Reallocation sector count dramatically increased

[djinn]

Well, finally I've decided to perform a secure erase on this drive. My question is if I can use the "enhanced procedure" or it's the "standard procedure" safer than the enhanced. Can "enhanced erase" kill the drive? (for example, firmware corruption in the service area). This drive supports enhanced erase. I want to perform the best erase to reinitialize the drive in order to mitigate current problems with bad and slow sectors.

I see MHDD does not run the enhanced method, but hdparm running under Parted Magic environment does.
Please reply.

Thank you for your patience.

[BlackST]

Must be a new technology : "enhanced erase"

Anyway, to cut a long story short, it should not kill the drive but won't fix your problem. If the drive gets killed , it's because of something else.
"Security erase" is an internal feature that is started from standard ATA level and is performed autonomously and internally. Some drives act on some way, other other way. Honestly, I don't know what this "enhanced method" is, sorry.

P.S. even if it APPARENTLY will fix your problem, expect for it to be back with a vengeance in some time.

[Vulcan]

@BlackST,

Completely agreed with your comments. Just FYI:

Honestly, I don't know what this "enhanced method" is, sorry.

For example see ATA-8 standard, section 7.43.2 - Security Erase Unit command - for some relevant explanation (some text removed for brevity):

When Normal Erase mode is specified, the SECURITY ERASE UNIT command shall replace the contents of LBA 0 to the LBA reported by the larger of READ NATIVE MAX or READ NATIVE MAX EXT with all binary zeroes or all binary ones.
[...]
The Enhanced Erase mode is optional. [...] When Enhanced Erase mode is specified, the device shall write vendor specific data patterns from LBA 0 to the Maximum LBA reported in DEVICE CONFIGURATION IDENTIFY data words 3-6. In Enhanced Erase mode, all previously written user data shall be overwritten, including sectors that are no longer in use due to reallocation.

Hope that info helps to explain the differences between "normal" and "enhanced" Secure Erase, according to the ATA standard (i.e. what is used to overwite the user LBAs, whether the reallocated sectors are also overwritten etc.). However not all drives strictly comply, in my experience. As you say, ultimately, the behaviour is a drive firmware decision.

[BlackST]

What's worst : not all drives DO perform correctly the SE procedure

I intended "don't know what HDDPARM "ENHANCED ERASE" is.

[Vulcan]

What's worst : not all drives DO perform correctly the SE procedure

Agreed.

I intended "don't know what HDDPARM "ENHANCED ERASE" is.

It is the hdparm option to select the Enhanced Erase bit (instead of Normal Erase) when sending the Secure Erase Unit command to the drive. Perhaps we are having a communication problem?

[djinn]

Ok, then either the normal erase or the enhanced will not harm my drive. It's good to know.

Thanks for the fast replies, much appreciated.

[Vulcan]

Ok, then either the normal erase or the enhanced will not harm my drive.

Personally I would not say that - your phrase "will not harm" is too definite IMHO.

Personally I would say: Secure Erase should not harm a normal, working drive. The result on any other type of drive (like yours) is unknown. (And, of course, there could be latent fault on any drive, which means that no drive is actually known to be definitely working!)

For example: I have seen faulty drives which started a Secure Erase and never finished (even after running overnight) - after resetting, they were not accessible via SATA at all (to be clear, I don't mean that they were password locked - there was no response at all, invisible to BIOS etc.). However the drives were faulty, so we knew that anything could happen...

[Edited to clarify.]

Last edited by Vulcan on June 29th, 2012, 15:39, edited 1 time in total.

[BlackST]

Should not in normal condition, but your drive is unstable and with non native firmware, so everything can happen...

[djinn]

I understand, but then do you think this person is wrong?

If it was me, I'd get all my data off and use the drive's secure erase feature and use it for a DVR or something where you'll not care if it goes belly up. Once burnt, twice shy with lost data. A useless backup taught me very well. ;(

This user recommended the secure erase for my drive. Hence, my decision to apply it to my drive. But now I'm scared...

[Vulcan]

I understand, but then do you think this person is wrong?

IMHO you are asking the wrong question - there is no "right" or wrong". All of this is opinion and guess, due to the unknown state (and strange SMART data etc.) of your drive.

Did JWCC guarantee there would be no problem after doing a secure erase? No! In fact JWCC even told you to use the drive for "something where you'll not care if it goes belly up". Therefore that agrees with all the comments you are getting from BlackST and me, about the risk in that plan.

Since your drive is sick, anything strange could happen at any time - just accept that situation, and stop trying to find certainties. There aren't any!

[djinn]

Ok, Ok, I understand your points, but FYI, I have just made both standard and ENHANCED secure erases to the drive with the hdparm utility. Before running the tests, there were 80 reallocated sectors and after that no additional sectors have been found. No problems when tests finished, the drive came back to ready state from busy and I think it remains in the same health state before performing the erasing.

However, I've ran the HDAT2 utility from a boot disk, and I've found SMART verification throws a "data checksum error". What does it mean? Also I'm running all tests offered by the utility.

And a last question, please, if there were REAL bad sectors on the drive (not wrongly recognised by the drive due to firmware mismatch), will ALL existent bad sectors appear after running the internal erasing or may they appear after some usage (because of wear)? That is, if on a healthy drive (not this one) bad sectors may appear after some time even when performing several check tests, including the SE.

Yes, this is not my main drive any more, it will be used as a temporary storage for movies (I watch and delete them later).

Thank you again for your patience and advises. Lesson learned, don't mess with firmware versions any more.

