Exercise 2

Modify scripts from Exercise 1 so they count following statistics for processes that are running in a system:

• number of attempts to open existing file;
• number of attempts to create a file;
• number of successful attempts.

At a period that is defined as command line arguments (specified in seconds) script should print:

• Current time and day in human-readable format.
• Table that contains gathered statistics per process along with that process name and PID. Numbers should be cleared during each iteration.

You can use module file_opener to demonstrate your scripts. This module uses working directory which is passed as root_dir parameter, fills it with some files that are created preliminary (their number is set by created_files parameter). While executing request, it uses file random variable (which range is cut to [1;max_files)) and either tries to create a file or open it depending on create parameter.

Run several experiments using TSLoad workload generator varying created_files parameter and compare the results:

# EXPDIR=/opt/tsload/var/tsload/file_opener
# for I in 1 2 3; do
    mkdir /tmp/fopen$I
    tsexperiment -e $EXPDIR run                             \
            -s workloads:open:params:root_dir=/tmp/fopen$I  \
            -s workloads:open:params:created_files=$((I * 160)) &
done

Try to explain differences you get from the nature of file_opener workload generator module.