Linux Process Management: ps, top, htop, kill, and Process Priorities

Process management is essential for system administration, troubleshooting, and performance optimization. Understanding how to view, prioritize, and control processes is fundamental to Linux expertise.

Viewing Processes

ps - Process Status

View snapshot of running processes.

# Simple process list
ps
# PID TTY TIME CMD
# 1234 pts/0 00:00:00 bash
 
# All processes (all users, all terminals)
ps aux
# USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
# root  1   0.0  0.1 19612 3556 ?   Ss  Jan18 0:05 /sbin/init
 
# List with full format
ps -ef
# UID  PID PPID C STIME TTY TIME CMD
# root  1   0    0 Jan18 ?   00:05 /sbin/init
 
# Find specific process
ps aux | grep nginx
# www-data 1234 0.0 0.2 12345 2048 ? S 10:00 0:00 nginx
 
# Process tree hierarchy
ps -ef --forest
# Shows parent-child relationships
 
# Show threads
ps -eLf
# Shows lightweight processes (threads)
 
# Process by user
ps -u john
# Only john's processes
 
# Processes using most CPU/Memory
ps aux --sort=-%cpu | head
# Sorted by CPU usage
 
ps aux --sort=-%mem | head
# Sorted by memory usage

ps Output Fields

# Common fields explained:
# USER: Process owner
# PID: Process ID (unique identifier)
# %CPU: CPU usage percentage
# %MEM: Memory usage percentage
# VSZ: Virtual memory size (KB)
# RSS: Resident set size (physical memory, KB)
# STAT: Process state (S=sleeping, R=running, Z=zombie, T=stopped)
# START: Process start time
# TIME: CPU time consumed
# COMMAND: Command executed
 
# Get custom output
ps -o pid,user,cpu,mem,comm
# PID USER %CPU %MEM COMMAND
# 1234 john 0.5 2.3 bash

top - Interactive Process Monitor

Real-time process monitoring and resource usage.

# Start top
top
# Press 'q' to quit
 
# Show processes for user
top -u john
 
# Batch mode (non-interactive)
top -b -n 1
# -b: batch mode
# -n: number of iterations
 
# Show specific process
top -p 1234
# Only PID 1234
 
# Delay updates (seconds)
top -d 2
# Update every 2 seconds
 
# Inside top, useful commands:
# q: quit
# f: change displayed fields
# F: filter by field
# k: kill process
# r: renice process
# u: filter by user
# M: sort by memory
# P: sort by CPU
# T: sort by time
# R: reverse sort

htop - Enhanced Process Monitor

Advanced interactive process monitor (install with: sudo apt-get install htop).

# Start htop
htop
 
# Show processes for user
htop -u john
 
# Show tree view
htop -t
# -t: tree view
 
# Search for process
# In htop: type /process_name
 
# Custom display
htop -p 1234,5678
# Show specific PIDs
 
# Columns:
# PID: Process ID
# USER: Process owner
# VIRT: Virtual memory
# RES: Physical memory
# CPU%: CPU usage
# MEM%: Memory percentage
# TIME+: CPU time consumed
# COMMAND: Command
 
# htop advantages:
# - Colored output
# - User-friendly interface
# - Tree view
# - Easy process killing
# - Memory monitoring

pstree - Process Tree

Display processes in hierarchical format.

# Show process tree
pstree
 
# Show with PID
pstree -p
 
# Show with CPU/memory info
pstree -p -u
 
# Show specific process tree
pstree -p 1234
# Shows descendants of PID 1234
 
# Show user's processes
pstree john
 
# Show numeric users (UID instead of username)
pstree -u
 
# Show process command line
pstree -a

atop - Advanced Top

Detailed system and process monitoring.

# Install atop
sudo apt-get install atop
 
# Start atop
atop
 
# Record system statistics
sudo atop -w /tmp/atop.log
 
# Replay recorded data
atop -r /tmp/atop.log
 
# Show specific interval
atop -r /tmp/atop.log -g

Process Signals and Control

kill - Send Signals

Terminate or signal processes.

# Send SIGTERM (graceful termination)
kill 1234
 
# Send SIGKILL (force kill)
kill -9 1234
kill -KILL 1234
 
# Send SIGSTOP (pause process)
kill -STOP 1234
 
# Send SIGCONT (resume process)
kill -CONT 1234
 
# Send SIGHUP (reload configuration)
kill -HUP 1234
 
# Kill all processes matching name
killall apache2
killall -9 firefox
 
# Kill process group
kill -15 -1234
# - (minus sign) = process group
 
# Common signals:
# SIGHUP (1): Hangup
# SIGINT (2): Interrupt (Ctrl+C)
# SIGTERM (15): Termination signal (default)
# SIGKILL (9): Kill (cannot be caught)
# SIGSTOP (19): Stop
# SIGCONT (18): Continue
 
# List available signals
kill -l
 
# Kill with custom signal
kill -USR1 1234
 
# Kill multiple processes
kill 1234 5678 9012
 
# Kill processes of user
sudo killall -u john

pkill - Kill by Pattern

Kill processes matching pattern.

