LVM Configuration error

[FATEH PAL SINGH MALHI]

We are using 2 B2600 workstations for plant operation.

One of these is giving problem and restarting again and again with following message.

SYSTEM HALTING during LVM Configuration.
Could not configure root VG.

Please help me in fixing the problem.

[Torsten.]

LVM: VG 64 0x000000: Quorum check failed!
LVM : Activation of root volume group failed
Quorum not present, or some physical volume(s) are missing


If the vg00 is mirrored, the alternative disk is gone.

Boot with "hpux -lq" from ISL>


If vg00 consists of 2 disks without mirror, you need to replace the bad disk and restore from backup.

[Mel Burslan]

It seems like some of the disks in your vg00 did not come back up after reboot, due to god knows what. It might be bad hardware, or just a faulty connection among many other things. Have a hardware service person evaluate and fix the problem before you can boot your system. I am under the impression that, this volume group was not mirrored, in which case yo might need to rebuild the OS (or if you have ignite tapes made) re-ignite it and then restore you application and data from the backups. (You have backup, don't you ?)

Good luck

[FATEH PAL SINGH MALHI]

Thanks for showing interest in fixing our problem

After interepting the system during booting
on executing SEA command
We find P0, P1, P2 drives
on executing BO P1 command
when it start booting

it gives the above messege.

We don't have the backup of the system.
Is it possible to retrieve the old data from UNIX machine before relaoding it.

[Torsten.]

No backup? Bad!

If the disk was mirrored, stop the automatic boot, boot from disk, interact with ISL and do

ISL> hpux -lq

[FATEH PAL SINGH MALHI]

If possible tell me the complete procedure.
One server is main (faulty), other is working ok.
Faulty is having ARCHIVE folder for external archiving and COL is running on both machines.
In normal when one machine is shutdown and restarted it is synchronised with the other and work as hot redundent.
Active-Red Passive-Red
My problem is that on reinstalling the main server it will delete all the ARCHIVE & COL or is it ok.
On my machine Maestro-UX software is running on UNIX.
If COL & Graphic get deleted from other machine I would lost everything.
How to take back up a UNIX machine.
Please guide us.

[Torsten.]

Stop at booting; here is a screen capture:

Information Menu: Enter command > ma

---- Main Menu ---------------------------------------------------------------

Command Description
------- -----------
BOot [PRI|ALT|] Boot from specified path
PAth [PRI|ALT] [] Display or modify a path
SEArch [DIsplay|IPL] [] Search for boot devices

COnfiguration menu Displays or sets boot values
INformation menu Displays hardware information
SERvice menu Displays service commands

DIsplay Redisplay the current menu
HElp [|] Display help for menu or command
RESET Restart the system
----
Main Menu: Enter command or menu > bo pri
Interact with IPL (Y, N, or Cancel)?> y

Booting...
Boot IO Dependent Code (IODC) revision 1


HARD Booted.

ISL Revision A.00.38 OCT 26, 1994

ISL> hpux -lq

[FATEH PAL SINGH MALHI]

Thanks I will try this in the moring.
Also tell me the wat to take backup of the other running UNIX machine

[Mel Burslan]

since you do not have a backup solution, at this point you can do one of the two things:

Method 1 (this requires both machines have identical tape drives attached and you have blank tapes) use fbackup (how to use is explained in man page. Run command 'man fbackup' to learn more about it.

Method 2 (this requires both machines are on the same network, which should be the case since you are copying data from one machine to the other daily)

reload hpux to the first machine in which way you can (from tape or cd), create the directory structure as it is on second machine

go to the second machine which is not broken, use scp or rcp to copy data over to the first machine. Since you have a setup to copy from the main machine to the second machine, you can use the same logic to copy it back.

Without knowing the exact data structure that you want to copy, it is hard to make an exact suggestion.

And, last but not the least, implement a BACKUP method to prevent this from happening again. If you are not experienced on HPUX, you can always ask additional service from HP or other consulting companies or independent consultants.

Cheers...

[Torsten.]

Old machines have most of the times an old OS - I would guess it is 10.20.

Please logon to the other identical system as root and do

# uname -a
# swlist
# vgdisplay -v
# lvdisplay -v /dev/vg00/lvol1 | head -n 30

[FATEH PAL SINGH MALHI]

Thanks everybody who has participated so far in the discussion in the forum.
I have gone through the machine today again through Hyper Terminal and collected all the display to help you guys to fix the problem, Hope it would be some useful to you guys.

[Torsten.]

The only thing I see is - it is 10.20 (unsupported for at least 6 or 7 years now).

How about the other commands in my post above?

1) Try to boot with "-lq" switch (see above)

2) Collect more information from your working server (see above).

If we find out the disks are mirrored we are near to the solution ...

[FATEH PAL SINGH MALHI]

Thanks for all the efforts being done by you to fix our problem.
I will check this thing tomorrow as revert back.
Can you tell me more about these commands?

What does ISL stand for?
What would happen by ISL> hpux -lq

What is the purpose to run following commands on other machine?
# uname -a
# swlist
# vgdisplay -v
# lvdisplay -v /dev/vg00/lvol1 | head -n 30

Last question
What does mirrored disk means?

We are running Maestro UX software on these machines.

Are you sure this a SOFTWARE problem and nothings to do with HARDWARE?
Have you gone through the attachment?

Please elaborate properly all the procedure, if possible for you.

[RajuD]

ISL Stands for initial system loader

isl implements the operating system independent portion of the
bootstrap process. It is loaded and executed after self-test and
initialization have completed successfully, To know more type man isl...

hpux -lq
Will boot the system without quorum restriction. Quorum check will take place before activating the vg.


What is the purpose to run following commands on other machine?
# uname -a
# swlist
# vgdisplay -v
# lvdisplay -v /dev/vg00/lvol1 | head -n 30

The purpose of checking the above command is to come to the conclusion whether the disk is mirrored and to find out how many disk are present in the vg.

What does mirrored disk means?
Copy of data will be present in another disk in the same vg.........

Cheer

[FATEH PAL SINGH MALHI]

Hi

Great answar.
Full marks to all guys.
This a great move to create a FORUM.
To solve the problem by sharing.
I am thankful to everybody who has participated in the discussion so far.

Few questions yet to replied
what is quorum?
& what is quorum failure.

what does hpux -lq cmd?
In this cmd lq stands for?

Do you know somebody having knowledge of Meastro UX software of ABB.
We have problem of ARCHIVING in the Maestro UX.
In ws01 & ws02(name of work stations) we have /arc folders.
/exarc folder is only in ws02, from where it goes to external ARCHIVE PC.
At present this data is not going to ARCHIVE PC. as a result of no noline trends comes & ws02 (work station) shutdown frequently.

Thanks a lot

[Torsten.]

If we assume the disks are mirrored, we can boot with only 1 disk.

All other commands (performed on your other working system) will give this information.

>> what is quorum?

By default, the system will check if you have >50% of all disks available. If you have 2 disks, 1 failed you have exactly 50%, but not >50% - the check will fail.

This is what you have now.

By booting with "-lq" this check will be disabled.