Nagios Support Forum

I have Nagios XI running on VMware Linux CentOS release 5.4 (Kernel 2.6.18-164.9.1.e15 on an i686)
I installed it from VMware image downloaded from Nagios download site.
It is running Nagios XI 2011R1.5.
The host OS is Windows 7.

NagiosXi is monitoring 62 hosts.
I created a service called, mpm for 915wib, which returns three performance data:
MPM_combined
MPM_Rx
MPM_Tx.

Initially, I included 16 hosts that the service should be run in October.
Then, I added 13 more hosts to the same service in Nov.

The Performance graphs of the 16 hosts that I included in the hosts list of the service to be run in October stopped displaying the charts for the three performance data in Nov. 4th.
The Performance graphs of the newly added hosts in Nov. are still displaying the graphs for those three performance data.

It took me several weeks to try out several variations of changing configuration parameters of those hosts that were configured initially in October.
Then, I took a look at /usr/local/nagios/share/perfdata directory and found that there is a directory created for each host being monitored.
I did 'ls -l' on each directory and found that mpm_for_915wib.xml and mpm_for_915wib.rrd files exist in each directory.
Ironically, those hosts that don't display Performance graphs have the mpm_for_915wib.rrd file size of 1534556 bytes
and those hosts that do display Performance graphs have the mpm_for_915wib.rrd file size of 1151284.

I am attaching a zip file which has two sub directories:
B11-BGW
B41-3FA-FR

B11-BGW is one of hosts that the mpm for 915wib service stopped displaying Performance graphs.
B41-3FA-FR is one of hosts that still displaying Performance graphs.
In each directory, mpm_for_915wib.xml and mpm_for_915web.rrd files are also included.

I made sure that the file mode is set to 777 for /usr/local/nagios/share/perfdata and all of its sub-directories.
Then, I restarted npcd and httpd as I have seen in one of the previous postings regarding Broken Performance graphs.
But, it did not change the problem that I am having.

Could you take a look at those files and let me know what I should do to make Performance graphs continuously display for those hosts that stopped displaying the graph?

It looks like there could be an issue with this particular plugin returning performance data correctly. It looks like at one point it may have been returning 4 sets of performance data, and now it's only returning 3. To my knowledge, plugins must always return the same number of performance data sets, otherwise the rrd's won't update correctly. Is this a custom check that you're using?

Message in the XML file:

/usr/local/nagios/share/perfdata/B11-BGW/mpm_for_915wib.rrd: expected 4 data source readings (got 3) from 1321994756

Perfdata Captured in the XML file.

MPM_Combined=132.716666666667 MPM_Rx=129.683333333333;280;300;0 MPM_Tx=3.03333333333333;280;300;0

Also the same in the other file you sent.

/usr/local/nagios/share/perfdata/B11-BGW/mpm_for_915wib.rrd: expected 4 data source readings (got 3) from 1321994756
MPM_Combined=132.716666666667 MPM_Rx=129.683333333333;280;300;0 MPM_Tx=3.03333333333333;280;300;0

Right now you could probably fix this by just deleting the rrd's and the xml files, and then allowing PNP to rebuild them, but I wouldn't do that until you have the issue sorted out with the data sets.

Thanks so much.
That's exactly what happened.
Initially, the customer plugin returned four performance data.
Then, I change it slightly to return only three performance data.
I did not realized that I had to delete the rrd file when the number of performance data changes.

That resolved my issue.

Happy Thanksgiving!

Good deal, you too!