Solaris

I read the sticky and thought of a script I use on a regular basis. Since unless you patch/upgrade the df command on solaris you have a very tought time teling how full the system truly is.

Output looks like

Code:

[xxxxx admin]$ biggest.sh /tmp
Filesystem            kbytes    used   avail capacity  Mounted on
/dev/md/dsk/d4       4133838  425419 3667081    11%    /tmp
05/13/06 10:45:27
find /tmp  -size +499999c -exec ls -lc {} \; -o -fstype nfs -prune -o -name cdrom\* -prune -print 2> /dev/null | nawk
   68,193,280 root     May 09 05:01 /tmp/05-09-06.dmp
   65,979,392 root     May 12 05:01 /tmp/05-12-06.dmp
   62,684,160 root     May 11 05:01 /tmp/05-11-06.dmp
   61,280,256 root     May 10 05:01 /tmp/05-10-06.dmp
   60,184,576 root     May 13 05:01 /tmp/05-13-06.dmp
   58,414,080 root     May 08 05:01 /tmp/05-08-06.dmp
   51,793,920 root     May 07 05:01 /tmp/05-07-06.dmp
    1,723,464 mysql    May 05 14:09 /tmp/temp.csv
                        00       
                        00       
                        00       
                        00       
                        00       
                        00

---------Start Script
#!/bin/sh
#
# SccsId[] = "%W% %G% (List 'biggest' files in filesystem)"
#
#----------------------------------------------------------------------#
# biggest.sh #
# -------------------------------------------------------------------- #
# Program documentation and notes located at the bottom of script. #
#----------------------------------------------------------------------#

#----------------------------------------------------------------#
# Script_name assignment is necessary if there exists the #
# possibility that this process may be run by the 'at' command. #
# Run via 'at' and $0 simply returns '/bin/sh' or 'sh' (hardly #
# desirable if you run that into basename). #
#----------------------------------------------------------------#
script_name="biggest.sh"
[ $0 = "/bin/sh" -o `dirname $0` = "." ] \
&& script_home=`pwd` || script_home=`dirname $0`
bin=/usr/bin # Default

#----------------------------------------------#
# Use awk, nawk or gawk, depending on the OS. #
#----------------------------------------------#
OZ=`uname -s 2> /dev/null | tr '[a-z]' '[A-Z]' 2> /dev/null`
if [ ."$OZ" = ."HP-UX" ]; then
AWK=awk
elif [ ."$OZ" = ."LINUX" ]; then
bin_dir=/bin
AWK=gawk
elif [ ."$OZ" = ."SUNOS" ]; then
AWK=nawk
else # Unknown OS, see if there's any kind'a Awk available.
if [ -f $bin/gawk ]; then AWK=gawk
elif [ -f $bin/nawk ]; then AWK=nawk
elif [ -f $bin/awk ]; then AWK=awk
elif [ `expr "\`awk 2>&1\`" : 'Usage: '` -gt 0 ]; then AWK=awk
else # This is really getting awkward :-o
echo "Unable to locate [gn]awk program! $0 terminating." 1>&2
exit 1 # Well behaved here
fi
fi


#======================================================================#
# L O C A L F U N C T I O N S #
# (in alphabetical order) #
#----------------------------------------------------------------------#
EXIT_USAGE()
#----------------------------------------------------------------------#
{
echo "Usage: biggest.sh -fHh -l <nn> -s <nnn> -t <dir> -v fs\n" 1>&2
echo " -f = follow links" 1>&2
echo " -H = Full documentation" 1>&2
echo " -h = Usage brief" 1>&2
echo " -l = Displays <nn> lines" 1>&2
echo " -s = Minimum file size is <nnn>" 1>&2
echo " -t = Temp/work directory, <dir>" 1>&2
echo " -v = Edit (vi) file list" 1>&2
echo " fs = Required filesystem argument." 1>&2
echo "" 1>&2
exit 1
}

