Shell Programming and Scripting

Hi,

I've been using the following commands in my automated scripts, to ensure that all text output is sent to a log file instead of to the screen:

Code:

exec 1>>$SCRIPT_LOG_FILE
exec 2>>$SCRIPT_LOG_FILE

However, I've now discovered that the system used for automating the script executions actually saves all screen output to its own log file anyway. It would be useful for me to write all future scripts so that all stdout and stderr text is sent to the screen (to be saved to the automation log) as well as to the $SCRIPT_LOG_FILE specified above.

I'm aware that I could use something like this on each command within my scripts from now on:

Code:

command 2>&1 | tee -a $SCRIPT_LOG_FILE

...and this would give me the desired result, but it's a bit tedious to have to do that after every command within the script.

I'm also aware that I the scripts could be called with a tee argument:

Code:

my_script.sh 2>&1 | tee -a $SCRIPT_LOG_FILE

...and this would also give me the desired result, but I don't want to complicate the automated commands used to call scripts.

Finally, I'm also aware that I could use a 'wrapper script' for each one, and put a command like that above within it, and this would also have the desired effect, but it's a bit of a messy solution and I'd rather not have to double up on the number of script files I need, just to accomplish this.

Apologies if this has already been answered before, but I've searched quite a bit and haven't been able to find exactly what I'm looking for yet.

Thanks for any help provided.

If you are using bash you can redirect to process.

Code:

exec > >(tee -a "$script_log_file") 2>&1

so everything from then on goes to tee

Hi neutronscott,

Thanks for the quick reply. Unfortunately I'm using ksh on Solaris 10 (probably should have mentioned that), it's the standard shell used for all of our scripts. I saw that method and tried it already, but it didn't work under ksh. A pity as it's quite elegant!

I think I might have found something that does work though:

Code:

#!/bin/ksh
 
LOGFILE="logfile.log"
PIPEFILE="pipefile"
 
mkfifo $PIPEFILE
 
# Start tee writing to a logfile, but pulling its input from our named pipe.
tee $LOGFILE < $PIPEFILE &
 
# capture tee's process ID for the wait command.
TEEPID=$!
 
# redirect the rest of the stderr and stdout to our named pipe.
exec > $PIPEFILE 2>&1
echo "Make your commands here"
echo "All their standard out will get teed."
echo "So will their standard error" >&2
 
# close the stderr and stdout file descriptors.
exec 1>&- 2>&-
 
# Wait for tee to finish since now that other end of the pipe has closed.
wait $TEEPID

Taken from here: Force bash script to use tee without piping from the command line - Super User

This seems to work ok under ksh, but I'm still testing it...

This is basically what that syntax does... Creates a piped background process. This code is good, except I always recommend to quote variable expansions (What if a filename ends up having a space in it?) and avoid all capital variables (What if we accidentally overlook its use as an environment/internal variable).

ie:

Code:

mkfifo "$pipefile"

Thanks for the feedback Scott.