Large EIDE Hard Drive Trouble

Large EIDE Hard Drive Trouble

Post by Steven Butterwor » Wed, 09 Jul 1997 04:00:00



Hello All,

I am looking for persons who have successfully installed ATA-2 (EIDE)
Hard Drives of capacity > 2.1 GB.

My brand spanking new 3.2 GB Quantum hard drive (C/H/S = 6323?,16,63)
is too big for my BIOS to handle (4092,16,63 = 2.1GB upper limit). To
get the machine to recognize its existence I lied and put in
4096/16/63 as the specs. I then booted linux using (from memory, the
1st number might be off here):

        linux hdc=noprobe hdc=6323,16,63

at the Lilo prompt. (Allowing probing makes the BIOS managed portion
of the boot process take a very long time).

During boot up time I get the following error messages (this may not
be be a perfect reproduction but it's very close)

==============

kernel: hdc: recal_intr: status=0x51 { DriveRead SeekComplete Error }
kernel: hdc: recal_intr: status=0x04 { DriveStatus Error }
kernel: hdc: read_intr: status=0x51 { DriveRead SeekComplete Error }
kernel: hdc: read_intr: status=0x04 { DriveStatus Error }
kernel: hdc: recal_intr: status=0x51 { DriveRead SeekComplete Error }
kernel: hdc: recal_intr: status=0x04 { DriveStatus Error }
kernel: hdc: read_intr: status=0x51 { DriveRead SeekComplete Error }
kernel: hdc: read_intr: status=0x04 { DriveStatus Error }
kernel: ide1: reset: success
kernel: hdc: set_geometry_intr: status=0x51
kernel: hdc: set_geometry_intr: status=0x04

kernel: hdc: recal_intr: status=0x51 { DriveRead SeekComplete Error }
kernel: hdc: recal_intr: status=0x04 { DriveStatus Error }
kernel: hdc: read_intr: status=0x51 { DriveRead SeekComplete Error }
kernel: hdc: read_intr: status=0x04 { DriveStatus Error }
kernel: hdc: recal_intr: status=0x51 { DriveRead SeekComplete Error }
kernel: hdc: recal_intr: status=0x04 { DriveStatus Error }
kernel: hdc: read_intr: status=0x51 { DriveRead SeekComplete Error }
kernel: hdc: read_intr: status=0x04 { DriveStatus Error }
kernel: ide1: reset: success
kernel: hdc: set_geometry_intr: status=0x51
kernel: hdc: set_geometry_intr: status=0x04

kernel: end_request: I/O error dev 16:00 sector

===============

Perhaps I need to let it probe first, and I will certainly try, but I
don't really think that is the problem. I can't try again until this
evening.

Also:  "dev 16:00" ???

Doesn't this mean major_device 16, minor_device 00 ?

hdc is major 22, minor 00.

My /dev entries look like the following:

brw-rw---- 1 root disk    3, 0 Dec 31 1969 hda
brw-rw---- 1 root disk   22, 0 Dec 31 1969 hdc
crw-rw-rw- 1 root sys    16, 0 Dec 31 1969 gs4500
brw-rw---- 1 root cdrom  16, 0 Dec 31 1969 gscd

Another result of noprobe ?

Other possibly relevant info:

ASUS P/I P55TP4N motherboard with BIOS ???
Intel P120 CPU

From boot messages:

Console: mono *MDA 80x25, 1 virtual console (max 63)
pcibios_init : BIOS32 Service Directory structure at 0x000fb810
pcibios_init : BIOS32 Service Directory entry at 0xfbcf0
pcibios_init : PCI BIOS revision 2.10 entry at 0xfbd20
.
.
.
hda: Maxtor 7546 AT, 522MB w/256kB Cache, LBA, CHS=530/32/63
hdb: WDC AC21000H, 1033MB w/128kB Cache, LBA, CHS=525/64/63
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14

NE*000 ethercard probe at 0x300: 00 40 33 2a 14 cb
eth0: NE2000 found at 0x300, using IRQ 3.
Partition check:
 hda:Started kswapd v 1.4.2.2
 hda1 hda2 hda3
 hdb: hdb1
VFS: Mounted root (ext2 filesystem) readonly.

