The first command we will examine is the ps command. This is probably the most basic process related command. It will provide you with a list of the current processes running on your system. With no options, ps will list processes that belong to the current user and have a controlling terminal. The ps command displays four items: Process ID (pid), Terminal (tty), Time (Total time used by the process), and the command (CMD). Output is unsorted by default. However, as with so many shell commands, this one has a number of options we can pass to it, in order to alter its behavior. Here are the more commonly used options:^[1]

You can see the ps option and the -e and -H options in Figure 4.1.

Figure 4.1. The ps command.

Notice that the -H option causes some of the processes to be shifted over to the right in the display, showing that they are part of a hierarchy. For example, you can see several processes under the init process. The ps command is a very useful administrative tool because it allows you to see everything running on a system or just for a specific user or terminal.

This command is very similar to the ps command, except it shows all the processes in the form of a tree structure (similar to how tree does it for directories). I personally find this view of running processes, shown in Figure 4.2, to be very useful, and I prefer it to the ps view.

Figure 4.2. The pstree command.

As with most commands, this one has several options you may need to use, including the following:^[2]

These options can be used to get additional information about the processes in the pstree.

The pgrep command will return the ID of a running process if you provide it with the name. This command is useful for finding the process id of a particular process when you know the name, or just part of the name of the process. This command can look for partial matches. The pgrep command looks through the currently running processes and lists the process IDs that match the selection criteria.

This particular command has many different options you can use. They are shown in this list:^[3]

In Figure 4.3, you can see pgrep first used with an entire process name (in this case, bash). Then you can see it used with just the first few letters. And, finally, it is used with the first few letters and with the -l option so that both the name and the process ID are listed.

Figure 4.3. The pgrep command.

The pgrep command will be very useful to you, particularly if you find that you cannot remember the exact name of a process you want to look up.

The top command is very much like the ps command, except it lists the processes in the order of how much CPU time the process is utilizing. This can be very useful for an administrator, particularly if you are finding that your CPU utilization is high and you want to identify the culprit. I personally also find the display to be far more informational than the ps display. I use the top command far more often than I do the ps command. You can see the top command in Figure 4.4.

Figure 4.4. The top command.

You should keep in mind that since this command lists processes by their CPU utilization, it is entirely possible to get different results if you run the command at different times. This command has a number of options that can alter the way it displays. In fact, it updates itself, so if you watch it you will probably see the items change.^[4]

You can see top with the -n 2 and -u root options shown in Figure 4.5.

Figure 4.5. The top command with flags.

This command is very simple. It will kill a running process by ID. It simply causes an application or a daemon to unload from memory. This command depends on the process ID, so you need to find the ID for the process. You can do this using a command like ps, as shown in Figure 4.6.

Figure 4.6. The kill command.

In this example, I used ps to find the ID of a process named cheese (cheese is a graphics program in GNOME), and then used the kill command to kill that process.

Contrary to what you might think, this is not used to kill all running processes. To do so would shut down Linux! This will kill a process by name rather than by ID.

So, as you see in Figure 4.7, I use the killall command to kill the cheese application using its name, rather than the ID.

Figure 4.7. The killall command.

pkill is used to kill processes according to some additional criteria (see Figure 4.8). Use the -u option to kill using a username and process name. You may want to kill the processes connected with a specific user, which is particularly useful in several situations. First, use it if you have a user that has launched several processes that are using up too many system resources. Second, use it if you suspect an intruder is on your server using a particular user account. If that is the case, you can kill those processes.

Figure 4.8. The pkill command.

Note that in this case, the username is test.

This command will display the currently running jobs you have executing. These are not the same as processes, although a job could involve a process. A job is some scheduled task. It is very likely that this task involves executing some process, but that does not mean the job and the process are the same thing. We will discuss how to schedule jobs later in this book. Obviously, if you have no scheduled jobs, this command will show nothing.

Note that the jobs command, the bg command, and the fg command are all related to jobs. We will discuss these again and in more depth later in this book.

This command sends a job to the background where it is running without interaction with the user.

This command sends a job to the foreground, so you can interact with it. The process will use your current terminal. If you omit the process you want to use with the fg command, it will simply bring the last process that was in the background to the foreground.

This command is very useful because it sets the priority of a given process. With priorities, the lower the number, the higher the priority will be. The minimum priority is 20 and the maximum priority is -20 and only a root user can assign negative priorities. However, any users can always lower the priorities of their own processes (see Figure 4.9).

Figure 4.9. The nice command.

Here we incremented the priority of the ls command by 10 and then ran that command.

The renice command will change the priority of an existing command (see Figure 4.10). You can use the option -u to change the priorities of all processes for a particular user name or the option -g to change priorities for all processes of a particular group. The default is to change via the process ID number.

Figure 4.10. The renice command.

Here we used ps to find the process IDs for running processes and then changed the priority level for the bash shell. There are only a few options for this command:^[5]

This process is somewhat similar to skill, except rather than kill the process, it simply changes its priority. This command is shown in Figure 4.11. This command has a few options you can use:^[6]

In Figure 4.11, you can see we are using the -v option.

This command will start a process; however, it is used primarily in scripts rather than from the shell, so we will explore this later when we discuss scripts.

For the purposes of this section, we are using GNOME 2.32 running on Fedora Linux. Using GNOME, if you select Applications → System Tools → System Monitor, as shown in Figure 4.12, you can accomplish many of these same tasks.

Figure 4.12. The GNOME System Monitor.

If you select the Processes tab, shown in Figure 4.13, you can view all processes and sort them by CPU usage if you want. This mimics the ps and top commands you might execute from the shell. Also, notice the End Process button; this will accomplish the same goal as the kill command. Unfortunately, there is no analog for the pstree command.

Figure 4.13. Viewing processes with GNOME.

One item that is very useful is the Resources tab shown in Figure 4.14. This gives you a clear view of what is happening on your system, CPU utilization, memory usage, and related data.

Figure 4.14. Resource allocation in GNOME.