[Nottingham] "ntpq -p" displayed flags

[Martin]

On 21/01/12 13:05, Camilo Mesias wrote:
> There are some hints under ntpdc man page, from memory * is a server
> you are synced to, + is a server you could sync to but it's not chosen
> currently. # and x are probably not good servers.
> 
> ntp is a bit of a dark art
> 
> -Cam

Thanks for that bit.


> On Sat, Jan 21, 2012 at 12:51 PM, Martin <martin at ml1.co.uk> wrote:
>> Folks,
>>
>> Anyone know what the characters mean in the first column of the
>> "ntpq -p" output?
>>
>> I've got a collection of: "x# +*"

The 'hint' from:

man ntpq
man ntpdc

is "Print a list of the peers known to the server as well as a summary
of their state."

For my example I have:


remote          st t when poll reach   delay   offset  jitter
=============================================================
 LOCAL(0)       10 l  14h   64    0    0.000    0.000   0.000
+ns2.adnscom.net 2 u 1041 1024  377   32.293   -1.253   6.789
+ns1.telecom.lt  2 u  941 1024  377   62.904   -0.332   1.524
xportal.kt.kg    2 u  119 1024  377  130.354  -59.117   5.163
 m07b            3 u 1036 1024  377    0.146   -3.015   1.468
#mdr             4 u  231 1024  377   36.435   -0.694   2.392
+m04             4 u 1041 1024  377   19.658   -2.064   5.741
#m06             4 u 1039 1024  377   28.774   -3.426   7.848
+m07             3 u  172 1024  377   24.623   -1.434   5.895
*chime1.surfnet. 1 u  394 1024  377   20.549   -3.125   4.740
+chime5.surfnet. 1 u  374 1024  377   24.066   -2.141   6.500
#net2.uni-paderb 1 u  167 1024  377   36.238  -14.957   3.072
+timehost.lysato 1 u   37 1024  377   50.186    0.358   3.075
+pony.magicalfor 2 u  176 1024  377   13.357   -1.912   5.552
#cheddar.halon.o 3 u  114 1024  377   16.272    8.119   6.789
 www-vhosts.spin 16 u   - 1024    0    0.000    0.000   0.000
#corrin.geekwu.o 2 u  195 1024  377   22.320    1.323   5.060
+mx1.xillhosting 3 u  149 1024  377   34.921   -1.331   3.244


There's only one "*" yet that doesn't have the smallest offset nor the
smallest jitter of the stratum 1 peers. The "x" does indeed look to be a
"don't use" for being too far off. That leaves the " #+"... The " "
perhaps means no info. I guess m07b is excluded because it has this
machine as its primary sync.

So, looking at offset and jitter, comparing mdr and m04 vs m07 and m06,
the story can go either way for the significance of "+" and "#"... So
how are they decided?


?

Cheers,
Martin



-- 
----------------
Martin Lomas
martin at ml1.co.uk
----------------