0

I have a process run by cron every hour. Such process calls another one (not sourced), ie: ./childProc.sh (Not sure if this detail is important)

Because of an endless loop in child process (which I'm still trying to find out why it does not end), I have to kill all processes related before run it again.

I've found this post and this article (Kill PPIDs section), but I don't understand the difference between: kill -9 -$PPID and kill -9 -- -$PPID.

My questions are:
What means here --? On the other hand, is -9 (SIGKILL) the right signal to kill a bash script process or is it too strong?

And lastly, my approach to check that behaviour (whether a process is still running) is creating a file that contains PID and PPID:

#!/bin/bash

# PID FILE 
PID_FILE="$BASE_BACKUP_DIR"/".pid.$HOSTNAME"; 

## If exists PID_FILE, kill 
if [[ -f "PID_FILE" ]]; then 
    pid_last=$(head -1 | cut -d " " -f1); 
    ppid_last=$(head -1 | cut -d " " -f2); 
    echoErr "Process is still running. Parent $ppid_last, child: $pid_last. "; 
    kill -9 -- -"$ppid_last"; 
fi;

## Create PID file
echo "$$ $PPID" > "$PID_FILE"; 

####### All process ####### 

# at the end, delete PID_FILE 
rm -f "$PID_FILE";
exit 0; 

What do you think? Any improvement?

1 Answer 1

2

There is no difference between kill -9 -$PPID and kill -9 -- -$PPID.

There is a difference between kill -$PPID and kill -- -$PPID. The -- marks the end of option arguments: any argument after --, even if it begins with a -, will not be treated as an option. Without the --, the -$PPID will be treated as the signal to be sent.

IMHO SIGKILL is too strong. Start with SIGINT or SIGHUP instead, or the default, SIGTERM.

1
  • I'll take a look at the differences of each signal, although I think they all will work because I do not control any signal, nor it is run with nohup or any other command that captures signals. Thanks.
    – Albert
    Dec 24, 2014 at 14:32

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .