[ntp:questions] Re: Refid?

David L. Mills mills at udel.edu
Fri Oct 22 17:00:54 UTC 2004


Richard,

If so configured a server will drop your packet and send a kiss-o'-death 
packet if the interval between packets is less than two seconds or the 
average interval is less than five seconds. Some servers are dropping 80 
percent of offered packets because of flood attacks. I'm talking about 
thousands per second here. There may be other instances where a KoD 
packet is sent; see the list of kiss codes.

Dave

Richard B. Gilbert wrote:
> David L. Mills wrote:
> 
>> Richard,
>>
>> If the association is a reference clock, the refid is the type of 
>> receiver or service. There is no official mapping, but you can guess 
>> what it is. If not a reference clock and unreachable, then the code 
>> should appear in the list of kiss codes in the documentation. The RSTR 
>> means you set off the call gap and the server is telling you to slow 
>> down.
>>
>> Dave
>>
>> Richard B. Gilbert wrote:
>>
>>> I've seen some refids in the output of nptq -p that I don't 
>>> understand.  I can't find any documentation that mentions any of the 
>>> refids.
>>>
>>> .GPS., .WWV., .CDMA. are all clear enough but what are:
>>>
>>> .PSC. and .RSTR.
>>>
>>> I actually found references to .RSTR. in the code but the 
>>> accompanying comments mentioned both "restricted" and "restrained", 
>>> neither of which are mentioned in the documentation.
>>>
>>> Two public servers that used to work are now displaying ".RSTR." as 
>>> refids.  Are they trying to tell me something?  One used to be a 
>>> pretty good server until something happened to it a couple of weeks 
>>> ago and it started using a server several thousand miles away.  The 
>>> other never kept very good time and I won't miss it much.
>>>
>>
> Hi Dave,
> 
> What's this "call gap"?   What might I do to not set it off?
> 
> As of about ten minutes ago one of the servers is still on the public 
> server list (stratum 2).  I haven't checked the others yet.
> 




More information about the questions mailing list