[mrtg] MRTG (or rateup?) silently pauses

[Niall O'Reilly]

Happy New Year.

Over the holiday period, a couple of our instances of MRTG silently  
paused.
These instances were on different boxes, but monitoring the same  
targets,
and stopped within 15 minutes of each other, soon after 08:30 on  
31-12-2007.

The MRTG processes appeared still to be running today (9 days later),  
but
none of the associated target log files had been updated since back  
then.
I still have to check several other boxes.

Restarting MRTG was sufficient to unjam whatever was stuck.  I'm  
baffled as
to the cause, and now hope someone on the list can offer an idea or two.

I'm suspecting that some resource (per-process file handles?) is being
consumed instead of re-cycled, but really have no idea.  It doesn't
seem to be memory, as far as our graphs show (I prefer not to advertise
the URL on the list).

I've tried trawling the Googlesphere already and have found only one
description of an apparently similar incident, but in different
circumstances (MS Win NT or 2000) from ours, and without any information
which seems to be useful to me.

So far, two boxes are involved, running different releases of RHEL, but
otherwise "equivalent".  We run MRTG in daemon mode, and use a locally
developed script in /etc/init.d which starts an instance of MRTG for  
each
configuration file found in the configuration directory.  The targets
involved are not SNMP targets, but use a locally developed script to
monitor DNS consistency and response time for a number of zones and  
servers.

As with a some other critical components, we install MRTG from the  
original
tarball, and not from some RPM.  We still use rateup, because we're  
used to
and like the way MRTG builds the graphs and annotation.

For one of the boxes, here are the details.

% uname -r; cat /etc/redhat-release; service mrtg status
2.4.21-27.ELsmp
Red Hat Enterprise Linux ES release 3 (Taroon Update 4)

   21632 running mrtg-dns from /usr/bin/mrtg (2.15.2)
   30392 running mrtg-local from /usr/bin/mrtg (2.15.2)

%

The other box runs ES release 4 (Nahant Update 3).

The log file for the mrtg-dns process had only the normally expected
contents, "Daemonizing MRTG ...".  I con't see anything relevant in
/var/log/messages for the "day the logging died" either.

Thanks in advance for any ideas.


	Best regards,

	Niall O'Reilly
	University College Dublin IT Services

	PGP key ID: AE995ED9 (see www.pgp.net)
	Fingerprint: 23DC C6DE 8874 2432 2BE0 3905 7987 E48D AE99 5ED9



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 186 bytes
Desc: This is a digitally signed message part
Url : http://lists.oetiker.ch/pipermail/mrtg/attachments/20080109/3e225aab/attachment-0001.bin