The new hard drive is number 3 (thus hdc no?). It is set to master and
attached to the secondary IDE interface on a cable with a single
connector.

Oh yes, the kernel is 2.0.23 . We haven't bothered to move towards
2.0.30 because the kernel has never failed between reboots (approx. 40 day
intervals, usually hardware change or power outage related).

*** later ***

I have now also tried it without the hdc=noprobe. I see the same set
of error messages at probe time and when running fdisk /dev/hdx
(x=b,c for various cases).  I have tried it as master on ide0, as
slave (with and without SP jumper set) on ide0 and as master on
ide1. I also tried it on my 486DX2/66 VLB motherboard at home and it
couldn't get past the POST unless I set the relevant drive setting to
[NONE]. Then at probe time I saw the same set of errors as with my
P120 setup.

This was supposed to be the end of my overlarge astronomical data set
problems. :(

Thanks for any suggestions.

--
Steve Butterworth
--
no .sig zone

 
 
 

Large EIDE Hard Drive Trouble

Post by Douglas E. Mitt » Wed, 09 Jul 1997 04:00:00


Hopefully someone will stop by and explain the PROPER way to do this.
I couldn't get it to work this way either!

I am using a 1.7G EIDE Harddrive on my IDE PC (Old 386DX-40)!
I told CMOS the truth about the drive, actual parameters.
I then created a "small" DOS partition (for old times sake).
I then booted Linux and created a "root" partition below the 1024
limit. Then a SWAP was created, then the remainder became my /usr
partition.

The bottom line being that Linux will only boot from below 1024
cylinders BUT will access the remainder of the drive just fine; after
Linux is executing it doesn't utilize the BIOS!  If you are using DOS
and LOADLIN, then even the boot limitation is removed (I think).

I've used this same scenario several times and all seems to work OK!
I read the HOW-TO but on my original Slackware 3.0 installation, I
couldn't get its ideas to work, so I experimented and determined my
own solution.

Good luck!


>Hello All,

>I am looking for persons who have successfully installed ATA-2 (EIDE)
>Hard Drives of capacity > 2.1 GB.

>My brand spanking new 3.2 GB Quantum hard drive (C/H/S = 6323?,16,63)
>is too big for my BIOS to handle (4092,16,63 = 2.1GB upper limit). To
>get the machine to recognize its existence I lied and put in
>4096/16/63 as the specs. I then booted linux using (from memory, the
>1st number might be off here):

>    linux hdc=noprobe hdc=6323,16,63

>at the Lilo prompt. (Allowing probing makes the BIOS managed portion
>of the boot process take a very long time).

>During boot up time I get the following error messages (this may not
>be be a perfect reproduction but it's very close)

>==============

------------------------------------------------
The FACTS are my Employers, OPINIONS are my own!

Sorry: SPAM reduction project in progress:
       Remove the "x." from my domain to reply!
------------------------------------------------

 
 
 

Large EIDE Hard Drive Trouble

Post by Frank Sweetse » Wed, 09 Jul 1997 04:00:00



Quote:

> Hopefully someone will stop by and explain the PROPER way to do this.
> I couldn't get it to work this way either!

> I am using a 1.7G EIDE Harddrive on my IDE PC (Old 386DX-40)!
> I told CMOS the truth about the drive, actual parameters.
> I then created a "small" DOS partition (for old times sake).
> I then booted Linux and created a "root" partition below the 1024
> limit. Then a SWAP was created, then the remainder became my /usr
> partition.

> The bottom line being that Linux will only boot from below 1024
> cylinders BUT will access the remainder of the drive just fine; after
> Linux is executing it doesn't utilize the BIOS!  If you are using DOS
> and LOADLIN, then even the boot limitation is removed (I think).

> I've used this same scenario several times and all seems to work OK!
> I read the HOW-TO but on my original Slackware 3.0 installation, I
> couldn't get its ideas to work, so I experimented and determined my
> own solution.

Fact is, this _is_ the Right Way To Do It. The problem is, the bios has no
way to access anything beyond the 1024th cylinder, and lilo is limited to
what the bios can access.  The best way is doing what you did, make a small
partition to make sure that the kernel is always belog the 1024th
cylinder.

