The server keeps losing/regaining sync. Here is the status from xntp showing
an interesting 30 minute pattern:

12 Jul 20:19:01:synchronized to x.x.x.9 stratum=3
12 Jul 20:19:01:synchronization lost
12 Jul 20:21:43:synchronized to x.x.x.9, stratum=3
12 Jul 20:22:13:synchronized to x.x.x.99, stratum=2
time reset (step) -8.991312 s
12 Jul 20:49:17:synchronized to x.x.x.9, stratum=3
12 Jul 20:49:17:synchronization lost
12 Jul 20:51:59:synchronized to x.x.x.9, stratum=3
12 Jul 20:52:29:synchronized to x.x.x.99, stratum=2
12 Jul 21:15:23 Writing -3.597 to drift file sys:\system\drift.ntp
time reset (step) -8.952683 s
12 Jul 21:19:32:synchronized to x.x.x.9, stratum=3
12 Jul 21:19:32:synchronization lost
12 Jul 21:22:14:synchronized to x.x.x.9, stratum=3
12 Jul 21:22:44:synchronized to x.x.x.99, stratum=2
time reset (step) -8.716475 s
12 Jul 21:49:49:synchronized to x.x.x.9, stratum=3
12 Jul 21:49:49:synchronization lost
12 Jul 21:52:31:synchronized to x.x.x.9, stratum=3
12 Jul 21:53:01:synchronized to x.x.x.99, stratum=2

Here is the ntp.conf file from the vm'd server:
# server
# fudge stratum 3
server x.x.x.9 minpoll 4 maxpoll 4 prefer
server x.x.x.99 minpoll 4 maxpoll 4
# noncp
# Step the time to the source clock for slewing
driftfile sys:\system\drift.ntp
logfile sys:\system\ntp.log

And the output of the peers from ntpq on the bad vm'd server:

remote refid st t when poll reach delay offset disp
================================================== ============================
*x.x.x.99 time1.srv.ualbe 2 u 4 16 377 0.44 -5526.1 226.50
+x.x.x.9 sylvester.sraq. 3 u 2 16 377 0.50 -5675.2 300.37

Those offsets seem rather large, but I don't know what is acceptable or
within tolerance; just that they are much larger than the other servers..and
it grows higher - I assume until the offset goes beyond some value which is
when the system reports that timesync is lost. Rinse repeat.

I've done some research into the time issues with vm's (not just NetWare)
and I think I've got it all setup per recommendations. But it seems that
this is mostly a black art rather than a science and basically folks are
just getting lucky with one config or another working - for most servers
most of the time. But this is not really an acceptable situation to me -
doesn't really hold up the Enterprise Software! banner too proudly...sigh.

So if any of you generous guru's/wizard's have a magic formula that you
would like to share with me... well I'd really appreciate it.



PS Yes the vmware tools are installed and the synctime option is set off.