Koozali.org: home of the SME Server

The boot sector and the map file have *NOT* been altered

John Davey

The boot sector and the map file have *NOT* been altered
« on: April 12, 2002, 03:23:25 AM »
I have sold SME to the management team here (moving away from NT 4.0!!).

I have obatained a new rack mounted dual P3 1GHz server with dual 40 MB ATA/100 drives attached to a Promise Ulta 100 TX2 controler. There is a IDE CD ROM attached to the onboard IDE control, and I can succesfully install from the image CD by specifying

                  Accept ide2=0xd000,0xd402 ide3=0xd800,0xdc02

However, after the install completes, when I attempt to modify /etc/lilo.conf with

                  append"ide2=0xd000,0xd402 ide3=0xd800,0xdc02"

and run either LILO or /SBIN/LILO with a -T option, I recieve the following message:

"The boot sector and the map file have *NOT* been altered"

Yikes! Attempting to boot from the Promise Raid controller results in lilo locking up at

LI

Any advise? I thought that I could get the Promise Contoller to work (after studying several threads on these forums), as it was the default controller provided by the manufacturer.

John.

Dan Brown

Re: The boot sector and the map file have *NOT* been altered
« Reply #1 on: April 12, 2002, 05:15:02 AM »
John Davey wrote:

> and run either LILO or /SBIN/LILO with a -T option, I recieve
> the following message:
>
> "The boot sector and the map file have *NOT* been altered"

    That's what you should get when you tell LILO not to write the changes with the -t flag.

Daniel van Raay

Re: The boot sector and the map file have *NOT* been altered
« Reply #2 on: April 12, 2002, 07:52:34 AM »
According to the man page for lilo:

       -t     Test only. Do not really write a new boot sector or
              map  file.   Use  together with -v to find out what
              lilo is about to do.

I've got SME working using an onboard Promise card using the append= bit (with slightly different numbers) in lilo.conf as you have.

Daniel van Raay

John Davey

Re: The boot sector and the map file have *NOT* been altered
« Reply #3 on: April 12, 2002, 09:03:58 PM »
I am an idiot - I know that the -t switch is for test...

I am still stuck trying to get this box to boot from the Promise Raid card however.

The machine will begin to boot and start to display the LILO prompt but stops after it displays LI

I can boot using the Boot Floppy, and at the LILO prompt typing in

linux ide2=0xd000,0xd402 ide3=0xd800,0xdc02

Am I plain out of luck with this add-on Promise raid controller card?

Filippo Carletti

Re: The boot sector and the map file have *NOT* been altered
« Reply #4 on: April 12, 2002, 09:20:25 PM »
Correct synatx:
append = "ide..."
in /etc/lilo.conf

John Davey

Re: The boot sector and the map file have *NOT* been altered
« Reply #5 on: April 12, 2002, 11:32:12 PM »
Filippo,

I have already modified the /ect/lilo.conf with:

append="ide2=0xd000,0xd402 ide3=0xd800,0xdc02"

then I ran LILO and also checked with LILO -T (and with -V)

I also ran /sbin/lilo (and also with the -T and -V switches).

When I reboot it stops as LILO is beginning (LI)

If I allow the boot floppy to proceed past the LILO prompt, it ends in a kernel panic. If I specify the ide2=0xd000,0xd402 ide3=0xd800,0xdc02 parameters at the LILO prompt, it boots successfully.

Any advise is appriciated - it looks like I might have to pull the Promise RAID controller and run it from the on board IDE contoller and just use the software RAID.

Filippo Carletti

Re: The boot sector and the map file have *NOT* been altered
« Reply #6 on: April 13, 2002, 02:55:19 PM »
> append="ide2=0xd000,0xd402 ide3=0xd800,0xdc02"

Under the right image section ? Could you please post your /etc/lilo.conf ?
Also output of dmesg would be useful.

