Harddisk detection failure

[sas0209]

Newbie Question: My harddisk is not detected properly by the bios. Is there any way to read the data anyway under Linux?? tx for any kind of help

[Stephen]

Newbie Question: My harddisk is not detected properly by the bios. Is there any way to read the data anyway under Linux?? tx for any kind of help

Knoppix (linux in general) does not use the BIOS for information on the HD you will be able to use the entire drive. My old computer had a 32gb limit on the HD and I had to jumper a 40gb HD for it even to boot however linux saw the extra 8gb and I was always able to use the full drive.

[sas0209]

Thanks for your explanation. Could you add a short description, how to get this harddisk kind of installed in Linux. I'm normally Windows user, and so I've got no idea how to handle Linux. I just tried to call the Harddisk partition information with "gpart -i /dev/hda" with the disk mounted on the Primary Master. But I didn't succeed. I would be pleased about any further help. TX

[Stephen]

Ok a few more details would be required. Is this a linux only machine or are dual booting windows? Did you have to use software to fool the BIOS into thinking it was a smaller drive for it to boot windows. What type of machine/HD? As many details as possible on what you are trying to accomplish would be helpful.

Something you can try when booted from the Knoppix CD open a console window (the clam shell icon on the taskbar) and sudo fdisk -l /dev/hda to see how the drive is partitioned also post that here as well. To see how knoppix detects the drive also from a console window dmesg | grep hda to search the boot messages for references to the hda post this also.

When I installed the HD (Maxtor 40gb) that I had in my old machine I had to set a jumper on the back of the drive to make the drive itself report that it only a capacity of 32gb therefore fooling the BIOS for the machine to boot and that is all windows would see however in linux there is IIRC an option in the kernel that enables large block devices support or something like that, that allows it to see the full capacity of the drive it does not need the BIOS it communicates directly with the drive itself and as such any of the disk tools should be able to see the full capacity of the drive.

[edit]
Another question are you using gpart to try and find out what the partitions are on the HD? That is what that tool is used for making a guess at the partition formatting on a drive that has corrupted partition tables.
[/edit]

[sas0209]

Well, the harddisk is a 40GB fujitsu IDE, originally mounted in a Win98 Domain Client PC, only one partition, FAT 32. One morning when the user tried to start it, the PC didn't recognize the Harddisk anymore, ergo it didn't boot no more. I tried to get access to the Disk with several tools like Partition Magic (Power Quest), but I didn't succeed. Then somebody gave me the advice to use Knoppix, because it doesn't need any harddisk and should have the ability to read out the data under very difficult circumstances. Fine, I thought, but as I have never used any Linux before, I'm just playin around and don't really know what to do.

By the way I don't think its a problem of a damaged partition or MBR(so gpart might be the wrong tool), I think the hardware interface of the disk is somehow damaged, although you can't see any defects visually. Could you describe me the command to get Linux recognize a harddisk plugged on a certain bus?

Here I tried the command you guessed:


knoppix@ttyp0[knoppix]$
dmesg|grep hdd
Kernel command line: lang=de ramdisk_size=100000 init=/etc/init noapic apm=power-off hda=scsi hdb=scsi hdc=scsi hdd=scsi hde=scsi hdf=scsi hdg=scsi hdh=scsi vga=791 initrd=miniroot.gz quiet BOOT_IMAGE=knoppix BOOT_IMAGE=vmlinuz
ide_setup: hdd=scsi
ide1: BM-DMA at 0xe008-0xe00f, BIOS settings: hdcMA, hddio

knoppix@ttyp0[knoppix]$
dmesg|grep hda
Kernel command line: lang=de ramdisk_size=100000 init=/etc/init noapic apm=power-off hda=scsi hdb=scsi hdc=scsi hdd=scsi hde=scsi hdf=scsi hdg=scsi hdh=scsi vga=791 initrd=miniroot.gz quiet BOOT_IMAGE=knoppix BOOT_IMAGE=vmlinuz
ide_setup: hda=scsi
ide0: BM-DMA at 0xe000-0xe007, BIOS settings: hdaMA, hdbio
hda: FUJITSU MPG3204AH E, ATA DISK drive
hda: 40031712 sectors (20496 MB) w/2048KiB Cache, CHS=2491/255/63
hda: hda1

knoppix@ttyp0[knoppix]$
dmesg|grep hdc
Kernel command line: lang=de ramdisk_size=100000 init=/etc/init noapic apm=power-off hda=scsi hdb=scsi hdc=scsi hdd=scsi hde=scsi hdf=scsi hdg=scsi hdh=scsi vga=791 initrd=miniroot.gz quiet BOOT_IMAGE=knoppix BOOT_IMAGE=vmlinuz
ide_setup: hdc=scsi
ide1: BM-DMA at 0xe008-0xe00f, BIOS settings: hdcMA, hddio
hdc: CD-ROM 52X/AKH, ATAPI CD/DVD-ROM drive
ide-cd: passing drive hdc to ide-scsi emulation.

