This documents shows a way to encrypt all data (incl. OMV) and unlocking it via ssh at boot time. This seems to be the only applicable way, as after boot and encrypted data drives only, this has had major influence on system usability (i.e. Docker mount on data lead to massive errors).

In general the approach is a fresh OMV install and a livecd with enabled ssh.

1. Install OMV as of documentation
2. Boot into live and make the OS drive enrypted with luks
3. Enrypt data drives and make the decryption derive from the OS drive

This approach seems to be scalable and doable, as the dropbear unlock approach is limited to one drive.

Create your usb stick with dd or apps like unetbootin

Follow the documentation: https://openmediavault.readthedocs.io/en/latest/installation/index.html

• Set admin ui password
• if needed change admin ui port
• copy your pub key for user root (ssh-copy-id root@<ip>)
• disable ssh password login
• update to latest state (apt update && apt dist-upgrade)

Make sure the following packages are present

apt install cryptsetup
apt install dropbear-initramfs

please refer to /usr/share/doc/dropbear-initramfs/README.initramfs and /usr/share/doc/cryptsetup/README.Debian.gz Section 8

I am on a local lan and with port forwarding I do not want to expose dropbear to the net. There I add DROPBEAR_OPTIONS="-p 2222" to /etc/dropbear-initramfs/config to change the listening port to 2222.

Now paste your public ssh key to the authorized_keys file.

echo "ssh-rsa ..." >> /etc/dropbear-initramfs/authorized_keys
chmod 600 /etc/dropbear-initramfs/authorized_keys

Update the boot image

update-initramfs -u -k all
update-grub

in case of perl language setting errors use dpkg-reconfigure locales

To later connect without host key error (you may of course use a specific ssh config)

ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -p2222 root@<ip>

Choose a live cd (or rescue system from your hoster) and enable ssh. This makes things easier by copy pasting or chilling on the couch.

I use (debian live)[https://cdimage.debian.org/debian-cd/current-live/amd64/iso-hybrid/] and enable sshd after boot manually (here)[http://www.zoyinc.com/?p=2510]

Create the live usb stick with

$ sudo dd if=./debian-live-9.3.0-amd64-lxde.iso of=/dev/sdb bs=4096

In order to encrypt the omv drive, we need to save the OS drive and repartition the device with luks.

To get an overview of the partitions to migrate use lsblk. In my case the OS drive is sde.

# lsblk
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda      8:0    0   3.7T  0 disk
sdb      8:16   0   3.7T  0 disk
sdc      8:32   0   3.7T  0 disk
sdd      8:48   0   3.7T  0 disk
sde      8:64   0 119.2G  0 disk
|-sde1   8:65   0 103.4G  0 part /
|-sde2   8:66   0     1K  0 part
`-sde5   8:69   0  15.9G  0 part [SWAP]

OMV creates those by default. We do want to change this to /boot (unencrypted), swap (later encrypted randomly), / (encrypted).

Select your usb stick at bios boot to use this instead of the hdd. Enable sshd, if not enabled.

apt update && apt install openssh-server
sudo passwd root
sudo su - root

/etc/ssh/sshd_config

PermitRootLogin yes
PasswordAuthentication yes

service ssh restart

Login with ssh, create the backup of the mounted drive.

$ ssh root@192.168.x.x

This should run without errors and copies /dev/sdx# to the ramdisk. As omv is rather small in size, this should not take too long.

mkdir /oldroot
mount /dev/sde1 /mnt
rsync -a /mnt/ /oldroot/
umount /mnt

Now that we have a local backup, it is possible to encrypt the drive.

make sure you select the correct device /dev/sdx#

cfdisk /dev/sde

A screen opens, where all partitions are deleted.

The new partitions look like this:

• sdx1: 1GB, type 83 Linux, bootable
• sdx2: 16GB (choose your desired swap size), type 83 Linux for later encryption
• sdx3: rest, type 83 Linux

Do not forget to write the changes

Check the partitions with lsblk and make the boot filesystem

mkfs.ext4 /dev/sde1

Now let's encrypt the root partition sde3, choosing a strong crypt key.

apt update && apt install cryptsetup
modprobe dm-crypt
cryptsetup --cipher aes-xts-plain64 -s 512 -h sha256 --iter-time 5000 luksFormat /dev/sde3

