It is recommended to install the other operating system first, before installing Linux. This allows the other OS to "get comfortable" with the hardware and possibly write values to the MBR that it would just over-write if Linux was installed first.
You will probably need to do this in a several step method, however. First start the install, but if the operating system partitions the entire drive for itself, see if you can "bail out" early and use the native fdisk to create a primary partition of the size you want to leave for the OS (150-500 megs seems to be usual depending on your needs). Then reboot and go through the install again and normally the OS will just use the space that you just set aside. Once the install is finished, you can begin the Linux installation.
If you are going to need to do a hard drive install due to some problem with your cdrom, you will need to follow these steps.
Have a DOS partition that is formatted in FAT16, and create a directory called \backslashRedHat. From there you will need to copy the items from the cdrom over to the hard-drive.
mkdir C:\RedHat mkdir C:\RedHat\base mkdir C:\RedHat\RPMS copy E:\RedHat\base C:\RedHat\base copy E:\RedHat\RPMS C:\RedHat\RPMS
If you do not have enough disk space for copying the entire RPMS directory tree over to your hard-drive, you will need to look in the file \backslashRedHat\backslashbase\backslashcomps file for the RPMS that are needed in the base and any other sections you feel you need.
Once you have done this, you can start the install and choose a Hard-Drive install. You will be asked to insert the supplemental floppy and a progress meter will "pop-up" to show you what is happening. Once the supplemental disk has been loaded, you will be presented with the next screen on the install.
We are sorry for the problem. The cdrom contains all the data to make new boot and supplemental floppies for your system. To make new floppies under DOS, Win95, or NT (over-writing any data on the 1.44 Megabyte floppies):
cd \dosutils
rawrite.exe
..\images\boot.img
..\images\supp.img
If you are in Linux or another Unix, you can mount the CD-ROM and use dd to write the data to floppy. With Linux, you could do the following:
dd if=/mnt/cdrom/images/boot.img of=/dev/fd0 bs=72k dd if=/mnt/cdrom/images/supp.img of=/dev/fd0 bs=72k
Linux sets up the drive system in a very different pattern than DOS, and this can be rather confusing. Instead of calling the first hard drive NT, it will be usually be a combination of letters signifying what kind of BUS (sd for SCSI, hd for IDE) and on which sequence it was detected. Finally a number is tagged onto the end to specify which partition on the drive is being referenced.
For IDE hard drives the layout depends on which IDE channel the drive is on and whether it is the master or slave on that channel.
Channel Jumper hdx =================================== ide0 master hda ide0 slave hdb ide1 master hdc ide1 slave hdd ide2 master hde ide2 slave hdf ide3 master hdg ide3 slave hdh ide0 = primary ide1 = secondary ide2 = tertiary ide3 = quarterary
The partition number follows an old PC standard that there are a limit of 4 primary partitions per hard drive, but one of those partitions can be designated as an extended partition. Inside of this extended partition, logical partitions can be specified (for most drives you can have 12 logical drives in the extended partition for 16 partitions all together).
The numbering scheme is broken into the following:
It has been found that the options for a cdu31a need some tweaking in the install. Instead of auto probing for the cdrom, select "specify options". A screen like the following should appear.
+---------------| Module parameters |---------------+ | | | Module options: | | | | IO base, IRQ, PAS?: cdu31a=_____________ | | | | Miscellaneous options: ____________________ | : : +---------------------------------------------------+
Erase the complete first input line (cdu31a=) and entered the following as Miscellaneous options:
cdu31a_port=0x360 cdu31a_irq=0 sony_pas_init=0
If your cdrom is on a Sony PAS
cdu31a_port=0x360 cdu31a_irq=0 sony_pas_init=1
With these input everything should work perfectly. Auto probing the CD-ROM drive will usually not be successful, but this might be caused by the not- too-usual port address setting of the CDU33A.
The following web pages contains a great deal of information regarding installing linux on a laptop computer:
http://www.cs.utexas.edu/users/kharker/linux-laptop/
The problem is that a very few earlier versions of rpm would write the database in a way that seems corrupted to later versions. Rebuilding the database fixes the install problems. We will need to upgrade rpm on your system to the one on the installation cdrom, and rebuild the databases.