#----------------------------------------------------------------------#
SHOW_DOCUMENTATION() # Function documentation located at bottom. #
#----------------------------------------------------------------------#
{
#----------------------------------------------------------------#
# If the following variables are not set, use these as defaults. #
#----------------------------------------------------------------#
: ${script_name:=`basename $0`}
: ${script_home:=`dirname $0`}
SD_script_home=`echo $script_home | sed 's/\/*$/\//'`

#------------------------------------------------#
# User wants help, so find the documentation #
# section and print everything from there down. #
#------------------------------------------------#
$AWK -v script_name=$script_name \
'BEGIN { n=0 }

{ #------------------------------------------#
# Until we find the documentation section, #
# keep looking at each line. #
#------------------------------------------#
if (n == 0)
{
if ($0 ~ /^# +D O C U M E N T A T I O N/)
{
n = NR
print line
print $0
}
else
{
line = $0
}

next
} #-------------------------------------#
else # Once we find it, print until EOF. #
{ #-------------------------------------#
print
}
}

END {
if (n == 0) # Means there is no documentation section.
{
"date +%Y-%m-%d" | getline yyyy_mm_dd
print yyyy_mm_dd" NO DOCUMENTATION",
"section found for "script_name".\a" | "cat 1>&2"
exit 1 # Exit failure
}
exit 0 # Else exit success
}' ${SD_script_home}$script_name

exit $?
} # "SD_" prefix identifies this function's variables


#======================================================================#
# I N I T I A L I Z A T I O N #
#======================================================================#
opt_v=0 # Default 'vi' option (0 = Do NOT vi the file list)
tmp=/var/tmp
follow=""
size="499999" # Default minimum filesize
lines="500" # Default maximum lines

while getopts fHhl:s:t:v opt 2> /dev/null
do
case "$opt" in
f ) follow='-follow' ;;
H ) SHOW_DOCUMENTATION;;
h ) EXIT_USAGE ;;
l ) lines="$OPTARG" ;; # Max number of lines to display.
s ) size="$OPTARG" ;; # Minimum file size.
t ) tmp="$OPTARG" ;; # Temp directory (if /var/tmp full)
v ) opt_v=1 ;;
* ) echo "Ignoring invalid option, $1.";;
esac
done
#----------------------------------#
# Shift past options to arguments. #
#----------------------------------#
shift `expr $OPTIND - 1`


#======================================================================#
# M A I N #
#======================================================================#

[ $# -eq 0 ] && EXIT_USAGE

#-----------------------------------------------------------------#
# Ensure we have write-access to temp/work directory. #
#-----------------------------------------------------------------#
if [ ! -d $tmp ]; then
echo "Temp/work directory, $tmp not found!" \
"\n$script_name terminated."
exit 1
elif [ ! -w $tmp ]; then
echo "No write access to temp/work directory, $tmp!" \
"\n$script_name terminated."
exit 1
fi

#----------------------------------------------------------------#
# File lists of remote filesystems is problematic, so we limit #
# our operations to local filesystems only. #
#----------------------------------------------------------------#
df -lk $1
if [ $? -ne 0 ]; then
echo "$1 MUST be a local filesystem--it is not!" \
"\n$script_name terminated."
exit 1
fi

#----------------------------------------------------------------#
# Build a 'find' command with the necessary options/arguments. #
# Be sure to exclude anything with cdrom in it and include -xdev #
# -xdev if the filesystem being searched is root (/). #
#----------------------------------------------------------------#
outfile=$tmp/$LOGNAME"_biggest.files" # Formatted 'find' output
include='-size +'"$size"'c -exec ls -lc {} \;'
exclude='-o -fstype nfs -prune -o -name cdrom\* -prune'
[ ."$1" = ."/" ] && find_opt="-xdev $follow" || find_opt="$follow"

date "+%D %T"
find_cmd="find $1 $find_opt $include $exclude -print"

