Dead SCSI devices / Media test Failure
 

Hi,

I have a problem that I am not quite able to understand...

I have a machine that has 3 SCSI hard drives... There was a power outage and since then it is not loading my Linux. I get the following error messages while booting:

"SCSI device at channel=0 Target=1 not responding!
WARNING: 2 system devices offline
WARNING: Dead SCSI devices (Channel:Target): 0:0 0:2
...
Initializing Intel Boot Agent Version 2.2...
Primary Master Drive - ATAPI Incompatible ... Press F1 to resume..."

After pressing F1, it gave me a table with all the system configurations and it again gave the following message:

"BSP CPU... Microcode OK
AP CPU... Microcode OK
Searching for Boot Record from Floppy... not found....
PXE-E61: Media test faliure, check cable"

Can anyone please let me know what might be the problem? I do not know much about SCSI drives.... But in the BIOS, I saw the Primary master to be set as CD ROM and all the other 3 SCSI HDDs are in pri-slave, sec-master and sec-slave...

Are the SCSI hard disks dead?

Thanks a lot for reading this big mail....
Manudath

Adaptec scsi controllers (the only ones i've had any experience with) have an inbuilt bios type utility that appears after the motherboard bios (usually by pressing cntrl+A or some other key combo) .

Do you have something similar which you can check the drives with?

I am not sure what my controller is since I havent worked with SCSI before, and this system is not working now...

But, I found the hard drives were "Quantum Atlas 3 series".

Also, during bootup, in the beginning, I can see that the system BIOS is AMIBIOS...
After checking for memory etc., the screen shows up "DAC960 BIOS version 4.10-38 Mylex Corporation"... and then displays the error SCSI device not responding...

When I get the "Press <Alt-M> for BIOS options, I get this message:
"2 SCSI drives installed"
If I do not press <Alt-M>, I get a message
"<DAC960 8GB Bisk BIOS is enabled>

Is this of any help ?

Thanks a lot,
Manu

FYI SCSI drives do not have a master / slave configuration. Each device on a channel has a number. Depending on the controller type anywhere from 0-15.

It appears that two of the drives have failed. What distribution are you running? The SCSI controller BIOS only sees two drives so one of the three is dead which might be affecting the other two.

Sounds like one of the drives is screwed up during the outage. Scsi drives do not like power off during a write especially if one is low level formating is even worst. They can be revived but can take a while and special software tools. Haven't seen any out for a long time. I had one do this to me a long time ago. It took 15 hours to get it back working. It was only a 2 gig drive.

Termination can be bad or set wrong on the end of the chain drive.

Try using one drive at a time till you find the bad one. Mark which one is 1, 2, & 3rd drive so as not to confuse the orginal order.

Brian1
" Google the Linux way @ http://www.google.com/linux "

- You shouldn't need to worry about the physical location of the drives, i.e. which connector on the cable they're connected to, etc.. All the adapter will care about is the SCSI ID of the devices. Unless one starts changing the SCSI ID jumpers on the drives -- which I wouldn't do until after I'd written the original SCSI ID on a piece of tape affixed to the top of the disk -- it's not a problem if the disks are reattached in a different order than they were originally.

- There's no requirement that SCSI IDs be sequential (or even start with "0" for that matter) so altering the SCSI ID shouldn't be necessary. Of course, if you're booting off the SCSI bus, there's probably an ID=0 disk out there since that's the most cmmon setup for booting. Adaptecs will let you specify the boot target but I'm not sure of about the Mylex adapters (I actually have a used one of those cards but I think the BIOS is too old for use with Linux :-( ).

- You ought to be able to attach a single drive at a time and see if the adapter's onboard tools can still see the drive. If there isn't an onboard set of tools, you may need to just watch for the messages on the monitor as the adapter initializes and reports on the attached devices. Don't blink; sometimes these messages go by quickly, especially if you only have a single device on the bus.

- Make sure to power down in between checking each drive. I.e., don't pull the cable off and attach another drive while the system's powered up.

- Hopefully the first drive only had the OS on it. You could then, at least, replace the drive, reinstall the OS and your data should still be out on the other disks.

- It is possible that problem is merely that the drive's electronics board is kaput. (It happens more often than I'd like to think is possible.) If this is the case, the contents of the first drive might actually be intact but if it's just that the drive's electronics are toast. Unfortunately, if you don't have a chunk of money to spend having the board replaced (and I'm not even sure if that's even done anymore with the throwaway mindset in computers) the data may as well be considered lost.

- I'd caution the original poster to be really careful running the adapter's onboard tests. They may do destructive write testing and if you don't have a recent backup... it'll absolutely ruin your day.

After some struggle, I managed to get a new SCSI adapter and put it in my machine... and then when I booted, it did not give any errors...

But, after all the boot checks, it halted after showing this message:

SCSI BIOS installed successfully!
.
.
.
Searching for boot record from Floppy... Not found
Searching for boot record from SCSI... OK

and halts here without any more messages/action...


Later, I used knoppix CD to boot and then tried mounting the scsi hard drives to check if data might be present in the hard drives... Unfortunately, I was not able to mount it at all... I get the error:
mount: wrong fs type, bad option, bad superblock on /dev/sda1 or too many mounted file systems...
I tried giving several different options in /etc/fstab, but none worked.

Can anyone suggest any method in which I can check if there is any data in any of the 3 hard drives?

Thanks a lot for your suggestions...
Manudath

I downloaded Debian and tried doing a 'rescue' operation on it... It is not able to see any of the hard disks at all...

But the BIOS shows 3 SCSI hard disks... What might be the reason why only BIOS is able to see the hard disks, but not the installation files? I am a little concerned about the data in those disks... Is there any other way I can get to the data?

Thanks,
Manudath