SUMMARY: NTP jitter (not parsed)

Thank you for the reples. Unfortunately, none of them
worked. It seems the problem was related to the nagios
or perl files on the monitoring server itself and not
NTP. Maybe they were corrupted as a result of the
power outage.

I can't say I found a solution to the problem, but I
was able to make the problem go away. With patience
getting short, I changed 3 things at once while
troubleshooting. As a result, I don't know excatly
which change eliminated my problem. I upgraded to the
latest perl distribution and recompiled nagios without
using the embedded perl option (a nagios configure
option).

Now everything works like normal again.

-John

> Hi gurus,
>
> Two very similar Solaris 9 hosts (both NTP clients)
> are keeping their clocks tuned just fine. My Nagios
> monitoring server polls them using a perl script
> (check_ntp) to verify their NTP stuff is in good
> shape.  This process was working great until a
sudden
> lack of electron movement (power outage) caused all
> hosts to shutdown very fast.
>
> Now that electrons are flowing again, the hosts are
> responding differently to the NTP test query. It
seems
> 'goodhost' has parsed its jitter (which is
apparently
> a good thing) while 'badhost' has not:
>
> foo:/usr/local/nagios/libexec # ./check_ntp -H
> goodhost
> NTP OK: Offset 0.000140 secs, jitter 0.31 msec, peer
> is stratum 3|offset=0.000140,
> jitter=0.00031,peer_stratum=3
>
> foo:/usr/local/nagios/libexec # ./check_ntp -H
badhost
> Argument "(not parsed)" isn't numeric in abs at
> ./check_ntp line 401.
> NTP OK: Offset 0.000327 secs, jitter (not parsed)
> msec, peer is stratum 3|offset=0.000327,
> jitter=0,peer_stratum=3
>
> I think I have 2 problems. One, the check_ntp script
> does not handle the string "(not parsed)" being
> returned when its expecting a numeric. I've
submitted
> that to the nagios-users list.
>
> I hope you can help me with problem Two: How do I
get
> the 'badhost' to parse its jitter?
>
> I cannot find any meaningful differences between the
> two hosts. I've used xntpdc to look a detailed
stats,
> loop filters, and more. Both NTP clients are using
the
> same NTP servers. Everything looks fine from an NTP
> perspective, but I think I'd like to force xntpd to
> parse its jitter even if it's zero. At least then,
it
> might return a numeric value instead of "(not
> parsed)".
>
> BTW: I'm referring to these two hosts for
illustrative
> purposes. This problem is actually affecting 80% of
my
> hosts while 20% are okay.
>
> TIA for any suggestions on how to troubleshoot this
> further! Will summarize.
> -John


		
__________________________________ 
Yahoo! Mail - PC Magazine Editors' Choice 2005 
http://mail.yahoo.com
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers