#!/bin/sh
#
# Program: Website content check <content-check>
#
# Author: Matty < matty91 at gmail dot com >
#
# Current Version: 1.2
#
# Revision History:
#
#  Version 1.2
#   Check for connection timeouts
#
#  Version 1.1
#    Added additional logging
#
#  Version 1.0
#    Initial Release
#
# Last Updated: 12-02-2006
#
# Purpose:
# The content-check script can be used to checksum a webpage returned from a server, and
# generate E-mail notifications if the content doesn't hash to a saved hash. This can
# help spot problems with servers that are accepting connections, but failing to return
# valid content (e.g., when a bad code deployment occurs).
#
# License:
#   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.
#
# Installation:
#   Copy the shell script to a suitable location
#
# Usage:
# To use content-check, you will need to first generate a checksum of the site
# you would like to monitor. You can acquire a checksum for a site by passing 
# the site to check to the "-g" option:
#
# $ content-check -g http://prefetch.net/index.html
# cd116f4c65952075ac3e374666f13c507685e5e7
#
# Once the hash is calculated, you can add the site and hash to a file:
#
# $ echo "http://prefetch.net/index.html cd116f4c65952075ac3e374666f13c507685e5e7" >> sites
#
# The file can contain one or more site / checksum key value pairs. This file
# can be passed as an argument to content-check's "-f" option. Content-check
# will retrieve the page listed in the file, and compare a current checksum
# to the one stored in the file. If content-check determines that the hashes 
# don't match, a syslog message will be logged, and an E-mail send to the
# website administrator (the E-mail address can be passed to the "-e" option).

PATH=/bin:/usr/bin:/sbin:/usr/sbin
export PATH

# Default user to send E-mail to
ADMIN="root"

# Paths to various binaries
FACILITY="daemon.notice"
LOGGER="/usr/bin/logger"
MAIL="/usr/bin/mail"
MKTEMP="/bin/mktemp"
OPENSSL="/usr/sfw/bin/openssl"
TIMEOUT=10
WGET="/usr/sfw/bin/wget"

# Useful functions
usage () {
	echo "Usage: $0 [ -e EMAIL_ADDRESS ] -g SITE || -f SITE FILE"
	echo "       -e address   : address to send alarms to"
        echo "       -f site file : file with list of sites and checksums"
        echo "       -g site      : generate checksum for site"
        echo "       -t timeout   : set a timeout value"
}

# If no options are provided, print a help screen
if [ $# -eq 0 ]
then
	usage
	exit 0
fi

# Parse the options that are passed to the script
while getopts e:g:f:t: option
do
        case "${option}"
        in
                e) ADMIN=${OPTARG} ;;
                f) SITES=${OPTARG} ;;
                g) GENERATE=${OPTARG} ;;
                t) TIMEOUT=${OPTARG} ;;
                \?) usage
		    exit 1;;
        esac
done

# Check to make sure the utilities are present
if [ ! -f ${WGET} ] || [ ! -f ${OPENSSL} ] || [ ! -f ${MAIL} ] || [ ! -f ${LOGGER} ] || [ ! -f ${MKTEMP} ]
then
	echo "ERROR: Cannot find the logger, mail, mktemp, wget or openssl utilities"
        usage
	exit 1
fi

# IF the "-g" option is present, print a checksum on the tty
if [ "${GENERATE}" != "" ]
then
	${WGET} -q -O - "${GENERATE}" | ${OPENSSL} sha1
	exit 0
fi

# Sanity check the arguments
if [ ! -f ${SITES}  ] && [ "${GENERATE}" != "" ]
then
	echo "ERROR: No options passed to script"
	usage
	exit 1
fi

# Create a temporary file to store the results of the wget
TMP=`${MKTEMP} /var/tmp/content.XXXXXX`

# Cycle through the file and send E-mail for each 
# checksum that doesn't match 
cat ${SITES} | while read SITE CHECKSUM
do
	# Make sure all entries are key / value pairs
	if [ "${SITE}" = "" ] || [ "${CHECKSUM}" = "" ]
	then
             ${LOGGER} -p ${FACILITY} "ERROR: Trouble processing ${SITE} in ${SITES}"
	     echo "ERROR: Trouble processing ${SITE} in ${SITES}"
	     exit 1
	fi

	# Grab the page from the website
	${WGET} -t 1 -T ${TIMEOUT} -q -O ${TMP} ${SITE}

        # If unable to resolve name or create a TCP connection, bail
        # out and log an error 
        if [ $? -ne 0 ]
        then
             ${LOGGER} -p ${FACILITY} "ERROR: Cannot establish connection to ${SITE}"
             echo "ERROR: Cannot establish connection to ${SITE}"
             exit 1
        fi

        # Generate a checksum
        CURRENTCHECKSUM=`cat ${TMP} | ${OPENSSL} sha1`

        # Make sure the checksums match
	if [ "${CHECKSUM}" != "${CURRENTCHECKSUM}" ]
	then
	    ${LOGGER} -p ${FACILITY} "ERROR: Content from ${SITE} did not hash to ${CHECKSUM}"
	    echo "ERROR: Content from ${SITE} did not hash to ${CHECKSUM}" \
                 | ${MAIL} -s "${SITE} did not pass health check" ${ADMIN}
        fi
done

# Remove the temporary file, but be paranoid and double check that it exists
if [ -f ${TMP} ]
then
     rm -f ${TMP}
fi

exit 0