Processes and Signals

When something is broken, one of the first questions is: what process is actually running, and what state is it in?

Core commands

ps aux                     # list running processes with CPU and memory usage
ps -ef | grep <name>
pgrep -a <name>           # find matching PIDs and print their command lines
top
htop                      # friendlier interactive process view
kill <pid>
kill -TERM <pid>          # ask a process to exit cleanly
kill -KILL <pid>          # force-kill a process
kill -HUP <pid>           # tell some daemons to reload config

Signal cheat sheet

Signal	Meaning	Typical use
TERM	Ask process to exit cleanly	Normal stop
KILL	Force immediate stop	Last resort
HUP	Hangup or reload	Re-read config for some daemons
INT	Interrupt	Similar to Ctrl+C

Useful reads

ps -p <pid> -o pid,ppid,user,%cpu,%mem,stat,start,time,cmd  # detailed view for one PID
cat /proc/<pid>/status                                       # kernel-exposed process status
tr '\0' ' ' < /proc/<pid>/cmdline                            # print the full command line cleanly
readlink -f /proc/<pid>/cwd                                  # show the process working directory

STAT matters. A process in D state is stuck in uninterruptible sleep, often on disk or kernel I/O. Z means zombie.

Practical rule

If systemd owns the process, use systemctl restart <service-name> instead of killing the PID directly. Otherwise systemd may just bring it back and you learn nothing.

Common mistake

Do not jump straight to kill -9. Try a normal stop first so the process can flush state, close files, and log a useful error.