John Davey

Re: The boot sector and the map file have *NOT* been altered
« Reply #7 on: April 15, 2002, 09:50:26 PM »
Filippo,

Here is the lilo.conf

#------------------------------------------------------------
# DO NOT MODIFY THIS FILE! It is updated automatically by the
# e-smith server and gateway software. Instead, modify the source
# template in the /etc/e-smith/templates directory. For more
# information, see http://www.e-smith.org.
#
# copyright (C) 1999, 2000 e-smith, inc.
#------------------------------------------------------------

boot=/dev/hde
map=/boot/map
install=/boot/boot.b
prompt
timeout=50
message=/boot/mitel.pcx
linear
default=esmith

image=/boot/vmlinuz-2.2.19-7.0.8smp
   label=esmith
   read-only
   root=/dev/hde5




image=/boot/vmlinuz-2.2.19-7.0.8
   label=esmith-up
   read-only
   root=/dev/hde5
   append = "ide2=0xd000,0xd402 ide3=0xd800,0xdc02"




#------------------------------------------------------------
# TEMPLATE END
#------------------------------------------------------------

I tried this with the append = "ide2=0xd000,0xd402 ide3=0xd800,0xdc02" line on each of the listed images, as well as switching the default image from esmith to esmith-up. Any thoughts?

Thanks for all the time you've spent looking at this.
John.

Charlie Brady

Re: The boot sector and the map file have *NOT* been altered
« Reply #8 on: April 15, 2002, 11:49:27 PM »
John Davey wrote:

> Any advise is appriciated - it looks like I might have to
> pull the Promise RAID controller and run it from the on board
> IDE contoller and just use the software RAID.

As far as I know the on board IDE controller and software RAID will be as reliable and will perform as well as the Promise RAID controller, so I don't see why you shouldn't do that.

Charlie

P.S. Be sure to let me know if you find any evidence that the Promise controller actually does a better job.

Filippo Carletti

Re: The boot sector and the map file have *NOT* been altered
« Reply #9 on: April 16, 2002, 01:22:27 AM »
You usually end up with 4 devices on the ide bus and it would be better to have one per channel.
My ide travan tape (I know, it's cheap) seems to hang the channel when rewinding.

If promise performances (or specs) aren't better than onboard controller I'd put hard drives on onboard controller and cd and tape on promise (if supported).

Filippo Carletti

Re: The boot sector and the map file have *NOT* been altered
« Reply #10 on: April 16, 2002, 01:26:53 AM »
But aren't you using software raid ?
Or hde is the promise device name for the mirror set ?
Could you please explain how that motherboard works ?
You could try the "pci=reverse" option, but I don't if it's still useful.

John Davey

Re: The boot sector and the map file have *NOT* been altered
« Reply #11 on: April 16, 2002, 02:31:45 AM »
Filippo,

Here's the Hardware config

Tyan MB with dual P3 running at 1GHz, with 1GHz of RAM.

Onboard IDE controller with nothing attached at IDE0, IDE1, or IDE4. There is a Sony 52x ATA/100 IDE CDROM attached to IDE3 (this is all reflected in the CMOS settings).

Video, and dual Nics onboard.

A Promise ATA Ultra100 TX2 Raid Controller card, with a 40Gig ATA/100 Maxtor Drive attached to both Channel One and Two (total of two 40gig drives in the box). These have been configured as a single 40 Gig RAID array.

When installing SME I selected the first option (Single Drive).

Booting from the Raid Card it locks up at LI

Using the boot floppy the SME install created, and specifiying linux ide2=*** ide3=*** at the LILO prompt will successfully boot the box.

Letting the box boot from the boot floppy, it seems to find the RAID array on the first Autodetecting pass (I see the hex details "0xd000-0xd407" etc... and it identifies the Maxtor drives. It then finds the CD ROM on the next pass and ends in a kernal panic with these details:

Uniform CD-ROM Driver Revision: 3.11
Autodetecting RAID arrays

autorun...
... Autorun DONE

VFS: Cannot open Root device 21:05
Kernal Panic: VFS: Unable to mount Root fs on 21:05

Thanks again, (and again, and again)...
John.

Filippo Carletti

Re: The boot sector and the map file have *NOT* been altered
« Reply #12 on: April 16, 2002, 06:52:30 PM »
> A Promise ATA Ultra100 TX2 Raid Controller card, with a 40Gig
> ATA/100 Maxtor Drive attached to both Channel One and Two
> (total of two 40gig drives in the box). These have been
> configured as a single 40 Gig RAID array.

And Linux views the mirror set as hde ?
Anyone can confirm Promise behaviour ?

I'd use software raid, I bet you'd better performances.
You'll certainly have better control.