check mountpoints like nfs, cifs, davfs, lustre, ocf2, etc.

Hello! Thanks for writing this.

I'm having an issue. When I run the plugin locally with -a, everything reports OK. When I run the command on nagios server via nrpe I get a CRITICAL error. Its almost like its checking for those paths on the nagios server instead of the remote host:

Nagios server:
check_nrpe -H myhost.mydomain.com -c check_mounts
CRITICAL: /mnt/testmnt doesn't exist on filesystem ;

Command from nrpe_local.cfg:
command[check_mounts]=/usr/lib/nagios/plugins/check_mountpoints.sh /mnt/testmnt

Result running command manually on remote host:

/usr/lib/nagios/plugins/check_mountpoints.sh /mnt/testmnt
OK: all mounts were found ( /mnt/testmnt )

Any ideas?

works with btrfs over iscsi

Was is wrong here?! :-O

root@nfsclnt ~ # grep -i nfsserver /etc/vfstab
nfsserver:/mnt - /mnt nfs - yes rw,xattr

root@nfsclnt ~ # grep -i nfsserver /etc/mnttab
nfsserver:/mnt /mnt nfs rw,xattr,dev=9300001 1507803587

root@nfsclnt ~ # ./check_mountpoints.sh /mnt
CRITICAL: /mnt is not mounted ;

Thanks for this plugin, it works great on local Nagios server.

When checking shares on a remote host I get :
"(Return code of 127 is out of bounds - plugin may be missing) "

The help file does not show there is a -H option, so was wondering if you plan to implement this, or if you can tell what I am doing wrong.

For me it works great with CIFS mount on CentOS host

I have a few servers that we ust rc.local to map NFS mounts (reasons beyound me). They look like this:

mount 1xx.26.xx.xx:/vol/e_oldlogs /geminilogs
sleep 2
mount 1xx.26.xx.xx:/vol/e_webdocs /webdocs
sleep 2
mount 1xx.26.xx.xx:/vol/e_weblogs /weblogs

I need to check agains this file for mounts is there an easy way or even a hard way to do this???

I guess this plug will not work under AIX.. :o(

Hi!

Tried out this and it seems to work fine. One problem though.. When I define the command like this:

command[check_mpoints]=/usr/lib/nagios/plugins/check_mountpoints.sh -a -w

And run:
/usr/lib/nagios/plugins/check_nrpe -H localhost -c check_mpoints

I'm getting "CRITICAL: /mnt/nfs/path/to/folder/.mount_test_from_hostname is not writable."

If I run it without nrpe, i.e.: /usr/lib/nagios/plugins/check_mountpoints.sh -a -w it's working fine.

Why is that? Issue with permission for the user Nagios?

Thanks,

Hello,

I just tested this plugin and I have to say: this script does not what it should do (on Ubuntu 12.04)

I have a mountpoint, say /media/user/share which is defined in /etc/fstab. I ran the plugin as root and it said: "OK: all mounts were found (/media/user/share)".

Then I unmounted this share from the client,re-ran the plugin and it still claimed to be mounted: "OK: all mounts were found ( /media/user/share)".

Since Ubuntu is binary-compatibel to Debian I do really wonder why this script/plugin gives totaly wrong information (moreover, Ubuntu also uses /proc/mounts and /etc/fstab).

To put it in a nutshell: I can't blame Ubuntu for the wrong result of this plugin. I wonder how that can be, since it was updated some weeks ago :(

I will use /bin/bash -x to see if I can "debug" this script (maybe just a wrong loop logic) but consider this the reason for a low rating.

Without the bug I see on Ubuntu this would be a great plugin!

I like this plugin. But I had to change LIBEXEC to include /opt/csw/libexec/nagios-plugins because i use OpenCSW-installed nagios in Solaris.

Any chance you could include AIX support?

AIX uses /etc/filesystems instead of /etc/fstab and its a completely different format.

