#!/usr/bin/perl -w
# df-analyze v0.10, a script to analyze log files of the format:
# df for MM.DD
# <df output follows>
# This is free software, released in the hope that it will be useful,
# but without even the WARRANTY OF MERCHANTABILITY or FITNESS FOR A
# PARTICULAR PURPOSE.  Let me say it again: NO WARRANTY AT ALL.  This
# script *should* be perfectly safe, but if something bad happens, then
# the person running this script is responsible for running it.
# You are free to use this software in any manner you see fit.  License
# to modify and redistribute this work is given under the GNU GPL,
# available at http://www.gnu.org/copyleft/gpl.html.
#
# In other words, if it breaks, you get to keep both pieces.
# Copyright 2000, Bill Jonas, with assistance from Chris Fearnley (text
# formatting and output).

# Right now, the only argument that this script takes is a filename, and
# that's required.
@ARGV || die "Specify a file\n";
@ARGV eq 1 || die "Oops!\nUsage: $0 <filename>\n";
open (FILE, "$ARGV[0]") || die "Can't open file: $!";

# Let's get the stuff that's not going to change out of the way, so we
# don't keep reassigning it over and over
<FILE>;
SETUP: while (<FILE>) {
    if (/^\/dev\/(\w+)\s+(\d+)\s+\d+\s+\d+\s+\d{1,3}%\s+(\/[^\s]*$)/) {
# (Device) (Total) Used Available Capacity% (Mountpoint)
        $dfmaster{$1}{total}=$2;
        $dfmaster{$1}{mountpoint}="$3";
    } elsif (/^df for \d\d\.\d\d$/) {
        last SETUP;
    }
}

seek (FILE,0,0); # Back to the beginning of the file

# Set things up so that the comparisons later will work
foreach $filesystem (keys %dfmaster) {
    $dfmaster{$filesystem}{maxused}=-1;
    $dfmaster{$filesystem}{minused}=undef; # Nicer than assigning it some
}                                          # arbitrarily large value.
                                           # Thanks to michael@toren.net.

# Set up a few housekeeping variables for gathering stats:
$totaldays=0; $today=0;

# Now, let's get our stats
while (<FILE>) {
    if (/^\/dev\/(\w+)\s+\d+\s+(\d+)\s+(\d+)\s+(\d{1,3})%\s+\/[^\s]*$/) {
# (Device) Total (Used) (Available) (Capacity)% Mountpoint
        $dfmaster{$1}{total_used}+=$2;
        $dfmaster{$1}{total_avail}+=$3;
        $dfmaster{$1}{total_used_pct}+=$4;
    if ($2 > $dfmaster{$1}{maxused}) {
        $dfmaster{$1}{maxused}=$2;
        $dfmaster{$1}{dayofmax}=$today;
    }
    unless (defined $dfmaster{$1}{minused}) {
        $dfmaster{$1}{minused}=$2;
        $dfmaster{$1}{dayofmin}=$today;
        } elsif ($2 < $dfmaster{$1}{minused}) {
            $dfmaster{$1}{minused}=$2;
            $dfmaster{$1}{dayofmin}=$today;
        }
    } elsif (/^df for (\d\d\.\d\d)$/) {
        $totaldays++;
        $today=$1;
    }
}

foreach $filesystem (keys %dfmaster) {
    foreach $item ("used","avail","used_pct") {
        $dfmaster{$filesystem}{"avg_$item"}=$dfmaster{$filesystem}{"total_$item"} / $totaldays;
    }
}

# foreach $filesystem (sort keys %dfmaster) {
#     print "$filesystem: ";
#     foreach $item (sort keys %{ $dfmaster{$filesystem}}) {
#         print "$item=$dfmaster{$filesystem}{$item} ";
#     }
#     print "\n";
# }

print "Mount    Device    Capacity       Range of      Daily Avg      % Used\n";
print "Point                             Space Used    Space Used           \n";
print "-----------------------------------------------------------------------\n";

foreach $filesystem (sort keys %dfmaster) {
        printf("%-9s %-8s %8sK %8sK-%-sK %8.0fK       %2.2f%%\n",
          $dfmaster{$filesystem}{'mountpoint'}, $filesystem, 
          $dfmaster{$filesystem}{'total'}, 
          $dfmaster{$filesystem}{'minused'}, 
          $dfmaster{$filesystem}{'maxused'}, 
          $dfmaster{$filesystem}{'avg_used'},
          (($dfmaster{$filesystem}{'avg_used'})/$dfmaster{$filesystem}{'total'})*100);
}

__END__

sda1: avg_avail=923993.583333333 avg_used=81930.375 avg_used_pct=9.125
  dayofmax=08.24 dayofmin=08.16 maxused=131055 minused=59671 mountpoint=/
  total=1018298 total_avail=22175846 total_used=1966329 total_used_pct=219 
sda6: avg_avail=2461132.25 avg_used=668371.916666667
  avg_used_pct=22.6666666666667 dayofmax=08.24 dayofmin=08.18 maxused=830988
  minused=610222 mountpoint=/usr total=3168186 total_avail=59067174
  total_used=16040926 total_used_pct=544