return code ??? out of bounds

[jssingh]

I've attached the plugin. It's for testing the read/write speed of a lustre filesystem. It also relies on a binary executable. I don't know how to get that to you. The binary is compiled for SLES 11.

[eloyd]

Did you try doing what it said?

Code: Select all

 author: XXXXXXXXX/NASA Advanced SuperComputing Facility
# email: [email protected]
#
# please report any bugs or other issues to the author, thanks!

[jssingh]

yep. he can't figure out what's going on either.

Those return codes don't make any sense in terms of the code. It can only return 0,1,2,3.

[Box293]

I suspect that you'll need to enable debugging for Nagios.

Code: Select all

sed -i 's/.*debug_level=.*/debug_level=-1/g' /usr/local/nagios/etc/nagios.cfg
service nagios restart

Make the unknown error occur.
Then look at the log /usr/local/nagios/var/nagios.debug

When you are finished this turns debugging off:

Code: Select all

sed -i 's/.*debug_level=.*/debug_level=0/g' /usr/local/nagios/etc/nagios.cfg
service nagios restart

FYI feature request submitted:
https://github.com/NagiosEnterprises/na ... issues/132

[eloyd]

I'm thinking that running it from the command line on the Nagios server, as the nagios user, will also shed some light on things.

[lmiltchev]

@jssingh Any updates?

[eloyd]

Code: Select all

cat > eric.pl
use lib oldPerlKnowledge;
use strict;
use vars qw($thisWillNeverWork, $disclaimEverything);
^D

./eric.pl -thisWillNeverWork -disclaimEverything=true

Okay, so my attempt at perl humor aside, debugging this remotely will never work without a lot of dev time trying to figure out exactly what it does, especially without the .../lib/cmst program to do the heavy lifting.

I'm _guessing_ that the various result codes are being caused by that cmst program. Look at lines 289 and 304 and you'll see that the perl code is blindly passing along the exit code of your cmst program without parsing it into Nagios 0/1/2 first. I'm guessing that there's something different/unexpected/changed/broken with that external program and/or how it's being activated through this (very complex) perl code that is now causing it to fail unexpectedly, and pass along seemingly random results, that the plugin is passing along without checking first.

[jssingh]

@eloyd thanks for catching that. I'll look into it. I've never had a problem with running it from the command line as the nagios user, but then again if you run this too often it hammers the filesystem, so I run it from the command line sparingly. We only run the nagios check every half hour.

@box293 I've turned debugging on. When the error happens again, I'll take a look. thanks.

[eloyd]

Only thank me if it's useful. Like I said, that's my quick examination of the code, as I was only looking for exit points.

[lmiltchev]

@box293 I've turned debugging on. When the error happens again, I'll take a look. thanks.

@jssingh, we will keep the thread open for the time being. Let us know if the issue resurfaces.