After extensively reading this forum and posting in various threads, I try to collect in this post what I have learned about a faulty Seagate 7200.14 HDDs so far:

- It has a low voltage TTL serial interface (1.8 V). A Silicon Labs CP2104 USB-UART converter works flawlessly. Alternative is a Prolific PL-2303H with a TTL level shifter (or modified accordingly).
- CTRL-Z gets me to the F3 T> command prompt. I found a list of commands for the F3 prompt of a 7200.12 HDD in this forum that I attach here. One user reported that the command library has not changed on the 7200.14, and another user started diagnosing the very same harddrive using these commands, so this encouraged me to do the same.

The only diagnostic commands that I used so far and that worked are (see screenshots of the terminal output attached):
Level T - V1, V3, V4, V40
Level 1 - N5
Level 2 - U, Z
Level 7 - X, D

I have not typed any other commands on the prompt yet.

The HDD reports extensive V1, V3 and V4 lists, but apparently no stored SMART data (all values are 0, except one that is 3). V40 gives an empty "non-resident G-List". X gives me the preamp head resistance, I can not assess if these values are healthy or not. D gives me a temperature of 35°C according to the resistance of the thermistor, neat but unuseful.

The HDD is not recognized by BIOS, which I do not understand. No strange sounds / clicking / etc.. Spins up / down just fine. No error messages on F3 prompt, terminal is responsive, does not hang up. If it is a bad sector problem why can I not get past BIOS?

One user in this forum (http://forum.hddguru.com/viewtopic.php?f=1&t=28447&p=203701#p203701) apparently recovered the very same HDD with the very same faulty behaviour by initializing / clearing the defect list and subsequently, cleared the SMART data:
F3T>i4,1,22
clear g-list
F3T>/1
F31>N1

The subsequent discussion criticises that this was very risky. If the non-resident G-list is not empty, a partial access problem can result. However in my case the non-resident G-list is empty as command V40 confirms. Is this an appropriate way to continue to try to get the disk recognized again by BIOS?

Another user in this forum reports that the command m0,5,1,10,3,,,22 has to be used on 7200.14 HDD. I understand from the command reference that this formats the hard drive, so this is not what I want .

The drive has an ext4 file system and my goal is to mount it again.

Thanks for your help. I hope this compilation helps others who are stuck with the same HDD series / model / situation.

This was a procedure for a 7200.12 found here: http://forum.hddguru.com/viewtopic.php?f=1&t=28523



I am hesitating to apply this to a 7200.14. Any comment would be appreciated. Thanks.

Read this first:
http://www.alexsoft.org/viewtopic.php?t=557

_________________
A backup a day keeps DR away.

hi frank.

link problems

It works for me.

Try http://malthus.mooo.com/viewtopic.php?t=557

_________________
A backup a day keeps DR away.

now works fine, but minutes ago not

@ fzabkar: Thanks for the link. However, I do not see InitiateMarkPendingReallocateRequest on the F3 prompt... Therefore I am not sure if disabling auto relocation helps...

@ hhddrec: I had lots of trouble in the past days to access the HDD Oracle forum... My firewall terminated the request due to a trojan on the webserver, or the server was not reachable at all... Seems to work now via the http://www.alexsoft.org URL...

hi Ghandi

sometimes can i open link and other times not

I had a look at the F command... Output is attached...

So as I understand, in my particular case / drive, it is byte 009A that stores RWRecoveryFlags.

So the appropriate command would be:
F009A,C1
to set disable_correction to 1.

So far so good. What would be the next step? Clone the device with MHDD? Could I maybe use Clonezilla?
Is there hope the device can be mounted in Linux without cloning?

Further investigating the topic, I found fzabkars post here:
http://www.alexsoft.org/viewtopic.php?t=694



Reading through the F output to find these flags right now...

Also very worth reading (presumably the same author):
http://lists.gnu.org/archive/html/bug-d ... 00002.html

Ok, I think I figured it out. To summarize, I would need to type:



Bit:0 needs to be set to 1, so:
1100 0000 = C0 changes to
1100 0001 = C1

Resulting command:
F009A,C1



Bit:0 needs to be set to 1, Bit:2 set to 0 and Bit:7 set to 1, so:
0001 0100 = 14 changes to
1001 0001 = 91
(As you can see, Bit:3 is not set. Using a '0' results in the correct hex code.)

Resulting command:
F00E2,91

What I do not understand yet:



This means, SpecialCacheFlags should be 1000 0000 = 80, not 90.
90 translates to
1001 0000
As Bit:4 has no use, it probably does not matter...

Considering that all bits except Bit:7 and Bit:5 are 0, I get the following solution:

Bit:5 needs to be set to 1, so:
1000 0000 = 80 changes to
1010 0000 = A0

Resulting command:
F00EC,A0

Comments are greatly appreciated!

MHDD does not have cloning function, at least not in the sense you need it.
Since it seems DIY is the route you intend, then clonezilla is not such a good option as the drive will more than likely struggle to some extent with bad sectors. Recommend practicing ddrescue on a non-critical drive. Also, can look into something like media tools. Your best chance of recovery is a hardware imager which is the type of equipment specialists use.

_________________
Hard Disk Drive, SSD, USB Drive and RAID Data Recovery Specialist in Massachusetts

I would try "F00EC,B0", ie leave bit 4 as is, but I don't have any way to test this.

