Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!usc!snorkelwacker!bloom-beacon!eru!luth!sunic!mcsun!cernvax!chx400!urz.unibas.ch!doelz
From: doelz@urz.unibas.ch
Newsgroups: comp.sys.sgi
Subject: Re: strange error : file system full
Message-ID: <1990Jun5.101622.767@urz.unibas.ch>
Date: 5 Jun 90 09:16:21 GMT
References: <1990Jun5.034812.762@urz.unibas.ch>
Organization: University of Basel, Switzerland
Lines: 27

In article <1990Jun5.034812.762@urz.unibas.ch>, kaul@arnor.watson.ibm.com (Anil Kaul) writes:
>    " no space on dev 406."


A core dump filled your file system. Create a soft link from your 
working directory to the /tmp partition if there is enough space. 
If there is none, create a file named core and give no permission to 
write to this file to noone. Additionally, because you need 
to do this in *any* directory you intend to dump core but you don't want 
to have it there, have a look into /usr/spool/cron/crontabs/root
ans see whether the gargabe collection is commented out there. Now, 
*copy the file 'root' into 'root.sav' as a backup.*   Then, 
remove the words -atime , -mtime and their arguments. If 
you are doing lots of core dumps, you could specify a shorter 
interval for removing cres, e.g. replace the 
0 5 * * * 
by 
0 5,10,15,20 * * * 
which will start the daemon at 5 am, 10 am, 3 pm and 8 pm. 
If you are neurotic, do 
0 * * * * 
end every hour all cores will be removed. Note that this approach will 
slow down performance because the 'find' will take some I/O resources.   



- Reinhard