Linux On A Disk

Quick Installation Notes:

[Install the Linux On A Disk 3½-inch hard drive as either the slave or master drive on the primary, secondary, 3rd or 4th IDE interface.]

[Setup CMOS BIOS to recognized the LOAD disk.]

Use LBA BIOS drive translation for all drives. DO NOT USE the "AUTO" setting.

8.4GB Linux On Disk

Western Digital WD84AA Geometry is 16383 Cylinders, 16 Heads, 63 Sectors

Its LBA translation is 1027 Cylinders, 255 Heads, 63 Sectors

Fujitsu MPE3084AE Geometry is 17301 Cylinders, 15 Heads, 63 Sectors

Its LBA translation is 1027 Cylinders, 255 Heads, 63 Sectors

18GB Linux On A Disk

Western Digital WD181AA Geometry is 35138 cylinders, 16 heads, 63 sectors

It's LBA translation 2204 cylinders, 255 heads, 63 sectors

27 GB Linux On A Disk

Western Digital WD273BA Geometry is 53040 cylinders, 16 heads, 63 sectors

It's LBA Translation 3328 cylinders, 255 heads, 63 sectors

Fujitsu MPE3273AH Geometry is 52953 cylinders, 16 heads, 63 sectors

It's LBA Translation 3322 cylinders, 255 heads, 63 sectors

[Ignore any settings for Landing Zone or Write Pre-Comp, or use 0 ]

[The Linux hard disk is pre-configured slave drive, if it is the only drive on a cable in your installation, the jumper on the drive has to be set to master drive position.]

SCSI Disks are not set in CMOS

[Boot with the LOAD installation floppy - Press <Enter> at the prompt]

[Remove the floppy, reboot the system, and type 'Linux' or 'Dos' at the boot prompt]
 
 

LINUX ON A DISK
Getting Started Guide

© 1996 Cosmos Engineering Co.
 
 

I. SYSTEM REQUIREMENTS

Linux On A Disk requires an Intel or compatible 32-bit CPU (486DX AMD 5x86, AMD K6, Athlon, Cyrix 6x86, Pentium, Pentium Pro, Pentium II or Pentium III). It will support multiple processors.

Eight megabyte RAM is the minimum required for proper operation; 16 megabyte RAM is needed to run the X-Windows graphical interface. 32 or more megabyte RAM is recommend for best performance.

An IDE hard drive interface, with an open port and space available inside the computer case to install the 3½" hard drive is required for the standard LOAD. A supported SCSI host adapter is required for the SCSI LOAD. Secondary IDE port adapters and recommended SCSI adapters are available from Cosmos Engineering Co.

A VGA or better video card is required for graphics. An accelerated SVGA card with at least 1 MB RAM is recommended.

Linux supports a wide variety of PC peripherals, including sound cards, CDROMs, video cards, tape backup systems, SCSI devices and hard drives. The Linux Documentation Project Hardware-HOWTO should be consulted. For a complete list of hardware supported by LOAD, see http://www.LinuxOnADisk.net/LOADhard.htm. The following is excerpted from Matt Welsh's excellent Installation-HOWTO:

8.4 GB LOAD

Western Digital WD84AA 3½ inch, EIDE Hard Drive

8.4GB, 9.5 ms seek, 5400 RPM, 256K buffer

UltraDMA66 , up to 66.6 MB/sec transfer rate - Dimensions: 3½" x 1"

Disk Geometry: 16383 cylinders, 16 heads, 63 sectors

LBA Translation 1027 cylinders, 255 heads, 63 sectors

Buffer size 2MB

Write Precomp = 0, Landing Zone = 0

Partition Table:

Partition  name	type	start cyl	end cyl	size	mount
/dev/hd?1	(83) linux native	1	383	3076416	/vol1
/dev/hd?2 (82)	(5) extended	384	1027	5172930
/dev/hd?5	(82) linux swap	384	400	136521
/dev/hd?6	(83) linux native	401	407	56196	/boot
/dev/hd?7	(83) linux native	408	722	2530206	/ (root)
/dev/hd?8	(83) linux native	723	1027	2449881	/home

'?' depends on installed drive number

8.4GB, 9.5 ms seek, 5400 RPM, 512K buffer

Ultra DMA Mode 4, up to 66.6 MB/sec transfer rate - Dimensions: 3½" x 1"

Disk Geometry: 17301 cylinders, 15 heads, 63 sectors

LBA Translation 1017 cylinders, 255 heads, 63 sectors

Buffer size 512 Kb

Writete Precomp = 0, Landing Zone = 0

Partition Table:

Western Digital WD181AA 3½ inch, EIDE Hard Drive

18GB, 9 ms seek, 5400 RPM, 2M buffer

Ultra DMA Mode 2, up to 66.6 MB/sec transfer rate - Dimensions: 3½" x 1"

Disk Geometry: 35138 cylinders, 16 heads, 63 sectors

LBA translation 2204 cylinders, 255 heads, 63 sectors

Drive size maybe reported as 8GB to bios

Write Precomp = 0, Landing Zone = 0

Partition Table:

27.3GB, 9.5 ms seek, 7200 RPM, 2MB buffer

Ultra DMA Mode 2, up to 66.6 MB/sec transfer rate - Dimensions: 3½" x 1"

Disk Geometry: 52953 cylinders, 16 heads, 63 sectors

LBA Translation 3322 cylinders, 255 heads, 63 sectors

Drive size maybe reported as 8GB to bios

Write Precomp = 0, Landing Zone = 0
 

Partition Table:

Partition  name	type	start cyl	end cyl	size	mount
/dev/hd?1	(83) linux native	1	600	4819468	/vol1
/dev/hd?2 (82)	(5) extended	601	3322	21864465
/dev/hd?5	(82) linux swap	601	616	128488
/dev/hd?6	(83) linux native	617	623	56196	/boot
/dev/hd?7	(83) linux native	624	2000	11060721	/ (root)
/dev/hd?8	(83) linux native	2001	3322	10618933	/home

'?' depends on installed drive number

IBM IBM Ultrastar 18ES 3½ inch, SCSI Hard Drive

9.1GB, 7.5 ms seek, 7200 RPM, 2MB buffer

Ultra2 SCSI, up to 80 MB/sec transfer rate - Dimensions: 3½" x 1"

Disk Geometry: 1111 cylinders, 255 heads, 63 sectors

Write Precomp = 0, Landing Zone = 0

IBM IBM Ultrastar 9ES 3½ inch, SCSI Hard Drive

18.2GB, 7.5 ms seek, 7200 RPM, 2MB buffer

Ultra2 SCSI, up to 80 MB/sec transfer rate - Dimensions: 3½" x 1"

Disk Geometry: 2222 cylinders, 255 heads, 63 sectors

Write Precomp = 0, Landing Zone = 0

Partition Table:

Partition  name	type	start cyl	end cyl	size	mount
/dev/hd?1	(83) linux native	1	600	4819468	/vol1
/dev/hd?2 (82)	(5) extended	601	2222	13028715
/dev/hd?5	(82) linux swap	601	616	128488
/dev/hd?6	(83) linux native	617	623	56196	/boot
/dev/hd?7	(83) linux native	624	1200	4634721	/ (root)
/dev/hd?8	(83) linux native	1201	2222	8209183	/home

'?' depends on installed drive number
 

PARTS LIST

1 - One of the above listed hard drives. Note: Although the Ultra DMA hard disk require proper support from the system to operate at it's full potential, it is completely backwards compatible with EIDE and IDE.

