Opps,
I figured it out. Turns out that I had a lousy
/etc/adjtime file. This file is used and set by "hwclock"
to track drift (inaccuracy) in the HW clock. Turns out
that I must have run "hwclock" twice in quick succession
durring a brief peroid of major drift in the clock.
"hwclock" assumed my HW clock always did this, and built
a adjtime file to compensate. After clearing that file,
doing an "rdate ntp2.usno.navy.mil; hwclock --systohc --utc"
followed by a reboot (to see if this all works) shows
a MUCH more realistic time.
Bryan
