vgimport

[Court Campbell]

> and this is faster than using the -s option. > what happens if I use the wrong device file

You want to use the -s on the vgexport so that it puts the vgid in the mapfile. then the vgimport will get the vgid and scan the disks looking for that vgid on the other host. the vgid is stored in the vgra is what will let each server know which disks belong to which volume groups.

[Steven Boyer]

Yes I have those instructions and executed them but, this process was taking to long and I was looking for an alternative to using the -s option. We have ~50 VG, the last guy to import a VG on our system said it took over 5hrs at that rate it will take over 3 weeks to import all our VG.

[Court Campbell]

Sorry, but I have never had a vgexport/vgimport take that long. But then again I have never had that many vg's. I assume the vgexport runs pretty fast, but the vgimport is what is taking so long. How many disks are presented to the host?

ioscan -fuC disk | wc -l

will give you a close figure. I guess if it has to scan a lot if disks that could take a while.

[john123]

What storage you are using.?
There would be some storage specific commands to find out the device name of the disks in the second node which corresponds the disk device name in the first node.

And then use -f option with disk file name while vgimport in the second node instead of using -s option
It will not scan all your disks and so it will be faster.



HTH
John

[Steven Boyer]

using the -f options on the partner isnt that using the device number as seen by the primary and not the partner.

Its my understanding that the partner needs to have the list of devices as seen by the partner not the primary.