1 - 3½" - 5.25" Hard Drive Adapter

1 - Dual Drive IDE Cable or Dual Drive SCSI-3 Cable

1 - Y-Power Adapter

1 - Boot Disk 1 - LOAD CD

1 - Installation Manual

III. DOCUMENTATION

Linux is nothing if it is not well documented. We have included much of this information in Linux On A Disk. Unfortunately it's on the disk. This short piece is designed to help you get the disk and Linux up and running in your computer so that you can get at that information

On-line help is provided by on-line manuals for virtually every command and program. By typing Menu at the shell prompt, and choosing Help, you will launch the text-base HTML browser lynx with the LOAD help system. Included are the Red Hat Linux Getting Started Guide , the Red Hat Reference Guide and and the http collection of Linux HOW TO papers from the Linux Documentation Project as well as the System Administrators Guide, Network Administrator's Guide, and Programmer's Guide. Detailed hypertext help is also provided for the desktop applications and the X Windows system.

The linux manual pages are the most basic and the most often used form of on-line help. These exist for virtually every command or program on the system. They are accessed at the shell prompt with the command man command. Remember, Linux is case sensitive - MAN won't work. For example man cp will give you several pages on the file copy command, and don't forget man man to find out more about the man command. Most commands also have a help option if the command is followed by a --help. ie. cp --help.

When you are on the X Windows Desktop, this group of documents can be accessed at anytime through the HELP Button. This control brings up a web browser with a menu of on-line documents.

Much more documentation is to be found down the path of /usr/doc, about 242 MB in text, postscript and http files.

Also included is the source code for virtually every program, from the kernel to the GNU utilities and the X server itself. Source code is the original text files written by a human, from which a program is generated. It is made available for your inspection, modification and even improvement. It is provided, together with the tools to build each package from scratch if you so desire. This openness, down to the source code level, is one of the things that has made Linux so advanced and so well debugged. You don't need to concern yourself with source code if you don't want to, but it's nice to know it's there.

Perhaps most useful to a new user in setting up a system are the HOWTO's from the Linux Documentation Project. These are relatively short pieces written on a particular subject. Below is a list of HOW TO's most relevant to installation and setup:
· Linux CDROM HOWTO, by Jeff Tranter <tranter@software.mitel.com>. Information on CD-ROM drive compatibility for Linux.

· Linux Ethernet HOWTO, by Paul Gortmaker <Paul.Gortmaker@anu.edu.au>. Information on Ethernet hardware compatibility for Linux.

· Linux Hardware Compatibility HOWTO, by Tawei Wan <frac@pobox.com>. A list of hardware known to work with Linux.

· Linux Kernel HOWTO, by Brian Ward <ward@blah.tu-graz.ac.at>. Upgrading and compiling the Linux kernel.

·· Linux NET-2 HOWTO, by Terry Dawson <94004531@postoffice.csu.edu.au>. configure TCP/IP networking, SLIP, PLIP, and PPP under Linux.

· Linux SCSI HOWTO, by Drew Eckhardt <drew@PoohSticks.ORG>. Information on SCSI driver compatibility with Linux.

· Linux Serial HOWTO, by Greg Hankins <greg.hankins@cc.gatech.edu>. Information on use of serial devices and communications software.

· Linux Sound HOWTO, by Jeff Tranter <Jeff_Tranter@Mitel.COM>. Sound hardware and software for the Linux operating system.

· Linux XFree86 HOWTO, by Matt Welsh <mdw@sunsite.unc.edu>. How to obtain, install, and configure XFree86 3.1.1 (X11R6).

IDE Cable and Connector on the controller card.

IDE Cable and Connector on the motherboard.

IV. INSTALLATION

1.) HARD DRIVE INSTALLATION

Quick Notes:

[Install the Linux On A Disk 3½-inch hard drive as either the slave or master drive on the primary or secondary IDE interface in your computer]

[Setup CMOS BIOS to recognized the LOAD disk. Use LBA BIOS drive translation for all drives.]
 
 

8.4GB Linux On Disk

Western Digital WD84AA Geometry is 16383 Cylinders, 16 Heads, 63 Sectors

Its LBA translation is 1027 Cylinders, 255 Heads, 63 Sectors

Fujitsu MPE3084AE Geometry is 17301 Cylinders, 15 Heads, 63 Sectors

Its LBA translation is 1027 Cylinders, 255 Heads, 63 Sectors
 
 

18GB Linux On A Disk

Western Digital WD181AA Geometry is 35138 cylinders, 16 heads, 63 sectors

It's LBA translation 2204 cylinders, 255 heads, 63 sectors
 
 

27 GB Linux On A Disk

Western Digital WD273BA Geometry is 53040 cylinders, 16 heads, 63 sectors

It's LBA Translation 3328 cylinders, 255 heads, 63 sectors
 
 

Fujitsu MPE3273AH Geometry is 52953 cylinders, 16 heads, 63 sectors

It's LBA Translation 3322 cylinders, 255 heads, 63 sectors

[Ignore any settings for Landing Zone or Write Pre-Comp, or use 0 ]

[The Linux hard disk is pre-configured slave drive, if it is the only drive on a cable in your installation, the jumper on the drive has to be set to master drive position.]

[SCSI Disks are not set in CMOS]
 
 

Installing the Power Cable

Installing the IDE Cable

Protect Your Data: You should make a backup copy of the data on your existing hard disk before installing the Linux On A Disk. Consult your operating system or utility documentation for backup procedures.

Handle the Linux Hard Drive Carefully:Hard drives can be damaged by electrostatic discharge (ESD), rough handing, or shock and vibration. Handle the hard drive by the sides only, and avoid touching the circuit board.

Make a Boot Floppy: If you don't have a boot floppy with at least fdiskon it, make one. You may need this to reboot your system if there is a problem. Under DOS you would create such a floppy with the command FORMAT A:/S , then you will copy the disk partition program to the floppy with the command COPY C:/DOS/FDISK.* A:

Consult your manual on the equivalent commands in another operating system.

Tools: You will need a small Phillips and a medium flat-blade screwdriver. You should also have the computer system manual and the operating system manual for your existing system.

Mechanically installing the Linux On A Disk is exactly the same as adding another hard drive to your system. Accordingly the best instructions for doing so would be those specific to your computer. If you have such material available, say, a good user guide or technical manual for your system, that material should be consulted. The following are the steps that need to be followed with all systems:

1.) Determine the drive number (first, second, third, forth, etc.) the LOAD Disk will be installed as. You will probably not have to change any drive jumpers, but you should determine if you need to first so that it may be done when most convenient. The LOAD Disk comes pre-configured as a slave drive. The jumper on it only need be changed if it will be the master drive in the system. Below are the most common installation situations.

