Kinoite, a KDE (and now XFCE) version of Fedora Silverblue

your totally missing the point, i followed your instructions, i created my repo and looking at yours
summary 28-Apr-2019 21:55 1838
summary.sig 28-Apr-2019 21:55 336

you show a summary file, however mine does not,
ls -al /srv/repo/
total 32
drwxr-xr-x 1 root root 4096 May 7 03:48 .
drwxr-xr-x 1 root root 4096 May 7 01:41 …
-rw-r–r-- 1 root root 38 May 7 01:42 config
drwxr-xr-x 2 root root 4096 May 7 01:42 extensions
-rw------- 1 root root 0 May 7 03:15 .lock
drwxr-xr-x 1 root root 4096 May 7 03:48 objects
drwxr-xr-x 1 root root 4096 May 7 01:42 refs
drwxr-xr-x 2 root root 4096 May 7 01:42 state
drwxr-xr-x 1 root root 4096 May 7 03:48 tmp

so whats created this summary file? i dont get one when i follow your guide.
sudo ostree remote refs myvariant
error: Remote refs not available; server has no summary file

and thats whats missing.

Indeed, I forgot to add a call to update the summary in the previous instructions. This should do it:

ostree summary --repo=... --update

I attempted installing the XFCE image on my Thinkpad, but it didn’t work. After the GRUB menu it booted to a blank screen. I had a mouse cursor and that was it. Thankfully I was able to roll back pretty easily. Any idea what could have gone wrong? Let me know if I can provide any additional information.

EDIT: I should mention I was able to do this in a virtual machine beforehand with no issues.
EDIT2: It also occurs to me I have a multi-monitor setup and I’ve had issues with DEs in the past not liking it upon first boot until after I set it up first. I’ll try again tonight with laptop undocked and see what happens and report back.

Try a journalctl -e -b-N, replacing N with the # of boots ago this happened (e.g. 1 is your previous boot, 2 is the boot before that, etc). Then, look for errors and see if anything sticks out.

Will do. Since I rolled back would those logs go away? Or are they persistent?

AFAIK not really, since they go in /var.

Ok, well if they aren’t there I’ll just rebase again and see if I can get to a TTY if undocking it doesn’t work. I’m at work right now but I’ll give all that a shot tonight.

UPDATE: So I was able to get it to work, it turns out something about the XFCE image (probably the display manager) doesn’t like my multi-monitor setup, and boots to a blank screen. When booting undocked it works fine, and I can dock it afterwards and set up the displays and it works fine, it just doesn’t like to boot up that way. Any ideas for fixing it? If not, I might just go back to the standard Silverblue image. I prefer XFCE to Gnome, but everything working with my monitors makes Gnome worth it at the present time.

Does this still get any updates? I decided to start using one monitor for now, and rebased to XFCE, but then I noticed that the version of Firefox included in the image is out of date. If it’s still being updated then I can be patient and wait for an upgrade, but if its not then I need to find another option.

I’m sorry, I am indeed a bit late with updates but I am still working on it. I will keep only the fedora 30 branch from now on until I have successfully automated everything. I will push an update soon.

4 Likes

All good man! Not trying to rush you or anything, I know this is all in your free time. Just curious if it was still being worked on at all. Good to hear it is! It’s the easiest way for a lot of us to get alternate desktops until official spins start popping up.

And here they come! I had issues with mirrors giving me 404 and the fedora-updates repo not being taken into account but I have finally resolved them (will make a PR soon). I have just pushed an update and it is a big one as it has all the updates since Fedora 30 came out! Enjoy!

As I said, I am only keeping the 30 branch for now with all variants. I fixed the versioning scheme so that it makes more sense and I will soon push a “base image” without any desktop environment for those that want a small image with a minimal DE overlayed on top (Sway, etc.).

I have pushed an update and a new ref named base (looking for a better name). This ref has no desktop environment installed by default so it is really small and lets anyone overlay a minimal DE with ease.

I have only done basic boot testing of this ref for now so it may be missing some essential packages otherwise available in other refs but that went missing in this one due to less indirect dependencies from the desktop environments.

Please report any issues you found. Thanks!

3 Likes

Where should I report an issue specifically with fedora-toolbox failing in Deepin?

$ toolbox enter 
/etc/os-release: line 22: 30.23: command not found
/etc/os-release: line 22: 30.23: command not found
toolbox: failed to start container fedora-toolbox-30

I suspect the first two errors are related to the formatting of the value ‘OSTREE_VERSION=Deepin 30.23’ in /etc/os-release. But I don’t know how that prevents the container from starting.

I get the same output while creating/entering toolbox but the commands work in the end. There is probably something else preventing the container from starting. Have you tried running a container directly with podman?

You may report the command not found errors upstream (toolbox) as this probably means there are quotes missing somewhere in the toolbox script.

I have found the origin of the command not found errors and they are mostly harmless (see this rpm-ostree PR). So there must be something else happening on your system.

