The scheduled bluebuild job that came with the template, starting failing on the files module on July 1. It looks to be because of a missing directory. Is this a Module error or my configuration? Job Error [06:23:15 g.i/c/cit-os:latest] => #44 0.093 ============================= Start 'files' Module ============================= [06:23:15 g.i/c/cit-os:latest] => #44 0.103 /tmp/modules/files/files.sh: line 15: cd: /tmp/config/files: No such file or directory [06:23:15 g.i/c/cit-os:latest] => #44 0.105 ============================ Failed 'files' Module ============================ [06:23:15 g.i/c/cit-os:latest] => #44 ERROR: process "/bin/sh -c /tmp/scripts/run_module.sh 'files' '{\"type\":\"files\",\"files\":[{\"source\":\"LockScreenUi.qml\",\"destination\":\"/usr/share/plasma/shells/org.kde.plasma.desktop/contents/lockscreen/\"}]}'" did not complete successfully: exit code: 1 [06:23:15 g.i/c/cit-os:latest] => ------

Cut

[06:23:15 g.i/c/cit-os:latest] => 0.093 ============================= Start 'files' Module ============================= [06:23:15 g.i/c/cit-os:latest] => 0.103 /tmp/modules/files/files.sh: line 15: cd: /tmp/config/files: No such file or directory [06:23:15 g.i/c/cit-os:latest] => 0.105 ============================ Failed 'files' Module ============================ The Files part of my config ......... modules: - type: files files: # - source: PearDark/* #If needed to redeploy PearCursors # destination: /var/home/.icons - source: LockScreenUi.qml #ScreenLock test destination: /usr/share/plasma/shells/org.kde.plasma.desktop/contents/lockscreen/ - type: dnf .........

i dont know why it isnt working, ive put the correct flatpak name [16:41:20 g.i/v/kami-os:latest] => 0.830 ERROR: This system install flatpak ID 'org.zen_browser.zen' doesn't exist in FlatHub repo, please check if you typed it correctly in the recipe. https://cdn.discordapp.com/attachments/1390735106913730570/1390735107387818124/ct4KWR5.png?ex=68695680&is=68680500&hm=5a49d2dd2449d0b7bba0377b990761148a3abc8855c159d9bfcba1608fea42da&

[15:45:35 g.i/a/morrigan:latest] => 75.17 Transaction failed: Rpm transaction failed.
[15:45:35 g.i/a/morrigan:latest] => 75.17   - file /usr/lib64/libastal-io.so.0.1.0 conflicts between attempted installs of astal-io-libs-0^20250629.ac90f09-1.fc42.x86_64 and astal-io-0~6.git7f2292f-1.fc42.x86_64
[15:45:35 g.i/a/morrigan:latest] => 75.17   - file /usr/bin/astal conflicts between attempted installs of astal-0^20250629.ac90f09-1.fc42.x86_64 and astal-io-0~6.git7f2292f-1.fc42.x86_64
[15:45:35 g.i/a/morrigan:latest] => 75.19 External command had a non-zero exit code

here's the full run: https://github.com/AbooMinister25/morrigan/actions/runs/15977252924/job/45062863072 and that repo

recipes/dnf.yml

has the relevant dnf stuff. https://github.com/AbooMinister25/morrigan/tree/221abb4268c80d277708987c514b010aca8e894f

Haven't changed my build recipe in quite a while so I expect some part of what I have isn't necessary. I'm installing the following which I did after looking at a couple other people's builds:

- type: dnf
      repos:
          cleanup: true
          nonfree: rpmfusion
      install:
          packages:
              - nvidia-container-toolkit
              - libva-nvidia-driver
              - nvidia-vaapi-driver

But results in

file /usr/share/doc/libva-nvidia-driver/README.md from install of libva-nvidia-driver-0.0.13^20241210git9c4a3c7-2.fc42.i686 conflicts with file from package libva-nvidia-driver-0.0.14-1.fc42.x86_64

https://github.com/CarrotManMatt/carrot-kinoite-nvidia/actions/runs/15952381683/job/44999687384 https://github.com/CarrotManMatt/carrot-kinoite-nvidia/blob/main/recipes/recipe.yaml