#----------------------------------------------------------------#
# Display find command before running it. Use [gn]awk to format #
# the output and sort it in descending order (biggest on top). #
#----------------------------------------------------------------#
echo "$find_cmd 2> /dev/null | $AWK"
eval $find_cmd 2> /dev/null | $AWK \
'BEGIN \
{
i = 0
own = 3
siz = 5
mmm = 6
day = 7
yyy = 8 # This may actually be yyyy or hh:mi
Mon = "^(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)$"
}
# "! /^[bcd]/" skips block and character devices and directories
! /^[bcd]/ \
{
#-----------------------------------------------------------#
# If it looks like owner and group fields are concatonated, #
# try backing up the field ($n) list and work with that. #
#-----------------------------------------------------------#
if ($siz !~ /[0-9]+/ && !match($mmm,Mon))
{
if ($(siz-1) ~ /[0-9]+/ && match($(mmm-1),Mon))
{
siz = 4 # 4th field
mmm = 5 # Etc.
day = 6
yyy = 7
}
}

gsub(/[\t ]+/," ") # Squeeze whitespace.
gsub(/./,"& ",$siz) # Isolate each digit,
q=split($siz,a," ") # then split the $siz into an array.
$siz="" # Clear $siz.
for (p=1;q>0;q--) # Insert commas into $siz.
{
$siz=a[q]""$siz
if ((p%3) == 0 && q != 1) $siz=","$siz # Insert commas here
p++
}

printf("%13s %-8s %s %02d %-5s %s\n",
$siz, $own, $mmm, $day, $yyy, $NF)

#-----------------------------------------------#
# if size value is not 5, then reset it, et al. #
#-----------------------------------------------#
if (siz != 5)
{
siz = 5
mmm = 6
day = 7
yyy = 8
}
}' | sort -r -k 1,2 | head -$lines > $outfile

#----------------------------------------------------------------#
# Unless 'vi' option was given, simply cat our file list. #
#----------------------------------------------------------------#
if [ `wc -l < $outfile` -eq 0 ]; then
echo "No files found in $1 > $size bytes in size."
else
[ $opt_v -eq 0 ] && cat $outfile || vi $outfile
fi

exit $?


#======================================================================#
# D O C U M E N T A T I O N #
#======================================================================#
# #
# Author: Bob Orlando (Bob@OrlandoKuntao.com) #
# #
# Date: April 8, 1995 #
# #
# Program ID: biggest.sh #
# #
# Code Contrl: aphrodite:~dmc/SCCS. #
# #
# Usage: biggest.sh -fHh -l <nn> -v -t <dir> -s <nnn> fs #
# #
# -f = Follow links #
# -H = Displays detailed documentation #
# -h = Provides usage brief #
# -l = Displays <nn> lines (default is 500) #
# -s = Minimum file size is <nnn> #
# (default is 500K) #
# -t = Use <dir> as temp/work directory #
# (default is /var/tmp) #
# -v = Edit (vi) file list #
# fs = Required filesystem argument. #
# #
# Purpose: List biggest files in a given filesystem (files #
# appear in descending order). #
# #
# Description: Using the find command, descend through the specified #
# file system (fs) listing all files whose sizes exceed #
# either the default minimum size (500K) or the minimum #
# value provided via size (-s) option. The filelist #
# is created in /var/tmp by default as it usually much #
# larger than /tmp. However, in the event that /var #
# is the filesystem that's full (or is not writable to #
# the user), the temp dir (-t) option is available to #
# redirect the output elsewhere. #
# #
# When root is the directory being searched, -xdev is #
# supplied as a find argument so only root, and none #
# of its subdirectories, is searched. #
# #
# With the 'vi' option (-v) the user can edit the #
# normally cat'd file list. #
# #
# Modified: 2005-03-02 Bob Orlando #
# v1.6 * Add code to parse the correct fields when #
# the owner and group fields are concatonated, #
# effectively making two fields, one (really #
# messes up AWK processing). #
# #
#----------------------------------------------------------------------#