First thing to do is mount the 5.0 CD on the system.
mount /mnt/cdrom
After doing this upgrade 'rpm' off the CD like so:
cd /mnt/cdrom/RedHat/RPMS rpm -Uvh --nodeps --force rpm-*rpm
When the new RPM is installed, rebuild the database.
rpm --rebuilddb
This will now put the database in a format that the installation RPM can use (since they are the same.)
The Adaptec 2920 does not use an Adaptec chip set, but actually uses the Future Domain TMC3260. If the card is not found with an auto-probe, you will need to specify options for it. When the installation menu asks you about SCSI, choose "options" and enter the following :
setup_called=1 port_base=<io base> interrupt_level=<irq>
An example of this would be
setup_called=1 port_base=0xd000 interrupt_level=9
In case you do not have the port_base or the interrupt_level of the card, you can do the following. If you have reached the 2nd stage of the install (meaning the install has found the cdrom), you can ALT-F2 over to the root console, and cat /proc/pci to see what Linux found on the PCI bus. Otherwise, you will need to boot into rescue mode, and say no to having any scsi devices in the system. Then at the # root prompt, cat /proc/pci and write down the values for the card. Then restart the machine with a CONTROL-ALT-DELETE and fill in the values for the card during the install.
If the above options do not work, your card may have a newer BIOS that we can't probe correctly. We are working on the problem, but do not have a time for when it will be resolved.
Unfortunately, Adaptec has not actively supported the development of a driver for the AHA2940 SCSI adapter and AIC7xxx SCSI chipset.
Specifically, Adaptec continues to release new BIOSes and chip set revisions without making the specifications available to the authors of the Linux driver. Consequently, some AHA2940 adapters work well with Linux, because they have compatible chip sets, while other AHA2940s have different chip sets that the developers haven't seen and do not work well with Linux. Until the authors of the Linux drivers are able to make the driver work with all the different AHA2940 BIOSes, we cannot put this hardware on the supported list, but instead provide a driver in an "AS IS" form. If it works, great, and if it doesn't, you can try the following:
Signal 11's and signal 7's are errors indicating a hardware error in memory or on the bus. This can be due to problems in executables or with the hardware of the system. The Linux kernel uses a lot more capabilities of the CPU, Cache, and memory, and is more prone to faulting on marginal hardware.
The first thing to do is check to see if you have the latest installation and supplemental floppies from Red Hat. Check the errata for updates and also the FTP site to see if newer versions are available. If the latest images still fail, it may be due to hardware. Common suspects are memory or CPU-cache. Try turning off the CPU-cache in the BIOS and see if the problem goes away. Also try swapping memory around in the motherboard slots to see if it is either slot or memory related.
More information is available at http://www.bitwizard.nl/sig11.
Ultra DMA support in the Linux kernel is rather limited in kernels before 2.0.34 For some motherboards it works well, and others not at all. When interfacing Ultra DMA drives with these motherboards, you may need to set the BIOS to use PIO mode 2 versus PIO mode 4 transfers.
Due to various problems, most Ultra DMA controllers that are not built onto the motherboard will not work with the stable Linux kernel series 2.0.xx. Due to these problems, these cards can be a pain to install Linux onto. The following mini-HOWTO describes has more information on these controllers and possible ways to install Linux on them.
http://www.redhat.com/linux-info/ldp/HOWTO/mini/Ultra-DMA.html
by Brion
Vibber, brion@pobox.com
The important parts are included below, but we recommend reading the rest of the HOWTO for why this works (and what to do if this doesn't work).
Under Red Hat 5.0, use the installation boot floppy, and at the boot prompt type rescue. It'll load some stuff, prompt you for the supplemental disk, ask you about your monitor and keyboard, and finally put you into a command prompt. Then, do cat /proc/pci, write down the numbers as above, and reboot from the boot disk. This time, type linux ide2=(this is where you put the numbers like shown above) ide3=(more numbers). It should now be able to install onto your hard disk without difficulty.