[17:24:33  INFO] => Retrieving OS version from ghcr.io/ublue-os/aurora-dx-hwe:42
[17:24:36 DEBUG] => Successfully inspected image ghcr.io/ublue-os/aurora-dx-hwe:42!
[17:24:36  WARN] => Failed to parse version:
Failed to deserialize version latest-42.20250627.2
[17:24:36  WARN] => Unable to get version via image inspection due to error:
  × Failed to parse version from metadata for ghcr.io/ublue-os/aurora-dx-
  │ hwe:42
[17:24:36  WARN] => Pulling and running the image to retrieve the version. This will take a while...
[17:30:16 DEBUG] => Successfully inspected image ghcr.io/blue-build/cli/build-scripts:7828e7872b342d05302417db1bd94bf93d347383!
[17:30:16 DEBUG] => Using build scripts image: ghcr.io/blue-build/cli/build-scripts:7828e7872b342d05302417db1bd94bf93d347383

Has anyone been able to avoid this 6 minute delay? It seems like it wants to avoid it, but fails to find what it is looking for. How can I help it to find the version it needs? Also reported here: https://github.com/blue-build/github-action/issues/86

I feel like I'm missing a simple step for enabling the

dnf

module (migrating away from

rpm-ostree

). My builds are failing due to:

chmod: cannot access '/tmp/modules/dnf/dnf.sh': No such file or directory

Anyone seen this kind of error before? - See https://github.com/michaelhaaf/bluebuild-custom/actions/runs/15807967892/job/44555389767#step:2:3440 for the failed build logs - See https://github.com/michaelhaaf/bluebuild-custom/blob/main/recipes/recipe-combined.yml for the entire recipe file - See https://github.com/michaelhaaf/bluebuild-custom/blob/main/recipes/common/packages.yml for an example dnf module

Dears, In a container file, I run a command to set the /etc/os-release file. I can do this with the files or script module of blue-build. If I want to set the VERSION to the currently build tag of blue build, I guess I have to use the script module. Can you tell me how I can access in such scripts the version tag? This is how the file looks like for fedora kinoite:

NAME="Fedora Linux"
VERSION="42.20250429.1 (Kinoite)"
RELEASE_TYPE=stable
ID=fedora
VERSION_ID=42
VERSION_CODENAME=""
PLATFORM_ID="platform:f42"
PRETTY_NAME="Fedora Linux 42.20250429.1 (Kinoite)"
...

I am having trouble booting my Aurora-DX: stable but it I cannot seem to load stable-20250608.1 or any future or past version. I have tried Latest, 42-20250606, and even on that is matching the current working one 42-20250601.1. Is anyone else having or had an issue booting? Here is a link to my image. Can someone please take a look and point me in the right direction or what to look for? Thanks. https://github.com/CITWil/wil-os/blob/main/recipes/recipe-work.yml

i have no idea if this is the right place to ask this type of question but after searching and seeing a lot of people asking the same thing on various different forums but all either have literally no solutions or are a different configuration and since i am using a bluebuild image, might as well ask here. i think yesterday i may have physically mishandled my laptop while it was still shutting down and may have made it crash mid shutdown process, and i woke up to trying to turn it on and being greeted to emergency boot. it took a whole morning and some bits of an afternoon to fix but i did manage to get booted again by directly modifying a grub entry from a live environment (the issue itself is that, i believe rpm-ostree or something was probably doing some background processing stuff while shutting down - i may have somehow managed to interrupt it in between that point, so the system got upgrade but none of the configurations did, so everything is trying to point and use directories such as

/ostree/boot.0/fedora/xxx

when they don't exist while

/ostree/boot.1/fedora/yyy

do) which thankfully got me booted, but sadly all rpm-ostree commands don't work at all. they all error out with

error: Loading sysroot: Failed to invoke RegisterClient: GDBus.Error:org.freedesktop.DBus.Error.NameHasNoOwner: Could not activate remote peer 'org.projectatomic.rpmostree1': startup job failed

, and poking at

systemctl status rpm-ostreed

reveals:

Reading config file '/etc/rpm-ostreed.conf'
error: Couldn't start daemon: Error setting up sysroot: loading sysroot: Parsing deployment /ostree/boot.0/fedora/625c979aaf057ab3b57b77c33c4157339dfd1cb4dd3c23734473ffb44337e724/0 in stateroot 'fedora': readlinkat: No such file or directory

[cont.]

