disks moved to different host without vgexporting first

[paul_xedos]

Hi, 

I'm not sure if this is terminal but disks from an hpvm were moved from one physical host to another without vgexport

One disk has made it across  from volgroup vg00 and so the machine boots

The remaning 2 disks have come across with what appears to be different target addresses and so is not mountable and I can not activate the volume group

Here's the disks I now see

bash-4.0# ioscan -fnC disk
Class I H/W Path Driver S/W State H/W Type Description
=====================================================================
disk 1 0/0/0/0.0.0 sdisk CLAIMED DEVICE HP Virtual Disk
/dev/dsk/c0t0d0 /dev/dsk/c0t0d0s2 /dev/rdsk/c0t0d0 /dev/rdsk/c0t0d0s2
/dev/dsk/c0t0d0s1 /dev/dsk/c0t0d0s3 /dev/rdsk/c0t0d0s1 /dev/rdsk/c0t0d0s3
disk 0 0/0/0/0.1.0 sdisk CLAIMED DEVICE HP Virtual Disk
/dev/dsk/c0t1d0 /dev/dsk/c0t1d0s2 /dev/rdsk/c0t1d0 /dev/rdsk/c0t1d0s2
/dev/dsk/c0t1d0s1 /dev/dsk/c0t1d0s3 /dev/rdsk/c0t1d0s1 /dev/rdsk/c0t1d0s3
disk 4 0/0/0/0.2.0 sdisk CLAIMED DEVICE HP Virtual Disk
/dev/dsk/c0t2d0 /dev/dsk/c0t2d0s2 /dev/rdsk/c0t2d0 /dev/rdsk/c0t2d0s2
/dev/dsk/c0t2d0s1 /dev/dsk/c0t2d0s3 /dev/rdsk/c0t2d0s1 /dev/rdsk/c0t2d0s3

Here's what lvm expects to see

bash-4.0# strings /etc/lvmtab
/dev/vg00
/dev/disk/disk3_p2
/dev/vg01
/dev/dsk/c0t2d0
/dev/dsk/c0t3d0

Here's what happens if I try to activate the volume group

bash-4.0# vgchange -a y vg01
vgchange: Warning: Couldn't attach to the volume group physical volume "/dev/dsk/c0t2d0":
The path of the physical volume refers to a device that does not
exist, or is not configured into the kernel.
vgchange: The path of the physical volume refers to a device that does not
exist, or is not configured into the kernel.
vgchange: Couldn't activate volume group "vg01":
Quorum not present, or some physical volume(s) are missing.

vgchange: Warning: couldn't query physical volume "/dev/dsk/c0t2d0":
The specified path does not correspond to physical volume attached to
this volume group
vgchange: Warning: couldn't query physical volume "/dev/dsk/c0t3d0":
The volume group is not activated
vgchange: Warning: couldn't query all of the physical volumes.

I've tried moving lvmtab out of the way and re-running vgscan -a in the hope that it recreates /etc/lvmtab to the new disk arrangement but no good

Any pointers would be appreciated

Regards

Paul

[Ralf Seefeldt]

Hi,

you could rename /etc/lvmtab and than rum vgscan.

Do you already have the devicefiles for the disks? (#ioscan -funCdisk)

I hope, this helps.

[paul_xedos]

Hi Ralf,
Thanks for the suggestion but I tried moving lvmtab and re-running vgscan. Device files exist for the disks ( c0t1d0 and c0t2d0)
Cheers
Paul

[Patrick Wallek · Accepted solution]

What you probably need to do is to go ahead and export VG01 to get rid of it.  The recreate the /dev/vg01 directory and group file and use vgimport to import the disks necessary.

For example:

# vgexport /dev/vg01

# mkdir /dev/vg01

# mknod /dev/vg01/group c 64 0x010000

# vgimport /dev/vg01 /dev/dsk/c?t?d? /dev/dsk/c?t?d?

# vgchange -a y /dev/vg01

Make sure you use the /dev/dsk files as appropriate for the VG.

[paul_xedos]

Hi Patrick,

Superb - thanks very much for the reply.

Can I just check - do I need to flip the disks back to the original box or is it safe to do these commands on the box where they have now been imported?

Thanks

Paul

[Patrick Wallek]

You can do this on the server that the disks now reside in.

[Steven E. Protter]

Shalom,

No its not a game ending issue. You just don't have a map file with which to import the disks to the new host.

If you have an old SysInfo report or other documenation you can still manually import the disks to the new host.

We are only talking about 4 disks or so here.

vgimport with no -m mapfile directive and indicating the name of the volume group and the physical volumes to be imported should work. There may be some trial and error involved.

Note that customized naming of logical volumes in the old vg01 volume group will be lost. If for example you had a logical volume named shmuelbagel it will be imported as lvol0 if its the first logcal volume you import.

http://nixdoc.net/man-pages/hp-ux/man1/vgimport.1m.html

Specifically

vgimport -v /dev/vg01 /dev/dsk/c0t1d0 /dev/dsk/c0t3d0

Regards,

SEP

[paul_xedos]

Thanks for the help with this one. Things have progressed. I'm almost there but ...

bash-4.0# vgimport /dev/vg01 /dev/dsk/c0t1d0 /dev/dsk/c0t2d0
Physical Volume "/dev/dsk/c0t2d0" contains no LVM information

bash-4.0# vgimport /dev/vg01 /dev/dsk/c0t1d0
vgimport: Warning: Volume Group belongs to different CPU ID.
Can not determine if Volume Group is in use on another system. Continuing.
vgimport: Warning: Volume Group contains "2" PVs, "1" specified. Continuing.
vgimport: Volume group "/dev/vg01" has been successfully created.
Warning: A backup of this volume group may not exist on this machine.
Please remember to take a backup using the vgcfgbackup command after activating the volume group.

I was wary of activating the volume group so decided to export the volume group again until I can understand what is going on
bash-4.0# vgexport /dev/vg01

[paul_xedos]

Sorry chaps, I'll try and make this the last question

If I import with just the one disk and activate the group is this situation reversible without too much damage. ie vgchange -a n vg01 and then vgexport again. For some reason it does not allow an import of 2 disks but can handle 1 with a warning that there should be 2

[Patrick Wallek]

Both of the disks you mention have had 3 slices carved out of them.  This usually indicates that they are boot disks.

Typically slice 2 (s2) is the slice that contains the LVM data.  I would try:

# vgimport /dev/vg01 /dev/dsk/c0t1d0s2 /dev/dsk/c0t2d0s2