Networking scripts

[Mystic_Slayer]

I want to create a script which verifies if my network is up every 5 minutes.

If it's not up, trying to start my network.

Does someone has that kinda script, or could someone tell me how to write that kinda script

[jasonlambert]

create directory:

mkdir /etc/cron.5

open /etc/crontab, add the following lines to the file:

5 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
10 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
15 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
20 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
25 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
30 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
35 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
40 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
45 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
50 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null
55 * * * * /usr/bin/run-parts /etc/cron.5 1> /dev/null

Create file /etc/cron.5/mynet.sh:

#!/bin/sh

/sbin/ifup eth0

make executable:

chmod 755 /etc/cron.5/mynet.sh

Done

[Mystic_Slayer]

I've had only one problem with this scripts. On of my network devices is DHCP. So i've killed 1500 processes from dhclient. Could you please advise me how to deal with this?
I know that dhclient has a pid file, but i don't know how to handle it.

[Mystic_Slayer]

does no one has a answer for this problem?

[Opnosforatou]

put the following into you rc.local

# Start the connection watcher
if [ -x /usr/local/bin/connection_watchdog.d ]; then
  . /usr/local/bin/connection_watchdog.d &
fi

Now bootup VI with vi /usr/local/bin/connection_watchdog.d and past the following into it.

#!/bin/bash
#variables
reconnect=0
maxreconnect=4
pingfailed=0
maxpingfailed=6
checkhost1="192.168.0.1" # An IP of one of your other machines will do.
buildconnection="ifup eth0"
stopconnection="ifdown eth0"
lockfile="/var/lock/NetCheck.lock"
logfile="/var/log/NetCheck.log" 
log="logger -s -i -t$0"
 
 
# this script may run only once ! (at a time)
if test -f $lockfile; then
      if ps -p `cat $lockfile` 1>/dev/null 2>&1; then
         $log "Already running, this instance will exit now..."; exit 1
      fi
   else $log "stale lockfile exists, I will remove it and continue..."
   rm -f $lockfile
fi
 
#create a new lockfile
echo $$ >$lockfile
 
#check connection forever......
while (true) ; do
        if ! ping -c 1 $checkhost1 2>/dev/null 1>&2
           then $log "ping to $checkhost1 failed, I'll try to ping $checkhost2..."
           if ! ping -c 1 $checkhost2 2>/dev/null 1>&2
               then $log "ping to $checkhost2 failed too... Assuming connection is lost.... (Re)connecting..."
               pingfailed=`expr $pingfailed + 1`
               $log "ping failed $pingfailed times in a row...."
               if test -$pingfailed -eq -$maxpingfailed; then
                  $log "ping failed maximum number of times ($pingfailed), waiting for 5 minutes before attempting reconnect..."
                  sleep 300
                  pingfailed=0
               fi
               # log this message in a special logfile to count disconnects and do this only once with every disconnect
               if test $connectfailed !="true"
                  then date >>$logfile; echo "disconnect">>$logfile
               fi
               # now try to reconnect....
               if ! $buildconnection
                  then $log "reconnect failed...retrying in about 20 seconds..." 2>>$logfile 1>&2
                  reconnect=`expr $reconnect + 1`
                  if test -$reconnect -eq -$maxreconnect ; then                    
                        $log "reconnect failed $reconnect times, waiting for 5 minutes before retrying.."
                        sleep 300
                        $reconnect=0
                  fi
               fi
           fi
           # ping success
           else
                        pingfailed=0
        #uncomment the followingline to flood your logfile.
        #               $log "Ping success. Waiting for 5min for next ping."
        fi
sleep 300
done

now make it executable with: chmod +x /usr/local/bin/connection_watchdog.d

Next time you'll bootup you'll see the program active with the command ps -ef

Enjoy.

Every time the eth0 is initialized it will create a DHCP call. That's why your other system is flooded with all those entries...
That's why i check if it's down by using ping instead of forcing the eth0 to go up every 5 min.

[Mystic_Slayer]

Thanks Opnosforatou

It works perfect. Because I'm still in learning phase, could you tell me where I can find information about scripting, etc?

[Opnosforatou]

Thanks Opnosforatou

It works perfect. Because I'm still in learning phase, could you tell me where I can find information about scripting, etc?

Try you local bookstore for books by publisher : O'Reilly.
And ofcourse google: Keywords scrip linux, or bash script.
If you are using a bash shell, usualy default in linux install.
There is tons of info available on the net...
I'm thinking about writing a starters guido to bash scripting.
Hope I can find the time...
Until then, goolge and Linux forum are your friends

[Mystic_Slayer]

Allright thanks. I will run to the bookstore, have some money left to spend.

[Mystic_Slayer]

Hmm after three days i was looking how good it works, but I found a problem again.

The dhclient still startsup everytime. I've had to delete 2300 dhclients. That's alot of work you know.

Did i something wrong?

I've looked in my NetCheck.log and see every 5 minutes:

Wed May 5 09:55:15 CEST 2004
disconnect
Wed May 5 10:00:18 CEST 2004
disconnect
Wed May 5 10:05:20 CEST 2004
disconnect
Wed May 5 10:15:24 CEST 2004
disconnect
Wed May 5 10:20:26 CEST 2004
disconnect
Wed May 5 10:25:31 CEST 2004
disconnect
Wed May 5 10:30:33 CEST 2004
disconnect
Wed May 5 10:35:36 CEST 2004
disconnect
Wed May 5 10:40:39 CEST 2004
disconnect
Wed May 5 10:50:43 CEST 2004
disconnect
Wed May 5 10:55:45 CEST 2004
disconnect
Wed May 5 11:00:46 CEST 2004
disconnect
Wed May 5 11:05:48 CEST 2004
disconnect
Wed May 5 11:10:51 CEST 2004
disconnect
Wed May 5 11:15:53 CEST 2004
disconnect
Wed May 5 11:25:57 CEST 2004
disconnect

[Opnosforatou]

# log this message in a special logfile to count disconnects and do this only once with every disconnect
if test $connectfailed !="true"
then date >>$logfile; echo "disconnect">>$logfile
fi

Remove these lines... These are flooding your log file.
It's a counter how often the checkfile runs...

Sorry :o