We have recently rolled out 4 Netware 6.5sp7 servers virtualized in XEN
with the dom0 being SLES10SP2 64bit. On each of the servers, we get a
hang about once per week that causes the server to lose communication
and make the Xen VNC console and NRM inaccessible. When this happens, we
can see the CPU usage on the Virtual machine manager on dom0 pegged at
25% (on a quad core machine.)

Some details:
- Each server runs basic file and print, plus Zenworks 7.5 and Groupwise
- These are on two different underlying hardware platforms. Each is on a
separate physical server.
- Three of the servers are being actively used, the third is being
prepped for a new deployment, so it has no clients connecting to it.
That seems to eliminate usage patterns as causes
- The hang seems to happen more often at night, but is not correlated to
any particular operations. Initially, I thought this might be due to
scheduled file compression, but that is now disabled and the issue
- The wayward VM cannot be shutdown from the virtual machine manager,
and we must kill it with "XM destroy"
- Two of these servers are totally new setups with eDir 8.8, two are
migrations running eDir 8.7.3 via the instructions in TID 7000096
- The servers have all of the most current NW patches listed on
downloads, except for iprntnw65sp7b, and the appropriate eDir 8.8 SP2
FTF2 or eDir 8.7.3 SP10b

Any help or suggestions would be appreciated. The lack of ability to
reach NRM or look at the console is making troubleshooting via the
standard methods pretty much impossible.