partitions not recognized.

On Sun, 05 Mar 2006 14:10:51 GMT, matty staggered into the Black Sun and
said:
> Attys Erebos wrote:
>> I have a problem when trying to install Suse, in fact any
>> distribution. I have two hard drives, hda (80 Gb) hdb (40 Gb), In
>> the first one I have a primary partition with XP ntfs and a extended
>> where I have Suse 9.3 ext2. I've tried recently to upgrade to Suse
>> 10, but when preparing for install it detects this hard drive as not
>> partitioned, and this happens either with Suse 10 or trying to
>> reinstall 9.3. The installed partitions are still recognized by the
>> installed system, and are perfectly bootable. The second hard drive
>> is perfectly recognized. Any clue? I'm out of ideas.
> I am not sure about this but try setting the second partition to a
> primary partition. I <think> that you cannot have an extend partition
> without first having a primary partition.

Er. Attys should post the output of "fdisk -l" from his already running
SuSE system, since his description is incomplete and the fdisk output
will have more information in it. There are a couple of things that
could be causing the symptoms that Attys describes, but I've never seen
them in the wild. It seems strange that a small 80G disk would have a
GPT partition table on it, and an 80G disk should be hooked to a
motherboard that's new enough to not require any drive overlay software.

> would try it though and see what happens. I have done a lot of dual
> boots off the same drive for many different OS's and have always used
> primary partitions.

There are only 4 of those things on an x86 partition table, so you're
really limited in how you can divide the disk up. Having extended and
logical partitions is pretty much forced on you if you use Windows tools
to do your partitioning.

--
Matt G|There is no Darkness in Eternity/But only Light too dim for us to see
Brainbench MVP for Linux Admin / mail: TRAP + SPAN don't belong
http://www.brainbench.com / "He is a rhythmic movement of the
-----------------------------/ penguins, is Tux." --MegaHAL

matty wrote:
> Attys Erebos wrote:
>
>
>>Hi guys, I have a problem when trying to install Suse, in fact any
>>distribution. Maybe it's hardware related, excuse me if this is not the
>>group to post it.
>>I have two hard drives, hda (80 Gb) hdb (40 Gb), In the first one I have
>> a primary partition with XP ntfs and a extended where I have Suse 9.3
>>ext2. I've tried recently to upgrade to Suse 10, but when preparing for
>>install it detects this hard drive as not partitioned, and this happens
>>either with Suse 10 or trying to reinstall 9.3. The installed partitions
>>are still recognized by the installed system, and are perfectly
>>bootable. The second hard drive is perfectly recognized.
>>Any clue on what happens? I'm out of ideas.
>>Thanks.
>
>
> I am not sure about this but try setting the second partition to a primary
> partition. I <think> that you cannot have an extend partition without
> first having a primary partition. That may way off though. I would try it
> though and see what happens. I have done a lot of dual boots off the same
> drive for many different OS's and have always used primary partitions.

I don't think that would work. The primary partition is the XP one. In
the extended partition I've been able to install suse more than once.
but I can't any more. Anyway even the primary partition is not detected,
the install acts as if the hard drive were empty. And I don't want to
screw the boot, unless absolutely needed.
I've noticed that the installed suse recognizes the ntfs as a linux
partition. But since this system is not initialized when installing, can
this affect in some way?

Dances With Crows wrote:
> On Sun, 05 Mar 2006 14:10:51 GMT, matty staggered into the Black Sun and
> said:
>
>>Attys Erebos wrote:
>>
>>>I have a problem when trying to install Suse, in fact any
>>>distribution. I have two hard drives, hda (80 Gb) hdb (40 Gb), In
>>>the first one I have a primary partition with XP ntfs and a extended
>>>where I have Suse 9.3 ext2. I've tried recently to upgrade to Suse
>>>10, but when preparing for install it detects this hard drive as not
>>>partitioned, and this happens either with Suse 10 or trying to
>>>reinstall 9.3. The installed partitions are still recognized by the
>>>installed system, and are perfectly bootable. The second hard drive
>>>is perfectly recognized. Any clue? I'm out of ideas.
>>
>>I am not sure about this but try setting the second partition to a
>>primary partition. I <think> that you cannot have an extend partition
>>without first having a primary partition.
>
>
> Er. Attys should post the output of "fdisk -l" from his already running
> SuSE system, since his description is incomplete and the fdisk output
> will have more information in it. There are a couple of things that
> could be causing the symptoms that Attys describes, but I've never seen
> them in the wild. It seems strange that a small 80G disk would have a
> GPT partition table on it, and an 80G disk should be hooked to a
> motherboard that's new enough to not require any drive overlay software.
>
>
>>would try it though and see what happens. I have done a lot of dual
>>boots off the same drive for many different OS's and have always used
>>primary partitions.
>
>
> There are only 4 of those things on an x86 partition table, so you're
> really limited in how you can divide the disk up. Having extended and
> logical partitions is pretty much forced on you if you use Windows tools
> to do your partitioning.
>

