check_disk on Solaris 8 and 9

[Frank, Jason]

From: nagios-users-admin at lists.sourceforge.net
[mailto:nagios-users-admin at lists.sourceforge.net] On Behalf Of Aaron
Carr
Sent: Tuesday, November 08, 2005 12:14 PM
To: nagios-users at lists.sourceforge.net
Subject: Re: [Nagios-users] check_disk on Solaris 8 and 9


At this point, I've tried copying the df binary from a known good
machine, with no effect.  I still get the "Disk "" not mounted or
nonexistant" error.

I also tried copying the df from /usr/xpg4/bin/ to /usr/sbin/, same
thing.  No difference.

Aaron


On 11/8/05, Luc I. Suryo <luc at suryo.com> wrote: 

	
	
	> Has anyone resolved the issue with the check_disk plugin on
Solaris
	> returning "Disk "" not mounted or nonexistant"?
	>
	> I have two Solaris machines that were here before I started
this position 
	> that both return the above error when check_disk is run. One
is Solaris 8,
	> one is 9, both are SPARC. The error occurs regardless of
whether I run it
	> via ssh (which is how all of my actual checks get run) or
locally. It's also 
	> the same result whether I leave it blank (to check all
partitions), use
	> mount points, or use disk devices.
	>
	> I searched the archives and found several emails about the
error, but none
	> seemt to successfully address the problem. 
	>
	
	the reasom, I believe is do the df. use/defined during
compilation..
	/usr/xpg4/bin/df -Pk
	and the -P is unknow to Solaris
	
	'hack' the sources with correct flags and you should be ok, I
had to do 
	that too, running on solaris 7, 8, 9(x86/sparc), and
10(x86/sparc) all working properly
	
	
	-ls
	


I think you're misunderstanding what Luc is saying.  The problem is that
the df command that you're running does not accept the -P option, so it
give up, and leaves everything in a funky state.  The version in
/usr/xpg4/bin does:
 
bash-2.05$ /usr/xpg4/bin/df -Pk
Filesystem           1024-blocks        Used   Available Capacity
Mounted on
/dev/md/dsk/d1          36357261    10268500    25725189    29%    /
/proc                          0           0           0     0%    /proc
mnttab                         0           0           0     0%
/etc/mnttab
fd                             0           0           0     0%
/dev/fd
swap                     3604088          24     3604064     1%
/var/run
swap                     3604080          16     3604064     1%    /tmp

 
but the version in /usr/bin does not:
bash-2.05$ /usr/bin/df -Pk
df: unknown option: P
Usage: df [-F FSType] [-abeghklntVv] [-o FSType-specific_options]
[directory | block_device | resource]

 
Copying the version from /usr/xpg4/bin to /usr/sbin won't fix anything,
since it's likely trying to run from /usr/bin.  If you copy the version
from /usr/xpg4/bin to /usr/bin, you're going to be disappointed in how
Solaris reacts.  The proper fix is to change the check_disk program to
use the version in /usr/bin/xpg4.  To be honest, I don't remember if the
path is set by your PATH during configure, a hard coded path, or your
PATH at runtime.  Perhaps someone can fill in the blank there.  It would
be easy enough to test with changing your path though.
 
Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.monitoring-lists.org/archive/users/attachments/20051108/668a9857/attachment.html>