<dustymabe>
I honestly think this would cut down on some "derivative projects" need to build "bootimages" for their offering
<jlebon>
if we do one, we could probably do the other since both involve calling rpm-ostree from the initrd
<dustymabe>
exactly
<jlebon>
dustymabe: did you see my msg higher up about restamp ?
<dustymabe>
jlebon: nope
<walters>
without a mechanism to apply configuration to the initramfs (repo files, TLS certificates, mirroring) that can't work
<dustymabe>
jlebon: my bouncer went down (new FCOS update - woohoo!)
<dustymabe>
let me check the logs
<dustymabe>
ok I see it now
<jlebon>
walters: i was thinking we'd chroot, not run in the the initrd userspace
<dustymabe>
jlebon: stamped
<jlebon>
we do that for a few things today
<jlebon>
dustymabe: thanks!
<dustymabe>
walters: thanks for the `--oscontainer` link - i might try to play around with it
ravanelli has joined #fedora-coreos
ravanelli has quit [Ping timeout: 268 seconds]
ravanelli has joined #fedora-coreos
aaradhak[m] has joined #fedora-coreos
jcajka has quit [Quit: Leaving]
<dustymabe>
hey aaradhak[m] - according to https://bugzilla.redhat.com/show_bug.cgi?id=2080043#c19 the issue is at least fixed in rawhide - so we can validate the test you wrote (on testing-devel the test should fail and on rawhide it should pass)
<aaradhak[m]>
HI dustymabe : I will need to work on writing the test case for checking the existence of /etc/issue.d/22_clhm_ens6.issue. If it is required to be verified manually in the meantime, I shall look into it.
<dustymabe>
guesswhat: hmm. I'm not sure that information is available at the time Ignition runs
jpn has quit [Ping timeout: 260 seconds]
<bgilbert>
guesswhat: you can use Ignition to write out systemd units that use the metadata attributes
<bgilbert>
but Ignition doesn't itself know about the attributes
<jlebon>
bgilbert: could we detect a previously installed CoreOS on firstboot with extra partitions and error out if it'll get wiped? that would break another set of reprovisioning users, but at least we err on the side of data loss prevention.
<jlebon>
s/it'll/the extra partitions will/
<bgilbert>
on firstboot? if we're in the reprovisioning case, coreos-installer has already wiped the partition table by then
<guesswhat>
bgilbert: write out systemd unit ? do we have any example? thanks
<bgilbert>
guesswhat: well, that's exciting. looks like we never documented this in FCOS docs
<jlebon>
bgilbert: was thinking about live env, but right yeah. coreos-installer would have to also purposely break for this where it didn't before
<bgilbert>
jlebon: yup
jpn has joined #fedora-coreos
<jlebon>
i guess one hack is to put the old partition table in a file in /boot and the first boot code analyzes that too
<dustymabe>
bgilbert: we never documented "write out systemd unit" example ?
<jlebon>
anyway, just brainstorming here. i feel like we should be able to find a path forward where we do require user changes without risking data loss.
<bgilbert>
I'm in favor of that if we can think of a path, but there are some fundamental constraints here
<dustymabe>
bgilbert: jlebon: I think this is what I was alluding to when I said it would be useful to enumerate the cases and identify ones where data loss could occur during a reprovision
<bgilbert>
the whole point of a user data partition is that we shouldn't decide later that it's actually ours
ravanelli has quit [Ping timeout: 268 seconds]
<bgilbert>
dustymabe: I'm open to that, sure
<dustymabe>
i.e. we can talk through the specific contstraints in each scenario and thus maybe come up with a possible solution that accounts for them
<guesswhat>
bgilbert: oh, afterburn is disabled by default, /run/metadata/afterburn env file is okish for me :) thanks
aaradhak has quit [Quit: Connection closed for inactivity]
<jlebon>
merged! thanks MichaelArmijo[m] for doing the releases!
<MichaelArmijo[m]>
jlebon: Of course! Thanks.
<lorbus>
Hey team, is there an ETA for the next rpm-ostree release cut? Having support for `ex module` in a container is needed for our story to migrate OKD/FCOS over to layered images.
<dustymabe>
not that I know of
<jlebon>
yeah, i'd say we're overdue at this point. i think walters may do one soon for other layering-related things.
nalind has quit [Quit: bye]
ravanelli has quit [Remote host closed the connection]
<lorbus>
thanks, that sounds promising :)
jpn has quit [Quit: Lost terminal]
ravanelli has joined #fedora-coreos
ravanelli has quit [Ping timeout: 260 seconds]
c4rt0 has quit [Ping timeout: 260 seconds]
crobinso has quit [Remote host closed the connection]