A. - One IDE Hard Drive in the system, no CD, or CD connected to it own interface and cable.

Installing the LOAD Disk as the slave drive on the primary IDE interface. Linux device :

/dev/hdb

This is the most common installation situation. On a newer system the CDROM is almost certainly IDE, and is connected to a second IDE interface. On older systems there may be only one IDE interface, and the CDROM, if any, is connected to the sound card or interface board. In this case you installed the LOAD Disk as it is, jumpered as the slave drive. If the drive data cable in your system does not have a connector for a 2nd IDE drive (cheap manufacturers!?), or if the cable isn't long enough, you may use the IDE data cable (It's the gray, flat thing. The red strip along one side indicates the pin 1 orientation.) supplied with the LOAD kit, otherwise you won't need it.

B. - Two Hard Drives or a Hard Drive and CDROM share the same IDE interface. Or you have an unused secondary IDE interface and you want to install the LOAD Disk as the master drive on its own interface to get better performance. (Note: Be sure your secondary IDE interface supports EIDE, some do not, even when the primary IDE does. If only one interface support EIDE, you will see the best performance if the LOAD Disk is connected to that interface.)

Installing the LOAD Disk as the master drive on the secondary IDE interface. Linux device :

/dev/hdc

On a machine with an unused secondary EIDE interface, this is the best solution. Jumper the LOAD Disk as a master. See drawing on inside cover or the label on the drive. Connect it to the open IDE interface with the data cable provided with the kit. If you don't jumper the drive as master, and it is the only drive on the cable, your system may not see the drive. Note that some drives, including all Western Digital drives have a special jumper setting called "single" if the drive is a master with no slave.

C. - Three Hard Drives or other IDE devices in the system. Now you will have four.

Installing the LOAD Disk as the slave drive on the secondary IDE interface. Linux device :

/dev/hdd

Basically the same as A.) above but you are using the secondary interface. You probably also know what you are doing, and you may be happy we included the Y-Power Cable Adapter in the kit. Linux is capable of supporting up to 8 IDE drives in a single system.

D. LOAD stands alone - You're building a web server, a router or a unix compatible box.

Installing the LOAD Disk as the master or only drive on the primary interface. Linux device:

/dev/hda

Jumper the LOAD Disk as the master and connect to the IDE interface with the cable. Note that if the LOAD disk is to be the only disk in the system, it is still possible to reformat the first partition, which is shipped empty, for DOS and Windows 98 and boot from it.

E. Linux on the 3rd of 4th IDE channel

LOAD will install and function completely in any one of eight (8) IDE positions, including as either master or slave on a tertiary IDE channel at 1E8H or the quaternary at 168H. The Linux device will be

/dev/hde /dev/hdf /dev/hdg or /dev/hdh

Your system CMOS BIOS must recognize the drive if lilo is to boot the Linux system, but it can be booted from a floppy even if there is no CMOS BIOS support.

In some cases, when you operate the LOAD Disk as a slave, you may also have to change a jumper on your existing master drive. Most modern IDE drives can operate as either the only drive in a single drive system, or the master drive in a dual drive system with no jumper change. With some older IDE drives, the original drive may need to be jumpered as a master drive for the system to work properly. You may need to consult technical documentation for your drive on how to do this, or you may call our tech support department for assistance. See also Appendix B. If you can't access either drive with the LOAD Disk installed and the cable isn't reversed, then this is most likely your problem.

2.) Turn the computer off, but leave it plugged in. This will keep the equipment grounded and give your computer some protection against static. Open the system box and locate an empty drive bay. Most computer systems today have four or more 3½" and 5¼" drive bays for installing hard drives, floppy drives, CDROM's, etc. You will use one of these to install the LOAD Disk.

The LOAD Disk may be installed in either a vertical or horizontal position. If mounted in the horizontal position, then it should be mounted with the top, that is the label side, up. 3 or 4 6/32 screws, no more than ¼" in length should be used to secure the LOAD Disk in place. These are provided in the plastic baggy with the 3½" to 5¼" drive adapter kit.

A 3½" to 5¼" drive adapter kit is provided so that the drive may be installed in a 5¼" bay. These two U-shaped channels are attached to either side of the LOAD hard drive with the provided screws. These brackets should be installed with the open side of the U facing the top side of the hard drive.

Running the ribbon cable to the original drive and the LOAD Disk is made easier if the drives are mounted next to each other. In any case, connecting a cable to the two drives is the overriding consideration in the choice of locations. For special situations 36" dual drive IDE cables are available from Cosmos Engineering Co. for $5.95. Part No. 45-IDE36

3.) DC Power must be connected to the LOAD Disk. The drive power cables comes from the power supply. The power supply is that big silver or gray (black in some cases) box that the AC power plugs into the outside of, and has a lot of (usually) black, red, and yellow wires delivering DC power to the various components of the PC. Looking at where these wires go into your existing hard drive or CDROM you will see the type of connector that is used. You should also see one or more spare drive DC power connectors in your system. Plug one of the spare power connectors into the power socket on the LOAD Disk. Don't worry, the power connectors are keyed so that you can't plug them in backwards. If there are no free drive power connectors on your system, use the Y-Power Cable Adapter to allow two drives to share one power cable. To use this Y-Power Cable Adapter unplug the power connector from one of your existing disk drives and plug it into the adapter. Then plug one of the female plugs into the drive from which the power was removed, and the other into the LOAD Disk.

4.) Connect a Data Cable between the LOAD Disk and the IDE adapter. If the LOAD Disk is a slave drive, that cable must also connect to the master. The Data Cable is a 40 wire flat ribbon cable that plugs into the 40 pin interface connector on the drive(s) and the IDE adapter or motherboard connector. If your computer was designed with expansion in mind, all this stuff is easy. Installing the second drive was simple, the power cable was right there, and the IDE cable has a connector for a second drive.

If you are not so lucky, you will need to use the supplied extra-long dual drive IDE cable. In that case replace your original single drive cable with this cable, using it in such a way that both hard drives and the IDE connector on the main board or adapter card are all connected.

When installing this flat Data Cable it is important to line up Pin 1 on the drives and adapter with Pin 1 on the cable. The Pin 1 position on most IDE hard drives, including the LOAD Disk, is on the side closest to the power connector. Pin 1 on the IDE adapter is usually denoted by a "1" or "^" silk-screened on the board. In any case all cables on the same board will usually have the same orientation. Pin 1 on the cable is usually denoted by a red stripe on one side of the cable. With IDE cables and connectors and such, you won't fry anything or lose data if you get it backwards. So don't worry about that, but it won't work till you get it right.

5.) Close the system case, turn on the system power and reboot the computer. It should boot the same as it always did. If it doesn't, go back and check your work. Something is wrong! You may find that you have better luck ( or in any case, less frustration! ) if you run this power-on test before you close the case. Just be careful, and be sure to turn the power back off before replacing the case.

