vgdisplay produce error

[Ahmad Munawwar]

Hi,

When running vgdisplay -v /dev/vg01, i got erorrs:

vgdisplay: Warning: couldn't query physical volume "/dev/dsk/c0t11d0":
The specified path does not correspond to physical volume attached to
this volume group
vgdisplay: Warning: couldn't query all of the physical volumes.
LV Name /dev/vg01/data2
LV Status available/syncd
LV Size (Mbytes) 1400
Current LE 1400
Allocated PE 2800
Used PV 4

vgdisplay: Warning: couldn't query physical volume "/dev/dsk/c0t11d0":
The specified path does not correspond to physical volume attached to
this volume group
vgdisplay: Warning: couldn't query all of the physical volumes.
LV Name /dev/vg01/fs1
LV Status available/stale
LV Size (Mbytes) 10500
Current LE 10500
Allocated PE 21000
Used PV 13

How should I look this problem in details.

br,
Munawwar

[Vincent Farrugia · Accepted solution]

Hello,

Seems like c0t11d0 is faulty. Try to dd on it:

dd if=/dev/dsk/c0t11d0 of=/dev/null bs=64k

If it does not work, most likely you have a faulty harddisk.

You can also try the mstm tools on it.

I see you have mirroring, so it shouldn't be too much of a problem for resyncing data.

HTH,
Vince

[Peter Kloetgen]

Hi Ahmad,

i think you have possibly a defect disk in your system, because of it's stale status in vgdisplay command. Check first the vgscan- command with -p option. You could also try the following command:

strings /etc/lvmtab

to find out, if there exists an entry for the disk. Please post the output of these commands here.

Allways stay on the bright side of life!

Peter

[Stefan Saliba]

What usually is the case is a failed disk /dev/dsk/c0t11d0.

try using dd if=/dev/dsk/c0t11d0 of=/dev/null bs=64k and see if you can read blocks from the disk or have the disk LED flashing.

Check the syslog for EMS messages and check the logtool for I/O errors under STM.

Usually though it is a failed disk which needs to be replaced. In your case you are mirroring so you can do it on-line without affecting operation if you have a hot-plug disk module.



Good luck
Stefan

[Michael Tully]

It certainly looks like a dead disk. If you have this mirrored with another disk, there should be no problem replacing it. Once you have it replaced you will need to run the 'vgcfgrestore' command. There are examples in the man page. If it is not mirrored, get your backup tape(s) ready.

[Tom Geudens]

Hi,
Looks like you lost a disk (fortunately you've got it mirrored). Verify this with "ioscan -fn". There should be a disk NO_HW. In order to replace the disk, take a look at the procedures described in http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x617e50011d20d6118ff40090279cd0f9,00.html

Basically it will come down to (VERIFY !) this :
1. Replace the disk
2. Resync the data
# vgcfgrestore -n /dev/v01 /dev/rdsk/c0t11d0
# vgchange -a y /dev/vg01
# vgsync /dev/vg01

Hope this helps,
Tom