Skip to content

linux-commands.md

Latest commit

 

History

History
242 lines (165 loc) · 5.21 KB

linux-commands.md

File metadata and controls

242 lines (165 loc) · 5.21 KB
Raw

Linux CPU Troubleshooting Commands

This document provides a comprehensive collection of commands for diagnosing and resolving CPU-related issues in Linux environments.

Table of Contents

Quick CPU Status

Basic CPU Information

# Show CPU model and basic information
lscpu

# Display number of processing units
nproc

# Show CPU information from /proc/cpuinfo
cat /proc/cpuinfo

CPU Load Average

# Quick overview of system load (1, 5, and 15-minute averages)
uptime

# Output:
# 14:23:45 up 23 days, 5:32, 3 users, load average: 1.05, 0.70, 0.48

The three numbers represent the system load average over 1, 5, and 15 minutes. Values below your CPU core count generally indicate the system is handling the load well.

Detailed CPU Information

CPU Architecture and Features

# Detailed CPU information including cache sizes and supported features
lscpu -e

# Show CPU topology
lstopo

CPU Frequency Information

# Check current CPU frequency
cat /proc/cpuinfo | grep MHz

# Check CPU scaling governors and available frequencies
cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_available_frequencies

CPU Load Monitoring

Real-time CPU Usage

# Monitor CPU usage in real-time (updates every 3 seconds)
top

# More user-friendly alternative to top
htop

# Focused CPU statistics with color highlighting
mpstat -P ALL 2

# CPU statistics with timestamp
sar -u 2 10

Graphical Output in Terminal

# Visual CPU load graph in terminal (may need installation)
s-tui

# Simple ASCII graph of CPU usage
dstat -c

Process CPU Usage

Finding CPU-Intensive Processes

# Sort processes by CPU usage
ps aux --sort=-%cpu | head -10

# Monitor processes and sort by CPU usage in real-time
top -o %CPU

# Check which processes are causing CPU load
pidstat 1

Process-Specific Investigation

# Analyze specific process CPU usage
pidstat -p PID 1

# Check what a process is doing
strace -p PID -c

# See which system calls are being made
strace -p PID

CPU Throttling

Check for Thermal Throttling

# Install and run thermald to check CPU temperature
sudo apt install thermald
systemctl status thermald

# Check CPU temperature
sensors | grep Core

# Check for throttling events
dmesg | grep -i throttling

Power-Related Throttling

# Check if CPU power management is limiting performance
sudo turbostat --Summary --quiet --show Busy%,Bzy_MHz,PkgWatt,PkgTmp,RAMWatt,GFXWatt

# Check Intel P-state driver status (Intel CPUs)
cat /sys/devices/system/cpu/intel_pstate/status

CPU Performance Tuning

CPU Governor Settings

# List available CPU governors
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors

# Set performance governor (temporary)
echo performance | sudo tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor

# Make performance governor setting permanent
sudo apt install cpufrequtils
sudo cpufreq-set -r -g performance

Process Priority

# Change process priority (niceness)
renice -n -5 -p PID

# Start a process with higher priority
nice -n -10 command

Common Issues and Solutions

High Load Average with Low CPU Usage

This typically indicates I/O bottlenecks or processes waiting for resources.

# Check for I/O wait
iostat -xz 1

# Check for processes in uninterruptible sleep (D state)
ps aux | awk '$8 ~ /D/ { print $0 }'

Single Core Overloaded

This often happens with single-threaded applications.

Solution:

# Set CPU affinity to distribute load
taskset -pc 0-3 PID

# Check current CPU affinity
taskset -pc PID

Zombie Processes Consuming CPU

# Find zombie processes
ps aux | grep Z

# Find parent of zombie process
ps -o ppid= <zombie_pid>

# Restart the parent process to clean up zombies
kill -HUP <parent_pid>

System Unresponsive Due to CPU Load

If the system becomes unresponsive due to CPU overload, you can use the Magic SysRq key to help recover:

# Enable SysRq if not already enabled
echo 1 | sudo tee /proc/sys/kernel/sysrq

# Kill the most intensive process
sudo alt + sysrq + f

# Emergency sync to disk
sudo alt + sysrq + s

# Emergency remount read-only
sudo alt + sysrq + u

# Emergency reboot
sudo alt + sysrq + b

Additional Resources

Note: Some commands may require installation of additional packages or root privileges. Always understand what a command does before executing it, especially in production environments.