| Age | Commit message (Collapse) | Author |
|---|
|
This ensures that there's empty space for arm-image-installer to
inject firmware blobs into images as part of preparing to install
for various single board computers.
|
|
nss-mdns is not included in RHEL, and mesa-vulkan-drivers is x86_64 only.
|
|
Signed-off-by: Troy Dawson <tdawson@redhat.com>
|
|
WSL does not support SELinux, so at best it's pointless to install. This
also actively breaks WSL 1 users, although I don't think we should
suggest that we support that, and if the policy on the builder doesn't
match the target things break during the build.
It also reduced the build time from ~80 seconds to ~70 seconds for me,
so that's nice.
See: 39b379f7d ("wsl: drop the SELinux packages from the WSL base image")
Signed-off-by: Jeremy Cline <jeremycline@linux.microsoft.com>
|
|
WSL does not support SELinux, so at best it's pointless to install. This
also actively breaks WSL 1 users, although I don't think we should
suggest that we support that, and if the policy on the builder doesn't
match the target things break during the build.
It also reduced the build time from ~80 seconds to ~70 seconds for me,
so that's nice.
Signed-off-by: Jeremy Cline <jeremycline@linux.microsoft.com>
|
|
shim-ia32 no longer exists at present. We can talk with pjones
about bringing it back, but for now, this is preventing Rawhide
composes from working.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
|
|
|
|
The anaconda-webui package is now installed by default for live
media, so these are no longer necessary.
|
|
The package called 'anaconda' is empty and only depends on two
things, anaconda-gui and anaconda-tui. Since anaconda 43.32,
anaconda runs the GTK UI if it's installed; it will only run the
web UI if anaconda-gui is not installed. So by including anaconda
we are forcing lives to run GTK UI and not really gaining
anything - we don't want anaconda-tui on lives anyway, I don't
think.
The anaconda-live package will pull in anaconda-webui, so all we
need to do is stop including anaconda. This will cause all live
images to use web UI, which was the intent of
https://github.com/rhinstaller/anaconda/pull/6530 . I guess if
any lives really want/need to use GTK UI, they can put
anaconda-gui into their profiles.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
|
|
|
|
This restores support for booting on 64-bit x86 systems that use
32-bit EFI platforms.
Reference: https://fedoraproject.org/wiki/Changes/32BitUefiSupport
|
|
This reverts commit 24c4fc39c2995e12610173acfdce689497a84120.
With a PR we can run the tests on openQA staging, identify any
problems with the change, and extend the tests to cover the
additional screens that show up with KDE.
|
|
Signed-off-by: Troy Dawson <tdawson@redhat.com>
|
|
This reverts commit e4d801674c86b0675f0ff6fed8c8fee0e5e47546.
openQA has not yet been prepared for this change and it is
causing all updates to fail gating. We need to prepare the tests
before merging this.
|
|
dedupe
Per erofs upstream, this results in faster and more efficient build times
with improved runtime performance.
Signed-off-by: Neal Gompa <ngompa@fedoraproject.org>
|
|
While the Web UI is still missing two features outlined in the Change Proposal [1], enabling it
early will help identify additional issues and improve integration. The missing features are:
- Date/Time Configuration Screen (currently implemented by GIS in Workstation)
- Configurable Browser Support (currently hardcoded to Firefox)
[1] https://fedoraproject.org/wiki/Changes/AnacondaWebUIforFedoraSpins#Detailed_Description
|
|
Fixes https://github.com/fedora-cloud/docker-brew-fedora/issues/119
Signed-off-by: Clement Verna <cverna@tutatnota.com>
|
|
The sshfs sharing backend doesn't work in containers where the default
one works just fine. To make the --isolated builds more accessible just
use the default.
The default in kiwi is currently: --9p-sharing
|
|
Signed-off-by: Troy Dawson <tdawson@redhat.com>
|
|
The serial device name on Arm is ttyAMA, not ttyS. This splits the
configuration for ELN images based on the architecture and adjusts the
kernel command line.
Note that prior to this change, aarch64 would still output kernel logs
to the ttyAMA0 console (on Azure, at least) due to the ACPI SPCR table.
