Script to wakeonlan server2 if server1 fails

Question

I need help to update this script to if ping fails it would send a wakonlan call to another host (besides the email which is sent now if ping fails). How can this be done from this script?

This is what i like to achive:

Server1 is up > PingComputer pings server1 = everything is fine. Server 1 fails > PingComputer sends mail AND sends wakeonlan call to server 2 = server2 starts up.

Not so important (only if its possible): After a fail on server1 i would like to get a mail when server1 is up again and then this scenario starts over.

    #!/bin/bash
HOSTS="IP ADRESS"
COUNT=4
for myHost in $HOSTS
do
    count=$(ping -c $COUNT $myHost | grep 'received' | awk -F',' '{ print $2 }' | a$
    if [ $count -eq 0 ]; then
        # 100% failed
        echo "Server failed at $(date)" | mail -s "Server Down" [email protected]
        echo "Host : $myHost is down (ping failed) at $(date)"
    fi
done

Im using ubuntu server on all 3 computers btw.

Answer 1

Specification

At least one host must be always pingable (ping success = UP).
When a pingable host becomes unreachable (ping failed) => mail and:
- if another host is already pingable => no problem => continue
- else => wakeonlan another host
When an unreachable host becomes pingable back => mail
When a host is still unreachable after wakeonlan it => mail

Script

The initial script was based on bash. But your Ubuntu Server seems to use a bash version too old without associative array. Fortunately, zsh supports associative array for a while. Therefore the following script has been rewriten for zsh. Moreover support for MAC ADDRESS has been also added.

Script one_host_should_ping.sh:

#!/bin/zsh

[email protected]
NEXTHOSTWOL=${1?Please provide hosts as arguments of the script}
PINGABLECOUNT=0

# use associative array to store the state of each host
unset      HOSTS
typeset -A HOSTS
unset      MAC
typeset -A MAC
while (($#))
do
  let PINGABLECOUNT++
  HOSTS[$1]=PINGABLE
  MAC[$1]=$2
  echo >&2 -e "Input #$PINGABLECOUNT:\tIP=$1\tMAC=$2"
  shift 2
done

nexthostwol() {
  for host in ${(k)HOSTS[@]}
  do
    if [[ ${HOSTS[$host]} == UNREACHABLE ]] 
    then
      NEXTHOSTWOL=$host
      break
    fi
  done
}

report() {
  for host in ${(k)HOSTS[@]}
  do
    echo "HOST IP=$host MAC=${MAC[$NEXTHOSTWOL]} state=${HOSTS[$host]}"
  done
  echo "Number of hosts Up = $PINGABLECOUNT"
}

pingable() {
  let PINGABLECOUNT++
  HOSTS[$1]=PINGABLE
  SUBJECT="[Server Up] ping $1 success at $(date +%c)"
  echo "$SUBJECT"
  report | mail -s "$MESSAGE" "$MYADDRESS"
  report
  [[ $NEXTHOSTWOL == $1 ]] && nexthostwol
}

unreachable() {
  let PINGABLECOUNT--
  HOSTS[$1]=UNREACHABLE
  SUBJECT="[Server Down] ping $1 failed at $(date +%c)"
  if [[ $PINGABLECOUNT -le 0 ]] 
  then 
    wakeonlan -i $NEXTHOSTWOL ${MAC[$NEXTHOSTWOL]}
    HOSTS[$NEXTHOSTWOL]=WAKEONLAN
    SUBJECT="$SUBJECT => wakeonlan $NEXTHOSTWOL"
  fi
  echo "$SUBJECT"
  report | mail -s "$MESSAGE" "$MYADDRESS"
  report
}

stillwol() {
  SUBJECT="[Server Down] ping $1 failed at $(date +%c)"
  echo "$SUBJECT"
  report | mail -s "$MESSAGE" "$MYADDRESS"
  report
}

# infinite loop => check every 5 minutes 
# use CTRL+C or kill to break it
while true
do
  for host in ${(k)HOSTS[@]}
  do
      STATE=${HOSTS[$host]}
      case $STATE in
         UNREACHABLE) ping -c 1 $host && pingable $host      ;;
         PINGABLE)    ping -c 1 $host || unreachable $host   ;;
         WAKEONLAN)   ping -c 1 $host && pingable $host || stillwol $host ;;
         *)           echo >&2 "ERROR: $host is in an unexpected state=[$STATE]";;
      esac
  done

  echo "will check again in 5 minutes"
  sleep $((5*60))
done

Usage

Give execution permission:

chmod +x one_host_should_ping.sh`

If you use two hosts:

./one_host_should_ping.sh  IP1 MAC1 IP2 MAC2

If you use five hosts:

./one_host_should_ping.sh  IP1 MAC1  IP2 MAC2  IP3 MAC3  IP4 MAC4  IP5 MAC5

Example

root@myc:/usr/local# ./one_host_should_ping.sh 192.168.0.197 00:1c:26:5c:7e:d5 192.168.0.187 f4:6d:04:e5:5c:4c 
Input #1: IP=192.168.0.197 MAC=00:1c:26:5c:7e:d5 
Input #2: IP=192.168.0.187 MAC=f4:6d:04:e5:5c:4c 
PING 192.168.0.187 (192.168.0.186) 56(84) bytes of data. 
From 192.168.0.168 icmp_seq=1 Destination Host Unreachable 
--- 192.168.0.187 ping statistics 
--- 1 packets transmitted, 0 received, +1 errors, 100% packet loss, time 0ms 
[Server Down] ping 192.168.0.187 failed at tor 26 jan 2012 12.44.29 
HOST IP=192.168.0.186 MAC=00:1c:26:5c:7e:d5 state=UNREACHABLE 
HOST IP=192.168.0.197 MAC=00:1c:26:5c:7e:d5 state=PINGABLE 
Number of hosts Up = 1 
PING 192.168.0.197 (192.168.0.197) 56(84) bytes of data. 
From 192.168.0.168 icmp_seq=1 Destination Host Unreachable
--- 192.168.0.197 ping statistics 
--- 1 packets transmitted, 0 received, +1 errors, 100% packet loss, time 0ms 
Sending magic packet to 192.168.0.197:9 with 00:1c:26:5c:7e:d5 
[Server Down] ping 192.168.0.197 failed at tor 26 jan 2012 12.44.36 
=> wakeonlan 192.168.0.197 
HOST IP=192.168.0.187 MAC=00:1c:26:5c:7e:d5 state=UNREACHABLE 
HOST IP=192.168.0.197 MAC=00:1c:26:5c:7e:d5 state=WAKEONLAN 
Number of hosts Up = 0 
will check again in 5 minutes