Genuen Excalibur 6 rocks!

I've updated the iso image after adding these virtio_* modules (plus virtio_gpu)

I managed installing the image with Qemu and the following steps in BusyBox worked for me:

Load modules

(initrmafs) modprobe  virtio_pci
(initrmafs) modprobe  virtio_blk
(initrmafs) modprobe  virtio_scsi
(initramfs) modprobe ext4

Check partitions:

(initramfs) cat /cryptroot/crypttab | cut -d: -f1
vda3_crypt
(initramfs) blkid | grep crypto_LUKS | cut -d' ' -f1
/dev/vda3

Unlock vda3_crypt (adapt it to your needs) typing your encryption password:

(initramfs) cryptsetup luksOpen /dev/vda3 vda3_crypt
Enter passphrase for /dev/vda3:

Activate LVM:

(initramfs) lvm vgchange -ay

Scan /dev/mapper with ls -l /dev/mapper and mount the *-root logical volume (don't forget to load the ext4 module first):

(initramfs) mount /dev/mapper/genuen--vg-root /root

At this point, the way to get out of Busybox isn't that straightforward. Like live-boot, most of initramfs scripts leave a trace when they run. To prevent BusyBox from asking you to unlock vda3_crypt again after exiting (a process that will fail because the device is busy), we manually create the said control file so that the system can find it and recognize that the mountpoint already exists:

touch /run/initramfs/fsck-root

Now we make use of switch_root in order to replace the current process with the real system:

exec switch_root /root /sbin/init

Good luck!

ALERT! /dev/mapper/genuen--vg-root does not exist. Dropping to a shell!

Should I edit the content of /proc/cmdline, or should /dev/mapper/ be populated in some way?

Second, I guess that some virtio modules are missing in the initrd. Can you try the following commands in the BusyBox console?

1) Check whether the nodes in /dev have be created:

(initramfs) ls -l /dev/vda*

For sure, you won't get anything, even thought vdev doesn't ignore them (lines 93-96 in vdevd/helpers/LINUX/disk.sh).

2) Ensure that vdev is still running (via `pidof vdevd`) as it ought to be because the init-bottom scripts have not been invoked yet. Then try to load the following modules:

(initrmafs) modprobe  virtio_pci
(initrmafs) modprobe  virtio_blk
(initrmafs) modprobe  virtio_scsi
(initrmafs) modprobe  virtio_ring
(initrmafs) modprobe  virtio_console

3) Check again the nodes:

(initramfs) ls -l /dev/vda*

Do they appear now? If so, you would be able to go ahead with the boot process by exiting:

(initramfs) exit

Thanks a lot for your patience, prospero!!

cryptsetup: Waiting for encrypted source device
UUID=123456x7-8y90-*...
Gave up waiting for root file system device.

Followed by the usual idle BusyBox.

# Explicitly force the use of LUKS2
d-i partman-crypto/crypto_type string luks2

# Pass the specific arguments to cryptsetup
# grub-2.12 is not compatible with argon2id algorithm; use PBKDF2 instead
d-i partman-crypto/passphrase/options string --pbkdf pbkdf2

and d-i with encryption worked. I updated the image.

My installation attempt was interrupted here:

grub-installer: grub-install: error: attempt to install to encrypted disk without cryptodisk enabled.
Set `GRUB_ENABLE_CRYPTODISK=y` in file `/etc/default/grub`

Vdev recognizes the /dev/mapper/gnuinos--vg-root volume from the original installation, despite the annoying mdadm-related messages.

I have been running the live system then text-installing on a VM, first without LVM then both with and without LVM.

At some point I tried the automated install with LVM, which is when I got the message about /dev/mapper/devuan--vg-root (instead of /dev/mapper/genuen--vg-root).