massive time drift. NTP help required

Mark Hutchinson markhsa at
Fri Jun 1 12:42:50 UTC 2007

Any idea what causes this?  Hardware issues?
Where does the 10195 number come from?

I just like to understand what the issue cause is and what the command
actually does.

Thank you for the assistance.


On 6/1/07, Dr. Michael J. Chudobiak <mjc at> wrote:
> Mark Hutchinson wrote:
> > I have a server that has a time drift of around a minute every hour.
> > Overnight, that is pretty extreme and as this is a PVR, it causes
> problems.
> > I have NTP enabled but am still getting this drift and NTP is not
> > correcting or putting much in the logs ( no exit messages )
> I had the same problem on a brand-new DC7700. ntp can only correct
> within 500 ppm (0.05%). For some reason, the DC700 clock was 1.95% too
> fast!
> To fix this, I added this to /etc/rc.local:
> /sbin/adjtimex --tick=10195
> This adjusts the "tick rate" at every boot-up. Once the basic accuracy
> is corrected by this, ntp can fine-tune it exactly.
> - Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the fedora-list mailing list