could we provide a similar maintenance/usage pattern for updating sysexts as what was described in t

could we provide a similar maintenance/usage pattern for updating sysexts as what was described in the video introducing logically bound images to bootc?

antheas•8/13/24, 6:21 PM

sysupd also uses squashfs images

j0rge•8/13/24, 6:21 PM

why is there conflict?

antheas•8/13/24, 6:21 PM

essentially the design of sysext is get your image with sysupd which is squashfs

antheas•8/13/24, 6:21 PM

then apply the sysexts which are also squashfs

Eethanjli could we provide a similar maintenance/usage pattern for updating sysexts as wha...

Noel•8/13/24, 6:21 PM

Yeah, this looked really interesting to me: 💾ublue-dev

antheas•8/13/24, 6:21 PM

the systems are designed for each other

antheas•8/13/24, 6:22 PM

e.g, you'd host both on an HTTP host

antheas•8/13/24, 6:22 PM

when using with ostree oci you get conflicting delivery methods for the base image and the extensions

antheas•8/13/24, 6:22 PM

and conflicting versioning

NNoel Is there a way to build tooling to utilize other existing tooling? Or is there n...

bketelsen•8/13/24, 6:22 PM

we can use sysupdate to download our extensions: https://github.com/flatcar/sysext-bakery see the bottom "consuming"

GitHub

GitHub - flatcar/sysext-bakery: Recipes for baking systemd-sysext i...

Recipes for baking systemd-sysext images. Contribute to flatcar/sysext-bakery development by creating an account on GitHub.

j0rge•8/13/24, 6:23 PM

yeah I don't get why people think we can't snag and update these things, the bakery has been doing it since day 1

bketelsen•8/13/24, 6:24 PM

we can do this, and we're marching slowly there

antheas•8/13/24, 6:24 PM

we'll get there

antheas•8/13/24, 6:24 PM

one important part of the puzzle was getting rechunk to work so that updates are manageable

antheas•8/13/24, 6:24 PM

now, slowly but surely extensions

Bbketelsen we can use sysupdate to download our extensions: https://github.com/flatcar/syse...

Noel•8/13/24, 6:26 PM

I guess what I mean is what @Kyle Gospo and I talked about earlier which is trying to use already existing packages and then converting them into system extensions transparently.

NNoel I guess what I mean is what @Kyle Gospo and I talked about earlier which is tryi...

bketelsen•8/13/24, 6:26 PM

ahh

antheas•8/13/24, 6:26 PM

thats also a possibility

Noel•8/13/24, 6:26 PM

Like, I don't need to create a custom formatted thing to work as a system extension, I already have an RPM package with all the information I need to get it installed on a Fedora system.

antheas•8/13/24, 6:26 PM

and a good idea

Noel•8/13/24, 6:26 PM

Then you don't need folks to repackage everything AGAIN

antheas•8/13/24, 6:27 PM

have a postprocessing hook that runs after rpmostree update and recreates the extensions in a parallel dir that matches the deployment

antheas•8/13/24, 6:27 PM

actually, probably a single extension

antheas•8/13/24, 6:27 PM

then does some selinux labelling

Noel•8/13/24, 6:27 PM

for sure. the other cool thing is if you had a tool like this, you could get it working for DEB packages as well.

Noel•8/13/24, 6:27 PM

RPM and DEB would be your 2 big ones you would want.

antheas•8/13/24, 6:28 PM

if i think about it the only problem with it is dependency management

Noel•8/13/24, 6:28 PM

Does Flatpaks serve a use case here or could you theoretically get rid of them now that you have systems extensions?

antheas•8/13/24, 6:28 PM

of course they do

antheas•8/13/24, 6:29 PM

if it can be a flatpak it should be

antheas•8/13/24, 6:29 PM

but since we're doing OCI

NNoel Does Flatpaks serve a use case here or could you theoretically get rid of them n...

ethanjliOP•8/13/24, 6:29 PM

system extensions don't do any sandboxing; the use-case is more for system files/programs

Eethanjli system extensions don't do any sandboxing; the use-case is more for system files...

Noel•8/13/24, 6:29 PM

That's a good point.

Noel•8/13/24, 6:29 PM

I was more thinking about how people complain about RPMs vs Flatpaks.

Aantheas and generating manifests with certain layers missing dynamically

antheas•8/13/24, 6:29 PM

im gonna be thinking about this

Noel•8/13/24, 6:30 PM

from a user experience perspective, I think the big thing is making whatever we do approachable.

antheas•8/13/24, 6:30 PM

solves every issue basically. Just need to lobotomize rpm-ostree to accept custom manifests

Aantheas if i think about it the only problem with it is dependency management

ethanjliOP•8/13/24, 6:30 PM

yeah, any dependency management/resolution would have to be implemented in the tool for downloading/managing local sysexts

Noel•8/13/24, 6:30 PM

whether that be plugging it into a store or making our own tool to install it for them.

Eethanjli yeah, any dependency management/resolution would have to be implemented in the t...

antheas•8/13/24, 6:31 PM

thats what rpm-ostree does already though

antheas•8/13/24, 6:31 PM

thats the whole point behind it, so you might as well use that

Bbketelsen we can do this, and we're marching slowly there

Kyle Gospo•8/13/24, 6:31 PM

Good addition to topgrade

antheas•8/13/24, 6:31 PM

if we are talking about rpms

antheas•8/13/24, 6:32 PM

generates commits for the deployed packages and merges them to the original ostree commit

bketelsen•8/13/24, 6:32 PM

don't forget this: https://blogs.gnome.org/chergert/2024/05/07/system-extensions-from-flatpak/

Happenings in GNOMEchergert

System Extensions from Flatpak

bketelsen•8/13/24, 6:34 PM

that's a significant chunk of the process we're working on here. Even introduces the idea of delivering a sysext via flatpak. Maybe that's not for us... maybe it is.

Bbketelsen that's a significant chunk of the process we're working on here. Even introduc...

antheas•8/13/24, 6:35 PM

article talks about GnomeOS

antheas•8/13/24, 6:35 PM

GnomeOS uses sysupd

Bbketelsen don't forget this: https://blogs.gnome.org/chergert/2024/05/07/system-extensions...

Noel•8/13/24, 6:36 PM

I did not know about this.

antheas•8/13/24, 6:36 PM

not only that, but the systemd dude pushed for ostree to be replaced for it

could we provide a similar maintenance/usage pattern for updating sysexts as what was described in t

Similar Threads

Similar Threads

Similar Threads