[Vulcan]

No problems when tests finished

As explained before, anything could happen. In your case, it seems the result was not bad.

I've found SMART verification throws a "data checksum error". What does it mean?

In short - the SMART data cannot be trusted (even less than usual!). See info in ATA-8 spec, section 7.53.6. IMHO this suggests drive firmware problems, as it is the drive firmware which maintains the SMART data checksum, when the recorded SMART data changes.

And a last question, please, if [...]

There are too many hypothetical parts in your question, for me to want to answer. Perhaps someone else will want to give you a guess.

Remember that "bad sectors" are a symptom, not a cause. The future behaviour depends (partly) on the actual cause, and you don't know that in your hypothetical situation! Any answer to your question (yes, no, maybe, 29.5 days later etc. ) might be correct.

Since you say that was your last question, I wish you good luck with your drive, and I'm glad you are not keeping any valuable data on it.

[djinn]

In short - the SMART data cannot be trusted (even less than usual!). See info in ATA-8 spec, section 7.53.6. IMHO this suggests drive firmware problems, as it is the drive firmware which maintains the SMART data checksum, when the recorded SMART data changes.

Wow, but I've realized that this SMART checksum error also appears in my new WD5000AAKX drive!!! What's this? And no, I've not overwritten its original firmware and it's a TRULY healthy drive (so far). I'm starting to think this is a bug of HDAT2, becase it's very unlikely both drives throw the same error, isn't it?

[Vulcan]

I'm starting to think this is a bug of HDAT2, becase it's very unlikely both drives throw the same error, isn't it?

I can think of an alternative explanation, but since I don't care about the answer, I'm not going to spend time investigating. Using MHDD or similar, your hypothesis could be tested by someone with the necessary skills to look at the raw 512 bytes of SMART data and calculate the checksum manually, for example.

[djinn]

Excuse me for bumping again this thread, but I've found my ST3500418AS was QUALIFIED for firmware update to CC49 version according to the serial number check of the Seagate web site. The Download Finder web page showed the following:

Code:

Serial Number: 5VML0RVx (last character hidden for privacy reasons)
Part Number: 9SL142-303
Model Number: ST3500418AS
Family: BARRACUDA 7200.12 FAMILY   
Firmware
--------
Name: Barracuda 7200.12 Firmware
Importance: Important
Version: CC49
Release Date: 11-Mar-11
Short Description: Barracuda 7200.12 firmware update More...
Downloads and Actions: Click here


As you can see, the first link is for redirecting to the download of the firmware updater, so I'm starting to think I could not update the firmware due to hardware incompatibilities with the updater (maybe the nVIDIA nForce 4 SATA controller). If this is true, then my drive was not bricked by me (by forcing the manual update) and this drive IS PHYSICALLY failing and, of course, I could RMA it, but according to HD Sentinel I've not reached the necessary very low health status (0%) to begin the RMA procedure. At this moment I have 129 reallocated sectors, 4 uncorrectable errors (increased by 2 yesterday) and 20 high fly writes. My health status is currently at 26%.

Any thoughts? Many thanks for your patience and replies.

[Keatah]

Ridiculous. The drive is obviously failing and is apparently under warranty, so replace it!

[SAjunky]

Please, I appreciate your answers but focus on my problem: if my drive is healthy and I can ignore those warning messages. I repeat that Victoria says SMART status is good (10 reallocated sectors are below the thresold, 54).

From user perspective this drive is not good:
- significant drop of seek error rate 79, worst 60. Close to threshold.
- even bigger drop in error rate 52, worst 35 (not tripping out).
The most interesting is minimum HDA temperature 13C. I is like it had been taken out from freezing bag and plugged in before it get warmed up.

I suggest to do more scan tests in MHDD/Victoria with erase delays and watch how SMART is updated after each test. But Bakup data first.

[Vulcan]

@SAjunky,

From user perspective this drive is not good:

As I've said previously in this thread, I'm also concerned about the likely longevity of this drive too, but not for the reasons which you have highlighted.

significant drop of seek error rate 79, worst 60. Close to threshold.

Considering what the starting value is, how this attribute is calculated, and what the threshold is (30), then I don't agree about the significant drop, nor being "close to [the] threshold". Similar values are not uncommon for these drives, in my experience.

even bigger drop in error rate 52, worst 35 (not tripping out).

FYI this SMART attribute is "not tripping out" because this one cannot cause a SMART "trip".

[SAjunky]

If this is true, then my drive was not bricked by me (by forcing the manual update) and this drive IS PHYSICALLY failing

Drive was failing and firmware updade coudn't make any better, only worse. No matter qualified or not qualified.
This is common user practice to update firmware on devices when something goes wrong. With hard drives it is in reverse. It prevents wrong things to happen, but it doesn't help when something goes wrong.

[djinn]

Well, Thank you again for your feedback.

I've contacted Seagate Support and, unfortunately, they DON'T replace my drive at this moment. They consider the following SMART values "normal":
Reallocated sectors count: 129
Uncorrectable Errors: 4
Uncorrectable CRC Error Count: 2
Command Timeout: 4325609
High Fly Writes: 20

They won't approve a RMA without a SeaTools testcode (when that tool fails) and they don't trust in third-party utilities like HD Sentinel (to see the SMART values). The case is SeaTools always passes, but this is not a guarantee, because I have new bad sectors as a regular basis, and high fly writes increase too.

This support is unacceptable, they know my drive is failing but they don't want to replace it. I'm glad I bought a WD drive which works perfectly (no errors in SMART so far).

I won't buy a Seagate drive any more. Too bad, oh well.