Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am trying to find the best way to migrate my VHD's to my DataCenter offsite for DR purposes.

Currently I have a number of Xen Stacks setup as the following:

  • Dual RAID ISCSI SAN multipathed to at least two Xen Servers

Below is an example diagram:

enter image description here

One option I have been testing and researching is:

  • Install CentOS onto the SAN's
  • Installing Gluster onto the SAN's
  • Using UCARP to create a VIP to use to attach the NFS storage to Xen
  • Using Gluster's geo-replication functionality to replicate this data to another Gluster node stored offsite.
  • Placing all the Xen Servers into a single stack in the office to give me alot more RAM to play with.

enter image description here

I have managed to get this setup and it is working great, however I am very concerned about Disk I/O. Testing a disk created on a NFS share in Gluster I seem to be getting 10MB/S write and about 5MB/S read, which seems quite slow. It is quite important to have high disk I/O as most of the VM's will be running SQL servers.

One important factor is that my Dev equipment only has slow disks (7k) so I am aware this might be effecting the speed a little bit.

What better methods might there be in achieving a redundant storage system for my VHD's?


Well, I migrated my gluster cluster onto some decent hardware with some decent disks and sequential write has increased to 100MB/S and read to 80 MB/S, I seem to loose about 30-40MB/S than running direct from iSCSI storage but I guess this is to be expected with NFS.

share|improve this question

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.