[ntp:questions] Re: time reset +492.958589 s

Jörgen Tegnér jorgen.tegner at telia.com
Fri Jun 17 17:23:54 UTC 2005


On Fri, 17 Jun 2005 07:47:18 +0100, David Woolley wrote:

> In article <pan.2005.06.16.20.20.29.18287 at telia.com>,
> jorgen.tegner at telia.com wrote:
> 
>> Jun 16 12:16:34 burken ntpd[1834]: time reset +492.958589 s
> 
>> Can anyone suggest possible causes?
> 
> Severe clock interrupt overruns (e.g. using IDE drivers heavily on a
> machine with HZ set to 1000 - see many recent articles).  Less likely,

At the time the kernel had some extra stuff compiled in. After
removing that, the large time resets have gone.

> given that it is Linux, some other software competing to set the clock.
> 
> You shouldn't be using a local clock driver on this machine as it will
> reduce the ability of the protocol to signal downstream that it has a
> problem.  Local drivers should always be added as the result of an
> informed choice, never as a default; it is usually better not to
> configure them.

OK, removed it from the config. No more hopping, but lost contact with the
server assigned by pool.ntp.org instead a few times. Replacing
server pool.ntp.org
with
server se.pool.ntp.org
server dk.pool.ntp.org
server no.pool.ntp.org
server de.pool.ntp.org
in the config helped. 

> 
> In any case, you should not be using this machine as a time server (and
> that is the only situation in which a local clock driver is sometimes a
> valid choice).
> 
> The hopping between remote and local clocks is perfectly reasonable,
> given that it's lost so much time.


Thanks,
Jörgen
-- 
I'm kinda looking forward to the alzheimer's... I'll get to meet new
people every day. I'll never have take the blame for anything. I'll get
to meet new people every day.




More information about the questions mailing list