So, GitHub actions is reporting that apparently from my [recipe file](https://github.com/cyrneko/ublue-kinoite-custom/blob/main/recipes/recipe.yml), the first column is invalid

× Recipe ./recipes/recipe.yml failed to validate
  ╰─▶   × Failed to deserialize recipe ./recipes/recipe.yml
        ╰─▶ found character that cannot start any token at line 10 column 1,

the first column is... a regular tab character this doesn't happen on my other image, [ublue-niri]() where instead of tabs I use spaces ...what's going on here?

how can i install bazzite kernel (and other kernels) in my image?

I have a home server and I would like to run my builds on that, I setup a proxmox lxc container, installed docker, and installed github action runner onto it and have it connected to my library. Keep getting errors during build, has anyone else got this setup and have any suggestions?

Hello, i've been running into this issue and its very annoying: https://github.com/NVIDIA/open-gpu-kernel-modules/issues/739 . I think I'm using the open kernel modules because

modinfo -l nvidia

says the license is Dual MIT/GPL. I wanted to switch to the closed source drivers, but I realized that the ublue-os/hwe repo I was using for my base was archived https://github.com/ublue-os/hwe. Is there a recommended way to install the closed source nvidia drivers? I noticed the thread about an nvidia installer module but thats not ready yet. Also, what should i use as my base image now that

ghcr.io/ublue-os/silverblue-nvidia

is archived?

Hi, I just wanted to restart my experiments with BlueBuild, today. One month ago, it worked: https://github.com/SteVio89/stevio-os/tree/b0a297007afc3d2ef1c6b03bf405aefb9e5a0f62, but I had some other issues and stopped it. Today I decided to give it another try. Locally, I can build the image, but it did not work with the GH-Action workflow. Always getting the same error: https://github.com/SteVio89/stevio-os/actions/runs/15099470535/job/42438517867 First, I thought it was because I now used the dnf module instead rpm-ostree to install packages. So I switched back to using the rpm-ostree variant I had 1 month ago, it also failed: https://github.com/SteVio89/stevio-os/actions/runs/15097940093/job/42435113436 It only works if I when I'm using eighter the dnf nor the rpm-ostree module (so just using the base-image): https://github.com/SteVio89/stevio-os/tree/153e19e77fa354a3d9f10923a67056825053e3e9

When I'm trying to build an image that uses the new dnf module's

replace

feature, I get the following result from my CI build.

[21:34:17 g.i/a/morrigan:latest] => 68.21 Packages for argument 'hyprpaper' available, but not installed.
[21:34:17 g.i/a/morrigan:latest] => 68.21 Packages for argument 'hyprlock' available, but not installed.
[21:34:17 g.i/a/morrigan:latest] => 68.22 External command had a non-zero exit code
BLAH BLAH BLAH
[21:34:17 g.i/a/morrigan:latest] =>  121 |     RUN \
[21:34:17 g.i/a/morrigan:latest] => --------------------
[21:34:17 g.i/a/morrigan:latest] => ERROR: failed to solve: process "/bin/sh -c /tmp/scripts/run_module.sh 'dnf' '{\"type\":\"dnf\",\"repos\":{\"cleanup\":true,\"files\":[\"vscode.repo\"],\"copr\":[\"solopasha/hyprland\",\"wezfurlong/wezterm-nightly\"],\"nonfree\":\"rpmfusion\"},\"install\":{\"packages\":[\"fish\",\"xdg-desktop-portal-gtk\",\"wf-recorder\",\"cmake\",\"meson\",\"gcc-c++\",\"hyprland-devel\",{\"repo\":\"code\",\"packages\":[\"code\"]},\"hyprland-qt-support\",\"hyprpolkitagent\",\"satty\",\"aylurs-gtk-shell2\",\"cliphist\",\"astal\",\"astal-gjs\",\"astal-gtk4\",\"hyprpanel\",{\"repo\":\"copr:copr.fedorainfracloud.org:wezfurlong:wezterm-nightly\",\"packages\":[\"wezterm\"]}]},\"replace\":[{\"from-repo\":\"copr:copr.fedorainfracloud.org:solopasha:hyprland\",\"packages\":[\"hyprland\",\"hyprutils\",\"xdg-desktop-portal-hyprland\",\"hyprpaper\",\"hyprlock\",\"hypridle\",\"hyprgraphics\",\"aquamarine\",\"hyprland\",\"hyprpicker\",\"hyprsunset\"]}],\"remove\":{\"packages\":[\"firefox\",\"firefox-langpacks\",\"waybar\",\"dunst\",\"swaylock\",\"swayidle\",\"swaybg\"]}}'" did not complete successfully: exit code: 1