6.) The system CMOS BIOS configuration must be told about the new drive. Most new systems come with the drive type set to AUTO. This will not always properly detect the Linux drive. You must set the drive type to USER DEFINED with the proper parameters. On most systems you can enter the CMOS Setup program by pressing the <Del> key at the appropriate time during the boot process. You should also look for a boot time message on your system telling you how to enter CMOS setup. On systems with an early Award BIOS try <Alt><Ctrl><Esc>, and <Atl><Crtl><S> for Phoenix BIOS. Install this drive as appropriate hard drive (1-8) as a user definable type using the parameters listed above. If the CMOS Setup program has a menu selection to "Automatically Detect Hard Drives", use it. If it finds all your hard drives, you know they are installed and connected properly. If 'Auto Detect' doesn't find your drive, check your work, something is wrong. Install the LOAD drive with LBA enabled. You can and should enable 32-bit block mode if your system BIOS supports it. If you are installing LOAD into an older (pre 1992) system that doesn't support large IDE drives, you can use Disk Manager software.

Notes On The SCSI LOAD

You install the SCSI LOAD disk is in your system as you would any SCSI hard drive. The power cable connection is exactly the same as for the IDE drive, so you may reference that discussion on page 16-3 for guidance here. The data cable connection is similar but different. The data connection is a 68 line SCSI-3/LVD interface that must be connected to a supported UltraWide SCSI or 2UltraSCSI controller. You may use the enclosed 68-pin SCSI-3 cable, an optional LVD cable or your existing SCSI-3/LVD cable to do this.

Before you install the drive, check that the drive SCSI ID# doesn’t conflict with anything else on your SCSI bus. The LOAD-SCSI is shipped with SCSI ID# 6 set. The label on the drive is generally your best source of information about the jumper setting for the drive.

You can disregard all that has been said about the IDE drives with regards to drive types and geometry in CMOS BIOS. It doesn’t apply to SCSI Drives. SCSI hard drives are generally unknown to the system bios directly. The SCSI drive is reported to the system through the SCSI controller. It is the SCSI controller that must understand the SCSI drive geometry.

IF YOU HAVE ONLY SCSI HARD DRIVES IN YOU SYSTEM, THE HARD DRIVE TYPES WILL BE SET TO ‘NONE’ IN SYSTEM BIOS.

The SCSI controller BIOS should be set to its defaults. The system BIOS may have a setting for the order of boot devices. Set the boot order so that SCSI comes before C: or IDE, if your LOAD-SCSI came with the Windows 98 or Windows 2000 option, or if you don’t want to have lilo written to the MBR of your first IDE drive.

Now that you have the LOAD SCSI installed and being reported to the system by the SCSI adapter, installation proceeds as with the IDE LOADs. On to 2) Booting Linux!

Note on System Commander:

lilo -v -C LOAD/lilo.conf.hd?

s will install a lilo as a secondary boot loader on the LOAD Disk. Then reboot and choose DOS, go to the \SC sub directory on your C:\ drive and run SCIN, to reconfigure System Commander. When you reboot, System Commander will find the LOAD Disk and add it to it menu.
If you are installing System Commander after you have installed the LOAD Disk, you should run lilo as above, and then reboot to DOS and install System Commander.

To install LOADLIN on your DOS C:\ , type Install_C at the Linux shell # prompt. It will create a subdirectory C:\LINUX on your C:\ drive and will install a copy of your Linux kernel vmlinuz, LOADLIN.EXE, README and LINUX.BAT to C:\LINUX. It will also copy MANUAL.TXT, a manual on LOADLIN and other relevant documentation to the subdirectory. The README file has further information on using LOADLIN as well as the latest information on Linux on a Disk

EZ-DRIVE is a disk manager program supplied with some large IDE drives to make them work with older (non-LBA) DOS systems. It is incompatible with the normal installation of Linux On A Disk because both use your C: drives master boot record (MBR). Here is the solution:
1.) Before using the LOAD installation disk, you must make a recovery disk with your EZ-DRIVE installation diskette. You will use this latter to reinstall EZ-DRIVE.
2.) Install Linux On A Disk with the LOAD Installation disk as described on page 18.
3.) Reboot and chose Linux at the LILO prompt.
4.) Run Install_C to install LOADLIN and the kernel in the C:\linux subdirectory on your C: drive. See box on page 15.
5.) Reboot to the EZ-DRIVE recovery disk and re-install EZ-DRIVE.
6.) Use LINUX.BAT at the C:\ prompt to load Linux.

Using Ontrack's Disk Manager with big LOAD's on old PC's

1.) Set CMOS (BIOS) drive type to Type 2 for the LOAD Disk.
2.) Put the Disk Manager Disk in drive A: and reboot the computer.
3.) When Disk Manager ask you to "replace and press any key when ready", put the LOAD Install Disk into drive A: and press a key. You may now resume normal installation by reading the warning message on the screen and pressing <Enter>. The installation program should find the LOAD Disk and complete normally. Next...
4.) Put the Disk Manager Disk back in drive A: and reboot the computer. When it again complains about a non-system disk, remove it, leave the drive empty, and press any key. The LOAD Banner and the LILO: prompt should appear. Press <Tab> and then type linux at the prompt. Login into linux as root. Good your system works with Disk Manager. But we have a problem. The default LOAD installation uses the MBR and so does DM.
5.) After you have logged into the LOAD system for the first time, you will need to edit the file /etc/lilo.conf so that the first line is changed from boot=/dev/hda to boot=/dev/hda1 if the LOAD is the first IDE drive in the system. You may use vi, TkDesk notepad in X Windows, or your favorite text editor. After the file is saved, you must run lilo -v. This will move lilo from the MBR to the 1st partiition of the 1st hard disk. You must also use fdisk to make that partition (/dev/hda1 under linux) active. If the LOAD is not the first IDE disk in the system, you will need to point boot= to an extended partition on the first IDE drive and then use fdisk to make that partition active. If you don't have an extended partition on the first drive, and you need to use Disk Manager, you may have to boot Linux from a floppy or get another boot manager, such as BOOTACTV or System Commander. You may also want to refer to the Lilo User's guide (/usr/doc/lilo*) for more information.
6.) After you have pointed boot= to someting other than /dev/hda, and ran lilo, reboot the computer with a DOS system floppy in drive A:. Always reboot from Linux with <Alt><Ctrl><Del> or the shutdown command.
7.) Remove the boot floppy and insert the Disk Manager disk in the floppy drive.
8.) At the A:> start Disk Manager in the BIOS Standard Format mode with the command DM /N
9.) From the top level menu chose the (M)aintenance Menu. From there, choice (U)pdate Dynamic Drive Overlay and answer 'yes' to the question.
10.) When you return to the Maintenance menu, chose (W)rite Boot Code in MBR, after that operation you may exit Disk Manager <Esc>, <Esc>,<Return>. You are now ready to run.

After you press enter, the installation floppy will probe for the LOAD Disk and make changes to its root device pointer (rdev) and the file system table (fstab). It will then configure lilo to boot either the system on the first partition of the first hard drive, (usually DOS or Windows) or the LOAD Disk. It will then modify the boot record of the first hard drive to point to the lilo dual boot (actually multiple boot) loader. If everything went as expected, you will see the message "Linux On A Disk has been successfully installed". It will then halt the system and prompt you to remove the Install floppy and reboot the system. You can safely ignore all the garbage messages that might appear after the halt.

