could be good in this case
17:17:57 <bgilbert> 3 months might be a nice middle ground
17:18:10 <jlebon> i'm good with that
17:18:16 <PanGoat> I could buy into 3 months
17:18:19 <dustymabe> WFM
17:18:53 <jlebon> ok, let me do a final amendment
17:19:00 <PanGoat> final he says
17:19:11 <PanGoat> oh, wait, we forgot
17:19:17 <bgilbert> we left just under 2 months for the rootfs migration
17:19:17 <travier> +1
17:19:32 <dustymabe> so we'd just leave it off until then and then enable it globally (installs and upgrades?)
17:19:52 <jlebon> #proposed we will enable countme support by default in upgrades and new installs during the f34 cycle. we will announce this on the mailing lists, twitter account, forum, and fedmag at least 3 months before with details about how to opt-out ahead of the rebase.
17:20:04 <travier> +1
17:20:05 <PanGoat> jlebon: I'm joking of course. Thanks for all these iterations.
17:20:14 <PanGoat> yes
17:20:28 <dustymabe> +1
17:20:39 <jlebon> PanGoat: :)
17:20:45 <PanGoat> travier: you said you are working on messaging?
17:20:46 <bgilbert> +1.  can we explicitly add F34 release notes also?
17:20:58 <lorbus> +1
17:21:05 <travier> PanGoat: Yes, working on drafting something to post as an article, etc.
17:21:07 <jlebon> #agreed we will enable countme support by default in upgrades and new installs during the f34 cycle. we will announce this on the mailing lists, twitter account, forum, and fedmag at least 3 months before with details about how to opt-out ahead of the rebase.
17:21:16 <jlebon> bgilbert: i'll add it when posting to the issue
17:21:17 <travier> PanGoat: feel free to join if interested
17:21:19 <PanGoat> excellent
17:21:28 <PanGoat> might. time permitting.
17:21:36 <travier> Will post the draft in the tracking issue
17:21:59 <travier> thanks jlebon
17:22:28 <bgilbert> before we move on: has there been any discussion of removing the pinger?
17:22:58 <walters> yeah I was going to ask about that too, keeping it confuses things here I think
17:23:01 <travier> would be in favor given that we did not make progress there
17:23:16 <lucab> bgilbert: not that I know, but I think that after the countme goes live, we can talk about doing that
17:23:29 <bgilbert> it had one job (providing a configuration knob for future metrics) and it failed at that (we didn't document the knob)
17:23:57 <jlebon> isn't it intended to have more metrics than just how long the node's been alive though?
17:24:25 <bgilbert> sure.  I'm not aware that it's on anyone's roadmap though.  we could always re-add later.
17:24:29 <jlebon> if we remove it, putting it back in when we actually want it will be harder
17:24:33 <bgilbert> jlebon: why?
17:24:47 <jlebon> because it's confusing messaging
17:25:11 <bgilbert> hmm, how so?
17:25:18 <dustymabe> bgilbert: is the bigger question: "do we still want to implement pinger?"
17:25:18 * jlebon sees the clock.  maybe let's discuss that in the ticket?
17:25:33 <bgilbert> #action bgilbert to file a new ticket to discuss removing pinger
17:25:38 <bgilbert> let's keep them separate
17:25:42 <jlebon> +1 thanks
17:26:00 <jlebon> ready to move to open floor?
17:26:15 <jlebon> #topic Open Floor
17:26:25 <jlebon> whoops, meant to just paste that but not press enter yet :)
17:26:42 <bgilbert> we'd like to rename FCCT, since it's starting to be used by distros other than Fedora CoreOS.
17:26:47 <lucab> I have the next set of releases in two weeks but I won't around at that time, can anybody else swap the slots with me?
17:27:09 <jlebon> lucab: can you check with the person who's after you?
17:27:13 <bgilbert> if you have naming ideas, you can put them here: https://github.com/coreos/fcct/issues/167
17:27:36 <jlebon> please bring all your buckets of paint :)
17:28:09 <lucab> according to https://hackmd.io/WCA8XqAoRvafnja01JG_YA it's Sinny
17:28:09 <jlebon> (but seriously, i agree naming is important)
17:28:25 <bgilbert> :-)
17:28:31 <jlebon> lucab: let's chat with ksinny in #fedora-coreos
17:28:49 <travier> :D
17:29:07 <jlebon> wanted to bring up the f34 rebase, but we don't have much time
17:29:17 <jlebon> beta go/no-go meeting is tomorrow
17:29:33 <jlebon> my proposal is to rebase next in two weeks if we're go
17:29:56 <dustymabe> +1
17:30:04 <bgilbert> +1
17:30:20 <travier> +1
17:30:26 <lucab> jlebon: targeting the upcoming 'next'?
17:30:45 <jlebon> lucab: yeah
17:31:03 <jlebon> bgilbert: just cut releases, so we can put it in next-devel this week
17:31:12 <jlebon> s/://
17:31:13 <dustymabe> looks like I'll be back right when the fun starts
17:31:20 <travier> And for cgroupsv2, feel free to raise concerns in the ticket, we will discuss that next week in the meeting with podman folks and finialize the week after
17:31:27 <jlebon> dustymabe: yay! :)
17:31:28 <lucab> (sorry for the duplicate question, I had some lag)
17:32:06 <jlebon> anything else for open floor?
17:32:24 <PanGoat> do folks mind if I spam the forum with a link to the OKD event this weekend?
17:32:37 <travier> PanGoat: go ahead
17:33:11 <jlebon> PanGoat: +1
17:33:32 <PanGoat> thanks
17:34:20 <jlebon> ok, ending it in 37.5 seconds
17:34:21 <lorbus> #info OKD Testing and Deployment workshop this weekend: https://www.okd.io/blog/2021/03/16/save-the-date-okd-testing-deployment-workshop.html
17:34:22 <jlebon> #endmeeting