root/alternc/trunk/tools/top_http_users

#!/bin/ksh
#
# $Id: top_http_users 22 2005-04-11 17:21:15Z jerome $
# ----------------------------------------------------------------------
# AlternC - Web Hosting System
# Copyright (C) 2003 by the AlternC Development Team.
# http://alternc.org
# ----------------------------------------------------------------------
# Based on:
# Valentin Lacambre's web hosting softwares: http://altern.org/
# ----------------------------------------------------------------------
# LICENSE
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License (GPL)
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# To read the license please visit http://www.gnu.org/copyleft/gpl.html
# ----------------------------------------------------------------------
# Original Author of file: Jerome Moinet
# Purpose of file: Parse the apache logs to give the n most active users
# ----------------------------------------------------------------------
#
PATH=""
PROG_NAME=top_http_users
PROG_VERSION=0.1.0
ALTERNC_ROOT=/var/alternc
ALTERNC_ETC=/etc/alternc
ALTERNC_LIB=/usr/lib/alternc
ALTERNC_CONF_FILE=$ALTERNC_ETC/local.sh
LOG_DIR=/var/log/apache
TMP_ROOT=$ALTERNC_ROOT/tmp
RES_FILE=$TMP_ROOT/$PROG_NAME.res.$$
INTERMEDIATE_FILE=$TMP_ROOT/$PROG_NAME.int.$$
LOCK_FILE=/var/run/$PROG_NAME
export TEXTDOMAIN=alternc-admintools
YES=yes
NO=no

# Be sure to use the right programs
# and be sure they are there
awk=/usr/bin/awk
grep=/bin/grep
cat=/bin/cat
zcat=/bin/zcat
head=/usr/bin/head
id=/usr/bin/id
sort=/usr/bin/sort
rm=/bin/rm
mysql=/usr/bin/mysql
sed=/bin/sed
cut=/usr/bin/cut
tail=/usr/bin/tail
ls=/bin/ls
gettext=/usr/bin/gettext
printf=/usr/bin/printf
lockfileremove=/usr/bin/lockfile-remove
lockfilecreate=/usr/bin/lockfile-create
lockfiletouch=/usr/bin/lockfile-touch

# Must have gettext first to display error messages
[ -x "$gettext" ] || { echo "Cannot execute $gettext"; exit 1 ; }

for i in $awk $grep $cat $zcat $head $id $sort $rm $mysql $sed $cut $tail $ls $printf $lockfileremove $lockfilecreate $lockfiletouch
do
        [ -x "$i" ] || { echo "$($gettext "Unable to execute") ${i}."; exit 1 ; }
done


#-------------------------
set_messages()
#-------------------------
{
        # Language-dependent messages
        # Uses gettext and mo files.
        # Don't change these messages, change the .po file instead.

        USAGE=$($gettext -e "Usage: top_http_users [ options ] number\n\ntop_http_users is a program that gives brief statistics\non apache usage by parsing the apache logs.\n\nOptions:\n  -h, --help           This help text.\n  -v, --version        Show version.\n  -z, --use-gz-logs    Parse gzipped and .1, ...n apache logs instead of just parsing the current log.\n  -s, --ssl            Parse the apache-ssl logs instead of parsing the non-ssl apache logs.\n  -n, --number=NUMBER  parse the NUMBER last lines of the current log.\n                     This option is not compatible with the --use-gz-logs option.\nSee the top_http_users(8) manual page for more information.")
        NOT_FOUND_MSG=$($gettext "does not exist.")
        NON_NUM_MSG=$($gettext "The \"number\" argument must be a number.")
        NON_COMPATIBLE_MSG=$($gettext "The -n and -z options are not compatible.")
        NON_NUM_MSG_FOR_N=$($gettext "The -n option requieres a number as argument.")
        LOCKFILE_CREATION_FAILED=`$printf "$($gettext "%s is allready beeing executed.")" $PROG_NAME`
        NON_ROOT_MSG=$($gettext "You have to be root (uid 0) to execute this program.")
        MISSING_PROG=$($gettext "Unable to execute")
        HIT_RES_MSG=`$printf "$($gettext "Top %s domains served by apache, sorted by number of lines in log (using gzipped logs: %s):")" $NB_USERS $($gettext "$USE_GZ_LOGS")`
        SIZE_RES_MSG=`$printf "$($gettext "Top %s domains served by apache, sorted by size (using gzipped logs: %s):")" $NB_USERS $($gettext "$USE_GZ_LOGS")`
        TIME_RES_MSG=`$printf "$($gettext "Top %s domains served by apache, sorted by execution time in seconds (using gzipped logs: %s):")" $NB_USERS $($gettext "$USE_GZ_LOGS")`
        ACCOUNT_HIT_RES_MSG=`$printf "$($gettext "Top %s AlternC accounts served by apache, sorted by number of lines in logs (using gzipped logs: %s):")" $NB_USERS $($gettext "$USE_GZ_LOGS")`
        ACCOUNT_SIZE_RES_MSG=`$printf "$($gettext "Top %s AlternC accounts served by apache, sorted by size (using gzipped logs: %s):")" $NB_USERS $($gettext "$USE_GZ_LOGS")`
        ACCOUNT_TIME_RES_MSG=`$printf "$($gettext "Top %s AlternC accounts served by apache, sorted by execution time in seconds (using gzipped logs: %s):")" $NB_USERS $($gettext "$USE_GZ_LOGS")`
        DEBUG_1_MSG=$($gettext "Parsing")
        DEBUG_2_MSG=$($gettext "Getting account for each domain")
        DEBUG_3_MSG=$($gettext "Printing results")
        MISSING_CONF_FILE=`$printf "$($gettext "Can't find %s. Are you sure AlterncC is properly installed?")" $ALTERNC_CONF_FILE`
        MYSQL_UNREACHABLE_DATABASE=`$printf "$($gettext "Cannot access accounts database. Please check either %s or Mysql state.")" $ALTERNC_CONF_FILE`
        SQL_ERROR_MSG=$($gettext -e "Sorry, an sql error appeared. The error message is:\n%s")
        UNKNOWN_OPTION=$($gettext "Unknown %s option.")
}


