These two patches should be what we need to put the check mark next to
ComputerSystemIndication. The first deals with a potential disaster that is essentially a
race condition somewhere in the libvirt/xend region. The second adds the Modified trigger
to VSMS and, since our usage model relies much more on the trigger than the polling loop,
increases the event loop wait time to three minutes (feel free to throw out any other
suggestions for that number).
Comments are welcome on the first patch; I think it's a pretty good solution but if
anyone thinks it can be improved we can certainly talk about it.