[ILUG] ntptrace strangeness

[Stephen Shirley]

Niall O Broin wrote:
> 193.120.10.3 is ntp0.esat.net, and I get that server as it's one of the  
> hosts which will be returned for ie.pool.ntp.org (ie.pool.ntp.org could  
> use some more machines, if you have a permanently connected time server  
> where you could spare maybe 10kbit/s)

The esat ntp servers have done this for at least a year or two now, i've 
never figured out why. Ntpdate still works, as does ntp itself, but 
ntptrace does not. From a tcpdump of ntpdate:

15:10:18.915104 IP 194.46.228.180.48748 > 193.120.10.3.123: NTPv4, 
Client, length 48
15:10:18.965562 IP 193.120.10.3.123 > 194.46.228.180.48748: NTPv4, 
Server, length 48

And a tcpdump of ntptrace:
15:14:41.184620 IP 194.46.228.180.48748 > 193.120.10.3.123: NTPv2, 
Reserved, length 12
15:14:46.184616 IP 194.46.228.180.48748 > 193.120.10.3.123: NTPv2, 
Reserved, length 12

Maybe the server just drops NTPv2 packets?

Steve
P.s. my personal ntp tip is to add the 'iburst' param to the server 
line, it means on startup ntpd sends 3 query packets in a short space of 
time, and voila it's synced. Having to wait 15 mins just to find out if 
ntp is working can be rather vexing.
-- 
"You are technically correct, the best kind of correct." -- the web