--
Windows: I can play Doom!              |RedHat Linux 2.0.31pre-2 i486
Linux: I can be a file server, be a Web|Because reboots are for upgrades!
server, run the accounting package with|http://www.wpi.edu/~rasmusin/pgp.html
twelve terminals AND play Doom!        |for pgp key.        frank sweetser

 
 
 

Large EIDE Hard Drive Trouble

Post by Steven Butterwor » Wed, 09 Jul 1997 04:00:00



>I am using a 1.7G EIDE Harddrive on my IDE PC (Old 386DX-40)!
<snip>
>The bottom line being that Linux will only boot from below 1024
>cylinders BUT will access the remainder of the drive just fine; after
>Linux is executing it doesn't utilize the BIOS!  If you are using DOS
>and LOADLIN, then even the boot limitation is removed (I think).


>>Hello All,

>>I am looking for persons who have successfully installed ATA-2 (EIDE)
>>Hard Drives of capacity > 2.1 GB.

I fear my problem is of a different nature. I'm not trying to boot off
the drive, and --- assuming I don't have a real hardware problem ---
the problem I THINK I'm having isn't encountered until > 2.1GB. I am
successfully using a 1.6GB drive on my old 486DX2/66 at home, and am
aware of the 1024 cylinder boundary for booting from Lilo.

I'm really beginning to suspect that the drive is just hosed. The
error messages look alarmingly similar to those from a drive --- a
1.08GB Western Digital Caviar --- that I just replaced under warranty
within the last month or so.

Regards,

Steve Butterworth
--
no .sig zone

 
 
 

Large EIDE Hard Drive Trouble

Post by Steven Butterwor » Wed, 09 Jul 1997 04:00:00


I said:

Quote:>I'm really beginning to suspect that the drive is just hosed. The
>error messages look alarmingly similar to those from a drive --- a
>1.08GB Western Digital Caviar --- that I just replaced under warranty
>within the last month or so.

Hi Folks,

Thanks to all who attempted to help out by usenet and e-mail.

The drive was indeed fubared and its replacement is humming along nicely.

Regards,

Steve Butterworth
no .sig zone

 
 
 

Large EIDE Hard Drive Trouble

Post by M. Buchenried » Fri, 11 Jul 1997 04:00:00



Quote:>Hopefully someone will stop by and explain the PROPER way to do this.
>I couldn't get it to work this way either!
>I am using a 1.7G EIDE Harddrive on my IDE PC (Old 386DX-40)!
>I told CMOS the truth about the drive, actual parameters.
>I then created a "small" DOS partition (for old times sake).
>I then booted Linux and created a "root" partition below the 1024
>limit. Then a SWAP was created, then the remainder became my /usr
>partition.

This is the correct way to do it.
The other possible way would be using LBA, which, however, isn't
available on machines with older BIOS versions.

Quote:>The bottom line being that Linux will only boot from below 1024
>cylinders BUT will access the remainder of the drive just fine; after
>Linux is executing it doesn't utilize the BIOS!  If you are using DOS
>and LOADLIN, then even the boot limitation is removed (I think).

[...]

Any OS has to boot from a kernel image below cyl. 1024 on a PC.
This is a BIOS limitation.

Michael
--

**************************************************************************
Satellite Safety Tip #4:
                If you see a bright streak in the sky coming at you, duck.

 
 
 

Large EIDE Hard Drive Trouble

Post by Cameron L. Spitz » Sun, 13 Jul 1997 04:00:00





>> I am using a 1.7G EIDE Harddrive on my IDE PC (Old 386DX-40)!
>> I told CMOS the truth about the drive, actual parameters.
>> I then created a "small" DOS partition (for old times sake).
>> I then booted Linux and created a "root" partition below the 1024
>> limit. Then a SWAP was created, then the remainder became my /usr
>> partition.

As long as the DOS partition isn't compressed, you can put
the LILO boot files (zImage, map, boot.b, chain.b, message)
there.  No need for a Linux partition to end below cylinder 1023.

However, there are other arguments in favor of a small root partition,
so you might as well have one that ends down there.
The reason to put the LILO files on the DOS partition is only if
you need the DOS partition as big as possible.
Some people actually use DOS, I guess.  :-)

Quote:

>> The bottom line being that Linux will only boot from below 1024
>> cylinders BUT will access the remainder of the drive just fine; after
>> Linux is executing it doesn't utilize the BIOS!  If you are using DOS
>> and LOADLIN, then even the boot limitation is removed (I think).

>> I've used this same scenario several times and all seems to work OK!
>> I read the HOW-TO but on my original Slackware 3.0 installation, I
>> couldn't get its ideas to work, so I experimented and determined my
>> own solution.

The LILO mini-HOWTO is seriously out of date.  I've got three more
scenaria to add, and instructions for E-Z Drive, OnTrack Disk Mangler,
and using the *real* documentation.  And I'm using the sgml-tools.
Later this summer.
Meanwhile, lilo-20 is out and the *real* documentation is good.

Quote:>Fact is, this _is_ the Right Way To Do It. The problem is, the bios has no
>way to access anything beyond the 1024th cylinder, and lilo is limited to
>what the bios can access.  The best way is doing what you did, make a small
>partition to make sure that the kernel is always belog the 1024th
>cylinder.

The semi Right Way To Do It is to use the "BIOS overlay" product
(OT DM or EZ Drive) that came with the drive.  You'll need to do it
that way if you want to use an MS OS with a partition over 528 MB anyway.
This has been working for a couple of years, now.

The other Right Way To Do It is if your BIOS is new enough to support
the so-called "Linear Block Addressing" which is nothing more than a
hack to let you use an absurdly large number of virtual heads
without breaking the drive.

The really right Right Way To Do It is SCSI (but it's overpriced). :-)

Cameron Spitzer in San Jose.
http://www.rahul.net/cameron/PC-info/truffula.html

 
 
 

Large EIDE Hard Drive Trouble

Post by Michael A. Lau » Sun, 13 Jul 1997 04:00:00



> Hello All,

> I am looking for persons who have successfully installed ATA-2 (EIDE)
> Hard Drives of capacity > 2.1 GB.

> My brand spanking new 3.2 GB Quantum hard drive (C/H/S = 6323?,16,63)
> is too big for my BIOS to handle (4092,16,63 = 2.1GB upper limit). To
> get the machine to recognize its existence I lied and put in
> 4096/16/63 as the specs. I then booted linux using (from memory, the
> 1st number might be off here):

>         linux hdc=noprobe hdc=6323,16,63

> at the Lilo prompt. (Allowing probing makes the BIOS managed portion
> of the boot process take a very long time).

In my admittedly limited experience, Linux takes instruction from the
BIOS rather seriously. I had some similar problems with hard drives
which were misconfigured in the BIOS affecting Linux's use of the
drives.

I suspect that you will either have to upgrade your BIOS, or find
some software which will fool your BIOS into dealing with the large
drive. My WD Caviar 3.1 GB HD came with special software which would
do such a thing (though fortunately I didn't need it!). Perhaps you
should check the Quantum web site for possible solutions.

        Cheers,

        Michael A. Laux

 
 
 

Large EIDE Hard Drive Trouble

Post by John Barr » Tue, 15 Jul 1997 04:00:00


Howdy.
There is possibly another way around it.  There are IDE adapters (e.g. some
Promise) that permit you to use their BIOS to extend main BIOS, to add LBA
addressing.  You just plug in the board, and configure existing drive(s) in
CMOS config as type "1" (as I remember).  Any hd will then be
auto-configured.  We did this first to install 4GB drive in Machine that
Micron said would never accept >2.1GB.
HTH, jb




> > Hello All,

> > I am looking for persons who have successfully installed ATA-2 (EIDE)
> > Hard Drives of capacity > 2.1 GB.

> > My brand spanking new 3.2 GB Quantum hard drive (C/H/S = 6323?,16,63)
> > is too big for my BIOS to handle (4092,16,63 = 2.1GB upper limit). To
> > get the machine to recognize its existence I lied and put in
> > 4096/16/63 as the specs. I then booted linux using (from memory, the
> > 1st number might be off here):

> >         linux hdc=noprobe hdc=6323,16,63

> > at the Lilo prompt. (Allowing probing makes the BIOS managed portion
> > of the boot process take a very long time).

> In my admittedly limited experience, Linux takes instruction from the
> BIOS rather seriously. I had some similar problems with hard drives
> which were misconfigured in the BIOS affecting Linux's use of the
> drives.

