[ntp:questions] Re: time reset, syncronisation lost and LargePPM values

David J Taylor david-taylor at blueyonder.co.not-this-bit.nor-this-part.uk.invalid
Thu Nov 3 09:00:03 UTC 2005


Danny Mayer wrote:
> Martin Burnicki wrote:
>> David,
>>
>> David J Taylor wrote:
>>
>>
>>> Tom Smith wrote:
>>> []
>>>
>>>> It's losing time, which could suggest lost interrupts. However,
>>>> the steps every half hour make me wonder if your server is doing
>>>> the stepping. The fact that 2 different clients have the same
>>>> problem in any case suggests a common cause. What is/are the
>>>> server(s)? It's not a Windows "NTP server", is it?
>>>
>>> There's nothing wrong with a properly configured Windows system
>>> running as an NTP server, with the distributed NTP software.  If
>>> you meant to say that running the Microsoft "W32time" software does
>>> not provide a fully compliant NTP service, then I would completely
>>> agree with your doubts!
>>
>>
>> Windows *is* a lousy timekeeper. We've had a customer where the
>> Windows system clock has lost about 30 seconds (!) whenever a
>> scheduled job was run to deal with their databases.
>>
>> Martin
>
> Is this with using the MM timers?
>
> Danny

If it is, I would have to ask what programmer wrote a database system that 
used MM timers!  What are they playing at?  If the databases were on PIO 
and not DMA disks, what does that say about the setting up of the system?

BTW: I've not seen the MM timers issue on Windows XP since (I think) 
Service Pack 2 was installed.

David 





More information about the questions mailing list