That said, I'm not a data recovery professional, but I expect that your solution is in that other thread:
viewtopic.php?f=1&t=28447&p=203701#p203701

BTW, ddrescue is generally regarded as the best DIY cloning tool.

As for the Oracle, see viewtopic.php?t=29510&p=203640#p203640

If you need professional services, I understand that forum member bcometa offers a fixed price recovery, using a hardware imager, for US$300.

_________________
A backup a day keeps DR away.

Thanks for the hint. I have an Ubuntu machine and will have a look at ddrescue then.

I somehow understand the disk times out / does not get ready when BIOS attempts to detect it - still there are two possible explanations / solutions:
a) Because it encounters unreadable sectors during start up and times out - so applying the Fxxxx,yy commands allows to bypass the time out in BIOS.
b) Because it takes the drive simply too long to go through the quite long V list of marked sectors during its initialization, so emptying the V lists using the i command and simultaneously applying the F commands (so the V list is not filling up again) seems to be the way.

Summarizing b), the correct command order would be



No m command e.g. for rebuilding translator etc....! Then power down, power up; hope that the drive is recognized by BIOS, mirror the drive with ddrescue, then look at the mirror for data recovery. I am hesitating to mount the original drive directly in Linux for data access (might not even work if one of the heads is bad - can somebody assess the resistance values in the screenshot above?).

Non-resident G-list is empty as confirmed by V40 command so the i4,1,22 command might be okay as in the other example here in the forum, but clearing SMART data wit N1 somehow does not appear meaningful to me as it is already empty...

BTW Dell sent us the same drive as a replacement. Given that the machine it was in runs its OS from a SSD and only accesses data on the HDD quite rarely (<40% of a work day), it is somehow embarrassing it died after roughly 12 months of operation. Of course you can pay extra for Seagate's data recovery service, if theeir own drive does not last even a year! Too bad we had quite new data on it that was not yet backed up.

Ok - I can get past BIOS now, although with some significant delay.
Testdisk shows me correctly a swap and an extended Linux partition. The logic partition inside the extended partition is not found, though. Mounting with backup superblock etc. not successful (well, how should it be, if logic partition does not get even discovered...).
I did a 3min file recovery attempt (total scan time was given with 1200h +) with photorec searching for text files and indeed recovered a text file that was on the disk. So something is indeed recoverable...

Guess I should attempt a mirror now with ddrescue to not use the device unnecessarily.

Ok - got a 3 TB HDD for cloning today.

I am running now ddrescue 1.18.1 as follows:


as my initial command

resulted in a readspeed of about 500 KB/s now, and I doubt the device will survive another 42 days...

After this I will try


Comments as always appreciated... Thanks!

Just to close this thread:
The recovery of the hard drive was partly successful. I played around with ddrescue's parameters for more efficient reading and while this caused larger gaps / jumps of apparently undreadable zones, I could successfully mirror about 60% of the device before it completely died. With this I mean that I could no longer read anything, although I still could enter the terminal. I managed to scan for file header signatures in the mirrored drive and recover exactly what I was looking for. So for my part, the recovery effort was a success, although I could only do a quite incomplete mirror of the drive. Aside, I learned quite a lot with this exercise (of course, also simple things like do to backups more often).

I hope this thread will be useful for other readers having the same problem.

I believe the lesson to be taken away is what everyone on this forum keeps telling the newer members: "if your data is important, take it to a professional with proper tools"

We aren't just saying this just to drum up work, we keep saying it because we all see too many DIY gone catastrophically wrong cases. At least you got 60% by taking the advice to use ddrescue (59% more than other programs would have read) but you could have had 99.9% instead.

_________________
Data Medics - Hard Drive, SSD, and RAID Data Recovery Service Company

Hello,
end of 2012 I ordered a Seagate 3TB USB 3.0 drive. Since then it was only used in average once a month for transfer
backup because internal drive is to small.

Last weekend my disaster happend, in middle of copying one file the HD stopped working, complete silience like it is of.
I unplugged it and plugged in again, computer wants to format disc to use it again.
I canceled and again plugged it out and in. After that I got the "click of death" sound from Barracuda 7200.11.

I ordered a CP2102 USB-UART converter and started a try to fix it after a manual for 7200.11 fix.
After connecting TX-TX and RX-RX I got messages in putty but I couldn't send CTRL+Z.

In this topic I found out that I need another converter with CP2104 chip and that I can't use the 7200.11 fix.
I ordered now a new converter and hope that I can fix the drive for some days to copy the data to my new buyed NAS.

So my question now, is there in the meantime a good way to get the drive back to life?
Otherwise I will give the "clear g-list" a try. The data is not important but it is annoying to lost them.

Thanks
Kai

Finished a head swap on ST2000DM001 yesterday. Imaged 99.9%. However after the swap before getting user area access i was forced to clear alt list ( V4 command gave TONS of hits and ended up spamming DIAG ERROR on the loop ) and force stop realocations and any idle activity. Used PC3K to do that ( however changes to sys files had to be done through terminal since for whatever reason they did not stick if i tried to apply them through UI )

With original heads the drive clicked with no ID and spun down. With donor heads in it would ID but go busy indefinitely after a split second of functionality. I had to spam CTR Z to get terminal access as soon as the drive IDs to get into terminal before busy raises.