[ale] NTP...

Joe Steele joe at madewell.com
Mon Apr 24 11:13:38 EDT 2006


Michael B. Trausch wrote:
> I tried using 'ntpq' from what I was able to infer for try9ing to 
> troubleshoot to see if I could figure out if it was talking and what I got 
> was this:
>
> root at cinnamon:~# ntpq -c pee time-a.nist.gov time-b.nist.gov time.nist.gov 
> time-nw.nist.gov
> server                remote           refid      st t when poll reach   
> delay   offset  jitter
> ===============================================================================================
> time-a.nist.gov  *LOCAL(0)        .ACTS.           0 l    3   64  377    
> 0.000    0.000   0.000
>   
[snip]
> I don't understand that output, however.  It looks as if it is saying that 
> my offset to each of these servers is 0.000; I can't hardly believe that, I 
> sync'd my clock *by hand* about two or three hours ago to another computer 
> on the Internet that I know has ntpd working.
>   
Your invocation of ntpq has told you nothing about the status of *your* 
computer.  Instead, you have asked ntpq to query 4 other servers and 
report on their status/peers.  Since they are all stratum 1 servers, 
they have no peers, their clocks are all sync'd to "LOCAL", and they 
have no delay, offset, or jitter.  Nonetheless, the one useful thing 
it's telling you is that you are able to talk with these remote servers, 
at least when the source UDP port is an unprivileged  port (which is 
what ntpq uses).  Therefore, someone else's suggestion that you try the 
"-u" option with ntpdate is a good idea.  You might also try using 
ethereal to confirm that bidirectional communication is occurring.

FYI, the status of your computer would be obtained with "ntpq -c pee" 
(the default host is then localhost).  However, this assumes that ntpd 
is already running on localhost (otherwise, ntpq will complain about 
"connection refused").

--Joe.



More information about the Ale mailing list