Terminology
One of the most important things to understand before using GRUB is how the program refers to devices, such as your hard drives, and partitions. This information is very important if you want to configure GRUB to boot multiple operating systems.
Device Names
The first hard drive of a system will be called (hd0) by GRUB. The first partition on that drive is called (hd0,0) by GRUB, and the fifth partition on the second hard drive is called (hd1,4). In general, the naming convention for filesystems when using GRUB breaks down in this way:
(<type-of-device><bios-device-number>,<partition-number>) |
The parentheses and comma are very important in the name. The <type-of-device> refers to whether a hard disk (hd) or floppy disk (fd) is being specified.
The <bios-device-number> is the number of the device according to your system's BIOS, starting with 0. The primary IDE hard drive is numbered 0, while the secondary IDE hard drive is numbered 1. The ordering is roughly equivalent to the way the Linux kernel arranges the devices by letters, where the a in hda relates to 0, the b in hdb relates to 1, and so on.
 | Note |
|---|---|
Remember that GRUB's numbering system for devices starts at 0, and not 1. This is one of the most common mistakes made by new GRUB users. |
The <partition-number> relates to the number of a specific partition on that device. Like the <bios-device-number>, the numbering of partitions starts at 0. While most partitions are specified by numbers, if your system uses BSD partitions, they are signified by letters, such as a or c.
GRUB uses the following rules when naming devices and partitions:
It does not matter if your hard drives are IDE or SCSI. All hard drives start with hd. Floppy disks start with fd.
To specify an entire device without respect to its partitions, simply leave off the comma and the partition number. This is important when telling GRUB to configure the MBR for a particular disk. For example, (hd0) specifies the first device and (hd3) specifies the fourth device.
It is very important, if you have multiple hard drives, to know their order according to the BIOS. This is rather simple to do if you have only IDE or SCSI drives, but if you have a mix of them, things get a bit more tricky.
File Names
When typing commands to GRUB involving a file, such as a menu list to use when allowing the booting of multiple operating systems, you must include the file immediately after the specifying the device and partition. A sample file specification to an absolute filename is organized like this:
(<type-of-device><bios-device-number>,<partition-number>)/path/to/file |
Most of the time, you will be specifying files by their path on that partition plus the file's name. This is rather straightforward.
You can also specify files to GRUB that do not actually appear in the filesystem, such as a chain loader that appears in the first few blocks of a partition. To specify these files, you have to provide a blocklist, which tells GRUB, block by block, where the file is located in the partition. As a file can be comprised of several different sets of blocks, there is a specific way to write blocklists. Each file's section location is described by an offset number of blocks and then a number of blocks from that offset point, and the sections are put together in order, separated by commas.
In other words, consider the following blocklist:
0+50,100+25,200+1 |
This blocklist tells GRUB to use a file that starts at the first block on the partition and uses blocks 0 through 49, 99 through 124, and 199.
Knowing how to write blocklists is useful when using GRUB to load operating systems that use chain loading, such as Microsoft Windows. You can leave off the offset number of blocks if you are starting at block 0. As an example, the chain loading file in the first partition of the first hard drive would have the following name:
(hd0,0)+1 |
You can also use the chainloader command with a similar blocklist designation at the GRUB command line after setting the correct device and partition as root:
chainloader +1 |
GRUB's Root Filesystem
Some users are confused by the use of the term "root filesystem" with GRUB. It is important to remember that GRUB's root filesystem has nothing to do with the Linux root filesystem.
The GRUB root filesystem is the root partition for a particular device. GRUB uses this information to mount the device and load files from it, among other things.
With Red Hat Linux, once GRUB has loaded its root partition that contains the Linux kernel, the kernel command can be executed with the location of the kernel file as an option. Once the Linux kernel boots, it sets its own root filesystem, and that is the one most people associate with Linux. The original GRUB root filesystem and its mounts are forgotten. They only existed to boot the kernel file.
See the notes given for the root and kernel commands in the section called Commands for more information.