here it goes:

Disk /dev/hda: 80.0 GB, 80026361856 bytes
255 heads, 63 sectors/track, 9729 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/hda1 1 5099 40957686 7 HPFS/NTFS
/dev/hda2 * 5100 9730 37198507+ 5 Extended
/dev/hda5 5100 8905 30571663+ 83 Linux
/dev/hda6 8906 9617 5719108+ 4 FAT16 <32M
/dev/hda7 9618 9730 902160 82 Linux swap / Solaris

Disk /dev/hdb: 40.0 GB, 40020664320 bytes
255 heads, 63 sectors/track, 4865 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/hdb1 1 1255 10080756 c W95 FAT32 (LBA)
/dev/hdb2 1256 4865 28997325 f W95 Ext'd (LBA)
/dev/hdb5 1256 2147 7164958+ b W95 FAT32
/dev/hdb6 2148 3737 12771643+ b W95 FAT32
/dev/hdb7 * 3738 4865 9060628+ 83 Linux

Disk /dev/sda: 128 MB, 128974848 bytes
16 heads, 32 sectors/track, 492 cylinders
Units = cylinders of 512 * 512 = 262144 bytes

Device Boot Start End Blocks Id System
/dev/sda1 * 1 492 125936 6 FAT16

Disk /dev/sdb: 14 MB, 14909440 bytes
2 heads, 32 sectors/track, 455 cylinders
Units = cylinders of 64 * 512 = 32768 bytes

Device Boot Start End Blocks Id System
/dev/sdb1 1 455 14531+ 1 FAT12

Dances With Crows wrote:

> On Sun, 05 Mar 2006 14:10:51 GMT, matty staggered into the Black Sun and
> said:
>> Attys Erebos wrote:
>>> I have a problem when trying to install Suse, in fact any
>>> distribution. I have two hard drives, hda (80 Gb) hdb (40 Gb), In
>>> the first one I have a primary partition with XP ntfs and a extended
>>> where I have Suse 9.3 ext2. I've tried recently to upgrade to Suse
>>> 10, but when preparing for install it detects this hard drive as not
>>> partitioned, and this happens either with Suse 10 or trying to
>>> reinstall 9.3. The installed partitions are still recognized by the
>>> installed system, and are perfectly bootable. The second hard drive
>>> is perfectly recognized. Any clue? I'm out of ideas.
>> I am not sure about this but try setting the second partition to a
>> primary partition. I <think> that you cannot have an extend partition
>> without first having a primary partition.
>
> Er. Attys should post the output of "fdisk -l" from his already running
> SuSE system, since his description is incomplete and the fdisk output
> will have more information in it. There are a couple of things that
> could be causing the symptoms that Attys describes, but I've never seen
> them in the wild. It seems strange that a small 80G disk would have a
> GPT partition table on it, and an 80G disk should be hooked to a
> motherboard that's new enough to not require any drive overlay software.
>
>> would try it though and see what happens. I have done a lot of dual
>> boots off the same drive for many different OS's and have always used
>> primary partitions.
>
> There are only 4 of those things on an x86 partition table, so you're
> really limited in how you can divide the disk up. Having extended and
> logical partitions is pretty much forced on you if you use Windows tools
> to do your partitioning.
>
Yea I don't do a lot with partitioning, at most I would have 1 windows, and
maybe one root and a boot partition. I usually run multiple hard drives
instead. I have 5 in one machine right now ranging from 2GB to 80GB. I
use Hard drive live partitions. I don't trust computers so I will set say
all my MP3's on a different disk than windows and maybe all my games on
another disk. This way if windows crashes, or I loose power and linux goes
down I don't, -as great- a chance of corrupting my more important files.
That is just my way of looking at it. The downside is my PSU, and I have
to run an ide card, or have external cd-roms and such.

