dslreports logo
 
    All Forums Hot Topics Gallery
spc
Search similar:


uniqs
3293

norwegian
Premium Member
join:2005-02-15
Outback

norwegian

Premium Member

Toshiba MK7575GSX 750gb HDD error count

Click for full size
This one is sitting here to be returned to a working order for a friend.

I have extracted the personal data finally.
The computer has had a standard return to factory default done via the Toshiba recovery tool supplied on one of the partitions. I have not bothered with the wipe/erase function on the HDD because I was hoping it was just a problem of partition table corruption, but now I've tried updating everything found some really serious issues you do not expect of a fresh install.

I ran the recovery tool, CHKDSK and SFC tools were used via a recovery disk and the CMD prompt, but errors suggested there was serious problems. The image is the first test results of SMART since getting it up and running. I'm imagining a new HDD is best, but can't help asking questions:
1. Could this be repaired via a full write to the HDD?
2. What damages a relatively new HDD to this extent - if I'm understanding there is indeed a larger than normal error count?

Thanks
norwegian

norwegian

Premium Member

I've gone off and started the erase with the option to write to every sector. Attached is a text file of the smart log above and the specific LBA's showing as "errors".

I will look at a log of HD Tune once the HDD is written to and check the sector count. Maybe I'll get lucky. But have big doubts of that possibility.

koitsu
MVM
join:2002-07-16
Mountain View, CA
Humax BGW320-500

2 edits

koitsu to norwegian

MVM

to norwegian
Going off of the screenshots:

No, the drive has already remapped a total of 2600 LBAs to spare sectors; all that data may have been lost (512*2600=1,331,200 bytes of data).

There are also an additional 3624 LBAs which are "suspect" (unreadable), regardless if those are determined (via writes) as truly bad or not. So, you have another 3624*512=1,855,488 bytes of data which has been lost as well. Edit: .txt file shows 3704, so make that 3704*512=1,896,448 bytes)

If this is a laptop hard disk, it looks like it's been heavily jostled around/bumped/aggressively placed on a desk while powered on a total of 65 times. If the owner of this device tends to operate like this, they should invest in an SSD.

A problem of this sort, given the large number of LBAs in sequential order, combined with a small power-on hours count (1093), could be caused by physical damage sustained during operation, could be the result of manufacturing mistakes, or could be the result of the platters themselves or the magnetic substrate flaking off or being sub-par-quality at specific areas on the platters. None of these things are fixable via software, nor via a data recovery company (especially the latter two).

1. Get whatever data you can off the drive using standard file copy operations and hope for the best. If the user has backups that were done when the drive was known to be in good condition (no remaps/issues) then use that instead.

I do not recommend using a sector-to-sector or LBA-to-LBA copying program in this situation; the advantage standard file copy operations give you is that if the file uses an LBA that cannot be read ("suspect" LBA), you'll get an I/O error telling you that -- this allows you to determine which files are impacted by the remaining "pending sector" count. Those are files which should be restored from backups, if applicable.

2. RMA the drive. It should not be used past this point. It's in very bad condition. You can zero if it you want (learning experience), but it should not be re-used.

3. Reinstall the OS or reimage the system from scratch. Do not try to "repair" the existing setup; there may be system files (DLLs, SYS files, etc.) that contain zeros, and like I said, there's no way to determine that.

norwegian
Premium Member
join:2005-02-15
Outback

norwegian

Premium Member


Thank you for the feedback. A few things:

1. G-Sense count of 65 relates to 'knocks' or 'bumps'?
I will talk about the SSD drive if that data is logged. That info will be a great way to "teach" them about laptops.
2. I had feedback to suggest one of the kids jumped on the computer that caused problems, your comments on the LBA's in order and what would cause this may highlight the heavy knock it may have incurred.
3. Personal data is already in a file on my desktop, there is nothing left to worry about with this as far as data goes.
4. So writing to every LBA then is out of the question. I will tell them to invest in a new HDD.
5. That unsafe shut down count of 529989704 seems excessively high? The user has been running this laptop without the battery and a flaky internal power plug, but a count that high? Surely that must not be a correct count?

Thanks again, I will organize a new HDD and install the O/S on that for them and hope they do not throw it about to create more work for me, I'll have to start charging.

koitsu
MVM
join:2002-07-16
Mountain View, CA
Humax BGW320-500

koitsu

MVM

1. Many 2.5" MHDDs have a shock sensor installed in them; that's what this attribute tracks. The sensitivity of this sensor varies; I've seen some environments where someone very lightly placing their laptop on a table (while powered on) caused this sensor to increment, while another laptop (same model/innards) treated identically had no effect. It's not a failsafe indicator that someone is being rough with their hardware, but it can be an indicator.

Let's put it this way: I have a Toshiba MK5055GSX in an enclosure which has been used off and on (power-on hours count of 179) over the past 3-4 years. Its attribute has a value of... 2. When I received it, the attribute had a value of... 1 (probably a result of factory testing).

2. It sounds like one or more of the heads may have made contact with the platter, causing slight damage. The two LBA ranges seen so far are sequential, and may be related to the same physical area on the platter -- it's impossible for me to tell at this level.

5. The count is correct; the error is in that you're reading the "Data" attribute wrong (you're treating it like a literal decimal number). SMART attributes can be vendor-encoded; T13/ATA specification does not define the format of the 6-byte data section. This is one such attribute. HD Tune Pro does not have the intelligence to decode this attribute, so all you can go off of is the adjusted value ("Current" in HD Tune Pro) and compare that to "Threshold" and "Worst". Other tools (like smartmontools) may be able to. If you want me to expand more on this I can, including providing screen shots and proof. My aforementioned Toshiba drive is a great example.

norwegian
Premium Member
join:2005-02-15
Outback

1 recommendation

norwegian

Premium Member


I don't think you need to expand on it thanks. I have enough info to know which path has to be taken.

If I have a play with the HDD after the full wipe I may have a question or 2, however I'm doubting I really have the time at present, with school for the kids coming back online there are other priorities.

Thanks for passing on your knowledge to help confirm my theories.
norwegian

norwegian to koitsu

Premium Member

to koitsu

FYI as a wind up.

A new HDD is on the way, just need to check prices tomorrow for SSD versus platter.

However for the end result of the wipe and write to every sector, the tool supplied on the laptop, this error occurred:

Toshiba Recovery Wizard

An error has occurred.
ERROR: 20-FFFE-0017
Data deletion error.
Please press [OK] to turn off the computer.

I had 4 hours left of a 31 hr process when the error happened.
Move on with new hardware.