My SSD hit zero free bytes on a Tuesday, and three unrelated things broke within about a minute of each other — a build, a database write, and my editor's autosave all failing at once. A full disk doesn't announce itself politely; every program that needs to write a single byte starts throwing errors, so the symptoms scatter and the cause hides. It took me an embarrassing while to run df, see / at 100%, and longer still to trace it back to one app I'd left running for a week that had been quietly appending to a log file the whole time.
Nothing was wrong with the app. It was doing what loggers do. The problem was that nothing was ever pruning what it wrote, so a directory I never looked at grew until it swallowed the last gigabyte on the drive. That's the trap with logs: they're individually tiny and collectively fatal, and they fail you silently right up until everything that touches the disk gives out at the same moment. The script below is what I now point at any directory that grows without bound, so old log files age out on a schedule instead of waiting to take the whole machine down — the part I wasn't going to sit through twice.
The cleanup script that saved my SSD
Paste this into a file called cleanlog.sh. Change LOG_DIR and DAYS to match your setup. Run with -print first before using -delete.
What this does, line by line
find "$LOG_DIR" searches your log folder. -type f limits results to files only (skips subdirectories). -name "*.log" matches only log files. -mtime +$DAYS filters to files last modified more than DAYS days ago. -delete removes them. Replace -delete with -print anytime to preview without deleting.
Setting it up without nuking the wrong files
Step 1 — Create the file
Paste the script, then Ctrl+X → Y → Enter to save.
Step 2 — Set your paths and age limit
| Variable | Example | What it means |
|---|---|---|
| LOG_DIR | /var/log/nginx | The folder containing your log files |
| LOG_DIR | /home/user/app/logs | A custom app log folder |
| DAYS | 7 | Delete logs older than 1 week |
| DAYS | 30 | Delete logs older than 1 month (default) |
| DAYS | 90 | Delete logs older than 3 months |
Step 3 — ALWAYS preview before deleting
Before you run the script for the first time, swap -delete for -print to see exactly which files would be removed:
Review the output. If it looks right, switch back to -delete and run for real.
Step 4 — Make it executable and run it
Schedule with Cron
Running this once manually is useful. Running it automatically every week is the goal.
Log output to a file
Add >> /var/log/cleanlog.log 2>&1 to the end of your cron line so you have a record of every cleanup run and any errors.
Versions for messier log directories
Clean multiple file types at once
Delete logs and report how much space was freed
Clean logs across multiple directories
How this deletes the wrong thing
Not previewing with -print first
The most common mistake. Always run with -print before -delete the first time on any new directory. One wrong path and you could delete the wrong files permanently.
Deleting system logs you shouldn't touch
Avoid pointing LOG_DIR at /var/log directly — that includes system logs that your OS needs. Always target a specific subfolder like /var/log/nginx or /var/log/myapp. Database transaction logs require a different approach — use the MySQL database backup script instead of deleting them with find.
-mtime counts in full 24-hour periods
-mtime +30 means strictly more than 30 full days ago, not "30 days ago today." A file from exactly 30 days ago won't match — it needs to be 31+ days old. Use -mtime +29 if you want to catch files from day 30.
Understanding the Commands
| Command / Flag | What it does |
|---|---|
| find $LOG_DIR | Searches recursively through the specified directory |
| -type f | Matches files only — ignores directories and symlinks |
| -name "*.log" | Matches files ending in .log — change to *.gz, *.tmp, etc. |
| -mtime +30 | Matches files modified more than 30 full days ago |
| Prints matching files to the terminal — safe preview mode | |
| -delete | Deletes matching files permanently — always preview first |
Frequently Asked Questions
How do I automatically delete old log files in Linux?
Use the find command: find /var/log/myapp -name '*.log' -mtime +30 -delete. This removes all .log files older than 30 days. Add it to cron with crontab -e to run it automatically on a schedule.
How do I safely preview what find -delete will remove?
Swap -delete for -print in your command. It prints the matching files to your terminal without touching them. Review the list, and when you're confident, switch back to -delete.
What does -mtime +30 mean in the find command?
-mtime +30 matches files whose last modification time was more than 30 full 24-hour periods ago. Use +7 for a week, +90 for three months. The + means strictly greater than.
How do I delete files older than 30 days in Linux?
Run: find /path/to/folder -type f -mtime +30 -delete. This finds all files modified more than 30 days ago and deletes them. Always test with -print before using -delete. For files that accumulate without an age-based pattern, the find duplicate files script identifies redundant copies by content hash.
Part of the Disk Management collection
