Intelligent Disk Cleaner

A zero-dependency Python script for Linux servers that prevents disk exhaustion by monitoring usage and intelligently cleaning up log files.

Key Features

1. Zero External Dependencies: Runs on standard Python 3 libraries. No pip install needed.
2. Dual Thresholds:
   • Warning (80%): Alerts admins to high usage and identifies the largest open file (suspected culprit) without modifying any files.
   • Critical (95%): Triggers active cleanup to recover space immediately.
3. Intelligent Truncation: Safely shrinks large active log files in-place by removing the oldest 50% of data, preserving the process file handle.
4. Fallback Cleanup: If active logs cannot be shrunk, iteratively deletes the oldest rotated log files first.
5. Spam Prevention: Rate-limits email notifications to once every 8 hours per disk volume.

Requirements

• Operating System: Linux (relies on /proc filesystem for process analysis).
• Runtime: Python 3.6+.

Configuration

Open disk_cleaner.py and configure the settings at the top of the file:

# --- Configuration ---
THRESHOLD_PERCENT = 95.0          # Critical cleanup trigger
WARNING_THRESHOLD_PERCENT = 80.0  # Warning email trigger
RATE_LIMIT_SECONDS = 8 * 3600     # 8 hours cooldown

# Email Settings
SMTP_SERVER = "smtp.example.com"
SMTP_PORT = 25
EMAIL_FROM = "alerts@example.com"
EMAIL_TO = ["admins@example.com"]

Usage

Manual Execution

Run with root privileges to ensure access to all system processes (required for accurate open file detection):

sudo python3 disk_cleaner.py

Automation (Cron)

To monitor continuously, add a cron job (e.g., run hourly).

1. Open root crontab:

   sudo crontab -e
   

2. Add the schedule:

   # Run every hour at minute 0
   0 * * * * /usr/bin/python3 /opt/scripts/disk_cleaner.py >> /var/log/disk_cleaner.log 2>&1
   

Logic Flow

1. Check Usage: Scans all mounted partitions (ignoring pseudo-filesystems).
2. Evaluate State:
   • < 80%: Healthy. Exit.
   • 80% - 94% (Warning): Start scan. Find the largest file held open by any process. Send a "WARNING" email with the file path and size. No action taken.
   • > 95% (Critical): Start scan.
     • Strategy A: additional checks confirm files are logs. Truncate the largest active log file by 50%. Check if space is recovered.
     • Strategy B: If A fails/insufficient, find rotated logs (e.g., .log.1, .gz) and delete the oldest ones until usage drops.
   • Action Taken: Send "URGENT" email detailing the cleanup.
   • Action Failed: If space cannot be freed, send "CRITICAL" email with the largest suspect file.
3. Rate Limit: Before sending any email, check the state file (/tmp/disk_cleaner_state.json). If an email was sent for this volume in the last 8 hours, suppress the notification.