I ran "rm -rf ~" in production

Earlier today, I was explaining to my dad why I don’t pay for managed cloud database services. “Why pay for infrastructure and backups when I can handle that myself for free?” Famous last words.

So naturally, I decided to set up a backup script for a NoSQL database I’ve been running. Using rclone seemed perfect for the job, and I quickly threw together this script over my sluggish SSH connection:

#!/bin/bash

# Set up some variables
BACKUP_PATH="~/backups"
DATE=$(date +%Y-%m-%d-%H-%M)
BACKUP_NAME="backup-$DATE"
REMOTE_NAME="your-rclone-remote-name"
REMOTE_PATH="your-remote-path"

# Create backup directory if it doesn't exist
mkdir -p $BACKUP_PATH

# Dump the database to the backup directory
mongodump --out $BACKUP_PATH/$BACKUP_NAME

# Copy it over
rclone copy $BACKUP_PATH/$BACKUP_NAME $REMOTE_NAME:$REMOTE_PATH/$BACKUP_NAME

Here’s where things went sideways. I forgot a crucial detail about bash: ~/backups doesn’t expand to /home/username/backups when it’s in quotes. Instead, it creates a literal folder named ~ in the current directory.

When I ran the script to test it, sure enough:

$ ls
job.sh  ~

Without thinking, muscle memory kicked in. I saw that weird ~ folder and instinctively typed:

rm -rf ~

The moment I hit enter, my stomach dropped. Even my dad noticed the look of horror that crossed my face.

The damage was swift and thorough. My entire home directory — gone. Thankfully, the MongoDB data lives elsewhere on the system, so the database survived. But everything else? Toast.

The immediate crisis: all my SSH keys in ~/.ssh were wiped, and password authentication was disabled on this server. I scrambled to generate new keys using my still-active SSH session before getting locked out entirely.

Once I regained access, the cleanup began:

• Reconfigure all my dotfiles (.bashrc, etc.)
• Reinstall my web server (thankfully just a git pull)
• Reinstall Node.js since my setup lived in ~/.nvm
• Verify my systemd services still worked
• Rewrite that backup script properly

The Node.js situation was particularly annoying. The npm and node commands defaulted to Ubuntu’s ancient system versions, but only when running through systemd. I patched it with symlinks and PATH fixes—not elegant, but functional.

So was my dad right about using managed services? Maybe. But honestly, I’m glad I went the DIY route and lived through this little adventure.

The real lesson here isn’t about cloud vs. self-hosted—it’s about being more careful with destructive commands and maybe implementing proper system backups. Also, it’s probably time to move critical code out of the home directory and into /usr/local where it belongs.

At least I’ve got a good story now.