[Nagiosplug-devel] what checks check_ntp?

[Gerhard Lausser]

Hi,
i am a bit confused on what check_ntp actually checks.
You give it a timeserver with the -H parameter and it determines the offset
between local time and the timeserver's time with the ntpdate command.
That's ok so far.
In a second step ntpq -np <timeserver> is called which gives you the
timeserver's daemon and peer status. The jitter we see here comes from the
communication between the timeserver and it's peers. Is this really relevant
for the time quality of our local host? Wouldn't calling  ntpq -np 127.0.0.1
give us the jitter information we're really interested in? 
In my opinion it would be better to first call  ntpq -np 127.0.0.1  which
returns information for both offset and jitter (for the current sys.peer or
candidates, not for the statically configured timeserver) and if that
doesn't work, use ntpdate -q <timeserver>  as a fallback.

Greetings from Munich,
Gerhard