One mishap aside the migration is steadily marching on. Last night the web server VM was cut up into a bulk file storage VM and a dedicated web VM. So far today DNS services have been migrated to their own virtual machines. By switching to sparse files for the disc images we reclaimed an extra 12G in storage. Now that DNS is being served from different VMs that can be placed on different physical servers its downtime is independant of anything that happens to the web server VM, helping to ensure consistent mail delivery even if the sites are down. The mishap was losing one of the physical servers. After performing upgrades to xen and the dom0 kernel it was rebooted and failed to come back up. It's the only server that does not have out of band management and isn't plugged into an IP KVM. Figures. At any rate I was planning ahead and no VMs were running on the machine at the time. Unfortunately until that server can be fixed our processing capacity is cut by about 25%

 

Anyway, this is an oldie but I just discovered it and I lubbers it http://www.youtube.com/watch?v=7bL-yULlNF8: