Installing Linux

Installing Linux involves several processes, beginning with creating Linux partitions, and then loading the Linux software, configuring your X Window System interface, installing the Linux boot loader (GRUB or LILO) that will boot your system, and creating new user accounts. The installation program used on Red Hat is a screen-based program that takes you through all these processes, step by step, as one continuous procedure. You can use either your mouse or the keyboard to make selections. When you finish with a screen, click the Next button at the bottom to move to the next screen. If you need to move back to the previous screen, click Back. You can also use TAB, the arrow keys, SPACEBAR, and ENTER to make selections. The installation screens will display a help panel explaining each step in detail. You have little to do other than make selections and choose options. Some screens provide a list of options from which you make a selection. In a few cases, you are asked for information you should already have if you followed the steps earlier in this chapter. Hardware components will be automatically detected and displayed as you progress. During installation, you will be able to perform administrative tasks such as configuring your network connections, creating users, and setting the time. Keep in mind that such administrative tasks can also be performed after installation. You are now ready to begin installation. The steps for each part of the procedure are delineated in the following sections. This should not take more than an hour.

Toward the end of the installation, you will be asked to create a boot disk. Be sure to make one. The boot disk will contain specific configuration information about your systems. Should anything go wrong with your boot loader, you will still be able to boot Linux using the boot disk.

Starting the Installation Program

You can start the installation using one of several methods. If your computer can boot from the CD-ROM, you can start the installation directly from the CD-ROM or the DVD-ROM included with this book. Just place the CD-ROM in the CD-ROM drive, or the book's DVD-ROM in the DVD drive, before you start your computer. After you turn on your computer, the installation program will start up.

Note

To boot from a CD-ROM or DVD-ROM, you may first have to change the boot sequence setting in your computer's BIOS so that the computer will try to boot first from the CD-ROM. This requires some technical ability and knowledge of how to set your motherboard's BIOS configuration.

If you have a DOS or Windows system installed on your hard drive, you can use the autoboot.bat command in the dosutils directory to start the installation, as shown here. You have to execute this command from a DOS system or a Windows DOS window.

e:\dosutils\autoboot.bat

If neither of these options is feasible for you, you can use the install floppy disk (see the preceding section on creating a boot disk). This is perhaps the most fail-safe method of installing Linux. Insert the Linux install disk into your floppy drive and reboot your computer.

The installation program will start, presenting you with an Introduction screen. After a moment, the following prompt will appear at the bottom of your screen:

boot:

Press ENTER. (If necessary, you can enter boot parameters as described in the installation manual.) Configuration information will fill your screen as the installation program attempts to detect your hardware components automatically.

Your system then detects your hardware, providing any configuration specifications that may be needed. For example, if you have an IDE CD-Write drive, it will be configured automatically as a SCSI drive so that CD writing software can make use of it (see Chapter 31). If you are installing from a floppy disk, it will detect your CD-ROM. If for some reason it cannot do so, your system will ask you to select your CD-ROM from a list. If you still have difficulty, you may have to specify the CD-ROM at the boot prompt.

Boot: linux hdx=cdrom

Replace the x with one of the following letters, depending on the interface the unit is connected to, and whether it is configured as master or a slave:

a First IDE controller master
b First IDE controller slave
c Second IDE controller master

d Second IDE controller slave

Note

As each screen appears in the installation, default entries will be already selected, usually by the autoprobing capability of the installation program. Selected entries will appear highlighted. If these entries are correct, you can simply click Next to accept them and go on to the next screen.

Partitions, RAID, and Logical Volumes

You will initially be asked to set up and designate the Linux partitions and hard disk configurations you want to use on your hard drives. Red Hat provides automatic partitioning options if you just want to use an entire single hard drive for your Linux system. To manually configure your hard disks, Red Hat provides a very detailed and graphic-oriented partitioning tool called Disk Druid. With Disk Druid, you can create specific partitions, or configure RAID devices, or set up logical volumes (see Chapter 31).