> I suspect that you will either have to upgrade your BIOS, or find
> some software which will fool your BIOS into dealing with the large
> drive. My WD Caviar 3.1 GB HD came with special software which would
> do such a thing (though fortunately I didn't need it!). Perhaps you
> should check the Quantum web site for possible solutions.

>    Cheers,

>    Michael A. Laux

 
 
 

Large EIDE Hard Drive Trouble

Post by Insectus Penti » Tue, 15 Jul 1997 04:00:00


: I fear my problem is of a different nature. I'm not trying to boot off
: the drive, and --- assuming I don't have a real hardware problem ---
: the problem I THINK I'm having isn't encountered until > 2.1GB. I am
: successfully using a 1.6GB drive on my old 486DX2/66 at home, and am
: aware of the 1024 cylinder boundary for booting from Lilo.

I have one computer with a 540 M and a 1.7 G drive. The Linux is on the
lower 1 G of the 1.7 G drive, leaving the remainder for D: as a place to
put stuff to burn to a CD. I use Loadlin, and have no problems. :)

On the other hand, my old computer has a 2.1 G Seagate EIDE hard drive,
and Linux doesn't like it. :( Has anyone found a way to get at least most
of the drive to run on Linux, like I did with the first computer's case? I
already know about the 1024 cans, and I use Loadlin anyways. (I partition
with the 1024 in mind anyways.)

--
CAUTION: Email Spam Killer in use. Leave this line in your reply! 152680

432590 bytes of spam mail deleted.        http://www.ripco.com/~pentius/

 
 
 

Large EIDE Hard Drive Trouble

Post by Sid Boyc » Wed, 16 Jul 1997 04:00:00




> : I fear my problem is of a different nature. I'm not trying to boot off
> : the drive, and --- assuming I don't have a real hardware problem ---
> : the problem I THINK I'm having isn't encountered until > 2.1GB. I am
> : successfully using a 1.6GB drive on my old 486DX2/66 at home, and am
> : aware of the 1024 cylinder boundary for booting from Lilo.

> I have one computer with a 540 M and a 1.7 G drive. The Linux is on the
> lower 1 G of the 1.7 G drive, leaving the remainder for D: as a place to
> put stuff to burn to a CD. I use Loadlin, and have no problems. :)

> On the other hand, my old computer has a 2.1 G Seagate EIDE hard drive,
> and Linux doesn't like it. :( Has anyone found a way to get at least most
> of the drive to run on Linux, like I did with the first computer's case? I
> already know about the 1024 cans, and I use Loadlin anyways. (I partition
> with the 1024 in mind anyways.)

> --
> CAUTION: Email Spam Killer in use. Leave this line in your reply! 152680

> 432590 bytes of spam mail deleted.        http://www.ripco.com/~pentius/

        Strange, I have a 2.5 Gig EIDE on one machine, a Cyrix p200+, which
before that was a DX4-133, HDA1 is 76Meg swap and hda2 is my Linux root,
on the other p200+ I have a 3.2Gig EIDE and a 1 Gig + AH2940AU SCSI,
with the 3.2 Gig set up like the one on the other machine and I've seen
no problems either at 2.0 or 2.1 latest kernels.
Disk /dev/hda: 128 heads, 63 sectors, 787 cylinders
Units = cylinders of 8064 * 512 bytes

   Device Boot   Begin    Start      End   Blocks   Id  System
/dev/hda1   *        1        1       18    72544+  82  Linux swap
/dev/hda2   *       19       19      787  3100608   83  Linux native
        That's the 3.2Gig drive. I'd expect your setup to work fine.
Regards
--
... Sid Boyce...Amdahl(UK)...44-121 422 0375
                   -----------------------------------
Any opinions expressed above are mine and do not necessarily represent
 the opinions or policies of Amdahl Corporation.

 
 
 

Large EIDE Hard Drive Trouble

Post by Roy Stogn » Wed, 30 Jul 1997 04:00:00






>>> > Hello All,

>>> > I am looking for persons who have successfully installed ATA-2 (EIDE)
>>> > Hard Drives of capacity > 2.1 GB.