#-------------------------
trap_EXIT()
#-------------------------
{
        # Does some cleaning
        $rm -f $RES_FILE
        $rm -f $INTERMEDIATE_FILE
        $lockfileremove $LOCK_FILE
        exit
}
trap trap_EXIT INT KILL TERM QUIT ABRT STOP HUP


#-------------------------
# Main
#-------------------------
set_messages
# Must be root
[ "`$id -u`" -ne 0 ] && { echo $NON_ROOT_MSG ; exit 1 ; }

# Parse args
IS_N_PARAM=false
USE_GZ_LOGS="$NO"
N_PARAM=""
COMMAND=$cat
for ARG in "$@" ; do
        [ "$IS_N_PARAM" = "true" ] && { shift ; IS_N_PARAM=false ; continue ; }
        [ "`$printf "$ARG\n" | $cut -c1`" != "-" ] && [ "$IS_N_PARAM" = "false" ] &&  break

        case $ARG in
                -h | --help        ) echo $PROG_NAME version $PROG_VERSION ; $printf "$USAGE\n" ; exit ;;
                -v | --version     ) echo $PROG_NAME version $PROG_VERSION ; exit ;;
                -s | --ssl         ) LOG_DIR=/var/log/apache-ssl ;;
                -z | --use-gz-logs ) USE_GZ_LOGS="$YES" ;;
                -n | --number=*    ) 
                        if [ "$ARG" != "-n" ] ; then
                                N_PARAM=`echo $ARG | $cut -d"=" -f2`
                        else
                                N_PARAM=$2
                                IS_N_PARAM=true
                        fi
                        [ `echo "$N_PARAM" | $grep -c [^0-9]` != 0 ] && { echo "$NON_NUM_MSG_FOR_N" ; exit 1 ; }
                        COMMAND="$tail -n $N_PARAM"
                        ;;
                *) $printf "${UNKNOWN_OPTION}\n" $ARG ; exit 1 ;;
        esac
        shift
done
LOG_FILE=$LOG_DIR/access.log

# -n and -z options are not compatible
[ "$USE_GZ_LOGS" = "$YES" ] && ! [ -z "$N_PARAM" ] && { echo $NON_COMPATIBLE_MSG ; exit 1 ; }
# Must have only 1 parameter
[ -z "$1" ] || [ "$#" -gt 1 ] && { $printf "$USAGE\n" ; exit 1 ; }
# Argument "number"  must be numeric
[ `echo "$1" | $grep -c [^0-9]` != 0 ] && { echo "$NON_NUM_MSG" ; exit 1 ; } || NB_USERS=$1
# These dirs/files must exist
[ -d "$LOG_DIR" ] || { echo "$LOG_DIR $NOT_FOUND_MSG" ; exit 1 ; }
[ -d "$TMP_ROOT" ] || { echo "$TMP_ROOT $NOT_FOUND_MSG" ; exit 1 ; }
[ -f "$LOG_FILE" ] || { echo "$LOG_FILE $NOT_FOUND_MSG" ; exit 1 ; }
# Have to get AlternC conf file :
[ -f "$ALTERNC_CONF_FILE" ] || { echo $MISSING_CONF_FILE ; exit 1 ; } && . $ALTERNC_CONF_FILE
# Must have access to mysql to retreive accounts owning domains :
mysql="$mysql --defaults-file=/etc/alternc/my.cnf -B -N -e"
$mysql "select count(*) from domaines_standby;" > /dev/null 2>&1
[ "$?" != 0 ] && { echo "$MYSQL_UNREACHABLE_DATABASE" ; exit 1 ; }
# Prevents executing more than one shell at the same time
$lockfilecreate --retry 1 $LOCK_FILE
if [ $? != 0 ]
then
        echo $LOCKFILE_CREATION_FAILED
        exit 1
fi
$lockfiletouch $LOCK_FILE &
BADGER="$!"


