Text Processing (Grep/Sed/Awk)

Delete a whole config block, not just a line

🧩 The Challenge

Ever had to rip out a whole block of old, crusty config from a server file? Like an obsolete Apache VirtualHost or a stale `Location` block from Nginx that’s just cruft now. Cutting and pasting lines always feels like you’re one wrong keystroke away from blowing up the whole service, and a simple search and replace just won’t cut it. You need the *entire block* gone, cleanly, context and all. Man, I’ve wasted hours cleaning up this kind of mess.

💡 The Fix

There’s a super precise `sed` trick that lets you target and delete entire sections of text between two patterns. It’s reliable, it’s surgical, and it saves you from carefully counting lines or creating a manual mess that’ll haunt you later. This is how you keep those configuration files tidy without introducing new errors.

# Example for Apache VirtualHost removal:
sed -i '/<VirtualHost \*:80>/,/<\/VirtualHost>/d' /etc/httpd/conf/httpd.conf

# Example for Nginx location block removal:
# sed -i '/^ location \/old-app {/,/^ }/d' /etc/nginx/sites-available/default

⚙️ Why It Works

This command taps into `sed`’s powerful range address feature. It tells `sed` to find the first line matching your start pattern, then delete every line from there until it hits the end pattern, including both boundary lines. That `-i` flag makes the change directly in the file, so maybe grab a quick backup before you run it, yeah?

🚀 Pro-Tip: Use a `grep -q` check first to confirm the block actually exists before `sed` runs.

Linux Tips & Tricks | © ngelinux.com | 9/25/2026