On an oes1 (vm guest on sles10 sp2 host), time is running fast. Hence, timesync is not happy.

Tried the clock=pit thing, no help. Have it looking to an external NTP source. That works, but immediately starts drifting. It puzzles me that "ntpq -p" shows a poll cycle of 64 seconds, but it never seems to reset the time, as seen by saying "date" at the console.

What else is there to try? How can I know if clock=pit is actually taking effect?