Shell Programming and Scripting

I got the solution,

In s1.sh script:
---------
capture pid of current script
S1PID=??

#call s2.sh in the background
s2.sh &

#capture the pid of the most recent background process
S2PID=$!

#wait for s2.sh to finish
#and capture the return code of s2.sh
wait $S2PID
RC=$?

Here, even if the s2.sh finishes in no time, S2PID=$! will still return it's pid and 'wait $S2PID' will finish with return code of s2.sh, even if s2.sh has finished long before the wait command was issued.

The reason why i want to get the pid, is to keep a list of kicked off process and cleanup any 'hung up' process in future. Even though the pid's are reused, the pid+time combination will be unique.

Earlier i said i don't want to run s2 in background, that's because i didn't want to continue s1 without s2 finishing first. However, if i can capture the pid of s2 only by running it in background, then i better run it that way and then wait for it to finish, before continuing with s1.


Thanks to all for your input.
-srinivas yelamanchili

Underneath *all* processes are really executed in the background, it's what the parent chooses to do between the fork and the wait that makes the difference.