Multiple IP Loop

Shell Programming and Scripting
1

Hey all,

I'm trying to do some very basic scripting. I been scouring the board looking at examples but I just can't seem to piece something together.

Edit: Now that I look at it, maybe my request is a little more complex and I'm biting off more than I can chew...

  1. Take a list of IPs/hosts (hosts.txt)
cat Hosts.txt
host1
host2
host3

For each host:

  1. Continually Ping with Timestamp
  2. Append each in their own log file (host1.log, host2.log, host3.log)
  3. Every 30 seconds perform an nmap -sP - Append another individual log file (host1-nmap.log, host2-nmap.log, host3-nmap.log)
  4. Run this script maybe in crontab so that it keeps running until you run a "service stop"?

Honestly, I'm not entirely looking just for the answer, would just appreciate some tips, examples and kinda where to look.

Thanks!

2

Let us start with 1.: Because commands in a script are only processed sequentially you will have to send the jobs in the background to have them run concurrently:

while read chHost ; do
     ping "$chHost" &
done < hosts.txt
  1. is quite simple, just add a redirection:
while read chHost ; do
     # ping "$chHost" > "$chHost".log &           # create new logfile every run
     ping "$chHost" >> "$chHost".log &            # append to existing log
done < hosts.txt

For 3. set up a second loop:

while : ; do
     while read chHost ; do
          nmap -sP $chHost >> ${chHost}-nmap.log
     done < hosts.txt
     sleep 30
done

That leaves the starting/stopping issue. I'd use a PID file in "/var/run" and use that to eliminate the process:

#! /bin/ksh

DispHostList ()
{
typeset chHost=""

while reach chHost ; do
     print - "$chHost"
done < /path/to/hosts.txt

return 0
}



pStop ()
{
typeset fPID="$1"

if [ ! -e "$fPID" ] ; then
     print -u2 "ERROR: there is nothing to stop"
     exit 1
else
     kill -15 "$(cat "$fPID")"
fi

return 0
}



pStart ()
{
typeset PID="$$"
typeset chHost=""
typeset fPID="$1"

if [ -e "$fPID" ] ; then
     print - u2 "ERROR: another instance seems to run."
     exit 1
else
     print - "$PID" > "$fPID"
fi

DispHostList |\
while read chHost ; do
     # ping "$chHost" > "$chHost".log &           # create new logfile every run
     ping "$chHost" >> "$chHost".log &            # append to existing log
done

while : ; do                                      # this will loop forever
     DispHostList |\
     while read chHost ; do
          nmap -sP $chHost >> ${chHost}-nmap.log
     done
     sleep 30
done

}


# main()
typeset fRun="/var/run/pinglog.pid"

case "$1" in
     start)
          pStart "$fRun"
          ;;

     restart)
          pStop "$fRun"
          pStart "$fRun"
          ;;

     stop)
          pStop "$fRun"
          ;;

     *)
          print -u2 "ERROR: do not know how to handle \"$1\""
          exit 1
          ;;
esac

exit 0

I hope this helps.

bakunin

3

Wow, my head is spinning a bit.

So, I had gotten close to the list thing stumbling through, but essentially took your advice verbatim and did the following:

#!/bin/bash

#while read chHost ; do
# ping "$chHost" &
#done < hosts.txt

while read chHost ; do
 # ping "$chHost" > "$chHost".log &             # Create new log ever run
 ping "$chHost" | xargs -n1 -i bash -c 'echo `date +"%m-%d %H:%M:%S"`" {}"' >> ./Ping/"$chHost".log &
done < hosts.txt

Which works great. But I noticed it was a bear to try and do any sort of PS since it spawned so many. The second PID portion of your suggestion REALLY has my head spinning and thus I'll need to re-read it.

So, in the meantime, I've just adjusted the script to do the following (which is probably a better I idea anyway)

ping -c 28800  "$chHost" | xargs -n1 -i bash -c 'echo `date +"%m-%d %H:%M:%S"`" {}"' >> ./Ping/"$chHost".log &

Thanks a ton!

4

As much as i would like to take the praise for this idea: it is common within many UNIX programs (daemons mostly) to create a file with the PID (process ID) in /var/run at the start and use the PID stored there to shut down the process.

I hope this helps.

bakunin