If you see the message "No Linux On A Disk found!", the Install program could not find the LOAD disk. Check your CMOS settings, cable connections, etc. Use the BIOS Auto Detect or other displayed features to make sure your system sees the drive, and then try the LOAD Installation disk again. Note also that when the kernel on the LOAD floppy is loading it should read out all the hard disks in the system, including the LOAD hard disk.

If lilo did not execute properly you will see the message "Linux On A Disk failed to install - lilo failed". You should be using a LOAD 6.2 or greater Install and Repair disk. You may get a correct version from our website at LinuxOnADisk.net, or call us at (323) 930-2540 and we will mail you a correct disk for your system.

When your system reboots, lilo will take charge of the process. You will see the Linux On A Disk banner, a message and a prompt. The prompt will hold for 15 seconds and then the system will continue to boot as it always did. If you enter LINUX at the prompt, it will boot the Linux On A Disk. If you enter DOS at the prompt, it will boot the system on drive C:. If you press the <Tab> key, you will be shown your boot choices on the screen.

lilo is a more advanced way to boot Linux, as well as other operating systems. It installs a new boot loader on the first hard drive that takes control of the computer before any operating system is loaded. It then prompts you for a choice of available operating systems, and loads the default OS if no choice is entered before the prompt times out. If you looked at the manual pages for lilo you already know that it can be configured many ways with the lilo.conf file. The standard configuration for Linux On A Disk is to load Linux if Linux is typed at the boot prompt and <Enter> is pressed, to load the operating system on drive C:\ if DOS is entered or <Enter> is pressed at the lilo: prompt. If nothing is done, it will time-out after 15 seconds and load the operating system on Drive C:. This may be DOS, Windows 95, or another OS or dual boot loader that loads from C:\. If the <Tab> key is pressed while the LILO prompt is displayed at the top of the screen it will wait for a keyboard entry. If <Tab> is pressed again it will display a list of available boot images. More on lilo can be found in with man lilo and man lilo.conf and the LILO User's Guide and the LILO Technical Overview both by Werner Almesberger <werner.almesberger@lrc.di.epfl.ch>. These two pieces, by the way, are among the most detailed and clear writing to be found on the PC boot process anywhere. They are to be found as files User_Guide.ps and Technical_Guide.ps under /usr/doc/lilo-0.2X/doc/.

This default configuration can be installed simply by typing lilo -v <Return> at the shell while logged in as root Other sample configuration files are available as /etc/LOAD/lilo.conf.hd*, or a new one may be created from the setup menu. It will guide you in the creation and installation of a custom lilo for your system.

The PC system BIOS normally looks to the boot sector of the first hard drive of the system for an Operating System to boot. This is how your system is currently booting DOS. In order to run Linux on the same system we use software to redirect the system bootstrap to the Linux partition. There are many ways to configure lilo, you may want to consult the manual pages and other documentation on lilo if you have special requirements.

We also provide an alternative way of booting the LOAD Disk. LOADLIN is a DOS executable file that unloads DOS and then loads Linux. This may be a better way to boot Linux in some situations. For example, if your system has a sound card or other device that must be initialized by a DOS program before it will work, or if for some reason you can't over write the boot sector of your first hard drive, LOADLIN may be your best bet. LOADLIN can also be use to run Linux on a drive not accessible by your system bios, like LOAD attached to the IDE interface of a Sound Blaster 16, for example.

To install LOADLINon your DOS C:\ , type Install_C at the Linux shell # prompt. It will create a sub directory C:\LINUX on your C:\ drive and will install a copy of your Linux kernel vmlinuz, LOADLIN.EXE, README and LINUX.BATto C:\LINUX. It will also copy MANUAL.TXT, a manual on LOADLIN and other relevant documentation to the sub directory. The README file has further information on using LOADLIN as well as the latest information on Linux On A Disk.

When Linux On A Disk is started for the first time you will be prompted to perform some configuration tasks.

First you will be asked to configure networking. This needs to be done only if there is an ethernet card on the system and it is attached to a network. If you are using a static IP address you will need at minimum the IP address and possibly also the netmask, the gateway and nameserver addresses. If you are using a dynamic IP address, you need to only press the <space bar> to select it and then <Tab> to OK and press <Enter>. If it doesn't take immediately <Tab> to BACK and press <Enter>.

Next you will be asked to set the time zone. You will also be ask if your hardware clock is set to Greenwich Mean Time (GMT), otherwise it is assumed that it is set to local time.

Then you must set the password security. NIS is a system for providing network wide passwords. If you don't know what it is, you probably don't need it. For best security, you should accept the defaults and use shadow passwords and MD5 passwords.

Finally, you will be asked what services should be started at bootup. Again the LOAD defaults is probably a good place to start. As you highlight each service, pressing <F1> will tell you more about each service. This is a good way to to learn what each service is.

After these tasks are completed Linux will detect the new hardware in you system and ask you to configure it. Among other things, it will detect your video card and run Xconfigurator to configure X Windows.

At the login prompt you can login as user with password cosmos.

For your reference, the administrative super user has the name root and password cosmos

As initially installed, the super user has awesome powers both for good and evil, and it is safest not to wield it (by logged in as "root"), unless you are doing things that require that power. You will need to be logged in as root to perform most configuration tasks for example.

To shutdown your Linux system and reboot use the familiar three finger salute <Atl>-<Ctrl>-<Del> or the halt command. To learn how to use the shutdown command type man shutdown at the command prompt. You should never kill the system by just turning off the power or pushing the reset button. Like most advanced operating systems, Linux caches data to RAM and writes to disk when it has time. If it loses power suddenly it may lose data. At worst you may lose important system data, and at best Linux will make you wait on the next boot up while it does an extensive file system check.

V. CONFIGURING LINUX

                    Quick Notes:

1.) Configuring the Mouse, Modem and CDROM

These devices are usually detected and configured by Red Hat Linux on boot-up. They can also be configured through the main menu or the GUI tools of the Control Panel under X Windows. Please note that the configuration of the modem may require the manual editing of configuration files in some cases. If the modem is on a serial port with a nonstandard interrupt (not irq 3 or 4) then the /etc/rc.d/rc.serial file will have to be edited for the interrupt. See the example in that file as well as the Serial-HOWTO for details.

In most cases, both the mouse and cdrom will be correctly detected and configured automatically on startup.

The setup menu and Control Panel mouse settings only associate the mouse with a specific linux device. It does not tell Linux the mouse protocol. The protocols recognized by Linux are Busmouse, Logitech, Microsoft, MMSeries, Mouseman, MouseSystems,PS/2 and others. If the mouse is not MS mouse compatible then the /etc/X11/XF86Config and a MOUSETYPE variable may need to be exported for your /etc/profile file. Busmouse should be used for the Logitech bus mouse, and while the older Logitech mice use Logitech, the newer ones use Microsoft or MouseSystems protocol. LOAD is shipped ready for a PS/2 mouse.

If the CDROM uses a SCSI interface, then the driver for the SCSI adapter will have to either be loaded as a module or built into the kernel before the CDROM will work under Linux. To setup SCSI devices, network devices and some other things, you will have to modify the /etc/modules.conf file to load the needed modules at boot time.