Attys Erebos wrote:
> Dances With Crows wrote:
> > On Sun, 05 Mar 2006 14:10:51 GMT, matty staggered into the Black Sun and
> > said:

> here it goes:
>
> Disk /dev/hda: 80.0 GB, 80026361856 bytes
> 255 heads, 63 sectors/track, 9729 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
>
> Device Boot Start End Blocks Id System
> /dev/hda1 1 5099 40957686 7 HPFS/NTFS
> /dev/hda2 * 5100 9730 37198507+ 5 Extended
> /dev/hda5 5100 8905 30571663+ 83 Linux
> /dev/hda6 8906 9617 5719108+ 4 FAT16 <32M
> /dev/hda7 9618 9730 902160 82 Linux swap / Solaris
>

I could be mistaken (Especially as this pertains to 2K/XP, NTFS 5.x) -
but if I understand your problem correctly (it's hda1 you're having
issues with?)

[ taken from OP's OP ]
> > I have two hard drives, hda (80 Gb) hdb (40 Gb), In the first one I have
> > a primary partition with XP ntfs and a extended where I have Suse 9.3
> > ext2. I've tried recently to upgrade to Suse 10, but when preparing for
> > install it detects this hard drive as not partitioned, and this happens
> > either with Suse 10 or trying to reinstall 9.3. The installed partitions
> > are still recognized by the installed system, and are perfectly
> > bootable. The second hard drive is perfectly recognized.
> > Any clue on what happens? I'm out of ideas.
> > Thanks.


I don't *think* you should not have the "Extended" (hda2) listed as
"Boot/Active" on /hda .
Perhaps for Win to boot you'll need hda1 as active -- I know *my* Linux
install resides on Logical Partitions (which are 0x83, and are
contained within a win95/FAT32 LBA/INT13 Ext'd partition -- ID=0f) and
boots from Grub without having a boot flag set for it (look at fdsik
output below -- the boot info/parameters for my install of Grub is in
/boot/grub/menu.lst -- some installes have or use a /grub.conf file
located somewhere).

With Grub or LiLo installed you should be able to set either NTFS or
LinuxAnd as others have stated, there are different types of
"Extended" - so the hex 0x00 assignment/designation should help
decipher the exact type of partition it is -- OH! - I see, the hda2 is
a DOS 3.3 FAT16 type Extended (0x05 - ID=5), not the typical FAT32
"Ext'd" that has LBA support and uses INT13 extensions (0x0f - ID=f).

Here's a typical (80GB drive) win98 + Linux readout;

$ sudo fdisk -l /dev/hda

Device Boot Start End Blocks Id System
/dev/hda1 * 1 1275 10241406 c W95 FAT32 (LBA)
/dev/hda2 1276 8036 54307732+ f W95 Ext'd (LBA)
/dev/hda5 1276 3188 15366141 b W95 FAT32
/dev/hda6 3189 4145 7687071 b W95 FAT32
/dev/hda7 4146 4814 5373711 83 Linux
/dev/hda8 4815 4996 1461883+ 82 Linux swap /
Solaris
/dev/hda9 4997 5604 4883728+ 83 Linux
/dev/hda10 5605 6820 9767488+ 83 Linux
/dev/hda11 6821 8036 9767488+ 83 Linux


more detailed info about partition "Types";
http://www.win.tue.nl/~aeb/partitions/partition_types-1.html


> Disk /dev/hdb: 40.0 GB, 40020664320 bytes
> 255 heads, 63 sectors/track, 4865 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
>
> Device Boot Start End Blocks Id System
> /dev/hdb1 1 1255 10080756 c W95 FAT32 (LBA)
> /dev/hdb2 1256 4865 28997325 f W95 Ext'd (LBA)
> /dev/hdb5 1256 2147 7164958+ b W95 FAT32
> /dev/hdb6 2148 3737 12771643+ b W95 FAT32
> /dev/hdb7 * 3738 4865 9060628+ 83 Linux
>
> Disk /dev/sda: 128 MB, 128974848 bytes
> 16 heads, 32 sectors/track, 492 cylinders
> Units = cylinders of 512 * 512 = 262144 bytes
>
> Device Boot Start End Blocks Id System
> /dev/sda1 * 1 492 125936 6 FAT16
>
> Disk /dev/sdb: 14 MB, 14909440 bytes
> 2 heads, 32 sectors/track, 455 cylinders
> Units = cylinders of 64 * 512 = 32768 bytes
>
> Device Boot Start End Blocks Id System
> /dev/sdb1 1 455 14531+ 1 FAT12