knoppix@ttyp0[knoppix]$
dmesg|grep hdd
Kernel command line: lang=de ramdisk_size=100000 init=/etc/init noapic apm=power-off hda=scsi hdb=scsi hdc=scsi hdd=scsi hde=scsi hdf=scsi hdg=scsi hdh=scsi vga=791 initrd=miniroot.gz quiet BOOT_IMAGE=knoppix BOOT_IMAGE=vmlinuz
ide_setup: hdd=scsi
ide1: BM-DMA at 0xe008-0xe00f, BIOS settings: hdcMA, hddio
knoppix@ttyp0[knoppix]$


I had a normal running Fujitsu Harddisk on the Primary Master. On the secondary slave I plugged the damaged one. as Linux didn't even recognize it (see section where "dmesg|grep hdd") I think there is no possibility to recover the data. If you've any idea, just post it. Anyway, thank you very much for your help. =) Stefan Schwarz

[Stephen]

I believe you are going to need a hardware recovery specialist fujitsu put out some really bad drives for a period and the one you speak of is know by my one of my friends in the hardware business as the drive of death due to its almost guaranteed failure rate. The controler chips on the drive have defects in the chipset at the time of manufacture and as such will fail you are going to have to find somewhere to have the drive disassembled and the data recovered by someone in that business. Also you should be contacting fujitsu and requesting a full refund of your purchase and recovery costs and if necessary taking them to court for having put out such a bad piece of hardware in the first place.

If you have any more drives of this type you should peform backups immediately you never know when the next failure will happen and they should be replaced at fujistu's expense with a drive from another manufacturer without these problems.

I just called my friend to see what they do when these drives fail he says the drives with the model numbers starting with mpg and some with the mpf are the drives that have these problems and fujistu does issue refunds at least here in Canada for a period of 3 years 3months from date of purchase even with no reciept.

He has a couple of procedures that he uses to get the information off the drives that you may want to try.

They both involve using a controller board from a similar drive by fujistu they do not have to have the same formatted capacity just identical controller boards eg. the screw holes must match up and have the same type of connectors. For the first procedure you would have to remove the controller board from the known good drive and put it on the drive with the bad controller and it will work to recover the data and you could continue to use the drive with the replacement controller.

The second is the same as the first procedure except if the bad drive is making clicking sounds after the good controller has been put on that indicates mechinical failure in addition to the controller failure he recommends putting the drive in a airtight freezer bag and putting it in the freezer for an hour before beginning the next step of data recovery. You only get one chance at recovering the data with this procedure so you must have already jumpered the drive you will be recovering the data from for a slave drive before putting it in the freezer. Next you would have a machine that you have removed the side of the case from and have access to the drive cable and power cable ready to be plugged in for a slave drive with a primary drive with enough spare capacity to copy the data to. Then it is a matter of connecting the slave drive after having taken it out of the freezer and coping as much data as you can before the drive actually fails.

I don't know if either option above appeals to you or not that is what they use when everything else fails whether you wish to attempt this is entirely up to you and your comfort level with taking a controller board off a working drive to attempt this the alternative as I stated above is a hardware data recovery specialist. Good luck hopefully you can get your data recovered.

NOTE!!!! IN re-reading your post I see the HDA in the system you plugged the bad drive in is a drive with the MPG model number and is one of the know bad drives of this type it is extremely important to get the data off this drive ASAP.

I CAN'T STRESS THIS ENOUGH BACKUP THE DATA NOW and as it has a working controller for now you may have nothing to lose by taking the controller board off it to use for recovery of the other drive if they are compatible.

[sas0209]

Well, after this nice news in your last post I think I'll double my caution in handling these drives. You have to know I'm network administrator in a small corporation, that has bought five or six Compaq Presarios. And now guess what kind of harddisk they've mounted????
But to see it from the positive side: I won't get out of work for the next time. Thank you very much for your help with my little problem in Linux. If I ever have a problem again, I will certainly post here again. Have a nice time, Stefan Schwarz