Nagios Support Forum

Hello,

I don't have any performance graphs, including for localhost.

I ran through the troubleshooting steps
https://support.nagios.com/kb/article.php?id=9
Those all looked ok.

I think it has to do with "rrds perl modules are not installed. falling back to rrdtool system call". I looked on a good system, it appears it's calling RRDs::update, which I think is part of RRDs.pm in "/usr/lib64/perl15/vendor_perl/RRDs.pm".

so, I have that perl module on the good & the bad system. But, on the bad one rrdtool "can't seem to find it".
Also, rra.cfg between a good & bad system are identical.

Please advise.

Are the /usr/local/nagios/libexec/process_perfdata.pl on both machines the same? The fall back message will occur if the USE_RRDs(line 36) is not set to 0.

That said, I'd still expect the fall back command to work - what happens if you try to run it manually? What do the permissions look like on and in /usr/local/nagios/share/perfdata/ ?

it's set to 1 but it's also set to 1 on a working system. /usr/local/nagios/share/perfdata/ ?
They're all set to nagios:nagios

The data is there so I'm assuming it's running fine.

The XML files in the directory don't contain the data. We would expect to see a .rrd file for each .xml file and the .rrd is where the data is stored.

Are both systems on the same OS version?

What happens if you try to run the "/bin/rrdtool create ..." command provided in the initial screenshot?

Hi,

They're not running the same OS version, but the one not working is running the latest version of Centos (7.5.1804). My other systems are using Centos (7.4.1708).

I'm sorry I'm not sure what the arguments to the rrdtool are. I tried passing it one of the xml files from the perfdata/host/ directory. This is the output.
If this is not correct please advise how to perform test.

Thanks.

ok, I see that I was supposed to execute that against the rrd file. Unfortunately none of the host folders in /usr/local/nagios/share/<host>/* have an rrd file. But as you noted the /usr/local/nagios/var/perfdata.log is passing to the rrdtool the rrd values from *.rrd files.

Does this make sense? Can you write up a quick synopsis to make sure we're on the same page?

The command is correct in using the .rrd extension since it is trying to create the rrd file with the command. It was too large for me to try and copy it verbatim from from screenshot, but you can try cutting an pasting it.

I think it is trying to run this command because like you said, it isn't able to find the RRDs.pm file. Run the following on both machines:

su - nagios
perl -le 'print for @INC'

This will print out where perl is looking for modules. Perhaps it isn't looking in the correct location.

Are RRD files persistent? Or are they like spooled files, disappearing after being executed? Two questions: where are the values coming from in perfdata.log when rrdtool executes? any guess why I don't have any rrd files (just the xml files) in the host directories?


Paths; grep RRDs.pm (it appears to be there); copied over a "good" RRDs.pm (diff shows no difference) between local pm & copied pm.

Can you try creating a test rrd file again by running the following commands as root on the server as well as the nagios user? Do you know if the system has rrdcached enabled?
If it is not running, that could cause the rrd files from getting created.

Check this file to see if the RRD_DAEMON_OPTS line is enabled.

both instances of the command yield /lib64/libpango-1.0.so.0: undefined symbol: g_log_structured_standard

grep of process_perfdata := RRD_DAEMON_OPTS = unix: /tmp/rrdcached.sock