What is a process in Linux and why do so many appear on a server?

Published on 2026-10-20 by frfrand

Procesos de Linux explicados: PID, CPU, memoria, ps, top, señales y diferencia entre un proceso y un programa.

A process is an instance of a program that is running or maintains a state managed by the operating system. In Linux, each process has its own information, such as an identifier, resources, and associated permissions.

That is why you can have an application installed only once and several instances running at the same time. The program file and the processes it executes are different concepts.

Why are there hundreds of processes even if I don't open anything?

Because the computer keeps working: it manages networks, logs events, runs services, and handles requests in the background.

On a server without a desktop, for example, there may be processes for SSH, databases, Nginx, monitoring, and system tools. You do not need to have a window open for a process to be active.

Some processes wait almost all the time without consuming too much CPU. Their existence does not automatically imply that the server is overloaded.

What is a PID?

PID stands for Process Identifier: a number that the system uses to identify a process during its existence.

In Linux you can consult processes with tools such as:

ps aux
top

ps shows information at a given moment; top allows observing activity changes. There are also tools like htop and special files under /proc.

The PID is not a permanent identity. When a process ends, the number can be reused at another time.

Who starts a process?

In Linux there are parent-child relationships between processes. One can start another and become its parent process.

For example, a service can launch workers that handle specific jobs. If the main process ends, the management of dependent processes will depend on how the program and the system are designed.

This behavior helps to understand why sometimes processes remain running after closing a terminal, or why restarting a service creates new instances with other PIDs.

What resources does it use?

A process can consume CPU time, memory, connections, and file descriptors. The kernel organizes access to those resources.

But seeing a high memory value in a tool does not necessarily mean that the process is wasting it. There is shared memory, caches, and other ways of accounting for resources.

To diagnose a problem, combine observation of CPU, RAM, disk, network, and actual application behavior.

What is a signal?

Signals are a communication and control mechanism between the system and processes.

For example, SIGTERM usually requests a termination that the program can handle, while SIGKILL causes a forced termination that does not allow the same cleanup process.

Using kill -9 as a first solution for any stuck process can leave operations halfway. Whenever possible, try an orderly stop first.

Is a process a service?

A service is usually a component intended to offer a function on an ongoing basis, often supervised by a manager such as systemd. It may consist of one or many processes.

A process, on the other hand, can also be a short command that you run once and then ends.

It is not advisable to confuse the .service file with the process that runs it: the unit describes how a service should be managed.

What happens when a process hangs?

It may be waiting for I/O, blocked by a resource, consuming CPU without progress, or may have suffered a logic error.

Restarting it may be a temporary solution, but it does not explain what caused the failure. To investigate, check logs, resource usage, connections, and recent changes.

If the service restarts continuously, it is advisable to resolve the cause before hiding the problem by increasing the number of attempts.

And on an ArduMaker VPS?

A deployment run on ArduMaker ends up becoming real processes on the server: backend, Nginx, workers, or databases. Understanding PIDs and consumption helps you distinguish whether the problem is in the application, in the configuration, or in the available resources.

In summary: a process is a concrete execution of a program under operating system control. Seeing many processes is normal; what matters is knowing which ones should be there and how they behave.

Sources: Proc on Linux · How child processes are managed.