Bug 653081

(In reply to comment #0)
> 1. have a collection of windows partitions on the drives
> 2. run setup and it reports incorrect

Incorrect in what manner? Could you attach /var/log/boot.msg (transferred e.g. via network or USB disk) and output of fdisk -l?

Could be a problem with dmraid. So the output of "dmraid -s -c -c -c" would
also help.

what sort of steps do i do to get out of setup to get this output or get a text output from setup?

i could try to photograph but i guess a text output is better...

i have a SUSE 11.2 Live cd and it sets up the disk correctly (with data accessable) (havnt tried installing it to double check) and i can use it to see what these utilities say in that.
(since the 11.3 installer on DVD is bad and i cannot use disks in the installed OS, hence Bug 640380)
but still, so how do i do it in the setup of 11.3... i do wish runlevel 3 was provided by the installation DVD! - i was told to type in "3" in the booter but this dont work here.... perhaps you can tell em to put it in 11.4...


SUSE11.2 dmraid output: 
isw_dhfiebjffi_2X200GB:781432320::256:stripe:ok:0:2:0
/dev/sdb:isw:isw_dhfiebjffi_2X200GB:stripe:ok:390716411:0
/dev/sda:isw:isw_dhfiebjffi_2X200GB:stripe:ok:390716411:0

as you can see it sort of sees my two drives as a unit "2x200GB" (forgot it was 200 - not 160)

and i have boot.msg but there is no way to upload it in this comments box (samba seems less troublesom in 11.2 too) there doesnt appear anything in here about mounting these partitions only finding the hard disks.

====
in the 11.3 setup i choose create partition setup, then the last option: custom partitioning for experts and i will tell how it sees the drives:
(the first option is this: 1.DM RAID, 372.62GB, /dev/mapper/isw_dhfiebjffi_2x200GB)

at the left it has linux then as a tree it has harddisks,raid,volumemg.., crpt..., devicem.., nfs, unused devices...

harddisks:
device (size) start (end)
/dev/mapper/isw_dhfiebjffi_2x200GB  (372.62gb) 0 (48640)
/dev/mapper/isw_dhfiebjffi_2x200GB_part1 (128.00gb) 0 (16708)
/dev/mapper/isw_dhfiebjffi_2x200GB_part2 (58.31gb) 16709 (24320)
/dev/sda (186.31) 0 (24320)
/dev/sdb (186.31gb) 0 (24320)

the two lines under the first item is the partitions it thinks are in the Raid stripe, and the ending 186GB (after these partitions) is not partitioned.

it reports no partitions in the individual drives because it thinks that i have a raid stripe, but as i said, i have a primaries on them. i did at one stage have the drives in stripe mode but one drive scratched its surface so i split them. i dont think i had 128gb as the first partition...

Charlie.

the output of Fdisk -l ON SUSE 11.2 follows (if i can do runlevel 3 on 11.3 setup i might be able to properly give its confused interpretation of my arrangement, if i install 11.3 it somewhat sees my partitions -- see Bug 640380) last drive is a usb.

linux:/home/linux # fdisk -l

Disk /dev/sda: 200.0 GB, 200049647616 bytes
255 heads, 63 sectors/track, 24321 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0xd646fea2

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1        2741    22017051    7  HPFS/NTFS
/dev/sda2            2742        3263     4192965    7  HPFS/NTFS
/dev/sda3            3264       24321   169148385    7  HPFS/NTFS

Disk /dev/sdb: 200.0 GB, 200049647616 bytes
255 heads, 63 sectors/track, 24321 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x00000001

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1               1       16709   134215011    7  HPFS/NTFS
/dev/sdb2   *       16710       24321    61143390    7  HPFS/NTFS

Disk /dev/sdc: 320.1 GB, 320072933376 bytes
255 heads, 63 sectors/track, 38913 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x0a8dd031

   Device Boot      Start         End      Blocks   Id  System
/dev/sdc1   *           1       38913   312568641    7  HPFS/NTFS
linux:/home/linux #

(In reply to comment #6)
> if i can do runlevel 3 on 11.3
> setup i might be able to properly give its confused interpretation of my
> arrangement

You can switch to console by ctrl-alt-f2 and do whatever you want.

