LILO with boot over 1024 cylinders?

LILO with boot over 1024 cylinders?

Post by Roger S Fluge » Sun, 31 Dec 1899 09:00:00



While setting up Red Hat 6.0 I was careful to make my root partition
(which contains boot) under 8.4 GB in size.

Unfortunately I set up my initial partitions with Disk Druid. Only
afterwards (when I ran fdisk at a later date) did I find out that my root
partition is 8063 MB but is also 1044 cylinders! fdisk spit back the usual
warning about LILO and my root (and boot) partition being greater than
1024 cylinders.

My questions are the following. Although my root (and boot) partition is
over 1024 cylinders, LILO works fine. My system works fine.

Should I leave things as they are? Or am I flirting with danger?

Should I wipe out my installation and repartition/reformat? I *really*
don't want to do this.

Or should I resize my partition with Partition Magic 4.0? Any advice on
this? After resizing my root (and boot) partition with PM 4.0, do I need
to pass any new cylinder information on to LILO for proper functioning?

TIA! Roger

 
 
 

LILO with boot over 1024 cylinders?

Post by Tom Pfeife » Sun, 31 Dec 1899 09:00:00



> While setting up Red Hat 6.0 I was careful to make my root partition
> (which contains boot) under 8.4 GB in size.

> Unfortunately I set up my initial partitions with Disk Druid. Only
> afterwards (when I ran fdisk at a later date) did I find out that my root
> partition is 8063 MB but is also 1044 cylinders! fdisk spit back the usual
> warning about LILO and my root (and boot) partition being greater than
> 1024 cylinders.

> My questions are the following. Although my root (and boot) partition is
> over 1024 cylinders, LILO works fine. My system works fine.

> Should I leave things as they are? Or am I flirting with danger?

It should be OK. As long as the files lilo needs (usually in the /boot
directory) are within 1024 cylinders, Lilo will work. If you are uneasy
about it, you can look into setting up Loadlin as a backup boot method.
Loadlin will boot Linux from anywhere on a disk, including from beyond
1024 cylinders. Loadlin boots Linux from a DOS prompt or a batch file,
so you can put it on a floppy disk or on a DOS partition. The only
things you need to use it are Loadlin.exe itself and a copy of your
kernel image file. Here's a typical loadlin command line:

loadlin  name_of_kernel_file  root=/dev/hdxx ro

Where hdxx is your Linux root partition - like hda1, hda2 etc. The "ro"
is to mount the root partition read-only until fsck runs during bootup -
this is normal procedure.

Quote:> Or should I resize my partition with Partition Magic 4.0? Any advice on
> this? After resizing my root (and boot) partition with PM 4.0, do I need
> to pass any new cylinder information on to LILO for proper functioning?

If you are using Lilo to boot Linux, and you resize a Linux root
partition with PM 4.0, be prepared to use an alternate boot method
afterwards - like Loadlin. PM 4.0 attempts to make the necessary changes
to the Lilo boot sector, but in my experience, it often (not always)
gets those changes wrong. This usually results in a "LI" when you try to
boot afterwards.

You will often need to boot Linux another way afterwards, and reinstall
Lilo by running the 'lilo' command as root. This will make the necessary
corrections and get you going again.

Tom

 
 
 

LILO with boot over 1024 cylinders?

Post by OldUncle » Sun, 31 Dec 1899 09:00:00


It was: Sat, 18 Sep 1999 07:49:55 -0400  and with STARTLING insight,  "Tom

  posted "Re: LILO with boot over 1024 cylinders?"
 to "comp.os.linux.setup" :

-->> Or should I resize my partition with Partition Magic 4.0? Any advice on
-->> this? After resizing my root (and boot) partition with PM 4.0, do I need
-->> to pass any new cylinder information on to LILO for proper functioning?

-->If you are using Lilo to boot Linux, and you resize a Linux root
-->partition with PM 4.0, be prepared to use an alternate boot method
-->afterwards - like Loadlin. PM 4.0 attempts to make the necessary changes
-->to the Lilo boot sector, but in my experience, it often (not always)
-->gets those changes wrong. This usually results in a "LI" when you try to
-->boot afterwards.

-->You will often need to boot Linux another way afterwards, and reinstall
-->Lilo by running the 'lilo' command as root. This will make the necessary
-->corrections and get you going again.

-->Tom

Thanks, this explains a lot.  

