Townsville Linux Users Group

Hello,

This script will do a bunch of things - it's main purpose is to help keep apt based systems up to date. Check out the top part of ths script for a more in depth explanation. 

Run this script periodically via cron.  

NOTE: This script has been tested on Debian (Sarge) and Ubuntu (5.10), your mileage may vary on other systems using apt.

#!/bin/sh
# Script Name: maint.sh
# Author Name: Keith Bawden
# Date: Wed May 17 15:40:32 JST 2006
# Description: This script will:
#   Clean up the local apt repository of retrieved packages (apt-get clean)
#   Resync the package index (apt-get update)
#   If called with AUTOUPDATE set to yes then updates will be downloaded and applied with no feed back (not recommended)
#   If called without AUTOUPDATE then packages are downloaded and an email is sent informing which packages are to be updated.
#   And more ;-)
# NOTE: Perl is needed for this script to work.

#
# Make user configuration changes in this section
#

export REPLYTO=bawdo2001@example.com
MAILTO="bawdo2001@example.com"
AUTOUPDATE="no"
LOGFILE="/var/log/server_maint.log"
THISSERVER=`hostname --fqdn`

#
# End of user configuration section
#

DASHES="---------------------------------------------------------------------------------"
DASHES2="================================================================================="

# Check if the script is being run as root exit if it is not.
if [ "$UID" -ne "0" ]
then
  echo "[ERROR] This script must be run as root"
  exit 1
fi

function startlogging {
  echo $DASHES2 >> $LOGFILE
  echo "$0 started running at `date`" >> $LOGFILE
  echo $DASHES2 >> $LOGFILE
}

function stoplogging {
  echo "`date` [MESSAGE] $0 finished runnning" >> $LOGFILE
  echo $DASHES >> $LOGFILE
}function check_return {
  if [ "$?" -ne "0" ]
    then
      echo "`date` [ERROR]   $1 failed to run" >> $LOGFILE
      send_error_email $1
      stoplogging
      exit 1
  fi
  echo "`date` [SUCCESS] $1 ran without error" >> $LOGFILE
}

function send_error_email {
  mail -s "[$THISSERVER] There was an error whilst running $0" $MAILTO <<EOF
Hello,

Whilst running the update script ($0) on $THISSERVER there was a problem.

[ERROR] "$1" failed to run

The server has the following network interfaces configured ${SERVERADDS[@]}.

Please log in via ssh (e.g. ssh root@${IPADDR[0]}) and check the log file:

vim $LOGFILE

Regards.
EOF
}

# IP Address stuff
declare -a IPADDR
declare -a NICINTERFACE
declare -a SERVERADDS
index=0

for i in $( ifconfig | grep 'inet addr' | awk '{print $2}'| sed 's#addr:##g' );
do
  IPADDR[$index]=$i
  let "index += 1"
done

index=0

for i in $( ifconfig | grep 'eth' | awk '{print $1}' );
do
  SERVERADDS[$index]="$i ${IPADDR[$index]}"
  let "index += 1"
done

# End IP Address stuff
 
startlogging

apt-get clean > /dev/null
check_return "apt-get clean"

apt-get update > /dev/null
check_return "apt-get update"

if [[ "$AUTOUPDATE" == "yes" ]]
then
  apt-get -yqq upgrade > /dev/null
  check_return "apt-get -yq upgrade"
else
  PACKAGES_TO_BE_UPGRADED=`apt-get -Vs upgrade | perl -ne 'print if /upgraded:/ .. /upgraded,/'`
  apt-get -yqd upgrade > /dev/null
  check_return "apt-get -yqd upgrade"
fi

if [[ -z $PACKAGES_TO_BE_UPGRADED ]]
then
  echo "`date` [MESSAGE] No packages need updating." >> $LOGFILE
else
  mail -s "[$THISSERVER] server may need some updates applied" $MAILTO <<EOF
Hello,

Packages have been downloaded onto $THISSERVER.

$PACKAGES_TO_BE_UPGRADED

The server has the following network interfaces configured ${SERVERADDS[@]}.

To update the server log in via ssh (e.g. ssh root@${IPADDR[0]}) and run the following command:

apt-get upgrade

See the logfile for more info: vim $LOGFILE

Regards.
EOF

  echo "`date` [MESSAGE] Packages need updating email sent to $MAILTO" >> $LOGFILE
fi

stoplogging
exit 0

An email detailing required updates looks like this:

Subject:[server.example.com] server may need some updates applied
Hello,

Packages have been downloaded onto server.example.com.

The following packages will be upgraded:
   exim4 (4.50-8 => 4.50-8sarge2)
   exim4-base (4.50-8 => 4.50-8sarge2)
   exim4-config (4.50-8 => 4.50-8sarge2)
   exim4-daemon-light (4.50-8 => 4.50-8sarge2)
   libc6 (2.3.2.ds1-22 => 2.3.2.ds1-22sarge3)
   libc6-dev (2.3.2.ds1-22 => 2.3.2.ds1-22sarge3)
   libsasl2 (2.1.19-1.5 => 2.1.19-1.5sarge1)
   locales (2.3.2.ds1-22 => 2.3.2.ds1-22sarge3)
   mutt (1.5.9-2 => 1.5.9-2sarge1)
   perl (5.8.4-8sarge3 => 5.8.4-8sarge4)
   perl-base (5.8.4-8sarge3 => 5.8.4-8sarge4)
   perl-modules (5.8.4-8sarge3 => 5.8.4-8sarge4)
   tar (1.14-2.1 => 1.14-2.2)
13 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

The server has the following network interfaces configured eth0 192.168.10.1 eth1 192.168.1.220.

To update the server log in via ssh (e.g. ssh root@192.168.10.1) and run the following command:

apt-get upgrade

See the logfile for more info: vim /var/log/server_maint.log

Regards.

Here is a sample of the log output:

=================================================================================
./maint.sh started running at Wed May 17 15:13:40 JST 2006
=================================================================================
Wed May 17 15:13:40 JST 2006 [SUCCESS] apt-get clean ran without error
Wed May 17 15:13:40 JST 2006 [SUCCESS] apt-get update ran without error
Wed May 17 15:13:41 JST 2006 [SUCCESS] apt-get -yqd upgrade ran without error
Wed May 17 15:13:41 JST 2006 [MESSAGE] No packages need updating.
Wed May 17 15:13:41 JST 2006 [MESSAGE] ./maint.sh finished runnning
---------------------------------------------------------------------------------

Please posts any fixes/suggestions here. In particular how to rewrite this without needing perl :-)

Regards, Bawdo2001