jstat is failing to show a proper result

[runzelpunzel]

This is so bizarre

It is indeed! That's why I'm kinda lost and seeking for your help

Code: Select all

ls -l /usr/lib/nagios/plugins/check_test.sh

-rwxr-xr-x 1 root root 199 Feb 21 08:19 /usr/lib/nagios/plugins/check_test.sh

Looks okay to me!?


PS: I will be AFK in the following week, so any further test / comment will be written after March, 9th.

[scottwilkerson]

I wonder if the problem is you need to add the full path to jstat in the script

Code: Select all

which jstat

When NRPE runs you don't get the full environment for the nagios user, and this could be causing it to fail.

[runzelpunzel]

It's the default /usr/bin/jstat.
Even when I run a /usr/bin/sudo /usr/bin/jstat -gc 16652, whereas 16652 is the PID of my java process, I get a proper result.
However, when I run configure

Code: Select all

command[check_sudo_test4]=/usr/bin/sudo /usr/bin/jstat -gc 16652

in /etc/nagios/nrpe.cfg and call check_sudo_test4 via NRPE from the nagios server, i get an "NRPE: Unable to read output"

[scottwilkerson]

What is the output you get when you run it directly?

[runzelpunzel]

the regular output from jstat, eg.

Code: Select all

 S0C    S1C    S0U    S1U      EC       EU        OC         OU       MC     MU    CCSC   CCSU   YGC     YGCT    FGC    FGCT     GCT 
45056.0 45568.0  0.0   24537.4 1306112.0 914401.5 1668096.0  1393081.2  574832.0 506528.5 79048.0 65307.1   2629  226.672  17     12.857  239.529

Looking into that, "unable to read output" might well be a vaild answer (as we've a multi-line-output, etc.).
This is a very basic testing-script, as the previous tests that were writing to a file an parse the output from jstat didn't work either.

[scottwilkerson]

I'm not really familiar with jstat but if it gives couninuous updates you may need to add the "count" number to the end "1"
Re: https://docs.oracle.com/javase/7/docs/t ... jstat.html

Code: Select all

command[check_sudo_test4]=/usr/bin/sudo /usr/bin/jstat -gc 16652 1

[runzelpunzel]

No, it really runs just one time:

Code: Select all

nagios@<MYMACHINE>[~]> /usr/bin/sudo /usr/bin/jstat -gc 16652
 S0C    S1C    S0U    S1U      EC       EU        OC         OU       MC     MU    CCSC   CCSU   YGC     YGCT    FGC    FGCT     GCT 
24064,0 38400,0 23785,2  0,0   1090560,0 323405,2 1622016,0  1587494,2  578800,0 509547,1 79432,0 65639,8   3436  288,142  18     13,699  301,841
nagios@<MYMACHINE>[~]>

[scottwilkerson]

Curious, what output do you get when you run

Code: Select all

/usr/bin/sudo /usr/bin/jstat -gc 16652 2> /dev/null
echo $?

[runzelpunzel]

"0", as you would expect.

Code: Select all

nagios@<MYMACHINE>[~]> /usr/bin/sudo /usr/bin/jstat -gc 16652 2> /dev/null
 S0C    S1C    S0U    S1U      EC       EU        OC         OU       MC     MU    CCSC   CCSU   YGC     YGCT    FGC    FGCT     GCT
17408,0 9216,0  0,0   9188,3 446976,0 325530,5 1664000,0  1205941,3  585968,0 515763,0 80200,0 66246,9   5507  437,466  25     20,204  457,670
nagios@<MYMACHINE>[~]> echo $?
0

[scottwilkerson]

I am really at a loss here why your check_sudo_test4 isn't working...

Looking around for similar usage I did find the following plugin that it looks like others are using that may do something similar
https://exchange.nagios.org/directory/P ... at/details