A Real Automation Script

This lesson pulls everything from the course together into one real script: a log-cleanup tool that takes arguments, validates them, loops over files, and reports what it did.

The goal

A small but genuinely useful script: given a directory and a number of days, delete any .log file in that directory older than that many days, and print a summary of what got removed. It uses arguments, a conditional, a loop, a function, file tests, and proper error handling — everything from the last eleven lessons, in one place.

The full script

#!/bin/bash
set -euo pipefail

log() {
    echo "[$(date +%H:%M:%S)] $1"
}

# usage: ./cleanup-logs.sh <directory> <days>
if [[ $# -ne 2 ]]; then
    echo "Usage: $0 <directory> <days>"
    exit 1
fi

target_dir="$1"
days="$2"

if [[ ! -d "$target_dir" ]]; then
    echo "Error: $target_dir is not a directory"
    exit 1
fi

log "Scanning $target_dir for .log files older than $days days"
removed_count=0

for file in "$target_dir"/*.log; do
    [[ -e "$file" ]] || continue

    if [[ $(find "$file" -mtime +"$days") ]]; then
        log "Removing $file"
        rm "$file"
        removed_count=$((removed_count + 1))
    fi
done

log "Done. Removed $removed_count file(s)."

./cleanup-logs.sh /var/log/myapp 7

[14:22:01] Scanning /var/log/myapp for .log files older than 7 days
[14:22:01] Removing /var/log/myapp/access-old.log
[14:22:01] Removing /var/log/myapp/error-old.log
[14:22:01] Done. Removed 2 file(s).

Walking through the pieces

• set -euo pipefail (lesson 11) turns on strict error handling from the start, so a mistake fails loudly instead of limping on.
• The log() function (lesson 6) centralizes how every status line is formatted, with a timestamp, so the whole script's output stays consistent.
• $#, $1, and $2 (lesson 3) read the directory and day-count arguments, with a usage message if the wrong number was given.
• The -d file test (lesson 4) confirms the argument is really a directory before doing anything destructive with it.
• The for file in "$target_dir"/*.log loop (lesson 5) walks every matching log file, and [[ -e "$file" ]] || continue skips the case where the glob matched nothing at all.
• find "$file" -mtime +"$days" checks each file's age, and rm removes it once confirmed old enough.
• removed_count (lesson 2) tracks a running total, printed once at the end.