[ntp:questions] Re: What accuracy should I expect on a LAN?

Per Hedeland per at hedeland.org
Wed Feb 8 23:04:54 UTC 2006


In article <1139412870.221176.43120 at g14g2000cwa.googlegroups.com>
skip at pobox.com writes:
>
>     remote           refid      st t when poll reach   delay   offset
>  disp
>==============================================================================
>*10.10.2.2       198.82.1.202     3 u  870 1024  377     0.40   -0.813
>  2.64
>+10.10.1.48      130.126.24.44    3 u  286 1024  377     0.50   -0.566
>  1.13
>+10.10.1.31      128.59.16.20     3 u  867 1024  377     0.34    3.490
>  1.37
>
>Does that indicate that the machine is within 3.5ms of the time of one
>of our servers and within 1ms of the time of the other two (offset
>column)?

Almost - it's the calculated offset, and can be wrong by upto half of
"delay". And the measurements were done 870/286/867 seconds ago.

>Another machine shows quite different offsets:
>
>     remote           refid      st t when poll reach   delay   offset
>  disp
>==============================================================================
>*10.10.2.2       198.82.1.202     3 u   30   64  377     0.44   67.072
>  3.16
>+10.10.1.48      130.126.24.44    3 u   63   64  377     0.44   70.433
>  3.02
>+10.10.1.31      128.59.16.20     3 u   62   64  377     0.40   75.967
>  3.37
>
>If the offset column is really the computed difference in the local
>and remote clocks 60+ms offsets seem huge to me for a LAN.

This system doesn't seem to have stabilized (yet), as evidenced by its
short poll interval - if it wasn't started recently, it probably at
least stepped the clock recently - check your syslog logs.

--Per Hedeland
per at hedeland.org




More information about the questions mailing list