Nagios Support Forum

Hello. I recently posted to the Nagios-users list about an issue that
seems to have cropped up since I moved to 2.9.

I use Nagiosgraph 0.8.2. My nagios.cfg file has the following
configuration options:

process_performance_data=3D1
service_perfdata_file_mode=3Da
service_perfdata_file_processing_interval=3D30
=09
service_perfdata_file_processing_command=3Dprocess-service-perfdata
service_perfdata_file=3D/var/spool/nagios/perfdata.log
perfdata_timeout=3D5

The process-service-perfdata command is "insert.pl" that comes with
nagiosgraph as modified to run for our environment. I have run with
this configuration successfully for many months now.

After I upgraded from 2.8 to 2.9, I discovered that my graphs (really
the underlying RRD files) were not getting updated. When I checked my
nagios.log file, I found quite a lot of:

Warning: Service performance data file processing command
'/usr/local/nagios/nagiosgraph/insert.pl' timed out after 5 seconds

I found that if I ran this command by hand, it correctly updated all my
RRDs, but it took (at that time yesterday) almost 5 minutes to complete.
The perfdata file had grown quite large and its oldest entries were from
the previous day.

I put a monitor on the file to see how large it grows and it's now on a
quite linear growth pattern and is now over 1.3 million lines (and
growing).

I never had to pay attention to this prior to Tuesday when I upgraded.
I'm assuming that Nagios somehow truncates or cleans the perfdata file?
It doesn't seem to be happening anymore.

At the moment I'm still sticking with 2.9, but I've increased my
perfdata_file_process_interval to 10 minutes and my perfdata timeout to
10 minutes while I watch it. Not ideal, but I can limp along for
another day or so.

Is this a bug introduced in 2.9 or am I suddenly doing something wrong?

thanks

Mark





This post was automatically imported from historical nagios-devel mailing list archives
Original poster: mark.frost1@pepsi.com