check_mount.pl

Hi

I’m strugging to set this plugin up. I can execute it correctly from the CLI but nrpe is showing NRPE: Unable to read output

[root@fc-web04 libexec]# /usr/bin/perl /usr/local/nagios/libexec/check_mount.pl -m /var/Performance
/var/Performance is mounted!

This is the error I’m getting

[root@fc-web04 libexec]# ./check_nrpe -H localhost -c check_df
NRPE: Unable to read output

Can you pls guide me how could I make it work ?

Thanks in advance Liju

Make it worked by adding nagios plug-ins directory in that script file like this use lib '/usr/local/nagios/libexec';

Hi Liju,

I expected to find the Nagios’ utils.pm in /usr/lib/nagios/plugins . I’ll update my articles to supplement it with a note about this variable ;-)

Thanks for your advice!

First off, sane exiting from the usage statement, in the event someone adds a check, but forgets the mount point..

sub how_to
{
	print "USAGE: $0\n\t-m MOUNTPOINT\twich mountpoint to check\n\t[-t TYPE]\toptionally check whether it's this kind of fs-type\n\n";
	print "OK: We really have no idea what you want to check, so we'll play it safe here, alright?\n";
	exit $ERRORS{'OK'};
}

In case of a mount point sitting under another like this:

/opt/something mounted from /dev/sdb
/opt/something/shared mounted from NFS

then you actually catch both in your $erg grep … To only catch the one filesystem you are looking for is doing something like this (ugly) hack ..

my $erg = `mount | grep $MOUNT | awk '{ print \$3\" \"\$5 }' | grep -v ^$MOUNT\/`;
chomp $erg; #I tend to get rid of newlines ...

Also CIFS mounts may lose connectivity to the server sharing the folder, but will recover on its own. In the meantime you’ll not be able (of course) to do anything with data in that mountpoint. What distinguishes things is simply a directory listing which will contain nothing but a number of question marks if you do a long listing ( ls -l ) ..

user@host $ ll /opt/some_dir
total 0
?--------- ? ? ? ?            ? another_dir

So, I’ve added a little bit of code, for this specific case (and changed a regexp along the way) …

if ($erg)
{
	if (defined ($TYPE))
	{
		if ($erg =~ m/$TYPE/x)
		{ if ($TYPE eq "cifs") {
			my $fs_listing = `ls -larth $MOUNT`;
			
			if ($fs_listing =~ m/\?/x) {
				print "WARNING: " . $MOUNT . " is mounted but we cannot access files!!\n";
				exit $ERRORS{'WARNING'};
			}
		}

			print "OK: " . $MOUNT . " is mounted! Type is " . $TYPE . "\n";
			exit $ERRORS{'OK'};
		}
		else
		{
			print "WARNING: " . $MOUNT . " is mounted! But type is not " . $TYPE . "\n";
			exit $ERRORS{'WARNING'};
		}
}
	print "OK: " . $MOUNT . " is mounted!\n";
	exit $ERRORS{'OK'};
}

Appears to be working nicely during initial tests. YMMV…

Thanks for your efforts, I’ll upgrade my version! ;-)

Hi. Thank you for this plugin. P.S. For getting this plugin work with SELnux you should set at least nagios_unconfined_plugin_exec_t type. for e.g.

semanage fcontext -a -s system_u -t nagios_unconfined_plugin_exec_t -f -- '/etc/nagios/plugins/check_mount.pl'

Thank you for this tiny but extremely useful script! Works like a charm for me and does what check_disk isn’t any use.

Hello. First of all thank you for supplying very nice plugin though.

I’m the newbie of using nagios. I’d like to suggest you how about to change the line number 14 from use utils qw(%ERRORS); to use Nagios::Plugin qw(%ERRORS);

Development doc says that utils.pm is gonna deprecated soon or later.

Where did you read that utils.pm is going to be deprecated? The website of that perl module Nagios::Plugin links to the guidelines in which they recommend the usage of utils.pm .