For those of you interested in why the server was down for so long yesterday, this is the explanation from our hosting company:
A one-off configuration change made previously on the host to increase disk performance inadvertently put the process that controls the starting and stopping of virtual servers into a perpetual state of uninterruptible sleep, so it was no longer listening for start or stop requests. Unfortunately after unsuccessfully trying many things, it was decided that the safest route was to reboot the problem host.
Given the nature of the problem and how we believe it was caused, I don't think this should be an issue again.
A one-off configuration change made previously on the host to increase disk performance inadvertently put the process that controls the starting and stopping of virtual servers into a perpetual state of uninterruptible sleep, so it was no longer listening for start or stop requests. Unfortunately after unsuccessfully trying many things, it was decided that the safest route was to reboot the problem host.
Given the nature of the problem and how we believe it was caused, I don't think this should be an issue again.