# Does the stuff
set_messages
# Have to parse files one by one or else system wil go on knees
[ "$DEBUG" ] && echo "$DEBUG_1_MSG $LOG_FILE" ; $COMMAND $LOG_FILE | $awk '{z=NF-1 ; domaine[$NF]++ ; if ($10 != "-") size[$NF]+=$10 ; time[$NF]+=$z} END {for (item in domaine) print item" "domaine[item]" "size[item]" "time[item]}' > $RES_FILE
for FILE in `$ls -1 $LOG_FILE.* | $grep -v "\.gz$"`; do
        [ "$USE_GZ_LOGS" = "$YES" ] && [ -f $FILE ] && { [ "$DEBUG" ] && echo "$DEBUG_1_MSG $FILE" ; $cat $FILE | $awk '{z=NF-1 ; domaine[$NF]++ ; if ($10 != "-") size[$NF]+=$10 ; time[$NF]+=$z} END {for (item in domaine) print item" "domaine[item]" "size[item]" "time[item]}' >> $RES_FILE ; }
done
if [ "$USE_GZ_LOGS" = "$YES" ] 
then
        for GZLOG in $LOG_FILE.*.gz
        do
                [ "$DEBUG" ] && echo "$DEBUG_1_MSG $GZLOG"
                $zcat $GZLOG | $awk '{z=NF-1 ; domaine[$NF]++ ; if ($10 != "-") size[$NF]+=$10 ; time[$NF]+=$z} END {for (item in domaine) print item" "domaine[item]" "size[item]" "time[item]}' >> $RES_FILE
        done
fi


# show results
$cat $RES_FILE | $awk '{domaine[$1]+=$2 ; size[$1]+=$3 ; time[$1]+=$4} END {for (item in domaine) print item" "domaine[item]" "size[item]" "time[item]}' > $INTERMEDIATE_FILE

[ "$DEBUG" ] && echo $DEBUG_2_MSG
> $RES_FILE
for i in `$cat $INTERMEDIATE_FILE | $sed s/" "/"@@@@"/g`
        do
                TMP=`echo $i | $sed s/"@@@@"/" "/g`
                DOMAIN=`echo $TMP | $cut -d " " -f1 | $sed s/\"//g | $sed s/"\\\\\\\\"/""/g`
                [ "$DEBUG" ] && echo DOMAIN : $DOMAIN
                ACCOUNT=`$mysql "select a.login, a.mail from membres a, sub_domaines b where a.uid = b.compte and concat(if(sub=\"\", \"\", concat(sub, \".\")), domaine)  = \"${DOMAIN}\";" 2>&1`
                [ "$?" != 0 ] && { $printf "$SQL_ERROR_MSG\n" "  $ACCOUNT" ; kill "${BADGER}" ; trap_EXIT ; }
                ! [ -z "$ACCOUNT" ] && [ `echo $ACCOUNT | $grep -c "^ERROR"` = 0 ] && echo "$ACCOUNT $TMP" >> $RES_FILE
        done


[ "$DEBUG" ] && echo $DEBUG_3_MSG
echo $HIT_RES_MSG
$cat $INTERMEDIATE_FILE | $awk {'printf ("%20.0f %s\n", $2, $1)'} | $sort -gr | $head -n$NB_USERS
echo ""
echo $SIZE_RES_MSG
$cat $INTERMEDIATE_FILE | $awk {'printf ("%20.0f %s\n", $3, $1)'} | $sort -gr | $head -n$NB_USERS
echo ""
echo $TIME_RES_MSG
$cat $INTERMEDIATE_FILE | $awk {'printf ("%20.0f %s\n", $4, $1)'} | $sort -gr | $head -n$NB_USERS

$cat $RES_FILE | $awk '{size[$1]+=$5 ; time[$1]+=$6 ; hit[$1]+=$4 ; mail[$1]=$2} END {for (item in size) print size[item]" "time[item]" "hit[item]" "item" "mail[item]}' > $INTERMEDIATE_FILE

echo ""
echo $ACCOUNT_HIT_RES_MSG
$cat $INTERMEDIATE_FILE | $awk {'printf ("%20.0f %s (%s)\n", $3, $4, $5)'} | $sort -gr | $head -n$NB_USERS
echo ""
echo $ACCOUNT_SIZE_RES_MSG
$cat $INTERMEDIATE_FILE | $awk {'printf ("%20.0f %s (%s)\n", $1, $4, $5)'} | $sort -gr | $head -n$NB_USERS
echo ""
echo $ACCOUNT_TIME_RES_MSG
$cat $INTERMEDIATE_FILE | $awk {'printf ("%20.0f %s (%s)\n", $2, $4, $5)'} | $sort -gr | $head -n$NB_USERS


# cleanly exit and remove temp files
kill "${BADGER}"
trap_EXIT