[NTLUG:Discuss] /var out of space....help please

[Christopher Browne]

On Tue, 17 Oct 2000 17:55:27 EDT, the world broke into rejoicing as
"./aal" <al_h at technologist.com>  said:
> Mandrake7.1
> K6III-400
> 192MB
> ALI AladinV
> 
> I have a strange problem...
> I am getting numerous errors about /var/(something) being out of space
> I can ls as any user and see plenty of space[250MB left on that part.]
> 
> I have looked on deja.com and MDK support histories to no avail
> I am getting strange errors from logrotate:
> /var/log/various_logs_.1.gz.1.gz.1.gz...... on and on : file not found
> 
> (10-16-2000)
> df:
> Filesystem            Size  Used Avail Use% Mounted on
> /dev/hda5             988M  238M  700M  25% /
> /dev/hda1              91M   10M   76M  12% /boot
> /dev/hda7             1.0G  740M  242M  75% /home
> /dev/hda8             2.9G  2.4G  297M  89% /usr
> /dev/hda9             3.3G  2.7G  511M  84% /opt
> /dev/hda10            494M  9.2M  459M   2% /tmp
> /dev/hda11            380M  258M  102M  72% /var
> /dev/hdb5             1.6G  1.4G  186M  88% /hdb/usr
> /dev/hdb6             1.4G  1.0G  318M  77% /hdb/opt
> /dev/hdb8             213M  129M   73M  64% /hdb/home
> /dev/hdd1             2.9G  1.3G  1.5G  47% /hdd
> 
> and if I try 'updatedb'...
> /usr/bin/updatedb: line 29: 25351 Segmentation fault      slocate $FROM
> $VERBOSE -l $SECURITY -d $DATABASE -f "$FSEXCLUDE" -e "$EXCLUDE"
> 
> 
> (10-16-2000)
> df:
> Filesystem            Size  Used Avail Use% Mounted on
> /dev/hda5             988M  238M  700M  25% /
> /dev/hda1              91M   10M   76M  12% /boot
> /dev/hda7             1.0G  737M  245M  75% /home
> /dev/hda8             2.9G  2.4G  297M  89% /usr
> /dev/hda9             3.3G  2.7G  514M  84% /opt
> /dev/hda10            494M  381k  468M   0% /tmp
> /dev/hda11            380M  370M     0 100% /var
> /dev/hdb5             1.6G  1.4G  186M  88% /hdb/usr
> /dev/hdb6             1.4G  1.0G  318M  77% /hdb/opt
> /dev/hdb8             213M  129M   73M  64% /hdb/home
> /dev/hdd1             2.9G  1.8G 1005M  64% /hdd  
> 
> I am at a loss
> what has happened?

Presumably there's some log file in /var that is not getting rotated
on any schedule.

I had a laptop recently run out of disk space; it wasn't rotating logs
for Postfix, and a bit of misconfiguration was leading to there being
a _lot_ of messages.

I'd suggest doing:
# cd /var
# du -x | more

... And look for the directory that's consuming A Lot Of the 370MB.

It will likely be /var/log.

Then look in the offending directory, to see what files are the lion's
share of the space.

My policy in that case is to then add in a cfengine rule to rotate
the log file in question; Your Milage May Vary, particularly if you
have not got cfengine installed/configured.

A short term answer is to (assuming this results from a big log file )
delete the offending file.

Longer term, two answers leap out:
a) If a log file is Growing A Lot, then there's probably a problem
   with the configuration of a service.  Fix the service, and the
   growth of the log goes away (mostly).

b) Improve log file rotation configuration.  I add cfengine rules;
   your favorite distribution may have its own devoted "logrotate"
   utility.
--
cbbrowne at acm.org - <http://www.hex.net/~cbbrowne/lsf.html>
When sign makers go on strike, is anything written on their signs?