Practical commands for working with Linux processes

Here are some of the commands you’ll use most often:

• ps – Shows processes on the system; you saw an example of this command earlier in the chapter. Flags modify which process attributes are displayed as columns. This command is usually used with filters to control how much output you get, for example, (ps aux | head –n 10) to cut your output down to just the top 10 lines. A few more useful tricks:
  • ps –eLf shows thread information for processes
  • ps -ejH is useful for seeing the relationships between parent and child processes visually (children are indented under their parents)

  

  Figure 2.2: Examples of outputs of the ps command with flags

• pgrep – Find process IDs by name. Can use regular expressions.

Figure 2.3: Examples of outputs of the pgrep command with flags

• top – An interactive program that polls all processes (once a second, by default) and outputs a sorted list of resource usage (you can configure what it sorts by). Also displays total system resource usage. Press Q or use Ctrl + C to quit. You’ll see an example of this command’s output later in this chapter.
• iotop – Like top, but for disk IO. Extremely useful for finding IO-hungry processes. Not installed on all systems by default, but available via most package managers.

Figure 2.4: Example of output of the iotop command

• nethogs – Like top, but for network IO. Groups network usage by process, which is incredibly convenient. Available via most package managers.

Figure 2.5: Example of output of the nethogs command

• kill – Allows users to send signals to processes, usually to stop them or make them re-read their configuration files. We’ll explain signals and kill command usage later in this chapter.