On aarch64 the console in this table, if it's present, is _always_
configured as the kernel's default console. However, systemd logs would
not be sent to the serial console because it seems it was honoring the
kernel command line configuration which had ttyS0 as the default.
Note: this is the exact same change as 4979662a547e ("Cloud: Adjust
aarch64 images to use the correct serial console"), if you're reverting
this, you probably want to also revert that one.
Also see: https://pagure.io/cloud-sig/issue/394
Also see: https://bugzilla.redhat.com/show_bug.cgi?id=1661288
Signed-off-by: Jeremy Cline <jeremycline@linux.microsoft.com>
|
|
The serial device name on Arm is ttyAMA, not ttyS. This splits the
configuration for Cloud images based on the architecture and adjusts the
kernel command line.
Note that prior to this change, aarch64 would still output kernel logs
to the ttyAMA0 console (on Azure, at least) due to the ACPI SPCR table.
On aarch64 the console in this table, if it's present, is _always_
configured as the kernel's default console. However, systemd logs would
not be sent to the serial console because it seems it was honoring the
kernel command line configuration which had ttyS0 as the default.
Also see: https://pagure.io/cloud-sig/issue/394
Also see: https://bugzilla.redhat.com/show_bug.cgi?id=1661288
Signed-off-by: Jeremy Cline <jeremycline@linux.microsoft.com>
|
|
This introduces extra validation for WSL image creation to ensure
the result is usable.
|
|
This introduces extra validation for WSL image creation to ensure
the result is usable.
|
|
This will be used by toolbox(1) to ensure that the certificates from
certificate authorities (or CAs) that are available inside a Toolbx
container are kept synchronized with the host operating system [1]. Any
program that uses PKCS #11 to access CA certificates should see the same
ones both inside the container and on the host.
This is the same approach taken by Flatpak [2].
[1] https://github.com/containers/toolbox/issues/626
[2] Flatpak commit 66b2ff40f7caf3a7
https://github.com/flatpak/flatpak/commit/66b2ff40f7caf3a7
https://github.com/flatpak/flatpak/pull/1757
https://github.com/p11-glue/p11-kit/issues/68
https://pagure.io/fedora-kiwi-descriptions/pull-request/189
|
|
The p11-kit RPM is neither explicitly mentioned in the list of default
RPMs on Fedora Silverblue and Workstation [1], nor was it explicitly
installed by the Container/Dockerfile equivalent of the fedora-toolbox
OCI images [2].
The Container/Dockerfile listed p11-kit as one of the RPMs that had to
be reinstalled, because it was inherited from the fedora base image, and
had its documentation and translations stripped out. The reinstallation
would be skipped if for some reason it was no longer part of the fedora
base image.
This was misinterpreted when creating the Kickstart equivalent of the
fedora-toolbox OCI images [3] and that's how it ended up in the KIWI
description.
In reality, p11-kit gets pulled in as a dependency of ca-certificates,
gnutls, etc. and that's sufficient.
Fallout from d27e4c3f6c3e13bcefe9d1d904386794140d8a91
[1] https://pagure.io/fedora-comps/
[2] https://src.fedoraproject.org/container/fedora-toolbox
[3] fedora-kickstarts commit 48e2c3b5598de32f
https://pagure.io/fedora-kickstarts/c/48e2c3b5598de32f
https://pagure.io/fedora-kickstarts/pull-request/964
https://pagure.io/releng/issue/11399
https://pagure.io/fedora-kiwi-descriptions/pull-request/189
|
|
Although we include curl, some folks expect wget. Also, one of the more
common uses of WSL is in combination with VS Code, and its setup script
on remote hosts uses wget.
Signed-off-by: Jeremy Cline <jeremycline@linux.microsoft.com>
|
|
Signed-off-by: Troy Dawson <tdawson@redhat.com>
|
|
This should improve the experience by having separate clickable
links for each test matrix item and see the pass/fail more early.
Also make Testing Farm scale better as it uses separate TF requests
for testing.