>>> > My brand spanking new 3.2 GB Quantum hard drive (C/H/S = 6323?,16,63)
>>> > is too big for my BIOS to handle (4092,16,63 = 2.1GB upper limit). To
>4092,16,63 is not 2.1GB.  4096,16,64 would be 2.0GB.
>I say this because my 2.11GB Quantum hard drive is 4092,16,63.
>When I by a 2.1GB Hard drive I expect it to be more than 2.0GB
>not less.  But I have to admit that Quantum make rather good
>hard drives even if they are a bit stingy on the capacity.

It depends.
Look at the math:
4092 cylinder * 16 heads * 63 sectors * 512 bytes/sector
= 2,111,864,832 bytes.

If you use the metric prefix Giga = 10^9, then that's 2.111 GB.
If you use the computer-friendly prefix Giga = 2^30, then that's 1.967
GB.
HDD manufacturers all pretty much stick with the metric prefixes,
since their hard drives sound bigger that way.

Quote:>>> > get the machine to recognize its existence I lied and put in
>>> > 4096/16/63 as the specs. I then booted linux using (from memory, the
>>> > 1st number might be off here):

>>> >         linux hdc=noprobe hdc=6323,16,63

>>> > at the Lilo prompt. (Allowing probing makes the BIOS managed portion
>>> > of the boot process take a very long time).

To whoever Stephen Butterworth is (sorry, I missed the original msg):

From what did you boot LILO?  I'm having trouble getting LILO to boot
from the MBR or a floppy: I'm currently trying to install Linux onto a
computer with a 3.8 GB (that's metric GB) HDD and a 4092 cylinder
limitation, and LILO never makes it past "LI".
---
Roy Stogner
If you can't cc: to my email, please leave my name in any quoted reply
so I don't miss it in a word search.  Thanks!

 
 
 

Large EIDE Hard Drive Trouble

Post by Ian COWEL » Thu, 31 Jul 1997 04:00:00







> >>> > Hello All,

> >>> > I am looking for persons who have successfully installed ATA-2 (EIDE)
> >>> > Hard Drives of capacity > 2.1 GB.

> >>> > My brand spanking new 3.2 GB Quantum hard drive (C/H/S = 6323?,16,63)
> >>> > is too big for my BIOS to handle (4092,16,63 = 2.1GB upper limit). To

> >4092,16,63 is not 2.1GB.  4096,16,64 would be 2.0GB.
> >I say this because my 2.11GB Quantum hard drive is 4092,16,63.
> >When I by a 2.1GB Hard drive I expect it to be more than 2.0GB
> >not less.  But I have to admit that Quantum make rather good
> >hard drives even if they are a bit stingy on the capacity.

> It depends.
> Look at the math:
> 4092 cylinder * 16 heads * 63 sectors * 512 bytes/sector
> = 2,111,864,832 bytes.

> If you use the metric prefix Giga = 10^9, then that's 2.111 GB.
> If you use the computer-friendly prefix Giga = 2^30, then that's 1.967
> GB.
> HDD manufacturers all pretty much stick with the metric prefixes,
> since their hard drives sound bigger that way.

> >>> > get the machine to recognize its existence I lied and put in
> >>> > 4096/16/63 as the specs. I then booted linux using (from memory, the
> >>> > 1st number might be off here):

> >>> >         linux hdc=noprobe hdc=6323,16,63

> >>> > at the Lilo prompt. (Allowing probing makes the BIOS managed portion
> >>> > of the boot process take a very long time).

> To whoever Stephen Butterworth is (sorry, I missed the original msg):

> From what did you boot LILO?  I'm having trouble getting LILO to boot
> from the MBR or a floppy: I'm currently trying to install Linux onto a
> computer with a 3.8 GB (that's metric GB) HDD and a 4092 cylinder
> limitation, and LILO never makes it past "LI".
> ---
> Roy Stogner
> If you can't cc: to my email, please leave my name in any quoted reply
> so I don't miss it in a word search.  Thanks!

To quote from the LILO User's Guide:
"LI The first stage boot loader was able to load the second stage
boot loader, but failed to execute it. This can either be caused
by a geometry mismatch or by moving the /boot/boot.b without
running the map installer."

