Master Linux Kernel Tuning: Boost Network, File Limits & Disk Scheduling for High‑Performance Servers

1. Optimize Kernel Parameters

Key kernel settings are adjusted via /etc/sysctl.conf. Add each parameter on a separate line and apply changes with sysctl -p.

sysctl -a   # view current settings
sysctl -p   # reload and check for errors

Network‑Related Parameters

net.core.somaxconn=65535

– maximum TCP listen queue length per port. net.core.netdev_max_backlog=65535 – upper bound of packets queued when arrival rate exceeds kernel processing speed. net.ipv4.tcp_max_syn_backlog=65535 – max SYN queue length; larger values help mitigate SYN‑flood attacks. net.ipv4.tcp_fin_timeout=10 – timeout for sockets in FIN‑WAIT‑2 state. net.ipv4.tcp_tw_reuse=1 – allow TIME‑WAIT sockets to be reused for new connections. net.ipv4.tcp_tw_recycle=1 – enable fast recycling of TIME‑WAIT sockets (disabled by default).

These values are suitable for hosts with 8 GB–16 GB RAM; adjust as needed.

Local Port Range

When the system runs out of local ports, errors like "Can’t assign requested address" appear because each socket needs a unique local port.

# View current range
cat /proc/sys/net/ipv4/ip_local_port_range
# Default example
32768 61000
# Change range
vim /etc/sysctl.conf
net.ipv4.ip_local_port_range = 1024 65000
sysctl -p

Note: the minimum must be ≥1024; raise it if your services use ports above 1024.

2. Increase Resource Limits

Configuration file: /etc/security/limits.conf

* soft nofile 65535
* hard nofile 65535

* – applies to all users.

soft – current effective limit.

hard – maximum enforceable limit.

nofile – maximum number of open files.

65535 – new limit value (requires reboot).

Case Study : Nginx experienced "Too many open files" errors during peak traffic. The ulimit -a output showed a low open‑file limit. Updating /etc/security/limits.conf and restarting Nginx resolved the issue.

root soft nofile 65535
root hard nofile 65535
* soft nofile 65535
* hard nofile 65535

In Nginx configuration add: worker_rlimit_nofile 65535; Verification shows the new limit applied:

3. Disk Scheduling Strategies

Parameter path: /sys/block/<em>devname</em>/queue/scheduler

noop – FIFO queue, favors writes, ideal for SSDs and embedded systems.

deadline – guarantees service within a deadline; default read deadline is shorter than write, best for database workloads.

anticipatory – similar to deadline but adds a short wait after reads to batch writes; good for write‑heavy environments but poor for databases.

cfq – Completely Fair Queuing algorithm.

Relevant kernel parameter directories: /proc/sys/abi/* – binary compatibility support. /proc/sys/fs/* – file system limits and quotas. /proc/sys/kernel/* – PID limits, shared memory, debug levels. /proc/sys/net/* – network optimizations for IPv4/IPv6. /proc/sys/vm/* – cache and buffer management.