03:36 <anthr76[m]> njha: those are some legitimate concerns. I've had as well. Especially if you're preferring bare metal environments. I would keep an eye on ostree native containers. I've managed to migrate some of my home k8s nodes from Talos to FCOS. I'm finally seeing it as a viable bare metal K8s platform (OKD/OpenShift/... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/4d8ea40afe4c5dcbdf1151601e2994654b43dea2>)

10:02 <lucab> Not sure if the author of that PR lurks here, but I've left my design feedback on https://github.com/coreos/zincati/pull/904

10:16 <cverna> travier: FYI there seems to be already osbuildImage tasks in fedora infra --> https://koji.fedoraproject.org/koji/tasks?state=all&view=tree&method=osbuildImage&order=-id

14:51 <dustymabe> spresti[m]: do you mind looking at the most recent rawhide failure in the pipeline? it's got a `Found systemd generator failure (/usr/lib/systemd/system-generators/coreos-installer-generator)` error

14:57 <spresti[m]> Sure, I can take a look

15:00 <dustymabe> step 1 - see if you can reproduce locally.. That makes root causing an issue 90% easier if you can :)

15:01 <spresti[m]> Yeah 100%

16:08 <dustymabe> jlebon: I need some advice on selinux/fcos

16:09 <dustymabe> I'm seeing a denial for systemd-timesyncd on FCOS that I can't observe on traditional Fedora Cloud image (same versions of systemd and selinux-policy rpms)

16:10 <dustymabe> I initially saw the denial on a system that I upgraded so I thought there was some handling of an "updated" selinux policy that could have been to blame... but I just booted fresh latest `testing-devel` and I see the problem there too

16:11 <dustymabe> which makes me think there is something we are doing during our build/or boot that is causing the issue

16:33 <jlebon> dustymabe: not sure offhand. can you file a tracker issue with the denial?

16:34 <dustymabe> yeah - i'll file an issue, but in case you're interested in the actual denial - `sudo systemctl disable --now chronyd && sudo systemctl enable --now systemd-timesyncd` should show the problem

16:40 <jlebon> yup, i see it. weird, not sure why that'd be different on FCOS

16:43 <dustymabe> maybe some oddities related to the selinux "helper" work we did recently for https://github.com/coreos/fedora-coreos-tracker/issues/701

17:10 <cverna> Colin Walters: I have seen that you have started to look at how to push the ostree container images in fedora's registry. Is the base ostree container image built in the FCOS pipeline? and would you push the images from the pipeline?

17:11 <cverna> I have added a small comment here https://pagure.io/releng/issue/11047#comment-831585, but generally how we push the base image now is far from ideal :(