Performance Tuning & Kernel Parameters (Sysctl)

Stop your network throughput from hitting a wall with buffer tuning

🧩 The Challenge

Dealing with 10GbE links that act like dial-up because the default kernel buffers are stuck in the mid-90s is infuriating. You watch your traffic crawl while the CPU sits there bored, and you know the hardware isn’t the problem.

💡 The Fix

Bump up your TCP window sizes to allow the kernel to keep more data in flight. It’s the easiest way to stop those long-fat networks from choking your throughput.

sysctl -w net.core.rmem_max=16777216
sysctl -w net.core.wmem_max=16777216
sysctl -w net.ipv4.tcp_rmem="4096 87380 16777216"
sysctl -w net.ipv4.tcp_wmem="4096 65536 16777216"

⚙️ Why It Works

Expanding these buffers gives the kernel enough headroom to handle higher latency paths without waiting for ACKs before sending the next packet. Your network stack finally stops idling and starts actually saturating the wire.

🚀 Pro-Tip: Always double check /etc/sysctl.conf so these settings survive the next time someone reboots the box.

Linux Tips & Tricks | © ngelinux.com | 10/26/2026