Can't boot from old system drive in new machine. / not mounted. X-server died.

JZL240I-U · 10-21-2010, 01:29 AM

I built a new machine with components as shown below:

Gigabyte GA-890GPA-UD3H (AMD 890GX)
Phenom II X6 1055T (6x 2800 MHz)
Seagate ST31000520AS 1 TB (1.000 GB) SATA

I added a 250 GB Maxtor IDE disk of my old box with all my data and operating systems (SuSE 11.2, SuSE 11.0, FreeBSD 8.0) and made it the boot device.

SuSE 11.2 says runlevel 5 has been reached but KDE (kdm, rather) does not start. I get a local host login an can log in as root. Issuing "startx" results in "X-Server 1.6.5 (EE) No devices detected. No screens found." "mount" shows no mounted partitions. The error output on <F10> shows: "kdmrc - kdm: X-Server died during startup. X-Server for display :0 cannot be started, session disabled". Booting in secure mode results in "Waiting for device to appear (root)" and halts there.

SuSE 11.0 stops with "USB 2.-1. New USB device string: Mfr=0, Product=0, SerialNumbers=0)" followed by a lot of dots and then: "Could not find /dev/sda7 -- can issue mount: no hd found." (/dev/sda7 is "/").

FreeBSD boots with the white lines (not light gray yet) and then says: "Root mount error. mountroot>".

I tried BIOS settings of the drives as AHCI and IDE with identical results (well, with IDE a little more output on the power on selftest screen, gone too fast to read).

What causes this and how can I overcome it? Thanks for reading.

jefro · 10-21-2010, 08:57 PM

I never do like moving drives unless I know it is an exact clone of the original computer. Too many variables that may or may not be known even.

I think you'd have much better luck by getting off the data to some storage and do a clean install.

You can try to fix each problem. Might use live cd's of each to compare or fix with. Might be able to use single user mode.

Some of your problem are how the drive is now seen by the OS's. It may be that a cd drive or usb something has changed the order of the partitions.

rigor · 10-21-2010, 09:48 PM

The PAUSE key can sometimes be helpful when trying to read boot messages.

I've had good luck with SuSE detecting broken AHCI implementations, so unless
it's complaining, I would suspect that's probably not the issue.

Beyond that we'd probably need to hear more details on the partitioning
and the device map file ( if using GRUB ) .

JZL240I-U · 10-22-2010, 01:16 AM

Right now I'm not in front of that machine but this much I can say:

@jefro I have a clean install on the 1 TB drive and can read the old 250 GB with the systems I want to boot. (I just switched the boot drive in the BIOS). My wife has our home banking program in the old systems, that's the only reason I try to get them to run.

The error is not the way the drive is seen: X is unable to start (no screens), but runlevel 5 is reached even though no partitions are mounted.

@kakaka Thanks for the tip of the pause key, I'll try that soonest.

Partitioning is for /dev/sdan:

Code:

1: /boot
2: /Opensolaris
3: /FreeBSD
4: /Extended
5: /swap
6: /data
7: / of 11.0
8: /home of 11.0
9: /boot of 11.2 (now abandoned for 1:)
10: / of 11.2
11: /home of 11.2
12: /data

As said above, GRUB has long since relinquished command when the problem(s) arise. It is during the start of X in runlevel 5.

This much I can do from memory. For device.map I'll have to look in /boot/grub...

JZL240I-U · 10-25-2010, 03:30 AM

Device.map assigns /dev/sda to hd(0) and /dev/sdb to hd(1) as it should. Here is some more information from the error screen:

Code:

localhost kernel: xhci_hcd
WARN: Stalled endpoint 0000:02:00.0

That of any help?

JZL240I-U · 10-26-2010, 03:17 AM

I must shamefacedly acknowledge that I blundered in the case of openSuSE 11.2. If I could kick my own butt...

The old box had a Riva TNT chip of NVidia. The new one has AMD/ATI graphics. Erm. Sorry. So I edited the old xorg.conf files and 11.2 can boot now.

When I try to boot 11.0 (with the edited xorg.conf file) the screen stays black. No idea why that is so

. And on 11.0 is our old homebanking program. *sigh*

FreeBSD 8.0 just carries on like before, i.e. it has mount-problems.