Increasing VM guest volume group

[Tim Nelson · Accepted solution]

John,

I have used / tested this in the past with non critical data and it works as advertised. Did this 5-6 times in a VM environment. In my case I did this two ways, increased the physical LUN to the VMHost and also just increased the LV backingstore to the VMGuest.


When working with production data the risk becomes increased and I guess until I have do this 20-30 times I will still be scared of it.

If you read through the man page and doc referenced you can get an idea of what it is doing.

1) modify the vg meta data to change the limiting parameters. In this case max_pe.

2) add the additional mappings or reservations to the VGDA, typically this is created with the size neccessary to support the initial sizing. In most cases when you try to grow the VGDA is not big enough and wants to extend itself into the next extent ( 0 ). Typically extent 0 is used by data, you then have to move that data to the end ( if you have room ), then all the extents need to be renumbered so the 1st extent is back at the pole position.

That is alot of juggling around and if something errors "where did that data go ?", I am sure you can find it and fix it but by now we are wishing we would have just migrated to a new vg.

There was also a post just yesterday with someone working on this. His last comment was it worked !! ( but I lost data, that is ok, it was test data )..??

For me the jury is still out for the vgmodify functionality.

[Torsten.]

This was the thread:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1269993

But the happy end was a bit different ;-)

[John Kelly_3]

Thanks everyone. I'll let you know how I get on once I have the additional LUNs.

[John Kelly_3]

I finally decided there was no advantage in going through the disruption required to grow the existing VM VG with vgmodify. So I just added the new LUN into the VG on the host. Created another LV and added it to the VM guest using hpvmmodify. Then I could vgextend within the VM guest and extend the filesystems that way.
Thanks to all for the many contributions.

JohnK