[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: QNAP 219 P II, Luks, Debian



On 31/08/14 22:11, Ian Campbell wrote:
On Sun, 2014-08-31 at 21:06 +0100, Tim Fletcher wrote:
On 31/08/14 20:40, Li, Peng wrote:
Hi Ian/Tim,

Trying to make sure I understand a correct steps. I'd like to re-install
the whole system again. I do have a serial console connection.

My plan is to have a /boot on a USB disk, with a 512bits data as
passphrase from usb key. whole two disk SDA/SDB will be RAID1 with LUKS
on top. Then LVM inside luks. Then, /root and /home.

  From the boot log you shared I'm almost certain the /boot on usb is not
needed for this to work, all you need to do to fix you current install
is to change the single setting in flash to make init look in the
correct place for the root disk.

The complex steps, ie finding and assembling the RAID, decrypting the
drive and finding the VG inside have all worked you have a single
problem left to solve which is the kernel commandline string.

Try on your serial console setting the following:

setenv bootargs console=ttyS0,115200 root=/dev/VG01/root
initrd=0xa00000,0x900000 ramdisk=34816

I think the root= here will be overridden by the flash-kernel hook in
the initrd (the script I pointed to earlier). A better bet might be to
hack that to override the root to /dev/VG01/root or to fiddle the
flash-kernel option which makes it do that
(Bootloader-sets-Incorrect-Root on Jessie, something else on Wheezy
which I can't remember)

Possibly, my thinking was that it wasn't been over written hence the boot failure. I'm pretty sure that even without the added complication of LUKS / RAID / LVM a normal debian install would fail with root=/dev/ram

--
Tim Fletcher <tim@night-shade.org.uk>


Reply to: