tee acts as a T-switch, as a duplexer. It takes an input stream and sends it to two output streams. The following example gets a directory listing from stdin and outputs this as well on stdout as into the file terminal.out:
ls -al | tee terminal.out
To verify, do a
This is very useful in trainings and debugging.
Using script, you can create a sub-shell and capture all input and output into a logfile.
Using exec within a script, you can redirect output to a logfile: exec 1>logfile 2>&1 But then you cannot see the output on the screen. You can use tee within a script, but it just works within that one line. So, the question is, how do we tell our script to put its output as well on the screen as into a file.
How to capture command line screen text output in a logfile and still have it show on the screen
This sample code can be included in an ash/bash shell script:
rm PIPE1 mkfifo PIPE1 #cat >logfile1 <PIPE1 & cat <PIPE1 & rm PIPE2 mkfifo PIPE2 cat >logfile2 <PIPE2 & #cat <PIPE2 & rm PIPE0 mkfifo PIPE0 tee PIPE1 >PIPE2 <PIPE0 & exec 1>PIPE0 2>&1 # now all STDOUT and STDERR is copied to logfile2 and is also still sent to the screen echo starting lsof -R | grep PI #sample output echo bye
This is a great way to really get to know mkfifo and tee. After you make a named pipe with mkfifo, you must connect the pipe to an output before you can connect it to an input. The & at the end of the line makes the script keep going while the spawned subtask is running.
Note: This is the hard way to do this. There is a program called script that is supposed to be able to capture all input and output in a spawned sub-shell. And it is also easy to run a script and capture the output:
sh script.sh 2>&1 | tee logfile
The above example is for situations where you want to be able to do these kinds of dynamic plumping redirections from within a script, operating at the current shell level.