Functional Scripting tricks

As I code and work more and more in this field of Information Technology that is now begin called DevOps, I begin to develop certain tricks that I form into a style.

Here are a few to play with:

- () { return; }

Now you can use the ‘-‘ at the beginning of a line as way
to leave comments. You should probably use this for structured comments that you might want to grab via reflection.

color() {
# The next line is functionally a comment too.
- alp color white red green blue
# If you overload the '-' function, then
# the color comment will become a function call
 if [ -n $1 ]; then
     case $1 in
       white)  FG='00m' ;;
       red)    FG='31m' ;;
       green)  FG='32m' ;; 
       blue)   FG='34m' ;;
      printf "\033[$FG";

You can experiment around with what symbols you can overload in this way. So far, I found:

, . - + @

In what my structured comment line, I begin with the word ‘alp’. I will now define what ‘alp’ does.

# alp creates a function for each option defined
alp ()
- alp $1 $2
echo "$1.$2() { $1 $2; }" > $T
source $T
rm $T
alp color red
alp color blue
alp color white
alp color green

Now you have four new functions, one for each color.

declare -f ()
    color red

So this is how I play with bash. It is very similar to how one might approach golang, python, or other functional languages.

Introducing sbstats (DevOps performance tracking)

For many years now, I have been capturing vmstat output and displaying it in a graphical LAMP solution.  It can load quickly into various graphing images that can be easily pulled down to your computer to annotate, or attached directly into emails.


The page above shows many servers, one after another, in an easy-to-view stream.  Each line of graphs represents one week, so that one can quickly learn “the story” of how a server has been performing over time.  Memory leaks and i/o waits are especially easy to spot.  But the interface is also flexible enough to handle other data, provided that it is saved in the correct format.