Slackware Startup (init) process

Introduction

At an IBM Linux seminar in Amsterdam, a teacher put it this way: "Startup is the hardest part of Linux." However, once the kernel has been loaded and taken over by Init, you are all right. From that moment on, all the things that have happened can be read in scripts and documents (Man page). ”

Of course he was right. There is no secret in the process of starting a Linux machine. You can read everything in the ASCII file. However, it's easy to get lost in the process of tracking these scripts. Slackware differs from other distributions in that it insists on using the BSD-style boot process, rather than using the SYS v startup process as they do.

In the boot system of SYS V, all start/stop scripts are placed in the/ETC/RC.D directory. At each run level (RunLevel) there is a directory (that is,/etc/rc.1,/ETC/RC.2, and so on), all of which are links to scripts that are required to run at that run level. When you enter a run level, there is a large script to process the links to start (or stop) the service at the run level.

The Sys v startup process is used in Redhat, SuSE, and I will leave it to the person who uses the system to explain it. Without actually using it, I'm afraid I'm going to get dizzy with details soon.

Most Slackware users (slackers) believe that the Sys v startup process is complex and difficult to maintain. To be honest, redhat and SUSE users have the exact opposite idea. But follow me first, and then make your own judgments.

In this document, I will try to prove that IBM's teacher is correct. I'll follow the Slackware (8.1) startup process, with script and Man page as a guide. Will quote a lot, and less (my own) text.

PS: You may notice that English is not my native language. I try to use the spelling checker as often as I can, but it's not very useful in grammar. Regardless of the content or grammatical errors, please do not hesitate to enlighten. Besides, I really like the way that people think about what I write.

Run level (RunLevel)

We'll talk about the run level a lot, so it's a good time to explain the run level. The run level is the *nix way to determine the service level of a machine. At each run level, one can define which services to activate and which to not. In principle, there can be many running levels. In fact, there are only 5:

Single user run level with minimal service running. This run level is often used for system maintenance.

Multi-user run level, all the services provided are running. These services may be such as HTTP servers, e-mail servers, SQL Servers, and so on. is what the system administrator (you) need.

Multi-user run level, the same as before, but this time there is a graphical interface login manager.

One of the multiple user run levels is the normal running state of the machine. You can decide on another run level. For example, a single user run level with network support for remote maintenance.

In addition to the above run level, there are two special run levels:

The run level of the Stop system (halt).

The run level of the reboot (reboot).