1 Like

I tried your repo out in a VM this weekend, and it works great and looks awesome. Thanks for your efforts!

2 Likes

@siosm ,

Thanks for all this work!

From a fresh install of SB with layered Chrome, vim-X11 and a couple of other things, using your instructions for XFCE all goes without errors but I still get Gnome coming up:

# rpm-ostree status
State: idle
Warning: failed to finalize previous deployment
         error: Bootloader write config: grub2-mkconfig: Child process exited with code 1
         check `journalctl -b -1 -u ostree-finalize-staged.service`
AutomaticUpdates: disabled
Deployments:
● ostree://fedora:fedora/30/x86_64/silverblue
                   Version: 30.1.2 (2019-04-25T23:13:10Z)
                BaseCommit: 982faf58087c9d020780b829d5f24b4e78bd40399b4e4769fc8bc2df9890e301
              GPGSignature: Valid signature by F1D8EC98F241AAF20DF69420EF3C111FCFC659B9
           LayeredPackages: mc mpv quodlibet vim-X11
             LocalPackages: google-chrome-stable-76.0.3809.132-1.x86_64 rpmfusion-free-release-30-1.noarch rpmfusion-nonfree-release-30-1.noarch
                    Pinned: yes

  ostree://fedora:fedora/30/x86_64/silverblue
                   Version: 30.1.2 (2019-04-25T23:13:10Z)
                BaseCommit: 982faf58087c9d020780b829d5f24b4e78bd40399b4e4769fc8bc2df9890e301
              GPGSignature: Valid signature by F1D8EC98F241AAF20DF69420EF3C111FCFC659B9
           LayeredPackages: mc quodlibet vim-X11
             LocalPackages: google-chrome-stable-76.0.3809.132-1.x86_64 rpmfusion-free-release-30-1.noarch rpmfusion-nonfree-release-30-1.noarch

# journalctl -b -1 -u ostree-finalize-staged.service
-- Logs begin at Sun 2019-09-08 00:18:55 AEST, end at Sun 2019-09-08 19:25:30 AEST. --
Sep 08 19:20:50 192-168-1-20.tpgi.com.au systemd[1]: Started OSTree Finalize Staged Deployment.
Sep 08 19:21:52 192-168-1-20.tpgi.com.au systemd[1]: Stopping OSTree Finalize Staged Deployment...
Sep 08 19:21:52 192-168-1-20.tpgi.com.au ostree[14096]: Finalizing staged deployment
Sep 08 19:21:56 192-168-1-20.tpgi.com.au ostree[14096]: Copying /etc changes: 12 modified, 0 removed, 22 added
Sep 08 19:22:01 192-168-1-20.tpgi.com.au ostree[14096]: error: Bootloader write config: grub2-mkconfig: Child process exited with code 1
Sep 08 19:22:01 192-168-1-20.tpgi.com.au systemd[1]: ostree-finalize-staged.service: Control process exited, code=exited, status=1/FAILURE
Sep 08 19:22:01 192-168-1-20.tpgi.com.au systemd[1]: ostree-finalize-staged.service: Failed with result 'exit-code'.
Sep 08 19:22:01 192-168-1-20.tpgi.com.au systemd[1]: Stopped OSTree Finalize Staged Deployment.

I have seen this error with another test install but I thought it was something I messed up then - maybe I have messed up something here too? The only other thing that I can think of is that I deleted all occurrences of “rhgb quiet” from grub2 scripts but I can’t see how that would upset anything?

UPDATE: I just tried the same thing on my laptop and got the same result . .

Thanks,
Phil.

To get the XFCE desktop, you need to switch to the XFCE branch with:

sudo rpm-ostree rebase kinoite:fedora/30/x86_64/xfce

If you get any errors at this point, you may have to look closer at your system logs to find the issue. Sometimes I had troubles with CUPS auto-generated files in /etc (that should really be moved to /var) and cleaning them up, rebooting and retrying worked as a workaround.

@siosm ,

Damn . . ICAEFRTFM . .

Anyway - I reinstalled the stock SB from scratch and without changing ANYTHING this time ran your XFCE rebase - and it worked first time! - fantastic! Thank you so much!

I had a PulseAudio issue on the previous setup on my workstation (but not on my laptop) - I will see if it first works and then fails again now.

Ideally, I would like to create my own, minimal OS without PA and then add XFCE, ALSA etc to it but now at least, I am well on the way to really exploring SB, FP, toolbox etc.

Regards,
Phil.

@siosm ,

Should be:
git clone https://pagure.io/forks/siosm/workstation-ostree-config.git

I know most of the docs are aimed at creating base OSes to be installed from a server but I need to make more progress with building my own base OS before I start serving anything - is there step-by-step info somewhere about building a custom base OS on a client first? If not, I guess I will just have to do the server thing . .

Thanks.
Phil.