Verify it worked with

cryptsetup luksDump /dev/sde3

Lets open the drive and create a filesystem

cryptsetup luksOpen /dev/sde3 root
mkfs.ext4 /dev/mapper/root

mkdir /newroot
mount /dev/mapper/root /newroot
mkdir /newroot/boot
mount /dev/sde1 /newroot/boot
rsync -a /oldroot/ /newroot/

Then bind folders needed to chroot

mount --bind /dev /newroot/dev
mount --bind /sys /newroot/sys
mount --bind /proc /newroot/proc
chroot /newroot

Next we will update everything

/etc/fstab:
UUID=<uuid of /dev/sde1> /boot ext4 defaults 0 1
/dev/mapper/root / ext4 defaults 0 2
# swap was on /dev/sde5 during installation
/dev/sde2 none            swap    sw              0       0

Use UUID to make sure the correct drive is mounted (blkid /dev/sde3)

/etc/crypttab:
# <target name>	<source device>		<key file>	<options>
root UUID=<uuid> none luks

RESUME is not wanted in initramfs

rm /etc/initramfs-tools/conf.d/resume

Update /boot

update-initramfs -u -k all
update-grub
grub-install /dev/sde

umount every device after leaving the chroot and reboot

logout
umount /newroot/boot
umount /newroot/proc
umount /newroot/sys
umount /newroot/dev
umount /newroot
reboot

make sure the livecd is not booted

log into the dropbear and run cryptroot-unlock

ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -p2222 root@x
Warning: Permanently added '[x]:2222' (ECDSA) to the list of known hosts.
To unlock root partition, and maybe others like swap, run `cryptroot-unlock`


BusyBox v1.22.1 (Debian 1:1.22.0-19+b3) built-in shell (ash)
Enter 'help' for a list of built-in commands.

~ # cryptroot-unlock
Please unlock disk root: <paste key>
cryptsetup: root set up successfully

Now you are able to login like normal and access OMV. Next swap will be encrypted. Currently free should show 0 swap.

Follow https://wiki.archlinux.org/index.php/Dm-crypt/Swap_encryption#UUID_and_LABEL e.g. using the LABEL method.

I setup for each data drive the same LUKS key and a second key that is derived from root decrypt. This way at boot only one key is necessary but "in case" the drives can be unlocked on their own.

Again, with lsblk show all drives.

Initial creation will be with an initial key, which is used to unlock in case of emergency.

cryptsetup --cipher aes-xts-plain64 -s 512 -h sha256 --iter-time 5000 --label storage luksFormat /dev/md127
cryptsetup --cipher aes-xts-plain64 -s 512 -h sha256 --iter-time 5000 --label backup luksFormat /dev/sdg1

Lets open the drives for further setup:

cryptsetup luksOpen /dev/md127 storage
cryptsetup luksOpen /dev/sdg1 backup

Now lets add the key from the derived root mapper.

cryptsetup luksAddKey /dev/md127 <(/lib/cryptsetup/scripts/decrypt_derived root)
cryptsetup luksAddKey /dev/sdg1 <(/lib/cryptsetup/scripts/decrypt_derived root)

cryptsetup luksDump /dev/md127 now shows 2 keys.

To create the data_key on the enrypted root device

This approach is chosen as i.e. decrypt_derived as keyscript= is ignored by systemd

mkdir -p /root/.luks
chmod 700 /root/.luks
/lib/cryptsetup/scripts/decrypt_derived root > /root/.luks/data_key
chmod 400 /root/.luks/data_key

Next append to crypttab, timeout is added to make it boot even in case of error.

/etc/crypttab:
# <target name>	<source device>		<key file>              <options>
root	        UUID=...	        none                    luks
swap	        LABEL=cryptswap     /dev/urandom            swap,offset=2048,cipher=aes-xts-plain64,size=512
# data drives
storage         UUID=...            /root/.luks/data_key    luks,timeout=10s
backup          UUID=...            /root/.luks/data_key    luks,timeout=10s

Reboot and check if everything is mounted correctly.

From here you may start to use OMV as you like. If a drive is appended, use the commands from above and add it to /etc/crypttab.

OMV drive encryption plugin is usable as well to add the key and backup headers.

Now whenever normaly refering to /dev/sdX just use /dev/mapper/x instead:

mkfs.ext4 -E lazy_itable_init=0,lazy_journal_init=0 /dev/mapper/storage -L storage
mkfs.ext4 -E lazy_itable_init=0,lazy_journal_init=0 /dev/mapper/backup -L backup

Now, using OMV, mount filesystems which will create records in /etc/fstab for you.