I'm not really sure why. This is the part of the recipe in question: https://github.com/AbooMinister25/morrigan/blob/main/recipes/dnf.yml And the CI run: https://github.com/AbooMinister25/morrigan/actions/runs/15031412187

Hey everyone, i added the intel vtune repo and installed the package, but its not anywhere on my machine. when i do

rpm -qa | grep intel-oneapi-vtune

, it shows up, and when i try to do

rpm-ostree install intel-oneapi-vtune

, it says

error: "intel-oneapi-vtune" is already provided by: intel-oneapi-vtune-2025.3.0-15.x86_64. Use --allow-inactive to explicitly require it.

But when I try to uninstall it, i get

error: Package/capability 'intel-oneapi-vtune' is not currently requested

. anyone know how to fix this?

Scenario is as follows, I have a simple build-script type situation going on for building something with

cargo

, then installing it and cleaning up build deps I'd like to make it more re-usable, so instead of hard-coding the git repo URL I wanted to take it in as

$argv[1]

(in fish) But I am unsure, if I do something like:

yml
- type: script
    scripts:
      - build.fish "argument"

will the

argument

be passed to the script? If not, I can deal with that

I downloaded the blue95 installer and loved it, wanted to try the blue7 module. I ran the

just

build for blue7 and it ran

bluebuild switch

-- the script said it was installing the new image but when I rebooted, the old image came back up. How might I go about diagnosing why it failed?

Hi there, I've recently run into some issues with the nvidia drivers and realised that because I'm using Ublue's base images I need to install the drivers myself like it's done for Bluefin and other images. My question is: now that I'm working on it, would people find it useful as a module? Can't promise immediate results, but will try to do my best 🥲 Cheers

I made some adjustments to my image recipe, such as using dnf instead of rpm-ostree, removing and adding some packages and enabling the negativa17 repo, but nothing, beyond that, that would cause boot problems. I'm suspicious of an option that is this:

yaml
install:
  install-weak-deps: false # Whether to install weak/recommended dependencies.

Here is my repo for anyone who wants to investigate the commits and files: https://github.com/Heus-Sueh/hyprland

while i found the tools for building - i'm wondering if there are any tutorials on doing automated tests for the intended behavior of a image and/or its attended/unattended installation i want to create images for my systems that are accompanied by code that shows they work

install:
  skip-unavailable: true # skip unavailable packages
  install-weak-deps: false # Whether to install weak/recommended dependencies.
  packages:
    # - repo: copr:copr.fedorainfracloud.org:custom-user:custom-repo
    - repo: copr:copr.fedorainfracloud.org:solopasha:hyprland
      packages: 
        # Window Manager
        - hyprland
        - hyprlang
        - hyprcursor
        - hyprpaper
        - hypridle
        - hyprlock
        - xdg-desktop-portal-hyprland

logs: https://privatebin.net/?0e100fa6f39bdece#DuQJDwX4umqnMiyTDjZ6hDDJTMyCsazTPqHV6cP9SAVr

I generated an ISO from a remote image and installed it onto a VM. I'm running against the below issue. Any idea how I can sort this out?

student@fedora:~$ rpm-ostree status
State: idle
AutomaticUpdates: stage; rpm-ostreed-automatic.timer: no runs since boot
Deployments:
● ostree-unverified-image:oci:/run/install/repo/joeri-os-latest
                   Digest: sha256:a40c9d665b55d0ad6bbd59d8f1f2074e203bccf8a3615606a8ea97c363f8d622
                  Version: 42.20250429.0 (2025-04-30T06:30:36Z)

student@fedora:~$ rpm-ostree update
note: automatic updates (stage) are enabled
Pulling manifest: ostree-unverified-image:oci:/run/install/repo/joeri-os-latest
error: Creating importer: failed to invoke method OpenImage: failed to invoke method OpenImage: lstat /run/install: no such file or directory
student@fedora:~$

AFAIK the remote image location is wrong:

oci:/run/install/repo/joeri-os-latest

😕

I've built my image with Podman. How can I generate the ISO from the local image?

bluebuild generate-iso

seems to use the Docker engine.