Once you have configured your hard disks, your partitions will be formatted. If you are formatting any old Linux partitions that still have data on them, a dialog box will appear listing them and asking you to confirm that you want to format them (new Linux partitions that you created will automatically be formatted). If you already have a Linux system, you will most likely have several Linux partitions already. Some of these may be used for just the system software, such as the boot and root partitions. These should be formatted. Others may have extensive user files, such as a /home partition that normally holds user home directories and all the files they have created. You should not format such partitions.

Recommended Partitions

If you are manually creating your partitions, you are required to set up at least two Linux partitions: a swap partition and a root partition. The root partition is where the Linux system and application files are installed. In addition, it is recommended that you also set up a boot partition that would contain just your Linux kernel, and a /home partition that would hold all user files. Separating system files on the root and boot partitions from the user files on the home partition allows you to replace the system files should they ever become corrupt without touching the user files. Similarly, if just your kernel becomes corrupt, you would have to replace only the kernel files on your boot partition, leaving the system files on the root partition untouched. This strategy of separating system directories into different partitions can be carried further to ensure a more robust system. For example, the /var directory, which now holds Web and FTP server files, can be assigned its own partition, physically separating the servers from the rest of your system. The /usr directory, which holds most user applications can be placed in its own partition and then be shared and mounted by other systems. One drawback to this strategy is that you would need to know ahead of time the maximum space you would want to use for each partition. For system and kernel files, this can be easily determined, but for directories whose disk usage can change dramatically, like /home, /var, and even /usr, this can be difficult to determine. As an alternative to creating separate physical partitions for each directory, you could use logical volumes (described later). A basic partition configuration is shown here:

Swap partition no mount point
/ Root partition for system files (and all other files if no other partitions are defined)
/boot Boot partition holding the Linux kernel (200MB approximately)
/home User home directories and files

Except for the swap partition, when setting up a Linux partition, you must specify a mountpoint. A mountpoint is a directory where the files on that partition are connected to the overall Linux file structure for your system. The mountpoint for your root partition is the root directory, represented by a single slash, /. The mountpoint for your boot partition is the path /boot. For a user's partition, it would be /home.

The size of the swap partition should be the same size as your RAM memory, with a recommended minimum size of 64MB. With 512MB of RAM, you could use a 512MB swap partition. If you have a large amount of RAM, you can make the swap partition the same size. If your disk space is limited, you should make your swap size at least 64MB.

To create the new partition, you will need to provide the mountpoint, the size (in megabytes), the file system type, and the hard disk on which you want to create the partition. For the size, you can select the Grow To Fill Disk option to have the partition automatically expand to the size of the remaining free space on the disk. For file system type, select ext3, the Linux native type for standard Linux partitions, and the Linux swap type for your swap partition.

RAID Disks

You also have the option of creating software RAID disks. First create partitions and select as their type Software RAID (see Chapter 31 for more details on RAID). Once you have created your partitions, you can create a RAID disk. Click the RAID button and then select the partitions you previously created that you want to make up the RAID disk, choosing also the type of RAID disk. See the Red Hat Customization Guide for detailed information on creating RAID disks.

Logical Volumes

Red Hat also supports Logical Volume Management (LVM), letting you create logical volumes to use instead of using hard disk partitions directly (see Chapter 30). Logical volumes are implemented by Logical Volume Management (LVM). They provide a more flexible and powerful way of dealing with disk storage, organizing physical partitions into logical volumes in which memory can be managed easily. Disk storage for a logical volume is treated as one pool of memory, though the volume may in fact contain several hard disk partitions spread across different hard disks. There is one restriction. The boot partition cannot be a logical volume. You still have to create a separate hard disk partition as your boot partition with the /boot mount point in which your kernel will be installed. See the Red Hat Customization Guide for detailed information on creating logical volumes.

