[ntp:questions] Flash 400 on all peers; can't get ntpd to be happy

unruh unruh at wormhole.physics.ubc.ca
Tue Mar 8 19:05:48 UTC 2011


On 2011-03-08, Miroslav Lichvar <mlichvar at redhat.com> wrote:
> On Tue, Mar 08, 2011 at 05:00:44PM +0000, unruh wrote:
>> >>>> filtoffset= 67671.8 66534.8 65931.3 65118.0 63317.3 63029.5 62216.4 58156.6,
>
>> Not at all sure how Mills comes into the picture. On a system where the
>> frequency fluctuates wildly, ntpd is not the right answer, nor is any
>> system. I suspect that the best you could do would be to run something
>> like ntpdate often and jump the clock around.
>
> The frequency offset in this case seems to be around 2% which is still
> well below the 10% maximum Linux can adjust. I'd try chrony before
> resorting to ntpdate, the timekeeping probably won't be very good, but
> at least the clock won't be stepped.

The problem on a VM system is that the frequency jumps around. Ie, when
the VM is running, its frequency should be very close to the fundamental
clock frequency, and when it is not running, its freq is 0. Thus the
time is a staircase, with the steps depending on how busy the VM is vs
how busy the other stuff on that computer is. This means that chrony's
frequency estimate also jumps around like mad (and means it is almost
always down around its "3data point" level in estimating frequency). Now
it certainly would be capable of adjusting a 2% frequency shift, if that
were consistant, but I am not sure how it would behave with the
inconsistant type of jumps you get in a VM. But I agree it would be
worth trying.


>




More information about the questions mailing list