2.) The Control Panel

The control panel is a launching pad for a number of different system administration tools. They make your life easier by letting you configure things without remembering configuration file formats and awkward command line options. The control panel is started from the System menu in Gnome or KDE. Below are some notes on the major control panel functions, extracted from The Redhat User's Guide.

File System Configuration

The file system configuration tool shown in Figure 8.3 allows you to easily examine and manipulate file system mount points, types, options, etc. If you do know what you are doing, it should be fairly straightforward as to how to mount, unmount, and add devices. The Reload entry in the FSM menu causes the file system configuration tool to re-load /etc/fstab from your hard drive. If you edit /etc/fstab by hand while the file system configuration tool is running (which you shouldn't do), you probably want to reload.

Info

Displays information on the file system, including the device, partition type, file system type, mount point and options, comment, size, percent used, etc.

Check

Performs a file system check (fsck) on the partition. You can do this only on unmounted partitions -- if the partition is mounted, you will get an error and you will have to unmount it first.

Mount

Mounts the selected file system.

Unmount

Unmounts the selected file system.

Format

Creates a new file system on the selected partition. This will erase all data on the selected partition! You can do this only on unmounted partitions -- if the partition is mounted, you will get an error and you will have to unmount it first.

Edit

Brings up a dialog box where you can edit the mount point, mount options, comment, etc.

Adding NFS Mounts

To mount a file system via NFS, select Add Mount from the NFS menu. A dialog box will appear and you will have to fill in the following values:

Device

Enter the host name and path, separated by a colon. For example, foo.bar.com:/usr/exported indicates the /usr/exported directory on foo.bar.com.

Mount Point

Enter the directory on your machine where you want to mount the NFS file system. For example, /mnt/foo.

Options

Enter the mount options for this file system. The default is soft,intr,rw. The rw means the file system is read-write, and soft,intr are options that make your system a little more resilient when the remote server goes down. See the mount man page for a complete list of available options.

Comment

This optional field can be used to store a small comment.

After filling everything out properly, click on OK. At this point the entry is made in your /etc/fstab, but the file system is not actually mounted. To mount it, select it in the main window and click on Mount.

Figure 8.3: File System Configuration Panel

Printer Configuration

The printer configuration tool maintains /etc/printcap, print spool directories, and print filters. The filters use Ghostscript (gs) and nenscript to allow you to print plain text as well as postscript files. While configuring your printers, keep in mind that lpr prints to the printer named lp by default. You will probably want lp to be one of your printer queue names. Needless to say, ghostscript and nenscript should be installed.

To add a new printer, click on Add. This will bring up a dialog box where you indicate if you are adding a remote printer or a local one. If your printer is connected to your machine's parallel or serial port, select local. If it is somewhere on the network, select remote.

Figure 8.4: Network Configuration Panel

Names

Enter the names you want to give to this printer (queue). You can have multiple names, separated by a "|" character, e.g., lp|lp0|PostScript.

Spool Directory

Enter a unique directory to hold documents queued for printing on this printer. For example, /var/spool/lpd/lp0.

File Limit

If you want to limit the size of documents printed, enter a size in KB here. A 0 indicates no limit.

Printer Device

This is the physical device to which your printer is connected. For example, /dev/lp1.

Input Filter

If you have a custom filter, enter the file name (full path) here. Otherwise, you should click on Select. This will bring up a dialog box where you select your Printer Type, Resolution, Paper Size, and you indicate if your printer requires an EOF character to be sent to it at the end of each job (many printers used under DOS require this).

Suppress Headers

If you select this, no header pages will be printed for each job.

For a remote printer, you will fill in the following values:

Names

See local printer.

Spool Directory

See local printer.

File Limit

See local printer.

Remote Host

Enter the host name of the machine that has the printer. For example, printer.foo.com.

Remote Queue

Enter the name of the queue on the remote machine for the remote printer. For example, Postscript. Click on Ok and your new printer is added. Some versions of the lpd printer daemon may need to be restarted before they will recognize the new printer. To do this select Restart lpd from the lpd menu.

Network Configuration

The network configuration panel shown in Figure 8.4 is designed to allow easy manipulation of parameters such as IP address, gateway address, and network address, as well as name servers and /etc/hosts. Network devices can be added, removed, configured, activated, deactivated and aliased. Ethernet, arcnet, token ring, pocket (ATP), PPP, SLIP, PLIP and loopback devices are supported. PPP/SLIP/PLIP support works well on most hardware, but some hardware setups may exhibit unpredictable behavior. When using the Network Configuration Tool click "Save" to write your changes to disk, to quit without making any changes select "Quit".

Managing Names

The "Names" panel of the Network Configuration tool serves two primary purposes: setting the hostname and domain of the computer, and determining which name server will be used to look up other hosts on the network. The Network tool is not capable of configuring a machine as a nameserver. To edit a field or add information to a field simply click on the field with the left mouse button and type the new information.

Managing Hosts

In the Hosts management screen you have the ability to add, edit, or remove hosts from the /etc/hosts file. Adding or editing an entry involves identical actions. An edit dialog box will appear, simply type the new information and click Done when you are finished.

Adding a Networking Interface

You can configure a networking interface with a few clicks of a mouse. Begin by clicking on Interfaces in the main panel. This will bring up a window of configured devices with a row of available options. To add a device, first click the Add button then select the type of interface you want to configure from the box that appears.

PPP Interface

Adding a PPP interface can be as simple as supplying the phone number, login name and password in the "Create PPP Interface" box. In many cases some degree of customization will be needed to establish a PPP connection. Selecting the Customize button will allow you to make changes to the hardware, communication, and networking settings for the PPP interface.

KPPP, the KDE PPP frontend is probably the easiest way to setup your dial-up Internet connection. It is much preferred over the PPP tools in the Control Panel. KPPP can be accessed from the start menu of either KDE or Gnome.

SLIP Interface

In order to configure a SLIP interface you must first supply a phone number, login name, and password. This will supply the initial parameters for the chat script needed to establish a SLIP connection. When you select the Done button an "Edit SLIP Interface" dialog will appear that enables you to further customize the Hardware, Communication and Networking parameters for SLIP.

PLIP Interface

To add a PLIP interface to your system you only have to supply the IP address, the remote IP address, and the Netmask. You can also select if you want to activate the interface at boot time.

Ethernet, Arcnet, Token Ring and Pocket Adaptor Interfaces

If you are adding an ethernet, arcnet, token ring or pocket adapter to your computer you will need to supply the following information:

Device

This is determined by netconfig based on the devices already configured.

IP Address

Enter an IP address for your network device.

Netmask

Enter the network mask for your network device.

Network

This is determined based on the values supplied for IP address and Netmask.

Broadcast

This is determined based on the values supplied for IP address and Netmask.

Activate at Boot

If you want the device to be configured automatically when your machine boots select this by clicking on the box.

Configure with BOOTP

If you have a BOOTP server on your network and would like to use it to configure the device select this by clicking on the box.