On another issue:  Lilo doesn't allow booting from hdc or hdd.  I read that
this is because bios won't 'handle' the second ide channel on its own.  It
seems that bios does detect the drives on boot that are on ide-1 tho, so
what is the basic difference?  Does bios have some routine, or driver built
in that operates devices on hda & hdb?  Are bios not large enough to
include the additional features for more of the system at boot?  Is this
the wrong track altogether?  Thanks for insights you might provide!  /ts


 
 
 

LILO with boot over 1024 cylinders?

Post by Bill Unr » Sun, 31 Dec 1899 09:00:00



]On another issue:  Lilo doesn't allow booting from hdc or hdd.  I read that
]this is because bios won't 'handle' the second ide channel on its own.  It

Don't know I suspect it dependson the bios.

]seems that bios does detect the drives on boot that are on ide-1 tho, so
]what is the basic difference?  Does bios have some routine, or driver built

Lazyness on the part of the bios manufacturers? Having better things to
do (like implimenting the latest ideas from Redmont)?

Anyway, lilo only brings up the bare system. Youcan put a minimal "boot"
partition on either hda or hdb and put everything else elsewhere (hdd
say)

]in that operates devices on hda & hdb?  Are bios not large enough to
]include the additional features for more of the system at boot?  Is this
]the wrong track altogether?  Thanks for insights you might provide!  /ts

 
 
 

LILO with boot over 1024 cylinders?

Post by L. Friedma » Sun, 31 Dec 1899 09:00:00



> It was: Sat, 18 Sep 1999 07:49:55 -0400  and with STARTLING insight,  "Tom

>   posted "Re: LILO with boot over 1024 cylinders?"
>  to "comp.os.linux.setup" :

> -->> Or should I resize my partition with Partition Magic 4.0? Any advice on
> -->> this? After resizing my root (and boot) partition with PM 4.0, do I need
> -->> to pass any new cylinder information on to LILO for proper functioning?

> -->If you are using Lilo to boot Linux, and you resize a Linux root
> -->partition with PM 4.0, be prepared to use an alternate boot method
> -->afterwards - like Loadlin. PM 4.0 attempts to make the necessary changes
> -->to the Lilo boot sector, but in my experience, it often (not always)
> -->gets those changes wrong. This usually results in a "LI" when you try to
> -->boot afterwards.

> -->You will often need to boot Linux another way afterwards, and reinstall
> -->Lilo by running the 'lilo' command as root. This will make the necessary
> -->corrections and get you going again.

> -->Tom

> Thanks, this explains a lot.

> On another issue:  Lilo doesn't allow booting from hdc or hdd.  I read that
> this is because bios won't 'handle' the second ide channel on its own.

Maybe if your BIOS is either old, or lesser quality.  Mine boots off of
hdc every time i use linux.

Quote:>  It
> seems that bios does detect the drives on boot that are on ide-1 tho, so
> what is the basic difference?  Does bios have some routine, or driver built
> in that operates devices on hda & hdb?  Are bios not large enough to
> include the additional features for more of the system at boot?  Is this
> the wrong track altogether?  Thanks for insights you might provide!  /ts

Its all based on how the BIOS was designed.  If you motherboard came
with more than one IDE slot, then the BIOS should be capable of booting
off of either.  If it can't, then you've either got a defect, or a
really poor quality system.  If you somehow managed to build a system of
random parts, then its possible that the BIOS is not really suited for
the motherboard that its on.

-L

 
 
 

LILO with boot over 1024 cylinders?

Post by Roger S Fluge » Sun, 31 Dec 1899 09:00:00



> It should be OK. As long as the files lilo needs (usually in the /boot
> directory) are within 1024 cylinders, Lilo will work. If you are uneasy
> about it, you can look into setting up Loadlin as a backup boot method.
> Loadlin will boot Linux from anywhere on a disk, including from beyond
> 1024 cylinders. Loadlin boots Linux from a DOS prompt or a batch file,
> so you can put it on a floppy disk or on a DOS partition. The only
> things you need to use it are Loadlin.exe itself and a copy of your
> kernel image file. Here's a typical loadlin command line:

> loadlin  name_of_kernel_file  root=/dev/hdxx ro

> Where hdxx is your Linux root partition - like hda1, hda2 etc. The "ro"
> is to mount the root partition read-only until fsck runs during bootup -
> this is normal procedure.

Thanks for the tip Tom. I will definitely give this a shot.

Alternatively, will the boot disk created during a Red Hat 6.0 install (or
by mkbootdisk) do the trick and avoid the "over 1024 cyls" problem? Does
such a boot disk use LILO? I would think not. I could not find any info
about this on the mkbootdisk man page. I would think such a boot disk uses
loadlin, as it already has the partition on the floppy to boot from!