Attys Erebos wrote:

> here it goes:

[list of IDE and SCSI hard drives deleted]
A simple querry: who/what of this collection is The Keeper of MBR
(hex=0x80) so the BIOS refers to it on completion of POST.
When mixing devices IDE and SCSI it pays to treat computer as a stupid
machine and spoon feed it the wanted steps.

I would remove temporarily the SCSI connection and see what gives.

Have fun

Stanislaw
Slack user from Ulladulla.

On Mon, 06 Mar 2006 00:37:09 GMT, Stanislaw Flatto staggered into the
Black Sun and said:
> Attys Erebos wrote:
>> here it goes:
> [list of IDE and SCSI hard drives deleted]
> A simple querry: who/what of this collection is The Keeper of MBR
> (hex=0x80) so the BIOS refers to it on completion of POST.

That'd probably be /dev/hda, but ICBW. Time to check that out inside
the machine's BIOS Setup. If it's set to boot from /dev/hdb , weird
things that you don't expect could be happening. If hdb is the boot
disk from the BIOS's point of view, that *shouldn't* lead to the
symptoms Attys reported (good partition table on hda, but a recent SuSE
wouldn't install, claiming the partition table was empty.)

Er... what's your bootloader config look like? It's possible to do
weird things with the table= command in LILO/GRUB.

> When mixing devices IDE and SCSI it pays to treat computer as a stupid
> machine and spoon feed it the wanted steps. I would remove
> temporarily the SCSI connection and see what gives.

Those disks were almost certainly not Real SCSI, but USB by their
capacities (128 and 14M). In most places, there's no reason to keep a
fixed disk of < 8G attached to a modern computer. Disks that old are a
lot slower than their modern counterparts, even if they *are* Real SCSI.
It's possible that Attys could be booting from one of these USB disks,
but that'd require explicitly setting "boot from USB" to Y in the BIOS.
And that still shouldn't lead to the symptoms Attys reported.

Gah. I'm confused by the reports here. Makes me wish I could look at
the machine in person....

--
Matt G|There is no Darkness in Eternity/But only Light too dim for us to see
Brainbench MVP for Linux Admin / mail: TRAP + SPAN don't belong
http://www.brainbench.com / "He is a rhythmic movement of the
-----------------------------/ penguins, is Tux." --MegaHAL

Attys Erebos wrote:

> This is lilo.conf:
> ---------------------
> # Modified by YaST2. Last modification on Mon Mar 6 17:49:29 GMT 2006
>
> menu-scheme = Wb:kw:Wb:Wb
> default = Linux
> timeout = 80
> message = /boot/message
> read-only
> prompt
> boot = /dev/hda
> change-rules
> reset
> disk = /dev/hda
> bios = 0x80
> disk = /dev/hdb
> bios = 0x81

dunno if this matters, but the bios= hex entries say Minix File Systems

iforone wrote:
> Attys Erebos wrote:
>
>
>>This is lilo.conf:
>>---------------------
>># Modified by YaST2. Last modification on Mon Mar 6 17:49:29 GMT 2006
>>
>>menu-scheme = Wb:kw:Wb:Wb
>>default = Linux
>>timeout = 80
>>message = /boot/message
>>read-only
>>prompt
>>boot = /dev/hda
>>change-rules
>>reset
>>disk = /dev/hda
>>bios = 0x80
>>disk = /dev/hdb
>>bios = 0x81
>
>
> dunno if this matters, but the bios= hex entries say Minix File Systems

No they don't. Those "bios=" lines are BIOS addresses/numbers for the
various physical disks. You seem to be confusing them with filesystem
types which aren't used in lilo.conf. The meaning of 0x80 depends
entirely on context, and in this context it does *not* mean Minix
filesystem type.

