Softpanorama (slightly skeptical) Open Source Software Educational Society

May the source be with you, but remember the KISS principle ;-) Softpanorama Search

Unix find tutorial

Prev | Contents | Next

Part 8: Finding Files based on size: largest, empty and within certain range

Among categories of files with specific sizes the most popular are probably zero-length file and "huge files".

Zero length files formally might be considered a subclass of abnormal files but in reality they are often used in software installations.  Do not try to delete them unless you understand what you are doing. 

Here is some examples:

• To find all zero length files:

  find / -type f -size 0 -ls

  You can also use special predicate -empty

  $ find . -empty -exec ls {} \;

  After finding empty files, you might choose to delete them by replacing the ls command with the rm command. But it's better to verify the list before jumping the gun...

• To find all files bigger then 100M :

  find / -type f -size +100M -ls

• You can combine several size predicated to get "range", which is useful if you forgot the name of the file, but know its approximate size and some other parameters (for example age of the file):

  find / -type f -size +12K -size -15K -mtime -3 -ls

• Find top 10 largest files. This is  useful command when you need to free some space when a UNIX file system has become full. For example, Oracle will hang whenever Oracle must expand a tablespace and Oracle cannot extend the UNIX filesystem.

  find /data -type f -size +100M -printf "%s:%p\n" | sort -nr | head -10

  or, disregarding the minimum threshold of 100M  just

  find /data  type f -printf "%s:%p\n" | sort -k1rn | head -10

  Notes:

  • We assume that you are using GNU find
  • To print the top 10 largest "files and directories" you can use du -a . | sort -nr | head
  • For more examples you can also look at
    • InformIT UNIX Disk Usage Identifying the Biggest Files
    • How To Find the Largest Files in your Unix system — UNIX Tutorial Learn UNIX
    • Linux Find Large Files

As with abnormal files discussed in previous section it is important to exclude certain parts of the directory tree from searching.  You can use all approaches discussed in section 6 but often more practical and more simple to use a small Perl script to weed "junk" directories.  For example:

#!/usr/bin/perl
open(SYSLOG,"find / -type f -size 0 -ls|");
@pattern=(
 "/var/spool/",
 "/var/tmp/",
 "/var/locks/",
 "/var/ct/",
 "/var/ifor/",
 "/usr/lpp/",
 "/usr/omni/log/debug.log",
 "/tmp/",
 "/proc/", 
 "/.swdis/",
 "/etc/Tivoli/tec/",
 "/usr/omni/lib/perl",
 "/opt/TMF/db/nti2171.db/",
 "/opt/tivoli/EPProxy/",
 "/opt/tivoli/TWS/stdlist/",
 "/var/adm/logs/",
 "/oracle/app/oracle/product/9.2.0/",
 "/oracle/app/oracle/product/9.2.0_OLD/",
 "/usr/opt/perl5/lib/5.8.2/aix-thread-multi/",
 "/usr/opt/perl5/lib64/5.8.2/aix-thread-multi-64all/",
 "/usr/opt/perl5/lib/site_perl/5.8.2/",
 "/usr/lib/objrepos/",
 "/usr/aix/_AIX_print_subsystem",
 "/etc/ipsec/inet/DB/",
 "/etc/objrepos/"
 );
while() {
   $data=$_;
   $line=substr($data,65);
   chomp($line);
   $found=0;
   foreach $string (@pattern) {
      if ( index($line,"$string") > -1 ) {
         $found=1;
         last;
      }
   }	
   if ($found == 0 ) {print "$data\n";}
}  

Prev | Contents | Next