Unable to add new disk to one VG, but able to add to other VG

[Florian Heigl (new acc)]

Don't tell anyone, but I've had 3 or 4 cases where pvcreate -f wouldn't 'wipe it all out'.

The error was the same You got.

I came to first doing a dd over the first 10 mb of the disk or so and doing a regular pvcreate after that.

You might want to give it a try.

[Stf]

# pvcreate -f /dev/rdsk/cxtxdx
# vgextend /dev/vgxx /dev/rdsk/cxtxdx
and then try lvextend -m 1 etc...

Stf

[Fred Ruffet]

What about using this :
vgscan -v

Regards,

Fred

[Todd McDaniel_1]

quote "i'm trying to add to the same old VG. but i've removed the entry related to this bad disk before the hdd replacement."

I believe this was your initial problem.

I would do as suggested with vgscan after moving your lvmtab out of the way to see if that clears your problem.



IF that doesnt work, you will need to use the VGREDUCE command with options related to the pvkey to fix your VG.


Here from the man page for VGreduce...
==========================================

Force reduction of missing PVs from volume group: vg01
vgreduce -f /dev/vg01
Removes the physical volume from the volume group when it is still in
the lvmtab file, but it is currently marked as missing from the volume
group: vg01
vgreduce -l /dev/vg01

The following messages will appear after missing PVS has been removed
successfully:

PV with key 0 successfully deleted from vg /dev/vg01

Repair done, please do the following steps.....:

1. Save /etc/lvmtab to another file.

2. Remove /etc/lvmtab.

3. Use vgscan -v to recreate /etc/lvmtab.

4. NOW use vgcfgbackup(1M) to save the LVM setup.

[Todd McDaniel_1 · Accepted solution]

First ignore my previous post, I failed to read the whole post first...

Next, This is NOT meant as a criticism, but In the future It is better not to remove any disks from the VG in cases like this. Whenever I have a bad disk I usu hotswap it and then run 3 commands. vgcfgrestore vgchange -ay vgXX and vgsync.


To your problem, I believe you need to use pvremove on the device to completely remove any lvm header from it. then issue pvcreate -f /dev/rdsk/cXtXcX on the disk.

also, I think it would be good to do some maintenance on the VG first. Get the VG stable if it is not already. Then make sure you have wiped the disk properly.

Then add the disk to the original VG.

[A. Clay Stephenson]

Your fundamental problem (if I am reading all this correctly) is that you should have never done a pvcreate at all. If this was a replacement of an existing PV in an existing VG, the correct initial command was vgcfgrestore. Then a "vgchange -a y /dev/vgxx" and finally (if the LVOL's were mirrored) a "vgsync /dev/vgxx". If the LVOL's were not mirrored then you should have created new filesystems, mounted them, and restored from backup.

[Madhu_17]

How to remove a PV from a disk.

here my problem is i could able to add the same disk to other VG successfully. but i'm facing problem with adding to one particular VG.

[Fred Ruffet]

You can use pvremove command to remove data structure from disk. This will make disk clean.

After doing that, other point is to know wether PV is still affected to VG. I believe it's already asked but I can't see it : Can you give output for vgdisplay -v vg02 (I believe it's vg02) ?

If you can't see this pv in vg structure, then it's OK. Restart from beginnning of process : pvcreate, vgextend, lvextend.

Regards,

Fred

[Madhu_17]

After adding this problematic disk to the vg02 vg, it's showing stale in lvdisplay 'n' pvdisplay also.
Allocated PE: 1000
Stale PE: 1000

but when i add thid disk to the other VG, it's working fine.

allocated PE: 25
stale PE: 0

[Madhu_17]

I am using HP-UX 10.20.

I can't see pvremove in my version.

[Fred Ruffet]

Even if you don't have pvremove, you will have ways to drop the structure (by recreation with pvcreate).

First of all, please provide output of "vgdisplay -v vg02"

and tell me if you issued the "vgscan -v" command. If yes tell me what is output.

Regards,

Fred

[Madhu_17]

now i detached back my problematic disk.

now the vgscan -v o/p is:


/dev/vg02
/dev/dsk/c2t2d0

vgdisplay -v shows only one working disk now.

if i add this problematic disk, it's also visible in vgdisplay.

[Fred Ruffet]

Do not add it "as is". First do a pvcreate on this problematic disk. If necessary, use -f option. Then try to vgextend vg02 with this new PV.

Rescan vgscan -v

Post output for all those commands.

Regards,

Fred

[Madhu_17]

#pvcreate -f /dev/rdsk/c3t4d0
Physical volume "/dev/rdsk/c3t4d0" has been successfully created.

#vgextend /dev/vg02 /dev/dsk/c3t4d0
Volume group "/dev/vg02" has been successfully extended.
Volume Group configuration for /dev/vg02 has been saved in /etc/lvmconf/vg02.conf

#vgscan -v

/dev/vg02
/dev/dsk/c2t2d0
/dev/dsk/c3t4d0

[Fred Ruffet]

And now, what about lvextend whith following command :
lvextend -m 1 /dev/vg02/u02 /dev/dsk/c3t4d0

What about output ?

Regards,

Fred

[Madhu_17]

Fred,

I could able to do that lvextend successfully.

Pls find below the o/p

lvextend -m 1 /dev/vg02/u02 /dev/dsk/c3t4d0
The newly allocated mirrors are now being synchronized. This operation will
take some time. Please wait ....
lvextend: Couldn't re-synchronize stale partitions of the logical volume: I/O error
lvextend: Couldn't resynchronize logical volume "/dev/vg02/u02".

lvdisplay showing stale for this disk.

[Fred Ruffet]

So all this thread to get to this conclusion : Your disk is down :-(

Regards,

Fred

[Todd McDaniel_1]

Madhu,

I dont see where you ever did a vgcfgrestore to restore the lvm info to the disk.

You MUST do this as a step to restore the mirror.

====================
#vgcfgrestore -n /dev/vg02 /dev/dsk/c3t4d0
====================

You cant resolve the stale extents till you do this. After which you can then vgsync the disk and then add it back in as the mirror.

[Todd McDaniel_1]

Oops I meant to say, vgsync vg02 then lvextend the mirror.

[Madhu_17]

aa