After providing the configuration information for your new device, click Done. The device should appear in your "Interfaces" list as an inactive device. (The active column should have a label of "no"). To activate the new device highlight it with a mouse click and then click on the Activate button. If it does not come up properly, you may need to reconfigure it by clicking on Edit.

Managing Routes

In the Routes management screen you have the ability to add, edit, or remove static networking routes. Adding or editing an entry involves identical actions, just like the Hosts panel. An edit dialog box will appear, simply type the new information and click Done when you are finished.

Time and Date

The time machine allows you to change the time and date by clicking on the appropriate part of the time and date display and clicking on the arrows to change the value. The system clock is not changed until you click on the Set System Clock button. Click on Reset Time to set the time machine time back to that of the system. CAUTION: Changing the time can seriously confuse programs that depend on the normal progression of time, and could cause problems. Quit as many applications and processes as possible before changing the time or date.

Kernel Daemon Configuration

Red Hat Linux includes "kerneld", the Kernel Daemon, which automatically loads some software and hardware support into memory as it is needed, and unloads it when it is no longer being used. The tool shown in Figure 8.10 manages the configuration file for kerneld. While kerneld can load some things, such as file systems, without explicit configuration,it needs to be told what hardware support to load when it is presented with a generic hardware request.

For instance, when the kernel wants to load support for ethernet, kerneld needs to know which ethernet card you have, and if your ethernet card requires special configuration.

Figure 8.10: Kernel Module Management

Changing Module Options

To change the options being given to a module when it is loaded, click on the line to select it, then click the Edit button. kernelcfg will bring up a window. The options which kernelcfg knows about then (normally all available options) each have their own window. Normally, you will want to ignore the "Other arguments:" window. Some modules normally take no arguments; just in case, they have an "Arguments:" window which allows you to enter configuration information.

Changing Modules

To change which module gets invoked to provide a generic service, such as an ethernet card or SCSI host adapter module, you need to delete the old one and add a new one. To delete a module, select it by clicking on it, then click on "Delete". Then click on "Add" to add the new module, as explained in the following section. If you have changed your SCSI controller (scsi_hostadapter), remember to make a new initial ramdisk with the /sbin/mkinitrd command.

Adding Modules

To add a module of any type, click on the "Add" button. You will be presented with a dialog box asking you to choose a module type. Ethernet is eth, Token Ring is tr, SCSI controllers are scsi_hostadapter, and so on. Click "OK" to continue to the next dialog box. If there is more than one module which can be used for the module type you have chosen, a dialog box which asks which module you want to use will be presented, and may also ask for specifics about the type of module; for ethernet, for example, you need to choose from eth0, eth1, etc. When you are done, click "OK" again to continue to specify any module options in the next dialog box, which is the same dialog box with which you are presented when you edit a module.

Restarting Kerneld

The changes that you make with the Kernel Daemon Configuration tool will be made in the file /etc/conf.modules, which kerneld reads whenever it is started. Once you have made changes, you can re-start kerneld by clicking on the "Restart kerneld" button. This will not cause any modules which are currently in use to be re-loaded, it will only notify kerneld to use the configuration when it loads more modules in the future.

3.) Using SCSI Devices

Before you can access your SCSI devices from Linux, you must have the proper driver for your SCSI host adapter loaded. There are two basic ways to do this. The first is by loading the proper modules in after booting, and the second is by building the proper drivers into the kernel.

To install a SCSI host module from the shell prompt, use the modprobe command, with the name of the module and any options, as in these examples:

modprobe aic7xxx for Adaptec 2940

modprobe buslogic for Buslogic

modprobe aha1542 for Adaptec 1542

Making an initrd image to load SCSI devices

An initrd image is required to load a SCSI module at boot time. The shell script /sbin/mkinitrd will build the proper initrd image provided the file /etc/conf.modules has a line for your SCSI adapter, such as this;

alias scsi_hostadapter aic7xxx

To build the new initrd image run the /sbin/mkinitrd command with parameters;

/sbin/mkinitrd /boot/initrd.img 2.2.12 ; where 2.2.12 is the version of the kernel modules to use. You also will have to add an initrd line to your /etc/lilo.conf under the linux image section as in;

image=/boot/vmlinuz

label=linux

initrd=/boot/initrd.img

4.) Customizing the kernel

The kernel is the heart of the OS, it is always loaded into memory and directly interfaces with the hardware on your system. Accordingly, the closer it fits your hardware, the more things will be done with less waste. Specifically, it may be necessary to change your kernel from the generic default one, which will run on a 486 processor, to one that is optimized for a Pentium, Pentium III, AMD K6 or Athlon processor. You may also want to build support for your SCSI interface, CDROM, sound card, tape backup, network card, mouse interface or other features into the kernel instead of as modules. It is almost always beneficial to customize your kernel. There are two ways that you can do this:

A. Install a Red Hat processor specfic kernel:

Red Hat Linux 6.X for Intel comes with a number kernel packages specifically compiled for the more advance X86 processor and for multiple processor systems. We have made it easy to upgrade to one of these kernels from the LOAD menu. You will also need the Linux ON A Disk CD. While logged in as root, enter menu at the shell (#) prompt, choose SETUP LOAD, then INSTALL KERNEL. Install the appropiate kernel for your system. If you have a Pentium, Pentium MMX, Cyrix 6x86, AMD K6, or AMD K6-2, chose the 586 kernel. If you have a Pentium II or a Pentium Pro, chose a 686 kernel. If you have a dual processor system, or if you have more than two processors, you will want to chose one of the SMP kernels.

B.Rolling your own kernel:

The very best kernel for your system is produced by compiling and linking one that is specific to your system. This is not as hard as it sounds, and Linux On A Disk is already set up for it. From /usr/src/linux type make menuconfig. Then you have to answer a lot of questions about your hardware, so be ready. When you are returned to the prompt, run make dep, then make clean, and finally make to produce the kernel. make bzImage compresses the kernel. You will then have to copy linux/arch/i386/boot/bzImage to /boot/vmlinuz

make modules builds the modules for your kernel and finally make modules_install installs the modules. You can also run this same sequence of steps from the menu->SETUP LOAD->INSTALL KERNEL->CUSTOM.

If you make a mistake while changing kernels, don't panic. Just boot from your Linux Boot Disk or the LOAD Install and Repair Disk login as root and cp /boot/vmlinuz.LOAD /boot/vmlinuz and rerun lilo to restore the original configuration. The original kernel configuration is stored as /usr/src/linux/LOAD.Basic. You may reload this into make menuconfig or make xconfig with the Load saved configuration menu option.

5.) Customizing X Windows

Linux On A Disk is shipped with a minimal X Windows setup that only supports 16 colors on standard VGA, and a Microsoft compatible mouse on COM1 (/dev/ttyS0).. You will want to change this to get the best performance from your video card and monitor. We provided two tools for doing this Xconfigurator and XF86Setup

Xconfigurator, which can be accessed through menu -> Redhat setup -> Xconfigurator, is a more advanced and automated approach to setting up X Windows, it will autoprobe the video card on a PCI bus, and will suggest the best resolutions.

XF86Setup, which can be accessed through menu-> LOAD Setup -> X Configuration, has a graphical interface, and more importantly, has been known to succeed when Xconfigurator has failed.

