You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Using Grml from today with grml-debootstrap 0.95, installing Debian Sid/unstable on a system, the EFI System Partition (ESP) was empty despite passing the switch --efi. (I remember this happening in the past.)
% sudo grml-debootstrap -c config --packages packages --efi /dev/nvme0n1p1 --chroot-scripts chroot-scripts
* Reading specified config file config.
* EFI support detected.
* grml-debootstrap [0.95] - Please recheck configuration before execution:
Target: /dev/mapper/nvme0n1p3_crypt
Install grub: no
Install efi: /dev/nvme0n1p1
Using release: sid
Using hostname: test
Using mirror: http://deb.debian.org/debian
Using arch: amd64
Config files: /etc/debootstrap
Important! Continuing will delete all data from /dev/mapper/nvme0n1p3_crypt!
* Is this ok for you? [y/N] y
* EFI partition /dev/nvme0n1p1 seems to have a FAT filesystem, not modifying.
* Running mkfs.ext4 on /dev/mapper/nvme0n1p3_crypt
mke2fs 1.46.1 (9-Feb-2021)
Creating filesystem with 249784657 4k blocks and 62447616 inodes
Filesystem UUID: bfaeb600-ff13-45c7-a18d-0af9652e04ed
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968,
102400000, 214990848
Allocating group tables: done
[…]
Adjusting /etc/localtime
Setting /etc/timezone to Europe/Berlin
Executing stage fstab
Setting up /etc/fstab
Executing stage hostname
Setting hostname to davitoria.
Executing stage initrd
Generating initrd.
update-initramfs: Generating /boot/initrd.img-5.10.0-3-amd64
Executing stage grub_install
Notice: $GRUB not defined, will not install grub inside chroot at this stage.
Executing stage passwords
Activating shadow passwords.
Shadow passwords are now on.
Executing stage custom_scripts
Executing script /etc/debootstrap/chroot-scripts/add-user.sh
[…]
(The notice is unfortunately hard to spot in the output.)
The method grub_install() returns if the configuration variable GRUB is unset.
echo"Notice: \$GRUB not defined, will not install grub inside chroot at this stage."
return 0
fi
efi_setup ||return 1
if [ -n"$EFI" ] ;then
GRUB_PACKAGE=grub-efi-amd64
else
GRUB_PACKAGE=grub-pc
fi
It looks like there is also a – at least for me – new configuration option GRUB_INSTALL (commit 0b1b798 (support skipping installation of grub using GRUB_INSTALL='no')).
So, should grml-debootstrap abort if installation of GRUB to the EFI System Partition (ESP) is requested, or should it be implied?
The text was updated successfully, but these errors were encountered:
I never tried it, but does installing EFI make sense without a boot loader (grub)? If yes, efi_setup should run, if not we should at least display a warning or so.
(Also, Linux built as EFI stub (EFI_STUB=y, which Debian does), the UEFI firmware can execute this Linux kernel. (Then you have “Linux as boot loader”.)
Anyway, and this is the point for my report, in grml-debootstrap’s case, it’s indeed a contradiction. Note, currently, the switch --grub takes a device, and is not a boolean if to install GRUB or not.
But until that is improved, the installation should abort in my opinion, if --efi is passed, but GRUB won’t be installed.
Using Grml from today with grml-debootstrap 0.95, installing Debian Sid/unstable on a system, the EFI System Partition (ESP) was empty despite passing the switch
--efi
. (I remember this happening in the past.)(The notice is unfortunately hard to spot in the output.)
The method
grub_install()
returns if the configuration variableGRUB
is unset.grml-debootstrap/chroot-script
Lines 602 to 617 in d9988a5
It looks like there is also a – at least for me – new configuration option
GRUB_INSTALL
(commit 0b1b798 (support skipping installation of grub using GRUB_INSTALL='no')).So, should grml-debootstrap abort if installation of GRUB to the EFI System Partition (ESP) is requested, or should it be implied?
The text was updated successfully, but these errors were encountered: