Production server stuck on EFI shell and cannot boot

[Taurian]

First a little background.

One of the mirrored disk on the system had failed, so we replaced it physically. For reinstating the data in replaced disk from the original, we followed the steps in the attached file.

We got up to step 9, and we had to reboot the machine to free up the vg00 so that it could be deactivated.

Now the machine won't boot and is stuck on EFI boot menu. I am guessing that it is trying to boot from the newly replaced disk, whose data was not yet fully restored.

Can anyone help us to boot from the original disk? Machine is rx4640, OS is HPUX 11.23. The system is part of an MCSG cluster. Let me know if any other info is needed.

[Torsten.]

Your procedure looks good.
Do you see any fs entries when you do a "map -r" from EFI, e.g. fs0: ?

[Taurian]

Yes there is an "fs0:" entry. It gives the following error when I try to load hpux from it.

--------------------------
fs0:\> hpux

(C) Copyright 2004 Hewlett-Packard Development Company, L.P.All rights reserved

HP-UX Boot Loader for IPF -- Revision 2.018

Press Any Key to interrupt Autoboot
\efi\hpux\AUTO ==> boot vmunix -lq
Seconds left till autoboot - 0
AUTOBOOTING...
AUTO BOOT> boot vmunix -lq
Could not boot maintenance mode
--------------------------

[Torsten.]

From the boot loader prompt, what do you get?

HPUX> ls -l

[Taurian]

The "ls -l" command output is:

--------------
HPUX> ls -l
no such file or directory
--------------


The "map -r" command output is:

--------------
fs0:\> map -r
Device mapping table
fs0 : Acpi(HWP0002,100)/Pci(1|0)/Scsi(Pun1,Lun0)/HD(Part1,SigC5E470F2-B558-11DD-8002-D6217B60E588)
blk0 : Acpi(HWP0002,0)/Pci(3|0)/Ata(Primary,Master)
blk1 : Acpi(HWP0002,100)/Pci(1|0)/Scsi(Pun0,Lun0)
blk2 : Acpi(HWP0002,100)/Pci(1|0)/Scsi(Pun1,Lun0)
blk3 : Acpi(HWP0002,100)/Pci(1|0)/Scsi(Pun1,Lun0)/HD(Part1,SigC5E470F2-B558-11DD-8002-D6217B60E588)
blk4 : Acpi(HWP0002,100)/Pci(1|0)/Scsi(Pun1,Lun0)/HD(Part2,SigC5E4711A-B558-11DD-8003-D6217B60E588)
blk5 : Acpi(HWP0002,100)/Pci(1|0)/Scsi(Pun1,Lun0)/HD(Part3,SigC5E47142-B558-11DD-8004-D6217B60E588)
--------------

[Torsten.]

Did you really sync the VG?

HPUX> ls -l
no such file or directory

This looks like your

/stand

is empty, hence no kernel.

[Taurian]

No, that is the problem. We didn't sync the VGs before restart.

[Torsten. · Accepted solution]

What is the meaning of

"We got up to step 9, and we had to reboot the machine to free up the vg00 so that it could be deactivated." ?

Looks like you don't have any OS data on your disk - the disk is empty. What happens to the other disk or the mirror pair?

There is only 1 disk with an EFI area.

Do you have an ignite backup?

You will probably need it now ...

[VK2COT]

Hello,

Your alternate disk does not have /stand
set up properly.

Just last week I was teaching HP-UX for Experienced Admins in Australia and
one of the students made a mistake to
created a mirror on the second disk but the first slice was not /stand :)

Obviously, it did not work. Luckily,
it was a virtual lab and I helped him.
Than I sat down, and for the sake of my students create a small Shell script to
mirror and/or unmirror LVM boot disks
for HP-UX (it even takes care of
legacy and agile DSF in HP-UX 11.31):

http://www.circlingcycle.com.au/Unix-sources/HP-UX-11x-LVM-mirroring.sh.txt

I am still testing the script...

Maybe similar mistake happened in your case?

Cheers,

VK2COT

[Taurian]

Thanks for the help guys.

Luckily we have ignite backup. We will be restoring from it shortly.