Closing due to lack of response.

i got the latest LIVE CD and problem is still there (plus a bunch more so i couldnt install). i can provide additonal info that this time i am using the same two drives in an older computer with a silicon image PCI sata raid card. i dont try to install to these, my datas on them.


it worked fine on 11.2 and now it calls it isw_dhfiebjffi_2x200GB -- it is not 2x200GB. im going to find another OS or stick to 11.2. so even if you fix it im not sure if i want to continue with SUSE.

despite chanaging to the silicon image controller it is still the above ID. isw_dhfiebjffi_2x200GB. i did double check 11.2 install. it too had the problem. i think why i said it was ok was because i can use the data in the OS and after updating to a later version i could not even do this. now the drives are seperate, i do not know how to seperate them and stop this 2x200gb business. if the installer gets fixed the rest of the OS should as well.

evidently the bug must be in mapper as that is where this 2x200Gb is found.
last comment:....i do not know how to seperate them in linux and stop ....

about boot.msg
=======
A-when CD boots up i choose "install" in the bootloader... after it is done loading i press CTRL-ALT-F2 and log in, and go to VAR LOG and i type dir...
this boot.msg file you speak of is 0bytes. i dont know the command to open it.

B-so i go into the live boot of 12.1, im at the desktop but nothing is happening after CTRL-ALT-F2 due to incessant CDROM activity (no numlock or mouse). i reported this as a bug but was thrown out. 

C-i used the burning software to reverify the disk but reports no problems (there is one speck of a scratch)

D-at the boot loader i choose to verify the disk.
the lines that are appear after "kernel logging enabled on" are these:
waiting for  cd dvd devices to appear
checking cddvd devices
usage: checkmedia iso
check suse installation media
press enter to reboot
so i did thinking its not going to check (because it did not) and then it boots into the OS and i get the thrown out bug!!

if this was the reason why there is 12.2, i did want to download it but the website directed me to 12.1.

Q-therefore i quit.

i probably should write what happens in the os/live cd. what do i get instead of seeing the files as i did in 11.2? well i had the patience to sit around the computer and i see 12.1's dolphin now works as it did in ll.2 however i did get errors because things were timing out due to CDROM access.

the last post was done with a new cd. it still wont self verify as in comment 12, but, it performs "better" and actually does stop the continual access. so maybe one drive is better at reading scratches than the other. the injury was on the back side BTW.

oh! and i did look for var/log/boot.msg with the new disk and it is still 0 bytes. i am getting the feeling that maybe its "yet to be written". so how do i get this information now?

i have more information.
i placed the drives into a dell however i only had one R/A sata connector so i only connected one. the raid firmware was active and it said my raid failed, it called it 2x200gb.

so i will explain how i de-raided my drives. on the gigabyte mainboard, all i did was move the plugs from the onboard SATA raid to the onboard intel sata controller. deleted all the partitions and repartitioned. i didnt delete via the raid controller. i looked in the dell controller setup and it does have an option to de-raid but it warns of loosing data -- i did so and my data was still accessible.
---
so its up to you if you want to persue this. drives were set to raid0.

so mapper read the disk raid data and presumed disks were still raided - and appears to be prepared to perform software raid.

Yeah, RAID controllers typically warn of losing all your data but they usually don't. It's a "just in case something happens, you should be prepared" kind of warning.

The dirty little secret about these "RAID" controllers is that that they don't actually do anything other than have a bit of nvram, some basic firmware that can write some metadata to the disks, and *sometimes* may have something basic like an xor accelerator that is nearly always slower than the subsequent generation of CPU doing it itself.

The main reason they're "RAID" controllers rather than a regular controller with a software raid front-end is so that generic ATA/AHCI drivers don't claim them as separate disks. For Windows, I suppose the driver probably claims the disk, reads the metadata and sets up software raid. I don't know for sure.

For Linux, dmraid claims the devices after reading the metadata off the disks. I'm not a dmraid expert but I expect that the 'dmraid -x' will also remove the metadata from the disks without having to involve the firmware. That's also why when you moved it to a different type of controller, it still came up.

I'm having trouble parsing the situation you're in now. Did going through the RAID controller firmware succeed in un-RAIDing your disks?

yes i did succeed, and i can now access the drives individually in the setup.

Thanks for the feedback. I've used it to write up an SDB article to describe the issue. It still needs work, but it's a start.

http://en.opensuse.org/SDB:DMRAID