Re: [tlug] RH9 install probelm: error enabling swap device

[Jean-Christian Imbeault <jc@example.com>]

Godwin Stewart wrote:
> 
> Actually they're more down-to-earth :)
> 
> If you can log into a console somewhere between the partitioning stage and
> the error message being produced - or just before rebooting - then try
> running "mkswap /dev/hda3".

Thanks for the tip. I tried that but the installer still failed.

I've tried to pin down the problem some more. When the installer starts
up and gives me a shell I immediately try and manually fdisk the HD
myself to see what happens. It seems fdisk can't get more than 2 logical
partitions installed on an extended partitions.

Maybe I my recollection of fdisk and partitions is wrong but I though
you could fit up to 4 logical partitions on every extended partition?

The steps I do are as follows:

# fdisk /dev/hda

1-  add a new extended partition that uses up the whole disk
2-  write partition table out
3-  fdisk again and add a new logical partition of +128M
4-  write partition table out
5-  fdisk again and add a new logical partition of +1024M
6-  write partition table out
7-  fdisk again and add a new logical partition of +1024M
8-  write partition table out
9-  fdisk again and add a new logical partition of +1024M
10- write partition table out
11- fdisk again, these errors pop up:

Warning: Ignoring extra data in partition table 8
Warning: Ignoring extra data in partition table 8
Warning: Ignoring extra data in partition table 8
Warning: Invalid flag 0xffffea55 of partition table 7 will be corrected
by w(rite)
Warning: Invalid flag 0x00f0 of partition table 8 will be corrected by
w(rite)

Printing out the partition table shows that partition 8's starting and
ending blocks have moved and have a ridiculously large range (25899 -
297775, my HD only goes form 1 - 6496) and the partition type has
changed to a BSDI swap.

Is my HD flaky? It's a 40GB drive so I'm wondering if there is BIOS
problem perhaps?

Should I just give up on this drive? I really would like to just go out
and get another one but I hate leaving this problem as is and not
figuring out what the root cause is. If I don't figure it out know it
will bite me later when I least expect it and when the server can't
afford to be down.

I'm not so sure it's the HD but feel it more of a BIOS/mobo problem
(this is that strange korean server with (hot-swapable ?) IDE drive
bays). On the motherboard there seems to be two IDE controller
connectors, each with a round type of IDE cable I haven't seen before.
These two cables then connect into a board that has four connectors that
accept one HD each. The IDE drive itself needs a special adaptor/board
plugged into it to be plugged into the board with the four connectors.

The board with the four connectors is also responsible for providing the
power. (i.e. the is no power cable going straight to the drive, it goes
into the 4-controller board).

Any advice?

Thanks,

Jean-Christian