Inconsistent vg00 after migration

[Mario_88]

Hello:

I don´t know if you remember me but recently I opened a post about migrating from 11i to 11.23. Finally
I make the following:

1º Split the mirror using lvsplit
2º Upgrade to 11.23 in one disk left the other untouched

Now everytime I try to create a lvol I get the following error:

ras1#cat pvdi isplay.e err
pvdisplay: Warning: couldn't query physical volume "/dev/dsk/c1t2d0":
The specified path does not correspond to physical volume attached to
this volume group
pvdisplay: Warning: couldn't query all of the physical volumes.
pvdisplay: Warning: couldn't query physical volume "/dev/dsk/c1t2d0":
The specified path does not correspond to physical volume attached to
this volume group
pvdisplay: Warning: couldn't query all of the physical volume

The same with vgreduce, if I move /etc/lvmtab and run vgscan -v I get the following message:

-----
ras1#vgscan -v
Creating "/etc/lvmtab".
vgscan: Couldn't access the list of physical volumes for volume group
"/dev/vg00".
Couldn't stat physical volume "/dev/dsk/c3t2d0"://*****???(DVD drive)
Invalid argument

vgscan: has no correspoding valid raw device file under /dev/rdsk.
Verification of unique LVM disk id on each disk in the volume group
/dev/vg00 failed.

Scan of Physical Volumes Complete.
----------------------

I think the best thing I can do is to recreate vg00
---------
. Boot up the server, and interrrupt the boot process at the first available chance.

2. Boot off your d2 disk - this will have the same hardware path as the old disk, except for ending in .2 instead of .1

3.When asked the question 'Interact with ISL?' answer Y.

At the ISL prompt type:

hpux -lm

4. Remove the existing root volume group configuration:

vgexport vg00

5. Re-create the root volume group:

mkdir /dev/vg00
mknod /dev/vg00/group c 64 0x000000

vgimport /dev/vg00 /dev/dsk/cXtYdZ

obviously with the correct path to the disk.

6. Activate the volume group and re-create links for boot:

vgchange -a y vg00
mount -a

lvlnboot -b /dev/vg00/lvol1
lvlnboot -r /dev/vg00/lvol3
lvlnboot -d /dev/vg00/lvol2
lvlnboot -s /dev/vg00/lvol2
lvlnboot -R /dev/vg00

------------------------------

Any other ideas?

** Obviously I have learned the lesson and I will cold install my production systems.

Regar

[Thayanidhi]

Your idea will work. Before that, how many disks vgdisplay -v reports (cur and act)?
Is PV are available?

Since you split the LV's the old PV is still part of the Vg configuration. see man page of vgreduce to remove the PV forcebly from a VG. The PV should be free or in-accesible first!

Regds
TT

[marie-noelle jeanson_1]

Mario,

I take this opportunity to advertise a brand new white paper published by LVM on docs.hp.com. There are actually two of them, one for LVM online disk replacement, one for dealing with disk failures and replacements under LVM. They are under http://docs.hp.com/en/oshpux11i.html#LVM%20Volume%20Manager
Or look directly at http://docs.hp.com/en/5991-1236/When_Good_Disks_Go_Bad.pdf

Hopefully you will find some good tips in there,

Marie-Noelle