I am moving my virtualisation from WS2003 / VS2005 / Dell 2950 to WS2008R2 /
HyperV / Dell T710 and looking to see what new features I should take
advantage of.
Servers have local disk for OS (RAID-1 + Swap), and FC SAN for guests. I am
not interested in clustering as the redundancy in the T710s coupled with the
use of MS-DPM for backup is enough for my SLAs.
I’m mainly interested in whether to use disk pass through. Currently I have
presented a reasonable sized disk to the host (eg 100-300GB) and then have a
number of VHDs on this one disk. I have used expanding disks where I might
setup the disk with a size of 40GB, but it only has a VHD of half that (i.e.
reasonably efficient). I monitor the free space on the guest and host and can
increase the size of the host disk on the fly using the SAN and DiskPart. If
a guest needed more space I would shut the guest, create a new larger VHD,
and use something like ShadowProtect to copy from the existing VHD to the new
VHD.
In HyperV I have the option of pass through. Today I experimented with
presenting a 20GB disk to the host and then created a guest and told it to
use the physical disk. I built this as 2008-x86 and all good. I then
increased the size of the disk to 25GB, but now in disk management I have a
25GB disk that shows as a 20 + 5. I tried extending with DiskPart, but I
wasn’t able.
So if I decided to have a separate pass through disk for each guest, how can
I allocate more space to them?
I gather there are performance benefits, but there seems to be too many
negatives to make this worthwhile? Does it sound wise to go down this path?
Thanks
Mick
--
Charlie.
http://msmvps.com/blogs/russel
"Mick" <Mi...@discussions.microsoft.com> wrote in message
news:96587744-90C6-4D44...@microsoft.com...
Charlie, thanks for your feedback. I don't think I will use pass through at
this stage. I do have a server with 1.3TB of data that down the track I might
do P2V and will consider it for that as it souns a bit scary to have such a
big VHD file.