AIX also does not have /etc/mtab.

Thanks for the consideration.

the best check for mountpoints, and actively maintained!
Thanks

Thank you for the plugin, which does exactly what I need. I find the '-a' option is particularly useful, but an exclusion list option to complement that would be super, as there may be some mountpoints I explicitly don't want to check.

One other more minor annoyance is that an unknown status is returned if there are no mounts to check. An option to silence that would also be super!

We run tons of nfs mounts across many machines and those mounts sometimes change based on needs so I made a slight modification to avoid the need to keep track of the mountpoints to provide as args. Especially since it's likely no one will tell me beforehand.

Before the while case loop I did:

MPS=`grep nfs /etc/fstab | grep -v '^#' | awk '{print $2}' | tr 'n' ' '`

then commented out the /*MPS line in the case block.

The check seems to work, but I get a warning for line 140, when I supply multiple mount points:

./check_mountpoints.sh: line 140: [: /srv/apache/etc: binary operator expected

I fixed it by quoting the variable:

if [ -z "${MPS}" ]; then

Thanks, that's what I needed!
To use it with NRPE, do the following:

1. add the following to commands.cfg on the Nagios server:
"
# check mountpoints
define command {
command_name check_mountpoints
command_line $USER1$/check_nrpe -H $HOSTADDRESS$ -p 5666 -c check_mountpoints -a $ARG1$ $ARG2$
}
"

2. Copy / move / download the plugin to the target server's plugin directory.

3. Adjust the NRPE config on the target server to be monitored:

#:nano /etc/nagios/nrpe.d/template.cfg
add:
"command[check_mountpoints]=/usr/lib/nagios/plugins/check_mountpoints.sh $ARG1$"

4. Restart the NRPE damon
#:/etc/init.d/nagios-nrpe-server restart

5. Add a service check on the nagios server..

If NFS is mounted, and connection to server is lost after, then script hangs.
NFS share is listed after mount command, but it also hangs after df -h.

We use Windows shared folders to mount as CIFS shares on our Linux servers for backups. The below modifications where made to get it working for our CIFS mount points. Overall, the script works as advertised, and I have tested (CentOS 5) to confirm the script properly errors when mount is not attached or if folder is moved.

CHANGES:
1. Since our mounts are CIFS, we changed all occurrences of "nfs" to "cifs".

2. Commented out section checking fstab. I would prefer to see this as a function in the script and have it be an optional check if the mount point is in fstab.
commented out:
# awk '{if ($3=="cifs"){print $2}}' /etc/fstab | grep -q ${MP} &>/dev/null
# if [ $? -ne 0 ]; then
# log "WARN: ${MP} don't exists in /etc/fstab"
# ERR_MESG[${#ERR_MESG[*]}]="${MP} don't exists in /etc/fstab"
# fi

Hi everyone, I am using this plugins and I got this error:
"Unknown service Usage: check_nfs_mounts.sh [-m FILE] $NFSmountpoint [$NFSmountpoint2 ...]"

Even though after a while it displays OK status, in the next schedule the same error appears again..

I don't know what could be happened.

Please, if anyone can help me...

In Debian Squeeze with NFS v4 support, partitions in /proc/mounts are show as
X.X.X.X:/path/to/remote /path/to/local nfs4 rw,...

In the script you must to change
grep -q /proc/mounts -e " ${MP} nfs " &>/dev/null
to
grep -q /proc/mounts -e " ${MP} nfs4 " &>/dev/null

And it'll work nice for you.

This was a good starting point, but I think I have improved on this template, while also simplifying it. Check out my modifications.

http://dpaste.de/nGQj/

One of the things I added was a reverse check, ot see if everything mounted on the system is also listed in fstab.

Nice works,
I did a modification on the file to works on debian lenny.

Comment
#LIBEXEC="/usr/local/nagios/libexec"
Replace by
LIBEXEC="/usr/lib/nagios/plugins"