Hi Per, We have ben using Xen for a couple of customer sites over the last 6 months (using SLES 10 sp1) and quite satisfied with it. For all these setups we are using EMC SAN's combined with Dell PowerEdge 2950 making the setup redundant. We don't use disk files but hand out LVM partitions to each guest. Also using heartbeat clustering to manage the Xen hosts with the guests defined as resources in the cluster. This document gave us a good head start : http://www.sandervanvugt.com/heartbeat/xenhadoc.pdf And also check this page for more info : http://www.sandervanvugt.nl/en/aankondigingen/heartbeat-documentation-0 The only real drawback we experienced is that Windows guests ACPI support is a bit unfinished on Xen hosts running 64 bit OS and Live migration is not supported for Windows. This is due to the fact that SLES is using a patched Xen v3.04 and Windows live migration is only supported starting Xen v3.1. A workaround is to pause the Windows guests, and resume them on the other host, resulting in a small delay. Other than that it's been very stable and the performance good. Also I would definitely recommend going for the 64bit host OS version as we found it gives better performance and memory handeling. If you are looking for a more manageable centralized interface and want to avoid the command line (vmware style) you could also have a look at XenSource. It is the most manageable interface I know but on the other hand I expect there will be some very useful management interfaces coming to the other distributions this year. If you are using Novell products you could also have a look at Zenworks Orchestrator , http://www.novell.com/linux/virtualization/ Cheers, Wj On Fri, 2008-04-18 at 18:50 +0200, Per Jessen wrote:
I noticed the creation of this group earlier today and signed up right away - I've been meaning to play around with xen and such for a while, but sofar I've only made it to vmware and virtualbox for playing a bit with Windows.
What I sometimes think about is - do any of you guys use virtualisation in production, and if so, are you using hardware with suitable redundancy? I can't help thinking that a machine with a 20 virtual systems will take town 20 servers when it suffers a hardware problem?
/Per Jessen, Zürich
-- To unsubscribe, e-mail: opensuse-virtual+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-virtual+help@opensuse.org