[Nagiosplug-devel] [ nagiosplug-Bugs-1481615 ] Problem with check_disk

[SourceForge.net]

Bugs item #1481615, was opened at 2006-05-04 08:04
Message generated for change (Comment added) made by ralph_roessner
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=397597&aid=1481615&group_id=29880

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: General plugin execution
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Tobias Mucke (mucke)
Assigned to: Ton Voon (tonvoon)
Summary: Problem with check_disk

Initial Comment:
Hi everybody,

yesterday I tried the new plugin version 1.4.3. I
experienced a problem with check_disk. Here is the
output. Filesystem /home is a reiserfs. I guess that's
the reason for inode=nan%?. But that's not the problem.
Look at the performance data. In the MIN field. It
should be 0 but it is -2147483648.

/opt/nagios/plugins/check_disk -w 20 -c 10 -p /home
DISK OK - free space: /home 480 MB (94% inode=nan%);|
/home=32MB;491;501;-2147483648;511

Thanks in advance.

Tobias


----------------------------------------------------------------------

Comment By: Ralph Rößner (ralph_roessner)
Date: 2006-05-08 20:08

Message:
Logged In: YES 
user_id=1515003

Hi everybody!

I have a similar problem on Solaris, only I get bitten by
validate_arguments(), which complains that "C_IDFP
(67065044...467904.000000) should be less than W_IDFP
(0.0)". It appears that the minimum free inode percentage
was not correctly initialized. This only appears when file
systems are selected with "-p".

The reason I suspect is that the c_idfp and w_idfp members
of the path list structure are never initialized in this
case, at least not at the point where c_dfp and w_dfp are
(check_disk.c lines 457ff). So random malloc()ed memory is
passed to validate_arguments(), and the results vary
according to architecture and phase of the moon.

I created a patch that treats the [cw]_idfp like [cw]_dfp at
the point where the "-p" argument is parsed. This makes the
problem go away for me. If there are other places where the
path list is added to then there is a good chance that the
same oversight happened there too.

Since I don't find a way to attach a file to a thread I did
not open, I will include the patch here:

--8><-----------------------------------
*** nagios-plugins-1.4.3/plugins/check_disk.c.orig      Mon
May  8 19:50:33 2006
--- nagios-plugins-1.4.3/plugins/check_disk.c   Mon May  8
19:49:48 2006
***************
*** 462,467 ****
--- 462,469 ----
        se->c_df = c_df;
        se->w_dfp = w_dfp;
        se->c_dfp = c_dfp;
+       se->w_idfp = w_idfp;
+       se->c_idfp = c_idfp;
        se->found = 0;
        se->found_len = 0;
        *pathtail = se;
--8><-----------------------------------

Hope this helps,
   Ralph Rößner

----------------------------------------------------------------------

Comment By: Tobias Mucke (mucke)
Date: 2006-05-05 16:20

Message:
Logged In: YES 
user_id=481892

Ton,

checked it, but same problem.

----------------------------------------------------------------------

Comment By: Ton Voon (tonvoon)
Date: 2006-05-05 10:37

Message:
Logged In: YES 
user_id=664364

Tobias,

Can you check the snapshot at http://nagiosplug.sf.net/snapshot? There was a 
fix reported for inode problems on Solaris, but probably applicable to reiserfs 
too.

Ton

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=397597&aid=1481615&group_id=29880