Creating logical volumes involves several steps. First you create physical LVM partitions, then the volume groups you place these partitions in, and then from the volume groups you create the logical volumes for which you then specify mount points and file system types. To create your physical LVM partitions, click New and select "physical volume (LVM)" for the File System Type. Create an LVM physical partition for each partition you want on your hard disks. Once you have created LVM physical partitions, you click the LVM button to create your logical volumes. You first need to assign the LVM physical partitions to volume groups. Volume groups are essentially logical hard drives. You could assign LVM physical partitions from different hard disks to the same volume group, letting the volume group span different hard drives. Once the volume groups are created, you are ready to create your logical volumes. You can create several logical volumes within each group. The logical volumes function like partitions. You will have to specify a file system type and mount point for each logical volume you create.

Boot Loaders

Once your partitions are prepared, you install a boot loader. Red Hat uses the Grand Unified Bootloader (GRUB). You use a boot loader to start Linux from your hard drive. The screen will display the partition to boot by default, listing all partitions with different operating systems installed on them, like one with Windows along with the one for your Linux system. Clicking the Advanced options lets you refine your boot procedure, enabling you to choose where to install the boot loader or determine the drive boot order. You have two choices for where to install the boot loader: the Master Boot Record (MBR) or the root partition. The recommended place is the MBR.

Network Configuration

The Network Configuration screen displays tabbed panes in the top half for the different network devices on your computer. The screen displays segments for your network devices, hostname, and miscellaneous settings. If you use DHCP to automatically configure your network connection, as most networks do, you will most likely not need to do anything on this screen.

The Network devices segment will list your network connection devices, such as your Ethernet device. For computers already connected to a network with an Ethernet card, the first entry is usually labeled eth0. If you need to manually configure your device, entering an IP address for it, you can click Edit to display the device's configuration panel. Here you can choose to either manually configure the device or use DHCP to automatically configure it. For manual configuration, you can enter the device's IP address (usually your computer's IP address) and your network's netmask.

In the Hostname segment, you can choose either to manually provide your network information or use DHCP to automatically obtain it. If you choose to manually provide your network information, you can enter a hostname and, in the Miscellaneous segment, enter your network's gateway and DNS servers' IP addresses.

Following network configuration, the Firewall Configuration screen then lets you create basic default levels of network security. You can opt to use default firewall rules or customize your configurations, specifying trusted devices and services to allow such as Web or FTP connections. The firewall configuration still implements the newer IP tables firewall rules (see Chapter 19).

System Configuration

The next screen lets you choose the language you want to use.

On the Time Zone Configuration screen, you have the option of setting the time by using a map to specify your location or by using Universal Coordinated Time (UTC) entries.

On the next screen, you can set the root password for the root account on your system. This is the account used for system administration operations, such as installing software and managing users. After installation, a similar screen will let you also add ordinary user accounts.

Software Installation

You can then select the packages you want to install on your system. Red Hat lets you choose a set of predetermined collections tailored to personal desktops, workstations, or servers. You can select one of these package configurations or choose a custom installation. On the custom installation, you can select groups of packages, such as those for GNOME, KDE, different servers such as Web or FTP servers, games, editors, and development support, which includes compilers and debuggers. To install all packages, select the Everything entry at the end of the list.

The Package selection interface is the same as redhat-config-packages, discussed in Chapter 4. To select individual packages, click the Details link to the right of a package category. This will list individual packages, which you can select or deselect by clicking their check boxes. Many software packages require that other software packages also be installed. This is called a dependency. If you don't have these already selected for installation, they will be selected for you. There will normally be two segments, Standard and Extra, though for some packages there will only be one or the other. The Editors category has only an Extra segment listing Vi and Emacs editors, whereas the Windows file server category has only a Standard segment listing the Samba server. The Extra packages are those not considered essential to the category described. In the Web Server category, the standard segment contains the Apache Web server, and the Extra segment contains added modules and servers such as Squid and Tux.