# Kill by process name
pkill nginx
 
# Kill by process name (exact)
pkill -x bash
 
# Kill by user
pkill -u john
 
# Kill by terminal
pkill -t pts/0
 
# Kill with signal
pkill -STOP nginx
pkill -CONT nginx
 
# Show what would be killed
pgrep -l nginx
 
# Full command matching
pgrep -f "python script.py"

Foreground and Background Jobs

Running in Background

# Run process in background
long_running_command &
# [1] 1234
 
# View background jobs
jobs
# [1]+ Running    long_running_command &
 
# Bring to foreground
fg
# long_running_command
 
# Control-Z to stop current process
# (Press Ctrl+Z)
# [1]+ Stopped    long_running_command
 
# Resume in background
bg
# [1]+ long_running_command &
 
# Resume in foreground
fg
# long_running_command
 
# Multiple background jobs
command1 &
command2 &
command3 &
jobs
# [1] Running    command1 &
# [2] Running    command2 &
# [3] Running    command3 &
 
# Bring specific job to foreground
fg %2
# Brings job 2 to foreground
 
# Terminate background job
kill %1
# Kills job 1

nohup - Immune to Hangups

Run process immune to terminal disconnection.

# Run with nohup (immune to terminal closure)
nohup long_running_script.sh &
# Appends output to nohup.out
 
# Redirect output
nohup ./script.sh > output.log 2>&1 &
 
# Check if running
ps aux | grep script.sh
 
# Kill nohup process
kill 1234
 
# View output
tail -f nohup.out

disown - Detach from Shell

Detach job from shell session.

# Start job
sleep 1000 &
# [1] 1234
 
# Remove from job table
disown %1
 
# View jobs (1 is gone)
jobs
# (no output)
 
# Process continues running
ps aux | grep sleep
# Process still exists

Process Priority and Resource Allocation

nice - Set Process Priority

# Run with nice value (lower = higher priority)
nice -n 10 command
# -n: nice value (-20 to 19)
 
# Higher nice value (lower priority)
nice -n 15 backup_script.sh
 
# No nice value (default 0)
nice command
 
# Negative nice (requires sudo, increases priority)
sudo nice -n -10 important_service
 
# View current nice value
ps -o pid,user,nice,comm
# PID USER NI COMMAND
# 1234 john 0 bash

renice - Change Process Priority

# Change existing process priority
renice 10 -p 1234
# Increases nice value of PID 1234 to 10
 
# Lower priority for process (increase nice)
renice 15 -p 1234
 
# Increase priority (requires sudo)
sudo renice -10 -p 1234
 
# Change priority for user's processes
renice 10 -u john
# All john's processes get nice 10
 
# Change priority for group
renice 10 -g developers
 
# View process with custom priority
ps -o pid,user,nice,cmd | grep script
# 1234 john 10 script.sh

Process Information and Status

Process States

# View process states with ps
ps aux
 
# Common states:
# S: Interruptible sleep
# R: Running
# Z: Zombie (finished but parent hasn't read status)
# T: Stopped
# W: Paging
# D: Disk sleep (uninterruptible)
# <: High priority
# N: Low priority
# L: Has pages locked in memory
# s: Session leader
# l: Multi-threaded
# +: Foreground process group

/proc Filesystem

# View process info in /proc
ls /proc/1234/
# cmdline: Command line arguments
# environ: Environment variables
# fd: Open file descriptors
# maps: Memory maps
# stat: Status information
# status: Detailed status
 
# View command line
cat /proc/1234/cmdline | tr '\0' ' '
 
# View environment
cat /proc/1234/environ | tr '\0' '\n'
 
# View open files
ls -la /proc/1234/fd/
 
# View memory maps
cat /proc/1234/maps
 
# View detailed status
cat /proc/1234/status

Practical Process Management

Monitor System Load

# View load average (1, 5, 15 minute)
uptime
# 10:45:33 up 5 days, 3:12, 2 users, load average: 0.5, 0.6, 0.4
 
# View CPU cores (compare with load)
nproc
# 4 cores
 
# If load > number of cores, system is overloaded
 
# Check CPU usage
top -bn1 | grep "Cpu(s)"
# Cpu(s): 15.2%us, 2.1%sy, 0.0%ni, 82.7%id, 0.0%wa, 0.0%hi, 0.0%si
 
# Monitor specific processes
watch -n 1 'ps aux | grep apache2'
# Update every 1 second

Troubleshoot Zombie Processes

# Find zombie processes
ps aux | grep Z
# USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
# john 1234 0.0 0.0 0 0 ? Z 10:00 0:00 <defunct>
 
# Zombie process parent
ps -o ppid= -p 1234
 
# Kill parent process (cleans up zombie)
kill -9 12345  # Parent PID
 
# Check if zombie still exists
ps aux | grep Z

Manage Resource-Heavy Processes

# Find memory hogs
ps aux --sort=-%mem | head -5
 
# Find CPU hogs
ps aux --sort=-%cpu | head -5
 
# Limit process resources
ulimit -u 100  # Max processes
ulimit -f 100M # Max file size
 
# View limits
ulimit -a
 
# Set nice priority for resource-heavy process
renice 15 -p 1234
 
# Use cgroups for resource control (Ubuntu/Debian)
systemctl --user-unit=process-group.scope \
  --MemoryLimit=500M \
  systemctl start myservice

Best Practices

Regular Monitoring - Use top/htop regularly to catch issues
Graceful Termination - Use SIGTERM before SIGKILL
Background Jobs - Use & for long-running tasks
Priority Management - Adjust nice levels for critical services
Log Monitoring - Check logs for process crashes
Resource Limits - Set appropriate limits to prevent resource exhaustion
Process Trees - Understand parent-child relationships
Cleanup - Kill defunct/zombie processes
Automate - Use systemd for service management
Document - Record why priority was changed

Summary

Process management is critical for system health:

ps, top, htop provide process information
kill and signals control process behavior
Foreground/background jobs enable multitasking
nice and renice control process priority
Monitoring enables early issue detection
Effective management ensures stable, responsive systems

Master these tools for professional Linux administration.

Linux Process Management: ps, top, htop, kill, bg, fg, and Process Priorities