performance data not processed for all services

[pepone]

Hi,

I had moved my nagios3 server from a ubuntu 10.04 machine to debian7. (now version 3.4.1)
So far everything works fine with nagios itself. Additionally I had moved nagiosgrapher also to debian,
I like to graph services, for example: http (time), cpuload and mem/swap.
For cpuload and swap I use check_snmp_load.pl and check_snmp_mem.pl.

Nagios monitoring is working fine itself, all the plugins seems to work.
But grapher makes some problems. I'm struggling already 3 days with the following problem:

The http-graphes worked out of the box after moving the config-files from old to new server.
But I get no graphs for the snmp plugins. I figured already some thing out, for example in the nagios
debug log, there are lines for processing the perfdate of http:

[1394805424.254960] [001.0] [pid=5219] run_service_performance_data_command()
[1394805424.254979] [001.0] [pid=5219] get_raw_command_line_r()
[1394805424.254990] [001.0] [pid=5219] process_macros_r()
[1394805424.254997] [2048.1] [pid=5219] **** BEGIN MACRO PROCESSING ***********
[1394805424.255003] [2048.1] [pid=5219] Processing: '/usr/lib/nagiosgrapher/fifo_write /var/lib/nagiosgrapher/ngraph.pipe '$HOSTNAME$\t$SERVICEDESC$\t$SERVICEOUTPUT$\t$SERVICEPERFDATA$\n' 3'
[1394805424.255025] [2048.1] [pid=5219] Done. Final output: '/usr/lib/nagiosgrapher/fifo_write /var/lib/nagiosgrapher/ngraph.pipe 'xxxxxxx.xxx.xxxxxxx.de\tHTTP-Dekanats-Website\tHTTP OK: Status line output matched OK,Authorization Required,301 Moved,302 Found - 26131 bytes in 0.091 second response time\ttime=0.091010s;;;0.000000 size=26131B;;;0\n' 3'

These lines are missing for the snmp services on the new server. At the old server they are there.

If I run some nagiosgrapher lines from the debug file on the old server at the new server, I get finally
graphs, for example:
# sudo -u nagios /usr/lib/nagiosgrapher/fifo_write /var/lib/nagiosgrapher/ngraph.pipe 'xxxxx.xxx.xxxxxx.de\tCPU\tLoad : 0.00 0.01 0.00 : OK\t\n' 3

I also compared the retention.dat of the old and new server, because I think these are the real values,
nagios3 takes out of the configuration.
The CPU service configuration of hosts for example is in both files identical.
I think important is:

plugin_output=Load : 0.00 0.03 0.03 : OK
performance_data=
process_performance_data=1

One point may by:
performance_data= is empty for the snmp-load in both retention.dat files. For http-services is not empty.
But anyway, on the old server it worked like that.
performance_data

Have you any idea what my problem could be?

Thanks,

Sebastian

[abrist]

Can you run the snmp checks from the cli and check for performance data (run it with -vvv for maximum verbosity)? If it returns perfdata, then it is a problem with nagios/configuration. If no perfdata is reported it is a problem with the device, the plugin, or potentially the snmp libs on the server.

[pepone]

Hi,

thanks for that hint. That gave me the solution. I checked the source of the check_snmp_*.pl plugins, they have a nice option:

sub check_options {
Getopt::Long::Configure ("bundling");
GetOptions(
...
'f' => \$o_perf, 'perfparse' => \$o_perf,
'T:s' => \$o_check_type, 'type:s' => \$o_check_type
);
Adding "-f" to all commands helped. THANKS
So it is not good, to copy only configuration and relay on it.

By the the way, now I can try also to graph my hard disk space, monitored with check_snmp_storage.pl
This was not configured on the old server.

Regards,

Sebastian