Install OS from ISO on UEFI boot OVH server

adns · November 2025

Hi LET Members!

I do not want to increase the traffic of the Kimsufi deals topic so I open a new for my question.

I install my OVH server with Debootstrap. Wipe disks, create partitions, create RAID array, debootstrap on it, install kernel, install grub, install DHCP on eno1, reboot and everything works well.

Fedora has not provide same tools as debootstrap so we need an alternative way.

A friend of mine would like to install a Fedora 43 on to this server. As no debootstrap, I suggested him to boot into rescue, install qemu-system and ovmf, download Fedora netinst ISO, and start a Qemu VM with VNC, attach disks, do the installation process from it.

apt update && apt install -y qemu-system ovmf && qemu-system-x86_64 -enable-kvm -nodefaults -cdrom /root/Fedora-Server-netinst-x86_64-43-1.6.iso -drive if=pflash,format=raw,readonly=on,file=/usr/share/OVMF/OVMF_CODE.fd -drive if=pflash,format=raw,file=/usr/share/OVMF/OVMF_VARS.fd -drive file=/dev/nvme0n1,index=0,media=disk -drive file=/dev/nvme1n1,index=1,media=disk -m 16G -net nic -net user -vga std -vnc 0.0.0.0:0 -usbdevice tablet -k en-us

He tried to install with only one / partition, or install the / and /boot onto two different partitions. The result always as same, initramfs can not find the disks to load the OS and can not continue from here.

So my question: How can we install an OS to the disks with a qemu VM when server boots in UEFI? On an old KS-LE-1 this method worked well without ovmf components, installed the OS, reboot and booted. However this server booted with legacy BIOS.

Any help appreciated.

fredo1664 · November 2025

No IPMI?
I did experiment with pushing a custom iPXE bootScript via the API to a server, that starts a pxeboot image for Centos, I imagine fedora have the same (?).
The problem was that I never managed to have it start the installer with the RDP option, so I was still doing the installation via the IMPI console.

Edit: https://lowendtalk.com/discussion/comment/4073024#Comment_4073024

Xedon · November 2025

There is an alternative to debootstrap called “febootstrap".

Xedon · November 2025

Ah I see, feebootstrap is EOL. The alternative is called "supermin".

Found that guide, should be possible with dnf, everything else is EOL!?
https://blog.mtaha.dev/linux/fedora_bootstrap

That was confusing...

adns · November 2025

@fredo1664 said:
No IPMI?
I did experiment with pushing a custom iPXE bootScript via the API to a server, that starts a pxeboot image for Centos, I imagine fedora have the same (?).
The problem was that I never managed to have it start the installer with the RDP option, so I was still doing the installation via the IMPI console.

Edit: https://lowendtalk.com/discussion/comment/4073024#Comment_4073024

Thanks, we'll give it a try.

@Xedon said:
Ah I see, feebootstrap is EOL. The alternative is called "supermin".

Found that guide, should be possible with dnf, everything else is EOL!?
https://blog.mtaha.dev/linux/fedora_bootstrap

That was confusing...

Thanks, Gemini suggested it too but I found information that supermin is eol too. However we'll try it.

bugabuga · November 2025

I followed this guide and was able to install proxmox with spice on OVH VPS.

https://blog.natalie.ee/posts/proxmox-on-ovh-vps-2025/

adns · November 2025

@Xedon said:
Ah I see, feebootstrap is EOL. The alternative is called "supermin".

Found that guide, should be possible with dnf, everything else is EOL!?
https://blog.mtaha.dev/linux/fedora_bootstrap

That was confusing...

Thanks for the tutorial. My friend tried it out and sent the following. I can not help to fix this issue, I installed high number of Debians with debootstrap but currently I can not find root cause of the error.

I've done it successfully with your guide: https://blog.mtaha.dev/linux/fedora_bootstrap
But I'd like to use the two disk in RAID1 with xfs filesystem. So I tried to transform your steps a little bit, but I encounter errors.

I perform the following steps (attached)

After restarting the server to boot from the disks the boot process stops at this point
Job dracut-initqueue.service/start running
Job dev-disk-by\x2duuid-444-cc4d7\x2da01f\x2d4189\x...\x2deb4d6f4fa440.device/start running

Then entering emergency shell and get the following errors too:
dracut-initqueue[389]: Warning: Not all disks have been found.
dracut-initqueue[389]: Warning: You might want to regenerate your initramfs.

Do you have any idea what might be causing this error?

