Page 1 of 1

Perfdata with incorrect timet attribute

Posted: Thu Apr 21, 2016 2:17 am
Good morning,

our nagios XI platform is drawing perfdata with incorrect time. Machine time is configured OK (clock, timezone,..), but perfdata files have NAGIOS_TIMED attribute is incorrect.

By using time conversion tool we found that it is about 3 days ago.

What can be done to fix this? I am uploading screenshots of everything.

Thank you.

Re: Perfdata with incorrect timet attribute

Posted: Thu Apr 21, 2016 12:20 pm
by lmiltchev
Are you having the same issue will all hosts/services?

Can you run the following commands, and show the output?

Code: Select all

strings /etc/localtime | tail -1
grep timezone /etc/php.ini
echo 'select now();' | mysql -t -pnagiosxi
Note: Modify the last command if you have mysql offloaded to a remote server.

Have you tried rebooting the server?

Re: Perfdata with incorrect timet attribute

Posted: Wed Apr 27, 2016 1:21 am
Hello,

yes its the same for all graphs.

I have tried to reboot server several times with no luck. Also cannot find any related information by searching it on google, seems to be a strange behaviour. I am considering to reinstall full nagios server but I will prefer not to do it...

Here I am uploading the output of these commands. Seems that sql one has failed.


Thank you.

Re: Perfdata with incorrect timet attribute

Posted: Wed Apr 27, 2016 1:29 am
by Box293
You did not type this command correctly:

Code: Select all

echo 'select now();' | mysql -t -pnagiosxi

Re: Perfdata with incorrect timet attribute

Posted: Thu Apr 28, 2016 5:04 am
Hello,

here it is the correct output.

Re: Perfdata with incorrect timet attribute

Posted: Thu Apr 28, 2016 5:08 am
And the current state of graphs.

Re: Perfdata with incorrect timet attribute

Posted: Thu Apr 28, 2016 12:04 pm
by tgriep
Could you post the following files so we can review them?

Code: Select all

/usr/local/nagios/etc/nagios.cfg
/usr/local/nagios/etc/pnp/process_perfdata.cfg
/usr/local/nagios/etc/pnp/config.php

Re: Perfdata with incorrect timet attribute

Posted: Mon May 02, 2016 10:14 am
Hello,

seems that I found the problem: the server was too overloaded (200+ charge). It had over 500 services configured to be checked every second. I had to configure it to 3 seconds and graphs are being displayed. Could this be the reason?


Thanks

Re: Perfdata with incorrect timet attribute

Posted: Mon May 02, 2016 11:52 am
by tgriep
Yes, that could be it. If the load is high for a period of time, the NPCD daemon and the process_perfdata script will stop graphing the performance data.
The settings can be changed to a higher value so the will continue to process the performance data. Here is how to do that.
Edit the following file on the XI server

Code: Select all

/usr/local/nagios/etc/pnp/process_perfdata.cfg
And increase the following setting. You can double or quadruple it.

Code: Select all

TIMEOUT = 5
Next, edit this file

Code: Select all

/usr/local/nagios/etc/pnp/npcd.cfg
And increase the following setting. You can double or quadruple it.

Code: Select all

load_threshold = 10.0
Save the files out and run the following to restart the processes so they will reload the changed settings.

Code: Select all

service npcd restart
service nagios restart
Post if you have any more questions.