The packages are then installed, showing each package as it is installed and the progress of the installation. (For the Red Hat DVD-ROM included with this book, the DVD-ROM will install in one continuous operation. The install program detects all packages as residing on the single DVD-ROM. There are no ejecting or inserting tasks as there are with CD-ROMs.) If you are using CD-ROMs instead (purchased from Red Hat or downloaded from the Internet), you will be prompted when to install the next CD-ROM. The current one will automatically be ejected for you. When the installation finishes, a postinstall process will complete. You are then usually given the option to create a boot disk. You can use this disk to access your Linux system should your hard disk boot somehow fail.

X Window System Configuration (Red Hat only)

For Red Hat, X configuration then completes the installation. (Fedora automatically detects your graphics card before installation.) This enables GUI interfaces, such as GNOME and KDE. X configuration uses the redhat-config-xfree86 utility to detect your video card and monitor automatically.

You first have to select and choose your video card on the Graphic Interface screen. redhat-config-xfree86 will initially probe your card and select it from a list of video cards shown on the screen, along with the amount of memory on the card. Check to see if your card is selected. If not, locate your card in the list and click it. You will need to know the brand and model number for your card. If your video card does not appear on the list, you can try specifying just the chipset. Find out the chipset used on your card and select it in the Generic Video Card entry. This entry will list several commonly supported chipsets. If your card or chipset is not present, you should skip the X configuration process. Just click the Skip X Configuration check box. You can later try to use X configuration utilities such as redhat-config-xfree86 or try to configure the /etc/X11/XF86Config configuration file directly. This requires technical expertise.

If you know how much video memory is on your video card, you can check for the appropriate entry for that amount located in the drop-down menu. If an incorrect amount is selected, you can select the appropriate one.

On the Monitor Configuration screen, you select your monitor. Initially, redhat-config- xfree86 will automatically probe your monitor and select a make and model from those listed on the screen. Check to see that the correct monitor is selected. If not, find its entry and click it. Monitor entries are organized by company in an expanding tree. For example, if you have a Hewlett-Packard monitor, click the triangle symbol next to the Hewlett-Packard entry to list all the monitors that Hewlett-Packard makes. Then click your model. When you select a monitor, the horizontal and vertical boxes at the bottom of the screen will display your monitor's horizontal sync range and vertical sync range (these values are generally available in the documentation that accompanies your monitor or from your monitor's vendor or manufacturer).

If your monitor does not appear on the list, select a Generic (usually Generic multisync) or Unprobed entry. redhat-config-xfree86 will then supply horizontal and vertical frequency values in the labeled boxes at the bottom of the screen. Check that the correct horizontal and vertical frequencies are entered. If you enter values that are too high, you could overclock your monitor and cause it to shut down. Do not select a monitor "similar" to your monitor unless you are certain the monitor you are selecting does not exceed the capabilities of your monitor.

On the Customize Graphic Configuration screen, you can specify your card's resolution and color depth by selecting the specification you want from their drop-down menus. Autoprobed entries will already be selected. You can change them to other entries if you wish: 1024 × 768 for smaller monitors or 1280 x 1024 for larger ones.

You can also choose to use a graphical login instead of the command line login. The graphical login check box will already be checked for you. If you choose this option, when you start, the GNOME Desktop Manager (GDM) will start and display a login screen where you can enter the username and password. A Sessions submenu in the Options menu enables you to choose whether to start KDE, GNOME, or a window manager. When you log in, your desktop automatically starts up. When you log out of your desktop, the GDM login window is redisplayed. Select Shutdown from the Options menu to shut down Linux.

If you do not choose this option, you will start up with the command line interface. Enter the username at the login prompt and the password at the password prompt to log in. Use startx to start GNOME or switchdesk to switch to KDE. The logout command logs out. You can enter the shutdown or halt command while logged in to shut down Linux.

When you finish, redhat-config-xfree86 will generate an X Window System configuration file called /etc/X11/XF86Config. This is the file the X Window System uses to start up.

If you are having difficulty, you can always click the Skip X Configuration check box to skip the X Window System configuration and perform it later, after you have installed your system.