The map installer is /sbin/lilo
If there is a geometry mismatch, then you have to tell lilo at
boot time the geometry that is to be used. The geometry used is
that used by fdisk in partitioning the system.
Two ways of doing this:
1) using the append statement in the lilo.conf file as in:
  image=/boot/zImage
        label=tlinux
        root=/dev/hda4
        append="hdb=973,64,63 mem=32M"
        read-only
2) using the parameter input at the LILO boot, as in
  LILO boot: linux hdb=973,64,63

Another way is to use the LBA option in BIOS. This sets the
number of cylinders to less than 1024. The 1024 limit is due to
BIOS. You still need to have fdisk and LILO using the same
geometry.

However there is another BIOS limit.
You cannot boot off the third disk. The message above states hdc
which is the third disk.
You will have to put the boot files on one of the first two
disks. How to do this is described in the LILO User's Guide
which is on the Redhat distribution Official cdrom at:
/live/usr/doc/lilo-0.19-1/doc/User_Guide.ps
and at:
/usr/doc/lilo-0.19-1/doc/User_Guide.ps

and available on the web at:
ftp://lrcftp.epfl.ch/pub/linux/local/lilo/lilo.u.19.ps.gz

The simple method is to put the root partition on one of the
first two disks. This partition contains all the boot files.

--
Ian Cowell

----------------------

 
 
 

Large EIDE Hard Drive Trouble

Post by Sigismondo Bosch » Thu, 31 Jul 1997 04:00:00



> > From what did you boot LILO?  I'm having trouble getting LILO to boot
> > from the MBR or a floppy: I'm currently trying to install Linux onto a
> > computer with a 3.8 GB (that's metric GB) HDD and a 4092 cylinder
> > limitation, and LILO never makes it past "LI".

> I recently encountered that problem.  When you do an fdisk in linux,
> after you install the new partition table, Linux informs you that LILO
> only works on a drive with fewer than 1024 cylinders.  Most large
> hardrives have more than 1024.

> If anyone has found a work around, let me know.  Otherwise, you could
> boot off of a floppy, or configure a smaller drive as your master, and
> install LILO manually on its Master Boot Record.

linux fdisk states the following:

# fdisk
Using /dev/hda as default device!
The number of cylinders for this disk is set to 1238.
This is larger than 1024, and may cause problems with:
1) software that runs at boot time (e.g., LILO)
2) booting and partitioning software form other OSs
   (e.g., DOS FDISK, OS/2 FDISK)
Command (m for help): p

Disk /dev/hda: 16 heads, 63 sectors, 1238 cylinders
Units = cylinders of 1008 * 512 bytes

   Device Boot   Begin    Start      End   Blocks   Id  System
/dev/hda1   *        1        1      183    92200+   6  DOS 16-bit >=32M
/dev/hda2          184      184     1238   531720    5  Extended
/dev/hda5          184      184      208    12568+  82  Linux swap
/dev/hda6          209      209     1238   519088+  83  Linux native

I have found that if the KERNEL image resides before the block 1024 it
works, else it gives me LI and nothing more. I have discovered it when
installing a new kernel: if I copy it over the previous one (I think) it
overwrites its blocks continuing to remain in the beginning of the
partition! How did I put the first kernel in the beginning? luck! :-)

I think you can make a small partition where to put kernel and lilo
files before block 1024 and then run lilo with:
lilo -r /lilofiles

where in /lilofiles you mount such a partition, and have:
/lilofiles/etc/lilo.conf
/lilofiles/vmlinuz

and so on.

BYE
--
                                        Sigismondo Boschi

PhD student at the Physical Chemistry Department
                   Bologna University (Italy)

 
 
 

Large EIDE Hard Drive Trouble

Post by Marc Holubo » Thu, 31 Jul 1997 04:00:00


Quote:> From what did you boot LILO?  I'm having trouble getting LILO to boot
> from the MBR or a floppy: I'm currently trying to install Linux onto a
> computer with a 3.8 GB (that's metric GB) HDD and a 4092 cylinder
> limitation, and LILO never makes it past "LI".

I recently encountered that problem.  When you do an fdisk in linux,
after you install the new partition table, Linux informs you that LILO
only works on a drive with fewer than 1024 cylinders.  Most large
hardrives have more than 1024.

If anyone has found a work around, let me know.  Otherwise, you could
boot off of a floppy, or configure a smaller drive as your master, and
install LILO manually on its Master Boot Record.

Marc