You can get more help on either of these programs by consulting the man pages. For exampletype man XF86Setup at the shell prompt. Also consult the XFree86 HOWTO for details.

6.) System Adminstration with Webmin

Webmin is an application we have included in Linux On A Disk that greatly simplifies the task of administering your Linux system. Webmin is a Unix configuration and system administration tool that runs on its own internet server in your system. Like other system services, its start on boot up is set by ntsysv. It can be accessed from any web browser that has network access to your system. It resides at port 10000 on the IP address of your system or of localhost (27.0.0.1) from a browser running on the system

Normally you will access it from Netscape Navigator in X-Windows. In the location bar, enter http://localhost:10000. You can also access it from the command line using the lynx browsers. At the root prompt # Enter "lynx localhost:10000". One of the major advantages of this web enabled administration is that you can configure your Linux system from any computer any where, even from IE on a Windows box in Alaska.

No matter how you access it. Webmin will ask you for a login name and password, using the web browser’s authentication procedure. As shipped, you can use the login name "root" and the password "cosmos". Once you have logged in you will be able to change passwords, create new users, setup samba, ftp, and virtual web servers and almost anything else you can think of.

It is possible to run Webmin as a secure server, and we suggest you do so if you plan to access it over the Internet. The Open SSL and Net SSLeay libraries needed for this are on the LOAD CD under \LOADtars. While logged in as root, mount the LOAD CD, and then unpack the files with the commands:

tar -xzvf /mnt/cdrom/LOADtars/openssl-0.94.tar.gz

tar -xzvf /mnt/cdrom/LOADtars/Net_SSLeay.pm-1.05.tar.gz

Next cd to the openssl dir and follow the instructions in the INSTALL file. Then cd to the Net_SSLeay dir and follow the instructions in the README file to install it. Finally you will have to delete the file /etc/webmin/config so that setup.sh will run as a new install. To do this:

rm /etc/webmin/config

cd /usr/libexec/webmin/

And run:

./setup.sh

If openssl and Net_SSLeay are installed properly the script will ask:

Use SSL(y/n):

You answer yes.

7.) Using the Install/Repair Floppy

The Install and Repair Disk starts with the LILO boot: prompt. This allows you to interrupt its normal boot procedure by pressing the [Tab] key while the prompt is displayed, and then typing in 'linux' followed by a legal kernel option such as 'root=/dev/hdb7' . This will boot a LOAD disk installed as the slave drive on the primary IDE channel. If no input is given, the floppy continues to boot from the kernel on the floppy. It then creates a ramdisk and loads a compressed file system into the ramdisk and mounts it. init and login are bypassed and the shell and a script are run directly by the kernel. This provides you with a very small but useful linux system that can operate independent of any hard drive.

When the Install and Repair Diskis initially used to setup the LOAD certain important information about the particular installation is recorded on the floppy. After this floppy has been use to install the LOAD, a new menu will appear when you boot with this disk.

This menu and the other files on the disk allow it to be used for a number of emergency repair and restoration functions. What follows is the brief description of each menu item:

1. Reinstall Linux On A Disk - This selection will run the installation program again and is used if the LOAD disk is moved to a different computer or to a different drive position after it is initially installed.

2. Run File System check for root partition - This selection will run fsck.ext2 on the LOAD root partition. This will generally repair all but the most serious file system damage. You should boot the Install and Repair Disk and run this if LOAD fails to boot and load with complaints about the root partition (/dev/hdx7 where x = a.....h)

3. Run File System check for another partition - This choice will run fsck.ext2 on any Linux partition. It will prompt you for the partition and expects it in the form of /dev/hdxn where x=a....h and n=1....8. You should boot the Install and Repair Disk and run this if LOAD fails to boot and complains of error on another ext2 partition.

4. Edit Configuration Files - This choice reveals a sub-menu and makes it easy to edit certain key LOAD configuration files with the vi editor. Here are some basic vi commands to get you started:

<i> puts you in the insert mode

<esc> put you in the command mode

<:> puts you at the command prompt

<wq> will save a file and quit vi

5. Restore vital LOAD files - In the event that important LOAD files are damaged or the entire drive must be replaced, this selection will bring up a sub-menu that will allow you to begin restoring the original LOAD files from CDROM.

6. Go to Shell Prompt - This selection puts you at the shell prompt with root privileges in the ramdisk file system with the LOAD root partition. The number of commands available is limited, but ls, vi, cp, mount, umount, fsck.ext2 fdisk, diff, dd and the ash shell built-ins are available. Remember that you are root and can delete or change anything, so be very careful.

8. Exit and Reboot - This selection will halt the system. You will have to reboot the system with Alt-Crtl-Del, or the reset button, or by powering off the system and powering it back on.
 
 

8.) Restoring Linux On A Disk from the CDROM

The entire LOAD system may be restored to the hard disk from the LOAD CDROM. Generally, this is not a good way to solve problems that don't rise to the level of a complete file system corruption. This is not Windows. You should never reload the whole system to fix configuration problems.

This is a two step process. In the first step you boot the system from either the floppy or the CDROM and install a minimal Linux system. In the second step you boot from the hard drive and install the rest of the system from the LOAD CDROM.
 
 

1.) Booting the LOAD Install and Repair Menu.

Recreating the LOAD disk is truly a boot strap operation. With the target hard drive properly installed in the system, the computer is booted with a small Linux system that has a bare minimum of tools to prepare the hard disk, access the CDROM , and restore a vital set of files to the hard disk.

The hard disk and CDROM must be either IDE or SCSI and only Adaptec 2940 Series, Buslogic and Tekram are supported as the SCSI controller for the hard disk. The floppy that ships with the IDE Linux On A Disk will only support installation to and from IDE devices.

Boot either from the Install and Repair Disk or from the LOAD CDROM - if your system supports booting from a CDROM.

If you boot from the Install and Repair Disk, it will detect the installation and automatically bring up the Install and Repair Menu. If it does not detect an installed LOAD Disk, it will bring up the Linux On A Disk lilo installation prompt, together with a screen full of warning and disclaimers. To enter the Install and Repair Menu from this screen, type EXPERT in all uppercase letters and press <Enter>.

If you boot from the CDROM it will first ask you if the target hard disk is IDE or SCSI. After you answer, it will hand you over to Install and Repair Menu. In any case you are here:

Linux On A Disk

Emergency Repair Disk

                                        1. Reinstall Linux On A Disk

                                        2. Run File System check for root partition

2) Preparing a new or unpartitioned hard drive

If the drive is a new drive, or if the partition table has been damaged, you will need to chose 6. Go to Shell Prompt. , and use fdisk to partition the drive. The LOAD system assumes five partitions on the drive, a first primary partition stores no files and can be used for DOS or Windows. The second partition is an extended partition and contains all the linux files. See the partition table examples earlier in this manual.

The Linux fdisk command is simple yet very powerful. If used improperly it can destroy any data on for systems hard disks. You must invoke it for the drive with which you want to work with a command like

fdisk <device file> Hard drive device files are as follows:

IDE Drives

Primary master /dev/hda