topic Re: system crash in Operating System - HP-UX

system crash

Umapathy S — Mon, 04 Mar 2002 06:03:23 GMT

hi
I am using a hpux10.2 on a k220 machine. My system crashed due to a power failure.
When I startup the system, after checking the Volume Groups the system is hanging for a long time (upto 15 mins). Is it internally checking the filesystems?

thanks for your time and help

cheers
Umapathy

Re: system crash

Animesh Chakraborty — Mon, 04 Mar 2002 06:08:04 GMT

Hi,
yes. It runs fsck which takes a while.
If not ok by now then boot in single user mode then check which file system causing problem.see /etc/rc.log file for errors

Re: system crash

Umapathy S — Mon, 04 Mar 2002 06:18:08 GMT

hi Animesh
It says /dev/vg02 successfully changed and it is hanging there.
It is yet to go into single user mode....

cheers
Umapathy

Re: system crash

Umapathy S — Mon, 04 Mar 2002 06:22:15 GMT

my machine state says that
RUN F12F
CPU01

what does this actually mean?
Could anyone explain me on this ?

thanks in advance

cheers
Umapathy

Re: system crash

Mary Ann Lipa — Mon, 04 Mar 2002 07:15:15 GMT

Hi,
The system could be still fsck'ing your disks due to the abrupt shutdown.
You can check for activities from the front panel display window.

-dan

Re: system crash

Balakumar M — Mon, 04 Mar 2002 07:24:32 GMT

Hi,
If its take's long time try to abort using ctrl + \.
U can try in two ways,
1) try to boot up in single user mode.
ISL> hpux -is
if its boots up run fsck manually to all the mount points.

If you could get through..
2) try to boot up in maintance mode.
ISL> hpux -lm
Run fsck to all the mount points.
Hope this will work.
-Bala

Re: system crash

Umapathy S — Mon, 04 Mar 2002 09:05:09 GMT

I think one volume group (vg04) got corrupted. I tried to vgchange all volume groups and tried to run fsck for all volume groups. All volume groups except vg04 went through.
I got this message when going through vg04

LVM : vg[0]: pnum=0 (dev_t=0x1f006000) is POWERFAILED

thanks in advance

cheers
Umapathy

Re: system crash

manoj_pu — Mon, 04 Mar 2002 09:05:57 GMT

Hello

How many volume groups do you have other than vg00 and vg02.

If system has additional vg's it may be possible that the system is trying to activate the next vg.

It will come out after some time and gives error also if it fails.

Rgds

Manoj

Re: system crash

Vijeesh CTK — Mon, 04 Mar 2002 09:09:15 GMT

hi,,

RUN F12F means

F - running HP-UX
1 - load level it varies from 0 to A accordinf to system load
2 - ur system having 2 processors
ie CPU01

Regards,

Vijeesh CTK

Re: system crash

Steven Sim Kok Leong — Mon, 04 Mar 2002 09:23:46 GMT

Hi,

>> LVM : vg[0]: pnum=0 (dev_t=0x1f006000) is POWERFAILED

Double-check that your disk 0x1f006000 (pvnum=0) in VG04 is powered up and working.

If you are using HA enclosures, check the contacts of the power cables as well.

Hope this helps. Regards.

Steven Sim Kok Leong

Re: system crash

Umapathy S — Mon, 04 Mar 2002 09:32:01 GMT

hi Steven
I am using a hp disk array enclouser. I have checked up the power supplies. I have 4 disks out of which one (LED) is blinking.
How I can find out which target is pointing to which disk array

thanks in advance

cheers
Umapathy

Re: system crash

Vijeesh CTK — Mon, 04 Mar 2002 10:20:59 GMT

hi,

go to /dev/dsk directory..find which disk is having minor no 0x000600 in ur case.. find out which device is that (cxt6d0)..and using ioscan find otu the disk is in disk array or not

Vijeesh CTK

Re: system crash

Luc Bussieres — Mon, 04 Mar 2002 13:08:11 GMT

Hi,

>> LVM : vg[0]: pnum=0 (dev_t=0x1f006000) is POWERFAILED

The disk that is causing you a problem is
c0t6d0 (from 0x1f006000 => 1f driver, 00 what follow the c, 6 what follow the t and 0 what follow the d, the last 2 digits should be ignored)

Luc

Re: system crash

Sanjay_6 — Mon, 04 Mar 2002 14:05:36 GMT

Hi Umapathy,

Looks like one of your disks have failed. Boot into single user mode using ISL and then active the vg without quorum, run a fsck and mount the filesystems. You will find one or more of the hard disks have failed. Replace those disks. You won't have to restore the data if the disks are mirrored, else you'll have to restore the data from the backups. After you replace the disk, do a vgcfgrestore to restore the vg info to the disk and then restore the data is the disk is not mirrored.

Try this link below on how to replace a lvm disk,

http://us-support3.external.hp.com/cki/bin/doc.pl/sid=914fb76903bc930c2d/screen=ckiDisplayDocument?docId=200000058036738

Hope this helps.

Regds

Re: system crash

Umapathy S — Tue, 05 Mar 2002 06:40:01 GMT

hi all
It was a classic cable connector problem. I took out the cable connectors, reconnected and asked sam to scan for disks. It was done successfully.
I activated the volume group, ran fsck on them and mounted the partitions. Its working fine. But I never thought of that big cable has a loose connection. I wonder how on a power failure the cable gets itself loosen.

Thanks for all the responses and invaluable time.

cheers
Umapathy