Hi, First post. I am not to experienced in the Unix or HP-UX world but I'm learning.
We have some problem with a HP rx4640 server. During boot it states the following:
Checking for LVM volume groups and Activating (if any exist)
Volume group "/dev/vg00" has been successfully changed.
vgchange: Warning: Couldn't attach to the volume group physical name "/dev/dsk/c5t6d0"
Device busy
Activated volume group
vgdisplay: Warning: couldn't query physical volume "/dev/dsk/c5t6d0"
The specified path does not correspond to physical volume attached to this volume group
vgdisplay: warning: couldn't query all of the physical voldeumes
Checking vxfs file systems
vxfs fsck /dev/vg01/lvol4/fsck read failure bno = 360, off = 0, len = 32768
vxfs fsck /dev/vg01/lvol5/fsck read failure bno = 360, off= 0, len ) 32768
couldn't fix vxfs filesystems, dropping into a shell ...
(bcheckrc)#Resynchronized volume group /dev/vg01

I tried a full fsck (fsck -F vxfs -o nolog, full /dev/dsk/c5t6d0) and just got that it failed because the disk is busy.

Any ideas how I shall proceed?

/Hoff

You cannot run fsck on a disk! You might have a disk failure here. Check on the server if there is any disk activity on that particular disk. Also check for any warning lights on the server itself.

Hm ok. The disk belongs to an external JBOD cabinet and all disks in that show no sign of error. All LEDs are green. The server machines LEDs are all green but the disk LED is off, not showing any disk activity.

Can I only run fsck on an volume group or what do you mean?

hoff, the physical volume /dev/dsk/c5t6d0 was added to volume group vg01. Possibly other disks might have been added to vg01 as well. Then vg01 is carved up into logical volumes such as /dev/vg01/lvol4. And then file systems were created on the logical volumes. It is possible to ignore LVM and install a file system directly on a physical disk, but few people would do that and the OS installation procedure would make it rather hard to do. So, yes, you need to fsck /dev/vg01/lvol4 and lvol5.

Hi, thx for all help. I think I got a hang of it now and found out what disk that is malfunctioning.
I did an diskinfo on the specified disk /dev/rdsk/c5t6d0 and it shows size: 0 Kbytes, which seems quite odd. All other disks seems fine and gives the correct size.

HI again, Really could use some more help.
Now then I have found out which disk that is malfunctioning. Can I somehow point out what disk it is?.
In the JBOD cabinet all disks are show green lights, so it seems like everything is OK, even if it isn't.
Is there any command to send out a signal to a specific drive (i.e. /dev/rdsk/c5t6d0) to start flashing yellow or shutdown?

Maybe a stupid question but I really can't see any other way to pick out the right disk from the cabinet

/Hoff

This command:
dd if=/dev/rdsk/c5t6d0 bs=4096 of=/dev/null
tries to read the entire disk and copy it to /dev/null. It may do nothing for the broken disk. But if it works, it usually causes some activity indication such as the green light flashing. You should draw a little picture of your JBOD cabinet and run the command for each disk to map them out. If you ID all of the good disks, the one that is left must be the bad disk. Once you ID a disk, use your interrupt character, usually cntl-c to kill the dd.

Thx for your help. This probably was the best solution.

Regards,
Christian