Does the floppy created by mkbootdisk boot the kernel right of my hard
disk or does it also copy my kernel to the floppy? I suppose I should make
a boot disk to find out for myself!

Lastly, I am running a dual processor machine and under /boot, my vmlinuz
is linked to vmlinuz-2.2.5-15 and *not* vmlinuz-2.2.5-15smp. Should I
re-link my vmlinuz to the SMP kernel instead? Does it make a difference
what vmlinuz is linked to, if my LILO configuration is set to boot the SMP
kernel?

Ultimately, I am not very worried. My / partition is actually 1028 cyls
(and not 1044 like I stated earlier). So I am only 4 cyls over the 1024
limit. I suppose this places the probability of my kernel somehow being
relocated to these "off limit" cylinders near 0.4%. I can live with these
chances, but I will make a bootable floppy just in case.

Quote:> If you are using Lilo to boot Linux, and you resize a Linux root
> partition with PM 4.0, be prepared to use an alternate boot method
> afterwards - like Loadlin. PM 4.0 attempts to make the necessary changes
> to the Lilo boot sector, but in my experience, it often (not always)
> gets those changes wrong. This usually results in a "LI" when you try to
> boot afterwards.

> You will often need to boot Linux another way afterwards, and reinstall
> Lilo by running the 'lilo' command as root. This will make the necessary
> corrections and get you going again.

Thanks for the info. I will tinker with partitions and PM 4.0 as a last
resort.

Roger

 
 
 

LILO with boot over 1024 cylinders?

Post by OldUncle » Sun, 31 Dec 1899 09:00:00


It was: Sat, 18 Sep 1999 16:15:46 -0400  and with STARTLING insight,  ""L.

  posted "Re: LILO with boot over 1024 cylinders?"
 to "comp.os.linux.setup" :

-->
-->
-->>
-->> It was: Sat, 18 Sep 1999 07:49:55 -0400  and with STARTLING insight,  "Tom

-->>   posted "Re: LILO with boot over 1024 cylinders?"
-->>  to "comp.os.linux.setup" :
-->>
-->> -->> Or should I resize my partition with Partition Magic 4.0? Any advice on
-->> -->> this? After resizing my root (and boot) partition with PM 4.0, do I need
-->> -->> to pass any new cylinder information on to LILO for proper functioning?
-->>
-->> -->If you are using Lilo to boot Linux, and you resize a Linux root
-->> -->partition with PM 4.0, be prepared to use an alternate boot method
-->> -->afterwards - like Loadlin. PM 4.0 attempts to make the necessary changes
-->> -->to the Lilo boot sector, but in my experience, it often (not always)
-->> -->gets those changes wrong. This usually results in a "LI" when you try to
-->> -->boot afterwards.
-->>
-->> -->You will often need to boot Linux another way afterwards, and reinstall
-->> -->Lilo by running the 'lilo' command as root. This will make the necessary
-->> -->corrections and get you going again.
-->>
-->> -->Tom
-->>
-->> Thanks, this explains a lot.
-->>
-->> On another issue:  Lilo doesn't allow booting from hdc or hdd.  I read that
-->> this is because bios won't 'handle' the second ide channel on its own.
-->
-->Maybe if your BIOS is either old, or lesser quality.  Mine boots off of
-->hdc every time i use linux.
-->
-->>  It
-->> seems that bios does detect the drives on boot that are on ide-1 tho, so
-->> what is the basic difference?  Does bios have some routine, or driver built
-->> in that operates devices on hda & hdb?  Are bios not large enough to
-->> include the additional features for more of the system at boot?  Is this
-->> the wrong track altogether?  Thanks for insights you might provide!  /ts
-->
-->Its all based on how the BIOS was designed.  If you motherboard came
-->with more than one IDE slot, then the BIOS should be capable of booting
-->off of either.  If it can't, then you've either got a defect, or a
-->really poor quality system.  If you somehow managed to build a system of
-->random parts, then its possible that the BIOS is not really suited for
-->the motherboard that its on.

-->-L

I don't know about component quality.  With these parts that's perhaps
debatable, perhaps not:  FIC pa2007 motherboard (VIA chipset) with Award
4.51g bios, on a 1MB flash rom.  Memory is Corsair pc100/2-cas 128MB, hdd's
are IBM deskstar 8.4GB.  Those are at least quality components for what
they are, super 7 board, sdram (close to top of the line here), very good
quality disks but not scsi.  Cpu in this case is AMD k6 233.  All is
recognized correctly on bios screens on boot.  Lilo returns LI when trying
to boot from hdc or hdd however.

