This message was deleted.
# citrix-vad
s
This message was deleted.
p
Clones are normal behavior.
j
So based disk is clone for each VDA ? Citrix has not optimized its own Hypervisor !?! I use 100GB on my master and 100GB for diff disk so each use 200GB ! Wow...
p
Not quite. It’s down to how you have your storage configured.
For Example.
That particular store has a couple hundred MCS win10 machines.
j
You use Thin Provisioning ? My storage is mount by hba fiber channel and does not allow thin provioning
p
Our storage is all hardware HBA’s, and for the MCS shared storage we are using GFS2 thin provisioning, yes. I’m interested in other replies, as i’m not sure we are doing it the best way. To your initial question, if you have HBA’s running LVM then the storage will appear to fill quickly - even if on the back end storage device you have thin provisioning enabled. Xen Server assumes each is a full copy, not a thinly provisioned clone.
j
As I understand, GFS2 is only available for cluster mode and not for Resource Pool. So I have to understand the pro and cons of cluster mode
Thanks for your help
p
FWIW we tried using NFS. It did not go well….😱 So GFS2 was the only viable alternative in our environment.
l
If you can, put some local storage in the hosts and just format them as EXT3 datastores. Using GFS2 will give you the same benefits of thin provisioning though. At that points it's more space efficient than vmware.
j
Thanks guys for your help. 😉
👍 1
j
We're using NFS with thin provisioning, but since the move to Windows 10, we have more than doubled our space requirements. It is now ~40GB per user. Even if thin provisioned works, and since Windows 10, it is not so great in term of space saved (5,298.7 GB used for 7,637.8GB allocated). I am interested in this GFS2, we have more fancy block storage available than NFS even if I have no HBA at the moment on my hosts. Last time I checked it was experimental feature (on 8.2 it seems production), and still in Citrix doc, they recommend NFS over GFS2. Anyway looks like some of you use it in production, would you mind sharing like nb of VMs, space utilisation / allocated, how it running? version of XS. Thanks!