Thanks to Gadi Oxman for the following information on getting the interface settings:
If we can access the console with the installation disk, we can also use "cat /proc/pci" to display the Promise interface settings: RAID bus controller: Promise Technology Unknown device (rev 1). Vendor id=105a. Device id=4d33. Medium devsel. IRQ 12. Master Capable. Latency=32. I/O at 0xe000. (a) I/O at 0xd804. (b) I/O at 0xd400. (c) I/O at 0xd004. (d) I/O at 0xc800. (e) and pass "ide2=a,b+2 ide3=c,d+2" as a command line parameter to the kernel.
Note that the numbers probably aren't the same as what you'll
have. Just as an example, the parameters to use for the above set of
numbers would be:
ide2=0xe000,0xd806 ide3=0xd400,0xd006. You may or
may not also need to include the IRQ, which would make it:
ide2=0xe000,0xd806,12 ide3=0xd400,0xd006,12. If you're only
using the first channel on the Promise controller (for instance, if you
only have one drive, or two if they're master and slave on the same
channel, on the Promise), then you won't need to specify ide3.
The best instructions on dual-booting NT and Linux are to be found in
the howto,
Linux+NT-Loader. The latest version can be found
at the HTTP site:
http://sunsite.unc.edu/LDP/HOWTO/mini/Linux+NT-Loader and a
possibly earlier one can be found on the cdrom in
/doc/HOWTO/mini/Linux+NT-Loader
You will need to have the kernel below this limit, and use a different bootloader than LILO. The BIOS may not be able to reach it for one of the following reasons:
You will probably need to use the LOADLIN boot loader that is
provided on the cdrom in
\backslashdosutils. You will need to copy
this directory over to your DOS hard drive partition and edit the
autoboot.bat file to point to its new position and remove the
initrd= line. If your system is a SCSI system, you will need
to go into rescue mode and copy the /mnt/boot/initrd.img from
the hard drive over to the dos partition, and use it for booting.
There are several methods to removing LILO from the master boot record of the machine. Inside of Linux, you can replace the MBR with an earlier saved version of the MBR using the /sbin/lilo command:
/sbin/lilo -u
In DOS, NT, and Windows 95 you can use the fdisk command to create a new MBR with the "undocumented" flag /mbr. This will ONLY rewrite the MBR to boot the primary DOS partition
fdisk /mbr
If you need to remove linux from a hard drive, and have attempted to do this with the default DOS fdisk, you are having the "Partitions exist but they don't exist" problem. The best way to remove non-DOS partitions is with a tool that understands partitions other than DOS.
You can do this with the installation floppy by doing the following. Start the installation, select install (versus upgrade) and when it comes to partitioning the drive, choose fdisk. In fdisk type p to print out the partition numbers, and remove the Linux partitions with the d command. If satisfied with the changes you have made, you can quit with a w and the changes will be saved to disk. If you deleted too much, type q and no changes will occur.
Once you have removed the Linux partitions, you can reboot the box
with
Control-Alt-Delete instead of continuing with the install.
If you have rebooted the system and have gotten a L, LI, or a combination of this and a lot of scrolling numbers, this indicates that LILO is having a problem bootstrapping itself due to one or more problems.
Write down the error codes that are being printed and what letter it stopped on (L, LI, LIL, etc). If you can access the cdrom (from DOS floppies or another machine), you can cd to (assuming DOS) \backslashlive\backslashusr\backslashdoc\backslashlilo-0.20\backslashREADME and check to see what the error seems to indicate.
Most of the time, the LILO failures are due to the BIOS and the hard-drive mismatching geometries or something similar:
If you need to gather more information for someone else (either for official support or from the mailing lists, news groups, friends, etc) you can use the rescue mode
Insert the installation floppy, and at the boot prompt type:
linux rescue
After a couple of screens that ask about hardware you will get a root prompt (#). You will need to mount the linux root partition like is done in this example which has the / partition as /dev/hda5
mkdir /mnt mount /dev/hda5 /mnt
Then do the following:
lilo -v -r /mnt
Record the output of the command. You can add more -v's if you need more information. If errors still occur, you can send that output to the appropriate support group.