Executed commands

mdadm --stop /dev/md0
mdadm --misc --zero-superblock /dev/nvme0n1p2
mdadm --misc --zero-superblock /dev/nvme1n1p2
wipefs -a /dev/nvme0n1
wipefs -a /dev/nvme1n1

apt update
apt install dnf arch-install-scripts
cfdisk /dev/nvme0n1
cfdisk /dev/nvme1n1

    ###### The result of formating ######

    Disk /dev/nvme1n1: 1.75 TiB, 1920383410176 bytes, 3750748848 sectors
    Disk model: SAMSUNG MZQLB1T9HAJR-00007
    Units: sectors of 1 * 512 = 512 bytes
    Sector size (logical/physical): 512 bytes / 512 bytes
    I/O size (minimum/optimal): 512 bytes / 512 bytes
    Disklabel type: gpt
    Disk identifier: F04472EB-44B0-3648-9244-E09F4BA76E71

    Device           Start        End    Sectors  Size Type
    /dev/nvme1n1p1    2048    1050623    1048576  512M Linux filesystem
    /dev/nvme1n1p2 1050624 3750748159 3749697536  1.7T Linux RAID


    Disk /dev/nvme0n1: 1.75 TiB, 1920383410176 bytes, 3750748848 sectors
    Disk model: SAMSUNG MZQLB1T9HAJR-00007
    Units: sectors of 1 * 512 = 512 bytes
    Sector size (logical/physical): 512 bytes / 512 bytes
    I/O size (minimum/optimal): 512 bytes / 512 bytes
    Disklabel type: gpt
    Disk identifier: 02C05F03-501C-D349-87C4-0616BCF38ECF

    Device           Start        End    Sectors  Size Type
    /dev/nvme0n1p1    2048    1050623    1048576  512M EFI System
    /dev/nvme0n1p2 1050624 3750748159 3749697536  1.7T Linux RAID

    #####################################

mdadm --create --verbose /dev/md0 --level=1 --raid-devices=2 /dev/nvme0n1p2 /dev/nvme1n1p2
mkfs.vfat -F32 /dev/nvme0n1p1
mkfs.xfs /dev/md0
mount /dev/md0 /mnt
mkdir -p /mnt/{boot/efi,opt,root,srv,usr/local,var,etc}
mount /dev/nvme0n1p1 /mnt/boot/efi
genfstab -U /mnt | sudo tee /mnt/etc/fstab

mkdir -p /etc/yum.repos.d
nano /etc/yum.repos.d/fedora.repo

    ###### File content ######

[fedora]
name=Fedora $releasever - $basearch
#baseurl=http://download.example/pub/fedora/linux/releases/$releasever/Everything/$basearch/os/
metalink=https://mirrors.fedoraproject.org/metalink?repo=fedora-$releasever&arch=$basearch
enabled=1
countme=1
metadata_expire=7d
repo_gpgcheck=0
type=rpm
gpgcheck=0
#gpgcheck=1
#gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-fedora-$releasever-$basearch
skip_if_unavailable=False

    ##########################

dnf --installroot=/mnt --releasever=43 --forcearch=x86_64 --setopt=fastestmirror=True group install "Core"

cd /mnt
rm -f etc/resolv.conf
touch etc/resolv.conf
for f in dev etc/resolv.conf proc sys sys/firmware/efi/efivars; do
  sudo mount --bind /$f $f
done

chroot /mnt
dnf --releasever=43 --setopt=fastestmirror=True --refresh update
dnf --releasever=43 --setopt=fastestmirror=True group install core standard
rpm -E %fedora



vi /etc/dnf/dnf.conf

    ###### File content ######

gpgcheck=True
installonly_limit=3
clean_requirements_on_remove=True
best=True
skip_if_unavailable=True
fastestmirror=True
max_parallel_downloads=10

    ##########################

dnf install -y langpacks-{en,tr}* glibc-all-langpacks
ln -sf /usr/share/zoneinfo/Europe/Budapest /etc/localtime
echo 'LANG=en_US.UTF-8' > /etc/locale.conf
echo 'KEYMAP="hu"' > /etc/vconsole.conf
echo 'FONT="eurlatgr"' >> /etc/vconsole.conf
echo 'raidtest' > /etc/hostname

rm -f /etc/dnf/protected.d/grub*
dnf remove grub*
dnf install -y systemd-boot-unsigned sdubby
bootctl install
dnf install -y kernel{,-core,-devel,-modules,-modules-extra,-headers}

mdadm --detail --scan | sudo tee -a /etc/mdadm/mdadm.conf
dracut --regenerate-all --force

tux · November 2025

Check the names of the RAID devices, as there might be an error in them.

Xedon · November 2025

@adns said:

dracut-initqueue[389]: Warning: Not all disks have been found.
dracut-initqueue[389]: Warning: You might want to regenerate your initramfs.

Do you have any idea what might be causing this error?

I think you have to add the mdm module in a dracut config ("add_dracutmodules+=" mdm ") and/or kernel parms:
1. mdadm --detail /dev/md0 | grep UUID
2. add kernel parm: md.uuid="your-raid-uuid"
3. Regenerate initramfs

Edit:
Adding just rd.md=1 and rd.md.conf=1 and rd.auto=1 would probably also work.

Howdy, Stranger!

Categories

In this Discussion

Install OS from ISO on UEFI boot OVH server

Comments

Howdy, Stranger!

Quick Links

Categories

In this Discussion

Install OS from ISO on UEFI boot OVH server

Comments