Process Management and Job Control

A process is any running instance of a programme on a Linux system. When you type a command like ls or cat, the system creates a process to execute that command. Each process is identified by a unique number called a Process ID (PID). The kernel, which is the core of the operating system, manages these processes. It decides which process gets to use the CPU next and for how long. This decision-making is called scheduling.

You can see all the processes running on your system using the ps command. Without any options, ps shows only processes associated with your current terminal. To see every single process on the system, you use ps aux. The a option shows processes from all users, u gives a user-oriented format (including who owns the process and CPU/memory usage), and x shows processes without a controlling terminal (background daemons). The output includes several columns:

Another popular tool is top, which gives a real-time, dynamic view of running processes. It updates every few seconds and sorts processes by CPU usage by default. You can press keys like P to sort by CPU, M to sort by memory, or k to kill a process by entering its PID. For a more modern alternative with better visuals, there is htop, but top is more commonly tested on the LPIC-1 exam.

Signals are the way the kernel communicates with processes. A signal is a software interrupt that tells a process to do something. The most common signal is SIGTERM (signal 15), which asks a process to terminate gracefully. If a process ignores SIGTERM, you can send SIGKILL (signal 9), which forces the process to stop immediately without any cleanup. The kill command sends a signal to a process by PID. For example, kill -9 12345 sends SIGKILL to process 12345. You can also use killall to send a signal to all processes with a specific name, like killall -9 firefox.

Job control allows you to manage multiple tasks from a single terminal. When you run a command and add an ampersand (&) at the end, the command runs in the background. For example, sleep 30 & runs the sleep command in the background. The shell gives you a job number and a PID. You can suspend a running foreground job by pressing Ctrl+Z. This stops the job and puts it in a suspended state. The jobs command lists all the background and suspended jobs. The bg command resumes a suspended job in the background, while fg brings it back to the foreground.

Foreground processes tie up your terminal – you cannot run another command until they finish. Background processes run independently, allowing you to continue using the terminal for other commands. This is useful for long-running tasks like backups or compiling code. Job control numbers start from 1, and you refer to them with a percent sign, like %1.

Zombie processes are a special case. A zombie process is one that has finished executing but still has an entry in the process table because its parent process hasn't read its exit status. Zombies are harmless but indicate a programming issue in the parent. They are identified in ps output with a state of Z. The only way to remove them is to kill the parent process so the init process (PID 1) can clean them up.

The nice and renice commands control the priority of a process. The Linux kernel gives more CPU time to processes with a lower nice value. The nice value ranges from -20 (highest priority) to 19 (lowest priority). By default, processes run with a nice value of 0. Using nice -n 10 ./myprogram starts myprogram with a lower priority. The renice command changes the priority of an already running process: renice +5 -p 12345 increases the nice value (lowering priority). Only the root user can set negative nice values (higher priority).

The pstree command shows processes in a tree-like structure, revealing which process is the parent of which. This helps visualise process dependencies.

Finally, the uptime and w commands show system load averages, which indicate how many processes are waiting to run on average over the last 1, 5, and 15 minutes. A load average above the number of CPU cores means the system is overloaded.

Understanding these concepts is crucial for the LPIC-1 exam because questions often ask you to interpret ps output, choose the correct signal to terminate a process, or explain the difference between foreground and background jobs.

Imagine you are the sole system administrator for a small online store that runs on a Linux server. The store software includes a web server (Apache), a database (MySQL), and a payment processing script. One day, you receive alerts that the website is responding slowly. Your first step is to log in via SSH and run top to see what is consuming CPU and memory.

You see that a process called php-cgi (a PHP interpreter) is using 99% of the CPU and over 2GB of memory. The store's checkout page calls this script, and it seems to be stuck in an infinite loop. You note its PID, say 5678. You decide to try a graceful shutdown first, because if you force-kill it, you might lose a customer's order in the middle of processing. You run kill 5678 (which sends SIGTERM). After a few seconds, the process is still there. You check top again – it remains at 99% CPU. That means the process is ignoring the gentle shutdown request.

Now you have no choice but to use SIGKILL. You type kill -9 5678. The process immediately disappears. The website speeds up. You then need to investigate why the script got stuck. You look in the Apache error logs and find a bug that causes an infinite loop under certain conditions.

Later, you need to run a full database backup that may take two hours. You don't want to sit and wait. So you start the backup script in the background: ./backup.sh &. The shell returns a job number like [1] and a PID. You can check its progress with jobs. If you need to, you can suspend it with Ctrl+Z and then resume it in the background with bg. This lets you continue fixing other issues while the backup runs.

Another common task is adjusting process priority. Suppose a data analysis script is running but it's making the web server lag. You can lower its priority so it doesn't steal CPU time from the web server. You find its PID with ps aux | grep analysis and run renice +10 -p 9876. Now the analysis script uses spare CPU only when the web server isn't busy.

Zombie processes occasionally appear. You notice a zombie process in ps aux with state Z. You find its parent PID (PPID) and use kill on the parent. If the parent is a service, you may need to restart it. This keeps the process table clean.

Real-world process management also involves monitoring. You might set up a cron job (a scheduled task) that runs ps and top at intervals and logs unusual activity. You could also use killall to stop all processes of a particular user if you suspect a security breach.

In short, an IT professional uses these commands daily to troubleshoot slowdowns, manage long-running tasks, and ensure critical services stay responsive. The LPIC-1 exam expects you to be comfortable with all of these tools.