[ntp:questions] Selection and Clustering of candidates

Mischanko, Edward T Edward.Mischanko at arcelormittal.com
Mon Jun 27 12:07:19 UTC 2011


State	Remote		Refid	Stratum	Type		When	Poll
Reach	Delay	Offset	Jitter
#	156.144.102.1		156.144.104.1	4	Unicast server
555	1024	377	23.192	1.465	1.760	
+	156.144.106.1		156.144.104.1	4	Unicast server
751	1024	377	29.431	0.030	1.339	
+	156.144.109.1		156.144.104.1	4	Unicast server
583	1024	377	22.822	-0.005	1.796	
#	156.144.110.1		156.144.104.1	4	Unicast server
701	1024	377	22.546	0.855	0.908	
+	156.144.112.1		156.144.104.1	4	Unicast server
699	1024	377	22.950	0.070	1.015	
#	156.144.120.1		156.144.104.1	4	Unicast server
646	1024	377	22.829	2.164	1.991	
#	156.144.121.1		156.144.104.1	4	Unicast server
547	1024	377	24.809	-0.417	0.742	
+	156.144.122.1		156.144.104.1	4	Unicast server
359	1024	377	22.891	0.153	1.089	
+	156.144.123.1		156.144.104.1	4	Unicast server
388	1024	377	23.672	-0.448	1.056	
+	156.144.124.1		156.144.104.1	4	Unicast server
306	1024	377	37.218	-7.239	7.356	
#	156.144.125.1		156.144.104.1	4	Unicast server
91	1024	377	22.781	0.236	1.016	
*	156.144.126.1		156.144.104.1	4	Unicast server
453	1024	377	22.692	0.357	0.525	
#	156.144.129.1		156.144.104.1	4	Unicast server
595	1024	377	23.330	1.231	1.114	
+	156.144.131.1		156.144.104.1	4	Unicast server
881	1024	377	23.196	0.637	0.763	
+	156.144.134.1		156.144.104.1	4	Unicast server
617	1024	377	23.003	0.057	0.668	

Why is 156.144.124.1 selected for clustering?  It has the the highest
delay and the highest jitter; it is not needed to synch the clock, as
you can see I have plenty of candidates to choose from.

-----Original Message-----
From: questions-bounces+edward.mischanko=arcelormittal.com at lists.ntp.org
[mailto:questions-bounces+edward.mischanko=arcelormittal.com at lists.ntp.o
rg] On Behalf Of David Woolley
Sent: Sunday, June 26, 2011 4:04 AM
To: questions at lists.ntp.org
Subject: Re: [ntp:questions] Selection and Clustering of candidates

Mischanko, Edward T wrote:
> Some of the servers
> NTP selects have offsets that are higher than the majority;

These might be the ones with the most correct time.  A high offset only 
indicates a poor time if you are certain that the local time is 
absolutely correct.

>  Why are they being selected?

Because they have times that are consistent with the times from all the 
other selected ones and NTP_MAXCLOCKS in ntpd.h hasn't been reached.

> Why are so many being selected when I really want the best 5 or 6.

How is ntpd to determine that they are the best ones.  A high offset may

simply mean the local time is wrong.

> 
> # Miscellaneous stuff
> tinker allan 7
> tos minclock 4 minsane 3

It is tos maxclock, but it doesn't seem to be documented in at least 
some versions of the HTML.

_______________________________________________
questions mailing list
questions at lists.ntp.org
http://lists.ntp.org/listinfo/questions



More information about the questions mailing list