Infrastructure & Operations › Linux & Servers

Managing Processes (ps, top, kill)

Seeing and controlling running programs.

Also known as: ps, top, kill, htop

A running program is a process with a PID. These tools show them and stop them.

ps aux                      # snapshot of all processes
ps aux | grep nginx         # find one by name
pgrep -f myapp              # just the PIDs matching a pattern

top                         # live view, sorted by CPU (q to quit)
htop                        # friendlier alternative, if installed

In top, watch the %CPU and %MEM columns to find what is hogging the machine. See also load average.

Stopping a process: use signals

kill doesn’t just kill; it sends a signal to a PID.

kill 4242         # SIGTERM (15): "please shut down", lets the program clean up
kill -9 4242      # SIGKILL: stopped by the kernel immediately, no cleanup
pkill -f myapp    # signal by name pattern

The classic mistake: reaching for kill -9

SIGKILL can’t be caught, so the program can’t flush buffers, close connections or remove lock files. That can leave corrupted files or half-finished work. Try plain kill first, give it a few seconds, and use -9 only when it ignores you.

Other points

• Check before you kill. ps and grep can match more than you expect; read the PID and command first.
• A process managed by systemd will be restarted if it dies. Stop it with systemctl stop name instead.
• You can only signal your own processes unless you use sudo.
• A process that exits reports an exit code.