Performance Tuning & Kernel Parameters (Sysctl)

Stop letting your web servers sit on a mountain of wait state sockets

🧩 The Challenge

Dealing with a massive spike in incoming requests usually means your server ends up with thousands of sockets stuck in TIME_WAIT. You end up watching your netstat output freeze while the kernel holds onto those file descriptors for way longer than they need to exist.

💡 The Fix

Enabling the TCP reuse setting lets the kernel recycle those dead connections for new requests, which is a lifesaver when you are running a high-traffic API. It basically stops your server from exhausting its local port range during a traffic burst.

sysctl -w net.ipv4.tcp_tw_reuse=1

⚙️ Why It Works

Setting this flag tells the Linux network stack that it is safe to reuse a socket in the TIME_WAIT state for a new connection if the timestamp sequence is valid. It bypasses the default two-minute waiting period, keeping your ephemeral port pool wide open.

🚀 Pro-Tip: Double check your load balancer config first, because sometimes the problem is just your own proxy hogging the connections before they even reach the app.

Linux Tips & Tricks | © ngelinux.com | 10/30/2026