Shell Programming and Scripting

Hi all,

I've been trying to get this script right for the past few days and just can't figure out where I'm going wrong. Since one of my internet connections is extremely flaky I want to run this script to ping every 5 minutes and then reboot the router via telnet if I don't get a response.

Here's the script currently. Notice I've commented out the script and replaced it with an echo. This is because the script was rebooting the router every 5 minutes regardless of if the ping returned a result or not. Please note I'm running this on cygwin (hence the slightly different ping command). Unfortunately that's the only choice I have since the hardware is rather limited and I can't run a VM with linux on it. However I'm also testing it on a remote debian shell and it does the same thing.

Code:

#!/bin/sh

count=$(ping ii.net 56 3 | grep 'received' | awk -F',' '{ print $2 }' | awk '{ print $1 }')
if [ "$count" -eq 0 ]; then ECHO "bsasdasd"
#  ./reboot2.sh
fi

reboot2.sh works just fine. It logs into the router and reboots. The ping returns a 0 when I've taken the connection down and returns 3 when it's up. I just can't figure out why the script ignores that and reboots the router anyway.

I've tried both;

Code:

if [ "$count" -eq 0 ];

This returned a "unary code expected" error. Changing it to;

Code:

if [ ${count:-0} -eq 0 ] || [ -z "$count" ]; then

gets rid of the error but the script will still just reboot the router.

I'm sure there's something simple wrong here but I'm new to all this and I've been learning as I go. Any help would be appreciated.

My ping is slightly different (requires -w for timeout and -n for count):

Code:

$ ping ii.net -w 56 -n 3
Pinging ii.net [203.0.178.114] with 32 bytes of data:
Reply from 203.0.178.114: bytes=32 time=72ms TTL=243
Reply from 203.0.178.114: bytes=32 time=72ms TTL=243
Reply from 203.0.178.114: bytes=32 time=73ms TTL=243
Ping statistics for 203.0.178.114:
    Packets: Sent = 3, Received = 3, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 72ms, Maximum = 73ms, Average = 72ms

This gives me the Received amount in count:

Code:

$ count=$(ping ii.net -w 56 -n 3 | grep -i 'received' | awk -F',' '{ print $2 }' | awk '{ print $3 }')
$ echo $count
3

and this test seems to do the trick:

Code:

$ if [ "$count" -ne 3 ] ; then ECHO "fjdkfjk" ; fi
$ if [ "$count" -eq 3 ] ; then ECHO "fjdkfjk" ; fi
fjdkfjk

Note here that tcpip is a little flakey and 1 lost packet will cause you to reboot, perhaps you should ping 10 times and ensure at least 7 get thru.

Nothing looks funny on the surface.

If I were debugging the problem, I'd add a set -x to the top of the script to see what actually was being evaluated in the if. Is it possible that there is more than one line coming back from the ping with the word 'received' in it? Don't have a cygwin box, so I cannot test the exact ping to see what the output is; something like that could really cause issues.

@Chubler_XL -- this is really a nit, but ping uses icmp and not TCP.

Thanks guys. That seems to have done it! The script works perfectly when I run it manually however the cron job returns "integer expression expected" for both of the if [ "$count" lines. Not sure why it would do that. Some bug in cygwin's cron maybe?

echo $count before going for the if condition. Guess the variable's value is not numeric hence you get that error.
Also as suggested by agama there are possibilities of getting more than one -word 'received' from grep which maybe the reason for this error.

Huge thanks to everyone who helped out. It's working now. Turns out I needed to put the full path to ping in there instead of just the command, so ping wasn't running and there was no numerical output.