[LinuxPPS] So far, so good

[Miroslav Lichvar]

On Mon, Aug 09, 2010 at 11:20:50AM -0700, Hal V. Engel wrote:
> On Monday 09 August 2010 09:49:44 am Paul wrote:
> > Thanks for this and from Hal. But still how can I see drift of less than
> > 1 microsecond? The peers command in ntpq is only measuring microseconds
> > - what else do you use?
> 
> You are correct the ntpq only shows offset down to microseconds ntptime should 
> show you nanosecond offsets.  

You can get the offset in nanoseconds shown by ntptime also remotely
with ntpdc -c kerninfo. But it's the remaining kernel PLL offset,
which will be slightly smaller that the original offset depending on
how long from the last clock update it is queried. The only way how to
get the offset exactly seems to be enabling loopstats in ntp.conf and
read the third column from the file.

> My system once it has settled down will typically 
> show offsets in the +-250 nano seconds range when I query it with ntptime.  If 
> I increase the system load (by running a large build for example) the offsets 
> will increase and then settle down again once ntp has adjusted for the higher 
> latency of the interrupt handler.

Isn't that a change in frequency caused by increased temperature
rather than increased latency?

A shorter polling interval might help in such situations, but it will
probably make things worse when idle.

-- 
Miroslav Lichvar