kdump-tools uses ucf for config file management and naively
modifying the config file meant for the target system directly
will cause the file hash to not get updated in the ucf database.
This will then cause later modifications to fail because
"there's nothing to do". Although actually doing the modification
to the ucf database is messy. Let's just modify the file in the live
layer to get the behavior we want there.
We install the kdump-tools package to minimal layer via inclusion in the
server-minimal seed, but it is enabled by default. Include a new chroot
hook to set USE_KDUMP=0 to make sure it's disabled by default and let
the installer decide to enable it or not.
By placing the kernel in minimal, we can achieve the following
improvements:
1. Space savings - there are redundant packages present in the ship-live
pool and in the live layer. Adding the kernel to minimal means that
the kernel is already in the live layer, and we don't then also need
it in the pool.
2. Time savings - informal vm testing suggests more than a minute
improvement to have the kernel preinstalled over installing it at
runtime.
As always, there is a cost tradeoff:
1. If a different kernel is desired, we need to be able to remove this
preinstalled kernel. Relevant curtin and subiquity changes are
already landed.
2. When installing that other kernel, it'll take longer than today due
to still needing to install a kernel at runtime + the time cost of
removing the preinstalled kernel.
unminimize is currently present at /usr/local/sbin/unminimize,
which is spit out by livecd-rootfs currently. We'd like to switch
that to use the packaged unminimize, which will be at
/usr/bin/unminimize instead.
There was a change made by me in https://code.launchpad.net/~philroche/livecd-rootfs/+git/livecd-rootfs/+merge/466388
as part of LP: #2066905 to remove references to LXD in the unminimize scripts
but I also removed the calls to `unminimize` in error.
This still needs to run but without any references to LXD which no longer
needs to be `unminimized` via snap installation.
The LXD snap is no longer seeded in any images since Noble+ so the LXD related unminimize logic in
./live-build/auto/build?h=ubuntu/noble and ./live-build/ubuntu-server/hooks/01-unminimize.chroot_early
is no longer required.
lxd-installer can remain installed.
The unminimize script will try to install the lxd snap using the shim script
`/usr/sbin/lxd` from the lxd-installer package.
Previously `unminimize` was using `snap`
to install `lxd` directly which was being diverted by diverting the `snap` command.
This is no longer the case so we can remove `/usr/sbin/lxd` from the lxd-installer package
if it exists and then redirect any calls to `/usr/sbin/lxd` to `/bin/true`
This is a cherry pick forward port from Jammy livecd-rootfs version 2.765.37.
(cherry picked from commit 8b83212372)
live-build/auto/config:
- for Ubuntu Server live images and the arm64+tegra full arch, build a
tegra variant with linux-nvidia-tegra as the flavor and
linux-nvidia-tegra as the kernel meta-package
- default to nvidia-$SUBARCH as the kernel flavor for all images using
arm64+tegra as full arch
hooks/03-kernel-metapkg.chroot_early:
- use linux-nvidia-tegra as kernel meta-package for the nvidia-tegra
flavor
* Fix some issues with the netboot tarballs:
- Include the signed shim (oops).
- Make the kernel path on disk and in the bootloader config match (more
oops).
- Make paths more architecture dependent as the code in grubnetXXX.efi to
probe a platform dependent path first doesn't work.
- drop ds-identify policy, not needed with improved cloud config
- drop disabling network, doesn't work with ip=
- fixup setting up the INSTALLER_ROOT mountpoint
- enable cloud-init
- make cloud-init handle the default/baked in networking configuration
- install and enable openssh-server for the installation only
- provide cloud.cfg that generates random installer user password
- disable subiquity on sclp_line0 line based console
Subiquity images mount /usr/lib/modules from a squashfs, which systemd
tries to unmount on shutdown, whilst they are still being in use. As
systemd-udevd kmod built-in's libkmod has modules.* files
memorymapped. This produces an warning on shutdown, flooding the
screen with messages as systemd switches to a more verbose
output. Specify LazyUnmount=yes on that mount, such that unmount call
succeeds without flooding the shutdown log.
ubuntu-cdimage/debian-cd/tools/add_live_filesystem helpfully adds a
'filesystem.' prefix to the squashfs, hence the name of the file
didn't match what's on disk.