[NTLUG:Discuss] Finding huge files...

[Kevin Brannen]

Steve Southwell wrote:

> Odd thing happened to me yesterday... One of my clients' Apache 
> servers just bombed, and SSL wouldn't run, but normal http would.  As 
> it turned out, the ssl_engine_log had hit the 2GB limit.  It took me 
> awhile to find out what happened, so I wrote this little script to 
> find the largest files on the filesystem.  I'd be interested in 
> feedback or improvements, or perhaps I'm missing some sort of built-in 
> command to do this...
>
> #!/bin/sh
> # findbig.sh - Find the largest files on the filesystem
> cd /
> for line in ` du -a -b 2>/dev/null| sort -g -r |head -n 200| sed 
> s/^[0-9]*s*//g$
> do
> if [ -f $line ]; then
>    echo `du -h $line`
> fi
> done


I just got this emailed to me the other day from "Unix Tip of the Day" 
(see http://www.ugu.com/sui/ugu/show?tip.today to check them out):

   find / -xdev -size +1024 -exec ls -al {} \; | sort -r -k 5

though I find -exec inefficient and prefer xargs to do the work, so:

   find / -xdev -size +1024 -print | xargs ls -al | sort -r -k 5

and in testing it out, I found files with spaces, I like the big files 
at the bottom, I wanted to start at the current dir, and I wanted only 
the 10 biggest, so:

   find . -xdev -size +1024 -print0 | xargs -0 ls -al | sort -k 5 | tail -10

Is Unix great or what?! :-)

HTH,
Kevin