Chapter 14 covered the boot process of a Linux system. This section discusses how to identify which step of the boot process has caused a problem.
Typically, a failure at any given stage of the boot process looks like the following:
Failure to read and load the boot sector
A failure at this stage occurs after the BIOS screen disappears and before any part of the operating system is run. A failure here means that the boot sector was not found or did not load. Typically, you will see a blank screen with just a flashing prompt or a message similar to "No operating system found.". Possible problems include:
The boot sector is trashed and must be reinstalled.
The partition listed as the active partition has no boot sector.
The wrong partition has been marked as "active."
Failure to load the boot loader
What you see will depend on which boot loader you have.
If you're using LILO, the LILO prompt itself will lead you to the problem (see the next section). If you're using GRUB, an error may appear, followed by the initial GRUB screen.
Failure to load the kernel
This can have several causes, including a bad hard drive sector, inability to find a particular kernel module, or inability to find the root device. Typically, any of these errors will result in a "kernel panic." At this point, you need to review the on-screen data to determine where the problem lies.
Failure to run the init daemon and enter a specific run level
If you get past the kernel load, the next step is running the init daemon (which almost always succeeds). Problems here are usually configuration errors in /etc/inittab such as setting the default run level to 0 (which halts the system) or 6 (which reboots the system).
Failure to start the processes in said run level
Failures at this level are usually in the applications themselves. For example, the Apache httpd daemon may fail to start because of a misconfigured httpd.conf file.
If a network process, such as Sendmail, takes several minutes to start up, it is almost invariably caused by a problem with name resolution.
Failure to present a login prompt
Depending on how your system is configured, this problem basically means the login program (such as gd,getty,kdm, or xdm) can't run for some reason.