Is it possible to build two images in the same repo? (one from bluefin and one from bazzite)? Thank you

Hello everyone, I am wanting to build my image locally instead of on the cloud like github. What would be the best way to go about this? I do not have a separate computer for me to use, only the one that is running discord right now

I want to enable transient root, to install nix (since I updated to fedora 42 with composefs) To do that, I need to edit the file at /usr/lib/ostree/prepare-root.conf and add a configuration entry. Do I need to do it "manually" as a script step, or is there some better way to do that ?

Hello, I am trying to build an ISO for an image (that is not generated with Bluebuild, but does have a working container image). I was having problems with bootc-image-builder which led me here. However, I am getting "error: pathspec 'tags/' did not match any file(s) known to git" when I run the iso build. Any ideas?

cameron@orton-10-219-200-45:~$ sudo ~/.cargo/bin/bluebuild generate-iso --iso-name stillos.iso image registry.gitlab.com/stillhq/stillos/bootc:0
INFO  => Retrieving OS version from registry.gitlab.com/stillhq/stillos/bootc:0
g.i/j/build-container-installer => + mknod -m 0660 /dev/loop0 b 7 0
g.i/j/build-container-installer => + true
g.i/j/build-container-installer => + [[ -d /cache/skopeo ]]
g.i/j/build-container-installer => + [[ ! -d /cache/dnf ]]
g.i/j/build-container-installer => + make VARIANT=Kinoite ISO_NAME=build/stillos.iso DNF_CACHE=/cache/dnf SECURE_BOOT_KEY_URL=https://github.com/ublue-os/bazzite/raw/main/secure_boot.der ENROLLMENT_PASSWORD=universalblue IMAGE_NAME=bootc IMAGE_REPO=registry.gitlab.com/stillhq/stillos IMAGE_TAG=0 VERSION=0
g.i/j/build-container-installer => make -w -C external lorax/branch-0
g.i/j/build-container-installer => make[1]: Entering directory '/build-container-installer/external'
g.i/j/build-container-installer => git config advice.detachedHead false
g.i/j/build-container-installer => cd lorax && git reset --hard HEAD && git checkout tags/
g.i/j/build-container-installer => HEAD is now at ffba3078 Add prefixdevname support to the boot.iso
g.i/j/build-container-installer => error: pathspec 'tags/' did not match any file(s) known to git
g.i/j/build-container-installer => make[1]: *** [Makefile:3: lorax/branch-0] Error 1
g.i/j/build-container-installer => make[1]: Leaving directory '/build-container-installer/external'
g.i/j/build-container-installer => make: *** [Makefile:148: external/lorax/branch-0] Error 2
ERROR => Failed:
  × Failed to create ISO

message has been deleted

When running

bluebuild build ./recipes/recipe.yml

locally, I get the following error:

l/morrigan:latest => 0.075 /tmp/scripts/exports.sh: line 48: awk: command not found
l/morrigan:latest => 0.078 /tmp/scripts/run_module.sh: line 55: jq: command not found
l/morrigan:latest => ------
l/morrigan:latest => Containerfile:49
l/morrigan:latest => --------------------
l/morrigan:latest =>   48 |     # Module RUNs
l/morrigan:latest =>   49 | >>> RUN \
l/morrigan:latest =>   50 | >>> --mount=type=bind,from=stage-files,src=/files,dst=/tmp/files,rw \
l/morrigan:latest =>   51 | >>> --mount=type=bind,from=ghcr.io/blue-build/modules/script:latest,src=/modules,dst=/tmp/modules,rw \
l/morrigan:latest =>   52 | >>> --mount=type=bind,from=ghcr.io/blue-build/cli/build-scripts:v0.9.10,src=/scripts/,dst=/tmp/scripts/ \
l/morrigan:latest =>   53 | >>>   /tmp/scripts/run_module.sh 'script' '{"type":"script","scripts":["installdock.sh"]}'
l/morrigan:latest =>   54 |
l/morrigan:latest => --------------------
l/morrigan:latest => ERROR: failed to solve: process "bash -c /tmp/scripts/run_module.sh 'script' '{\"type\":\"script\",\"scripts\":[\"installdock.sh\"]}'" did not complete successfully: exit code: 127

My CI build fails for the same reason: https://github.com/AbooMinister25/morrigan/actions/runs/14477719246/job/40607276868. It had just worked a few moments ago when I ran the build, but when I try to run the build again (without anything changing), it seems to have broken. Any idea why?