Dances With Crows wrote:
> On Mon, 06 Mar 2006 00:37:09 GMT, Stanislaw Flatto staggered into the
> Black Sun and said:
>
>>Attys Erebos wrote:
>>
>>>here it goes:
>>
>>[list of IDE and SCSI hard drives deleted]
>>A simple querry: who/what of this collection is The Keeper of MBR
>>(hex=0x80) so the BIOS refers to it on completion of POST.
>
>
> That'd probably be /dev/hda, but ICBW. Time to check that out inside
> the machine's BIOS Setup. If it's set to boot from /dev/hdb , weird
> things that you don't expect could be happening. If hdb is the boot
> disk from the BIOS's point of view, that *shouldn't* lead to the
> symptoms Attys reported (good partition table on hda, but a recent SuSE
> wouldn't install, claiming the partition table was empty.)
>
> Er... what's your bootloader config look like? It's possible to do
> weird things with the table= command in LILO/GRUB.
>
This is lilo.conf:
---------------------
# Modified by YaST2. Last modification on Mon Mar 6 17:49:29 GMT 2006

menu-scheme = Wb:kw:Wb:Wb
default = Linux
timeout = 80
message = /boot/message
read-only
prompt
boot = /dev/hda
change-rules
reset
disk = /dev/hda
bios = 0x80
disk = /dev/hdb
bios = 0x81


image = /boot/vmlinuz
###Don't change this comment - YaST2 identifier: Original name: linux###
label = Linux
initrd = /boot/initrd
root = /dev/hda5
append = "resume=/dev/hda7 selinux=0 splash=silent "
vga = 0x317


image = /boot/vmlinuz
###Don't change this comment - YaST2 identifier: Original name: failsafe###
label = Failsafe
initrd = /boot/initrd
root = /dev/hda5
append = "ide=nodma apm=off acpi=off vga=normal noresume selinux=0
barrier=off nosmp noapic maxcpus=0 3"


image = /boot/memtest.bin
###Don't change this comment - YaST2 identifier: Original name: memtest86###
label = Memory_Test


other = /dev/hda1
label = Ventanas
table = /dev/hda
-----------------------------------------
But the machine is booting perfectly from the installed system, this
configuration file it's not loaded when trying a new install. BTW there
is some way to do this, i mean, to indicate the installation to take the
information from some file, instead of probing the hard drive?
>
>>When mixing devices IDE and SCSI it pays to treat computer as a stupid
>>machine and spoon feed it the wanted steps. I would remove
>>temporarily the SCSI connection and see what gives.
>
>
> Those disks were almost certainly not Real SCSI, but USB by their
> capacities (128 and 14M). In most places, there's no reason to keep a
> fixed disk of < 8G attached to a modern computer. Disks that old are a
> lot slower than their modern counterparts, even if they *are* Real SCSI.
> It's possible that Attys could be booting from one of these USB disks,
> but that'd require explicitly setting "boot from USB" to Y in the BIOS.
> And that still shouldn't lead to the symptoms Attys reported.
>
Yes, they are a USB key and a sd card reader. Anyway, booting from USB
is disabled, and there is no boot in the USB key.

> Gah. I'm confused by the reports here. Makes me wish I could look at
> the machine in person....
>
When I tried to install, Suse gives an option to recover a lost
partition table, It didn't work. Does someone know of some program to do
this outside installation? though I am not really sure if it can be
broken and still booting.

Thank you for helping, BTW.

John-Paul Stewart wrote:
> iforone wrote:
> > Attys Erebos wrote:
> >
> >
> >>This is lilo.conf:
> >>---------------------
> >># Modified by YaST2. Last modification on Mon Mar 6 17:49:29 GMT 2006
> >>
> >>menu-scheme = Wb:kw:Wb:Wb
> >>default = Linux
> >>timeout = 80
> >>message = /boot/message
> >>read-only
> >>prompt
> >>boot = /dev/hda
> >>change-rules
> >>reset
> >>disk = /dev/hda
> >>bios = 0x80
> >>disk = /dev/hdb
> >>bios = 0x81
> >
> >
> > dunno if this matters, but the bios= hex entries say Minix File Systems
>
> No they don't. Those "bios=" lines are BIOS addresses/numbers for the
> various physical disks. You seem to be confusing them with filesystem
> types which aren't used in lilo.conf. The meaning of 0x80 depends
> entirely on context, and in this context it does *not* mean Minix
> filesystem type.

Thank you for that correction and my apologies to the OP -- now I
know...I should've (known that), atleast investigated further before
posting.