Problem Graphing Host with 4hr Check Interval

[mrochelle]

Keep in mind, the only difference between the hosts with graphing problems and the others that are fine is the check interval for the hosts with problems is 4 hours.

[abrist]

Lets see what is in that file:

Code: Select all

rrdtool dump /usr/local/nagios/share/perfdata/ET_50063/_HOST_.rrd

Is this a host or service check that has the problem?

[mrochelle]

Ok, looks like my check interval is every 6 hours instead of 4. Which is OK.

<!-- 2013-04-25 01:00:00 CDT / 1366869600 --> <row><v> NaN </v></row>
<!-- 2013-04-25 07:00:00 CDT / 1366891200 --> <row><v> NaN </v></row>
<!-- 2013-04-25 13:00:00 CDT / 1366912800 --> <row><v> NaN </v></row>
<!-- 2013-04-25 19:00:00 CDT / 1366934400 --> <row><v> NaN </v></row>
<!-- 2013-04-26 01:00:00 CDT / 1366956000 --> <row><v> NaN </v></row>
<!-- 2013-04-26 07:00:00 CDT / 1366977600 --> <row><v> NaN </v></row>
<!-- 2013-04-26 13:00:00 CDT / 1366999200 --> <row><v> NaN </v></row>
<!-- 2013-04-26 19:00:00 CDT / 1367020800 --> <row><v> NaN </v></row>
<!-- 2013-04-27 01:00:00 CDT / 1367042400 --> <row><v> NaN </v></row>
<!-- 2013-04-27 07:00:00 CDT / 1367064000 --> <row><v> NaN </v></row>
<!-- 2013-04-27 13:00:00 CDT / 1367085600 --> <row><v> NaN </v></row>
<!-- 2013-04-27 19:00:00 CDT / 1367107200 --> <row><v> NaN </v></row>
<!-- 2013-04-28 01:00:00 CDT / 1367128800 --> <row><v> NaN </v></row>
<!-- 2013-04-28 07:00:00 CDT / 1367150400 --> <row><v> NaN </v></row>
<!-- 2013-04-28 13:00:00 CDT / 1367172000 --> <row><v> NaN </v></row>
<!-- 2013-04-28 19:00:00 CDT / 1367193600 --> <row><v> NaN </v></row>
<!-- 2013-04-29 01:00:00 CDT / 1367215200 --> <row><v> NaN </v></row>
<!-- 2013-04-29 07:00:00 CDT / 1367236800 --> <row><v> NaN </v></row>
<!-- 2013-04-29 13:00:00 CDT / 1367258400 --> <row><v> NaN </v></row>
</database>
</rra>
</rrd>

[abrist]

It looks like your check is not returning graphable data - "NaN".
Can you post the script that is generating the perfdata?

[mrochelle]

#!/usr/bin/perl
#
use Nagios::Plugin;
# Define the plugin and usage.
$np = Nagios::Plugin->new(
shortname =>"mpingv2",
timeout => 60,
usage => "Usage: %s [--device device] [--hostaddr hostaddr]",
);

# Add command line arg 1
$np->add_arg(
spec => "device=s",
help => "--device=DEVICENAME\n MCC device to be tested.",
required => 1,
);
# Add command line arg 2
$np->add_arg(
spec => "hostaddr=s",
help => "--hostaddr=HOSTADDR\n Host system to SSH for checking MCC device provided.",
required => 1,
);

$np->getopts;

$device = $np->opts->device;
$hostaddr = $np->opts->hostaddr;
$CMD = "/home/nagios/bin/check_mcc_remote.csh $device";
$SHELL = "/usr/bin/ssh";

# Ping a single Concentrator / MCC device on the remote host via SSH.
my $alive = `$SHELL $hostaddr $CMD`; chomp $alive;

if ( $alive =~ m/Finished reading|Got response/ ) {
$pluginstate = "OK"; $msg = "$alive, | Data=1;;;0;1";
} else {
$pluginstate = "CRITICAL"; $msg = "$device request failed: $alive, | Data=0;;;0;1";
}

$np->nagios_exit($pluginstate, $msg );

[abrist]

Could you compare the output of the plugin when run against the host in question and when run against one of the working hosts? Post both outputs here.

[mrochelle]

Ok, in order to compare apples to apples, I'm going to configure another similar host with shorter check intervals rather than compare with different hosts that work. I should have results within the hour.

[abrist]

Great. We await the news.

[mrochelle]

Host: ET_50063 with 6 hour interval. Last update 07:24 with following data.
mpingv2 OK - Tue Apr 30 00:03:06 2013 Finished reading MCC 50063.1 type cum snap 1 4/4/14, | Data=1;;;0;1

Host ET_50015 with 5 minute interval. Last update 10:50 with following data.
mpingv2 OK - Tue Apr 30 08:20:38 2013 Finished reading MCC 50015.1 type cum snap 0 352/352/308, | Data=1;;;0;1

I'm already starting to graph on ET_50015 which was changed to 5 min. interval updates. Do not pay too much attention to the data since I do not plot with. Nagios is provide a 1 for valid data was found and 0 if no data is found.
Based on these results, I'm leaning in the direction that there is some limit with the interval duration?

[abrist]

Those outputs do not match the format of the performance data your script returns. With a shorter interval it is now working though?