Shell Programming and Scripting

Hi Guys,

I found this script for monitoring the status of a services:

for i in syslogd cron; do
if [ `ps -eo comm | grep -v grep | grep -c ^$i` -gt 0 ] && [ `ps -eo comm,s|grep -v grep|grep ^$i|awk '{print $2}'|sort -u` = "A" ];
then
printf "%-8s" "$i";printf " is alive=A\n"
else
printf "%-8s" "$i";printf " is not alive\n"
fi

The script is working fine except if either syslogd or cron will have a defunct on the memory. It will have an error like this:

/usr/local/bin/checkserver.sh: line 9: [: too many arguments
syslogd is not alive

But of course, syslogd is running if I check with ps -ef | grep syslogd:

[itik@north001 ~]$ ps -ef|grep syslogd
root 1234 1 0 May04 ? 00:11:42/usr/sbin/syslogd
root 18555 5525 0 May04 ? 00:00:00 [syslogd] <defunct>
itik 22417 19840 0 13:23 pts/1 00:00:00 grep syslogd

Without error the output is like this:

cron is alive=A
syslogd is alive=A


Can someone please let me know where to change the error?

Thank you in advance.

Best Regards,
itik

The script doesn't quote the output from the backticks properly, and also contains a good sample of Useless Use of grep -c and Useless Use of grep | awk. Here's a quick attempt at straightening it out.

Code:

for i in syslogd cron; do
  if ps -eo comm | grep -v '[g]rep' >/dev/null && 
     ps -eo comm,s | awk '$1 == "'"$i"'" {if ($2 == "A") next; exit 1}'
  then
    printf "%-8s" "$i";printf " is alive=A\n"
  else
    printf "%-8s" "$i";printf " is not alive\n"
  fi
done

I hope I didn't miss anything important while refactoring this. For a less intrusive change, perhaps it would be enough to add double quotes around the backquotes.

I didn't change the logic, so it might still claim there is an error if there is a defunct process. Perhaps the awk script should be changed. I also don't really understand the usefulness of the remaining grep. If you could show the output of ps -eo comm and ps -eo comm,s when there is a defunct, it would be easier to see what the script is really supposed to do.

Code:

for p in syslogd cron; do 
  pgrep >/dev/null "$p" && printf "%s is alive\n" "$p" || 
    printf "%s is dead\n" "$p" 
done

What's =A ?

On ERA:

The response with your new script is not alive which is wrong because the service have two existence,one is the service and the other is a defunct. It should be alive.

One more thing, it should be alive if the service exist (and if service exist and a defunct exist). Otherwise if service doesn't exist, then is not alive

On ROD:


I think A is a tag so that one service can be trimmed to unique.

Thanks and more power.