Promise TX4000- two striped drives a goner or....

[fotoobscura]

Hi.

I have a Promise TX4000 and had a drive that went dark (intentionally,
unplugged) for a few months. One day I went into the bios and noticed
I only had one drive and remember I unplugged it. I had also since
removed the drive.

Here's the dumb part:

Forgetting this wasn't a mirrored array for some reason I "initialized"
or maybe better put "made online" but did NOT format one of the two
drives in the stripe. I did *not* do anything, at least knowingly that
would hurt anything. No to all "are you sure you want to do this?"
sort of thing.

I then realised the err in my ways and plugged the second drive in. I
went back into the Promise BIOS and noticed both drives recognized as
striped and the right size. I booted into Windows and went to computer
manager to "mount" it. It was there but was acting as if both drives
were empty and unformatted.

This means, obviously, I did something stupid.

The question is- the data on both drives is technically still there but
of course written across separate disks in 64byte blocks. Is there a
way to "save" this stripe or rebuild the MBR for the stripe so I can
get this data back?

It seems to me that even a program like Ontrack to recover data (raw,
formatted, etc) may not work if it doesn't recognize the stripe.
Without recognizing the stripe I presume Ontrack will not know that
64bytes is on one drive, 64 on the other (or so I believe the striping
works).

I have a feeling that i'm not SOL providing I can re-create the stripe.
It seems to me that because the striping is a reoccurring process that
you can infer the stripe that needs to "span" the files on the drive.

So, before I wipe the two drives and make a new stripe and let things
fall as they may, can anyone suggest a way to recover this lost stripe?

Thanks for all info.

p.s. and no I don't have a backup but its not end-of-the-world if I
lose the data. (hence the stripe in the first place)

 

[Svend Olaf Mikkelsen]

Hi.

I have a Promise TX4000 and had a drive that went dark (intentionally,
unplugged) for a few months. One day I went into the bios and noticed
I only had one drive and remember I unplugged it. I had also since
removed the drive.

Here's the dumb part:

Forgetting this wasn't a mirrored array for some reason I "initialized"
or maybe better put "made online" but did NOT format one of the two
drives in the stripe. I did *not* do anything, at least knowingly that
would hurt anything. No to all "are you sure you want to do this?"
sort of thing.

I then realised the err in my ways and plugged the second drive in. I
went back into the Promise BIOS and noticed both drives recognized as
striped and the right size. I booted into Windows and went to computer
manager to "mount" it. It was there but was acting as if both drives
were empty and unformatted.

This means, obviously, I did something stupid.

The question is- the data on both drives is technically still there but
of course written across separate disks in 64byte blocks. Is there a
way to "save" this stripe or rebuild the MBR for the stripe so I can
get this data back?

It seems to me that even a program like Ontrack to recover data (raw,
formatted, etc) may not work if it doesn't recognize the stripe.
Without recognizing the stripe I presume Ontrack will not know that
64bytes is on one drive, 64 on the other (or so I believe the striping
works).

I have a feeling that i'm not SOL providing I can re-create the stripe.
It seems to me that because the striping is a reoccurring process that
you can infer the stripe that needs to "span" the files on the drive.

So, before I wipe the two drives and make a new stripe and let things
fall as they may, can anyone suggest a way to recover this lost stripe?

Thanks for all info.

p.s. and no I don't have a backup but its not end-of-the-world if I
lose the data. (hence the stripe in the first place)

I do not understand in details what happened, but recovery software
exist that will simulate a stripe set, and copy the data.

You can get Findpart at

http://www.partitionsupport.com/utilities.htm

do:

findpart all fp.txt

and mail me the file fp.txt.


If as example two disks in a stripe set are visible as disk 2 and 3,
and the stripe size is 2^6 KB (64 KB), the Findpart command to search
the set would be:

findpart R236 fp-a.txt

 

[fotoobscura]

Thanks Svend! I didn't realise you wrote the program

I have located the "disk"..here is output..when its done looking for
partitions (so far nothing) i'll write back with what you wanted me to
email you...I think I may be without luck but i'm hoping

This is my "one" disk (2 drives)

Disk: 3 Cylinders: 30034 Heads: 255 Sectors: 63 MB: 235593

--PCyl N ID -----Rel -----Num ---MB --Start CHS- ---End CHS-- BS CHS
0 1 06 63482496147235593 0 1 1 30033*254 63 00 OK

This is output of findpart 3

C:\_temp>findpart 3

Findpart, version 4.71 - for Windows 95/98/ME/NT/2000/XP.
Copyright Svend Olaf Mikkelsen, 1999-2006.