This is not a real big issue as the boot section under 1024 cyl trick will
probablyl work well no matter Linux is installed or how many devices are on
the system, tho it does bear further testing.  But I am interested to know
because sooner or later I'll be installing Linux for others as a second or
greater o/s, primarily so they can experience it and decide.....        /ts


 
 
 

LILO with boot over 1024 cylinders?

Post by D-Ma » Sun, 31 Dec 1899 09:00:00


I have also had some problems booting with Lilo.

My system contains:

hda -- 8GB, Win32
hdc -- 10GB, Linux/Win32 (multiple partitions with /boot under 1024
cyls)
hdb -- cdrom
hdd -- zip
kernel 2.2.11 (can handle large disks)

With Lilo on a floppy, I can boot Linux with no problem, but there is no
/dev/hda on the floppy preventing booting of Windows.  I tried
installing Lilo on hda (the MBR), but can't boot at all -- "LI" error.

Would this be a crappy BIOS problem (on a Compaq Presario 5035 system),
or is there a solution to the problem.

I don't have any non-destructive partitioning software to put /boot on
/dev/hda.

Thank you for any assistance!!

_D-Man


> It was: Sat, 18 Sep 1999 16:15:46 -0400  and with STARTLING insight,  ""L.

>   posted "Re: LILO with boot over 1024 cylinders?"
>  to "comp.os.linux.setup" :

> -->
> -->

> -->>
> -->> It was: Sat, 18 Sep 1999 07:49:55 -0400  and with STARTLING insight,  "Tom

> -->>   posted "Re: LILO with boot over 1024 cylinders?"
> -->>  to "comp.os.linux.setup" :
> -->>
> -->> -->> Or should I resize my partition with Partition Magic 4.0? Any advice on
> -->> -->> this? After resizing my root (and boot) partition with PM 4.0, do I need
> -->> -->> to pass any new cylinder information on to LILO for proper functioning?
> -->>
> -->> -->If you are using Lilo to boot Linux, and you resize a Linux root
> -->> -->partition with PM 4.0, be prepared to use an alternate boot method
> -->> -->afterwards - like Loadlin. PM 4.0 attempts to make the necessary changes
> -->> -->to the Lilo boot sector, but in my experience, it often (not always)
> -->> -->gets those changes wrong. This usually results in a "LI" when you try to
> -->> -->boot afterwards.
> -->>
> -->> -->You will often need to boot Linux another way afterwards, and reinstall
> -->> -->Lilo by running the 'lilo' command as root. This will make the necessary
> -->> -->corrections and get you going again.
> -->>
> -->> -->Tom
> -->>
> -->> Thanks, this explains a lot.
> -->>
> -->> On another issue:  Lilo doesn't allow booting from hdc or hdd.  I read that
> -->> this is because bios won't 'handle' the second ide channel on its own.
> -->
> -->Maybe if your BIOS is either old, or lesser quality.  Mine boots off of
> -->hdc every time i use linux.
> -->
> -->>  It
> -->> seems that bios does detect the drives on boot that are on ide-1 tho, so
> -->> what is the basic difference?  Does bios have some routine, or driver built
> -->> in that operates devices on hda & hdb?  Are bios not large enough to
> -->> include the additional features for more of the system at boot?  Is this
> -->> the wrong track altogether?  Thanks for insights you might provide!  /ts
> -->
> -->Its all based on how the BIOS was designed.  If you motherboard came
> -->with more than one IDE slot, then the BIOS should be capable of booting
> -->off of either.  If it can't, then you've either got a defect, or a
> -->really poor quality system.  If you somehow managed to build a system of
> -->random parts, then its possible that the BIOS is not really suited for
> -->the motherboard that its on.

> -->-L

> I don't know about component quality.  With these parts that's perhaps
> debatable, perhaps not:  FIC pa2007 motherboard (VIA chipset) with Award
> 4.51g bios, on a 1MB flash rom.  Memory is Corsair pc100/2-cas 128MB, hdd's
> are IBM deskstar 8.4GB.  Those are at least quality components for what
> they are, super 7 board, sdram (close to top of the line here), very good
> quality disks but not scsi.  Cpu in this case is AMD k6 233.  All is
> recognized correctly on bios screens on boot.  Lilo returns LI when trying
> to boot from hdc or hdd however.

> This is not a real big issue as the boot section under 1024 cyl trick will
> probablyl work well no matter Linux is installed or how many devices are on
> the system, tho it does bear further testing.  But I am interested to know
> because sooner or later I'll be installing Linux for others as a second or
> greater o/s, primarily so they can experience it and decide.....        /ts