Signed-off-by: Miroslav Vadkerti <mvadkert@redhat.com>
|
|
This ensures that the menu does not pop up for the first boot on
variants that leverage boot success tracking.
|
|
This became necessary with 10.0.13 / after merging
https://github.com/OSInside/kiwi/pull/2535 but doesn't seem to
noticeable effects besides miisin "GRUB_ENABLE_BLSCFG=true" in
/etc/default/grub.
|
|
menu_auto_hide=1 is required for the boot_success provided by Fedora's
grub2 package. This will fall back to the previous kernel and show the
usually hidden grub menu when booting fails.
This makes it unnecessary to override the timeout_style and using a
reduced timeout of 1 second for disk images.
|
|
This drastically reduces the content included in the embedded ESP
for live media, and avoids weird side-effects caused by extra EFI
binaries being present.
|
|
This package only exists on x86 architectures.
Fixes: 06ef485a89eeb552ab84c322ddbb622e65772c20
|
|
We do not need the full livecd-tools on the SoaS image, only the
scripts needed to put the ISO on a USB stick with persistence.
|
|
This ensures that the initial-setup system is present on the image,
which is required for disk image builds.
|
|
|
|
This is not exhaustive, but does get users into the right environment
to build images. We also indicate that this is only tested on Fedora Linux.
|
|
Fixes https://pagure.io/fedora-kiwi-descriptions/issue/169,
https://bugzilla.redhat.com/show_bug.cgi?id=2355948.
|
|
This is needed for XUL-based applications to work (for example, Zotero).
Signed-off-by: Davide Cavalca <dcavalca@fedoraproject.org>
|
|
It can be useful to have access to `setfiles` inside a `toolbx`
container. Especially when using them as bootstrap roots to setup
other trees.
Signed-off-by: Simon de Vlieger <supakeen@redhat.com>
|
|
This ensures that the FEX image build does not fail.
|
|
When the kiwibuild_extra_args variable is not defined, we do not
want it to fail with "unbound variable" error.
Fixes: efc71e6d8629f0e3b5bd2541650266d632a2fe68
|
|
The ldconfig cache generator runs on first boot and attempts to
read the whole image into memory, which eats gigabytes of RAM and
slows the boot down by half a minute. This is unnecessary for
images being booted unmodified.
Thus, create the necessary dot files to inhibit their runs.
|
|
level 8 LZMA compression"
This did not help anything and it makes image builds slower.
This reverts commit b57a77e5717e31914faddf5c258d74a0db5e988f.
|
|
This simplifies supporting one-off custom actions without having to
define options in the helper script all the time.
|
|
If $PATH is not set, we get the default from crun [1]:
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
This is not useful, because in the image, /bin, /sbin, /usr/sbin are all
symlinks. /usr/local/sbin does not exist.
When /usr/sbin is a symlink to /usr/bin, and we use have a $PATH with
/usr/sbin before /usr/bin, various tools will discover /usr/sbin/foo when
looking for 'foo' [2], which is confusing and ugly.
[1] https://github.com/containers/crun/blob/a980c89665bb488ff206e47e9083f75b48528714/src/libcrun/container.c#L183
[2] https://src.fedoraproject.org/rpms/setup/pull-request/18#comment-253719
Co-authored-by: Simon de Vlieger <cmdr@supakeen.com>
|
|
compression
It seems the larger chunk size seems to slow the start of images in
low memory environments due to the ldconfig cache generator unit
eating gigabytes of memory from mapping the chunks into memory.
Attempt to work around this by lowering the chunk size while raising
the compression level to attempt to maintain image sizes.
|
|
|
|
Signed-off-by: Ryan Brue <ryanbrue.dev@gmail.com>
|
|
Signed-off-by: Ryan Brue <ryanbrue.dev@gmail.com>
|
Neal Gompa
Yaakov Selkowitz
Troy Dawson
Jeremy Cline
Adam Williamson
Dan Book
Katerina Koukiou
Clement Verna
Jiri Konecny
Debarshi Ray
Miroslav Vadkerti
Janne Grunau
Philip Molloy
Jan Pazdziora
Davide Cavalca
Simon de Vlieger
Zbigniew Jędrzejewski-Szmek
Allan
Ryan Brue