OS: Windows 5.1.2600 Service Pack 2

Disk: 3 Cylinders: 30034 Heads: 255 Sectors: 63 MB: 235593

None found.



Still waiting for it to end...its been a while...i'm about to break the
program in about 20 minutes if it doesn't exit on it own.

Thanks!

 

[Svend Olaf Mikkelsen]

Thanks Svend! I didn't realise you wrote the program

I have located the "disk"..here is output..when its done looking for
partitions (so far nothing) i'll write back with what you wanted me to
email you...I think I may be without luck but i'm hoping

This is my "one" disk (2 drives)

Disk: 3 Cylinders: 30034 Heads: 255 Sectors: 63 MB: 235593

--PCyl N ID -----Rel -----Num ---MB --Start CHS- ---End CHS-- BS CHS
0 1 06 63482496147235593 0 1 1 30033*254 63 00 OK

This is output of findpart 3

C:\_temp>findpart 3

Findpart, version 4.71 - for Windows 95/98/ME/NT/2000/XP.
Copyright Svend Olaf Mikkelsen, 1999-2006.

OS: Windows 5.1.2600 Service Pack 2

Disk: 3 Cylinders: 30034 Heads: 255 Sectors: 63 MB: 235593

None found.



Still waiting for it to end...its been a while...i'm about to break the
program in about 20 minutes if it doesn't exit on it own.

Thanks!

If the search takes too long to finish, you in stead can do:

findpart tables fp-a.txt

and mail me the file fp-a.txt as attached file.

 

[fotoobscura]

Svend,

Ok potentially bad news.

I removed the drives from the array and plugged them in separately with
a ide/usb connector and one drive isn't spinning up. I hope its a
fluke but I doubt it. I think the array correctly is interpreting the
second drive but it isn't testing that it spins up. I suspect I may be
SOL. Unless of course for some very odd reason that one drive won't
spin up with the IDE/USB connector i'm using (odd, same drives).

These are indeed IBM 120gb drives which aren't exactly workhorses....

I may have to call quits on this one

 

[Svend Olaf Mikkelsen]

Svend,

Ok potentially bad news.

I removed the drives from the array and plugged them in separately with
a ide/usb connector and one drive isn't spinning up. I hope its a
fluke but I doubt it. I think the array correctly is interpreting the
second drive but it isn't testing that it spins up. I suspect I may be
SOL. Unless of course for some very odd reason that one drive won't
spin up with the IDE/USB connector i'm using (odd, same drives).

These are indeed IBM 120gb drives which aren't exactly workhorses....

I may have to call quits on this one

In mail I had:

Disk: 3 Cylinders: 30034 Heads: 255 Sectors: 63 MB: 235593

Start cylinder: 30030 End cylinder: 30033 Index records not shown.

--------- CHS ----- LBA T -Record Cluster Name
30033 252 61 482496081 Boot or backup
Sectors per cluster: 8
MFT cluster: 786432
MFT mirror cluster: 30156009
Partition sectors: 482496147

Sector 482496081 for the backup boot sector is 64 KB lower than
expected, meaning that the disks in the array were in the wrong order.

Assuming a 64 KB stripe size (and sector size 512 bytes)

482496081 mod 256 = 81

meaning that sector 482496081 is on the same disk as the MBR we have
seen, so there is no proof that the second disk was actually read by
Findpart.

I would expect the TX4000 BIOS and drivers to read some stripe set
data at the end of the disks, so I then do not understand how the
235593 MB disk was reported.

Theoretically, IBM disks has a "no spinup" feature that can be set and
removed.

 

[fotoobscura]

So after all this does this mean i'm out of luck or its still worth
trying?

I'll do whatever it takes- but the fact the the drive doesn't spinup at
all (and I haven't done anything to cause this) I suspect its a goner?

 

[Svend Olaf Mikkelsen]

So after all this does this mean i'm out of luck or its still worth
trying?

I'll do whatever it takes- but the fact the the drive doesn't spinup at
all (and I haven't done anything to cause this) I suspect its a goner?

I still do not understand how the stripe set was reported, if one of
the disk did not spinup. If a disk does not spinup, I cannot help.


Aside from that:

It may be a very small chance, but as mentioned some IBM disk models
can be set to "no spinup". The purpose is to avoid too much power use
at system power on.

If a parallel ATA disk is set to no spinup, the Findpart for DOS
commands to remove the setting would be (example for primary slave):

set findpart=edit
findpart feature ps spinup


or to spinup the disk at once:

set findpart=edit
findpart feature ps spinupnow


It is required that a BIOS is present, that does not hang if a disk
does not spinup.