Mount business copy LUN on same host as parent LUN

[Deborah Davies]

I'm having problems setting up business copy - I'm trying to mount the business copy luns on the same host as the parent luns.

I have done the following to setup the business copy volume group :

mkdir /dev/vg20
mknod /dev/vg20/group c 64 0x1a0000
vgchgid /dev/rdsk/c22t5d1
vgimport /dev/vg20 /dev/dsk/c22t5d1
vgchange -a y vg20
vgcfgbackup vg20

When I create the volume group I can vgchange it and it works, however, when I then deactivate it, copy the parent data to the BC LUN (armcopy -s -a true Array1), then try to activate the volume group it fails with the error :

vgchange: Warning: Couldn't attach to the volume group physical volume
"/dev/dsk/c22t5d1":
Cross-device link
vgchange: Warning: couldn't query physical volume "/dev/dsk/c22t5d1":
The specified path does not correspond to physical volume attached to
this volume group
vgchange: Warning: couldn't query all of the physical volumes.
vgchange: Couldn't activate volume group "vg20":
Quorum not present, or some physical volume(s) are missing.

Lvmtab has the correct disk info for this volume group.

What have I missed? Any pointers greatly appreciated.

[GPMS - Xansa support]

You have to do a vgchgid on the BC VG after every copy as it recopies the vginfo from the parent VG. Then export & re-import

If you have a vgcfgbackup, as you mentioned, you could do a vgcfgrestore instead.

Works ok if you have the same vgid on a different node, the problem only comes about on your example of the same node.

If you are using cluster machines, might be worth doing a vgchange -a e vg20 on the way up.

Hope that's some kind of help.

[Bernhard Mueller]

Guess you should do a vgcfgrestore after every re-sync of the BC LUN before you activate it because the same LVM header which required youre vgchid will be back upon sync.

Regards,
Bernhard

[Sundar_7]

I would NOT do a vgcfgrestore after syncing with the primary.

vgcfgrestore will pick up the old LVM configuration file /etc/lvmconf/.conf which may out of wack with the source volume group.

After syncing the BCV

# vgexport /dev/vg20
# mkdir /dev/vg20
# mknod /dev/vg20/group c 64 0x1a0000
# vgchgid /dev/rdsk/c22t5d1
# vgimport /dev/vg20 /dev/dsk/c22t5d1
# vgchange -a y vg20
# vgcfgbackup vg20

[Prashant Zanwar_4]

We had some similar setup where everytime we used to import the business copy and split again. I remember it very well that everytime we used to do a vgchgid on the physical volume in question then do a vgimport. Usually when you sync the primary with business copy, it's a track to track to track copy of the disk. In my view your primary VGRA & PVRA is getting overwritten when you sync with the Business copy. So option is to change VGID and do a import.
Hope it helps, correct me if I am wrong somewhere.

Thanks and regards
Prashant

[Deborah Davies]

I've set this up a few years ago on another server with a va7100 and don't get this problem, which leads me to believe I've missed something out in the setup.

Does anyone have a guide or manual which details the steps required when setting up business copy - the Business Copy User Guide isn't very helpful/detailed.