Bad Seagate SSHD. While prepping for RMA, it got better?!

[GTMoraes]

Hey

I have an Seagate SSHD that gone bust on me. Many data lost, slow I/O, drive corruption everywhere, bad blocks, rats, cockroaches, all nasty stuff.
Ran the SeaTools and it identified my drive as BAD. SMART also reported it as FAIL. Went to the RMA page and it was accepted. Printed all the stuff and got everything set to send it away, but I had to copy my files. So I copied all my data (a 48 hours copy, through SATA to USB 3.0. Goddarnit question prompts) to another drive (trusty old Samsung from 2007). Many random stuff were lost, as I could see during copy. Nothing important, but lost nonetheless.

So well, after copying, I thought it would be a good practice to "low level format" it, and so I did it. After a couple of hours, it was done, I was just removing the hard drive and I thought It'd be nice to write the SeaTools error code on the shipping box, so I ran the DST test once again..

It gave me a green light. It says it's all good now?? Is that right, or can't it see the bad sectors out of a bunch of zeroes?
If I send it over, will they send it back or replace it? SMART still reports it as BAD (in one field, though: Uncorrectable Error Count = 1)


Thanks

 

[WyomingKnott]

If it failed once, never trust it again. If you already have an RMA, return it without second thoughts.

 

[GTMoraes]

If it failed once, never trust it again. If you already have an RMA, return it without second thoughts.

Thanks for the reply!

That's what I want to do, but the Seagate strictly says:

Please note that evidence of the following will result in rejected warranty claims:
[..]
- No Trouble Found (NTF). Before returning your Product, you may use the Seagate SeaTools diagnostic tool to determine the condition of your disk product and whether it is eligible to be returned under warranty.

Thing is, SeaTools reports it now, by any means, as a good drive. The drive is empty and zeroed. The LLF tool I've used also reported no bad blocks, but before doing a LLF, SeaTools was reporting it as a BAD drive in any test I could run

Short DST - Iniciado 02/04/2015 10:39:25
Short DST - Reprovado 02/04/2015 10:42:24
SeaTools Test Code: A7999FEE
Teste generico Completo - Iniciado 03/04/2015 03:44:32
Teste generico Completo - Reprovado 03/04/2015 03:45:28
SeaTools Test Code: A7999FDE
Apagamento completo (SATA) - Reprovado 03/04/2015 03:47:08
Reparação Rápida - Reprovado 03/04/2015 03:52:18
SeaTools Test Code: A7999FCE

and now:

Short DST - Iniciado 03/04/2015 07:07:25
Short DST - Aprovado 03/04/2015 07:08:28
SMART - Aprovado 03/04/2015 07:09:03
Short DST - Iniciado 03/04/2015 10:24:22
Short DST - Aprovado 03/04/2015 10:25:21
Teste generico Completo - Iniciado 03/04/2015 10:27:27
Teste generico Completo - Aprovado 03/04/2015 13:33:14

But SMART still thinks it's a problematic drive

I'm worried that if I send it over, they'll reject it (because it reports as good in any SeaTool test), void my warranty and have to pay the shipping back. What do you think? Can the drive still show errors when zeroed, or do I have to fill it up with data to give me errors? At the time it was corrupting data, it had 22GB left of 1TB

Thanks again!

 

[WyomingKnott]

The answer is - I don't know. If I were in your shoes, I'd call Seagate, explain the situation, get them to agree to the RMA, and have them send you an email confirming. Decent vendors will usually make exceptions to strict rules; the only problem I've had doing this is that Microsoft shipped my too many copies of XP before they got it right and sent 7.

Then again, if I were in your shoes my feet would probably hurt. They are size 13.

 

[GTMoraes]

The answer is - I don't know. If I were in your shoes, I'd call Seagate, explain the situation, get them to agree to the RMA, and have them send you an email confirming. Decent vendors will usually make exceptions to strict rules; the only problem I've had doing this is that Microsoft shipped my too many copies of XP before they got it right and sent 7.

Then again, if I were in your shoes my feet would probably hurt. They are size 13.

Thanks again for the answer!

I'm filling the drive as we talk, so I can test how it performs with a drive full. I'd imagine there's an software for this purpose, but I couldn't think of a name to google it, so I'll just do the old fashioned way. So far, I managed to copy a 590GB folder after several hours (many tiny files), but after it was done I rushed to remove the drive from the PC HDD drawer and I guess I corrupted the folder... Or it was the hard drive failing.. unsure. It reported that there were 590GB used in that drive, but there was no folder at all. Couldn't even find the folder cd'ing through command line. I know for sure that when I ran chkdsk /f, it "fixed" a LOT of errors and the 590GB usage was wiped from the drive. I'll try to copy it one more time to see what'll happen

Calling them is a good idea that I didn't think of. Perhaps we could get into an agreement and not only send it over and hope for the best. We're on a holiday here, so I'd better wait for monday

Heh, perhaps! I don't know my US size, but they're 28.5cm!
Happy easter!

 

[Paperdoc]

I fully agree with WyomingKnott - discuss with Seagate.

So how can this happen? Well every time you read a Sector the HDD internally checks for a problem and, if necessary, does a fix of sorts. You read all of the sectors that were occupied in copying everything. Then you did a "Low Level Format" - I don't know what you mean by that, but maybe it was a Full Format, which writes zeroes to all available sectors. This also triggers the self-testing and fixing routine, but this time for every Sector on the HDD.

What is this "fix"? It is not really a full repair. What it does if the HDD believes a Sector is faulty is to recover what data it can from there, then retire the Sector and replace it from a stock of known-good spares. When this is completed, the HDD no longer has an "Bad Sectors" it can expose to the outside world, and it looks almost perfect. BUT there is a limit to the number of good spares it has, so it keeps track of how often this is done. Eventually through the SMART system it sends out a warning that the Reallocated Sectors Count is too high. That means two things: (a) you are running out of spares and the process will fail sometime in the future; and, (b) a lot of Bad Sectors have been found, which probably indicates that more will be found, perhaps even more frequently than in the past. Net result: replace this HDD NOW before you find it impossible to get good data from failing Sectors.

So, if that really is what has happened, the Reallocated Sectors Count from the HDD's SMART data may be enough to satisfy Seagate that the unit is failing and needs replacement. It depends on what the alarm limit for this parameter is for your HDD unit.

 

[GTMoraes]

I fully agree with WyomingKnott - discuss with Seagate.

So how can this happen? Well every time you read a Sector the HDD internally checks for a problem and, if necessary, does a fix of sorts. You read all of the sectors that were occupied in copying everything. Then you did a "Low Level Format" - I don't know what you mean by that, but maybe it was a Full Format, which writes zeroes to all available sectors. This also triggers the self-testing and fixing routine, but this time for every Sector on the HDD.

What is this "fix"? It is not really a full repair. What it does if the HDD believes a Sector is faulty is to recover what data it can from there, then retire the Sector and replace it from a stock of known-good spares. When this is completed, the HDD no longer has an "Bad Sectors" it can expose to the outside world, and it looks almost perfect. BUT there is a limit to the number of good spares it has, so it keeps track of how often this is done. Eventually through the SMART system it sends out a warning that the Reallocated Sectors Count is too high. That means two things: (a) you are running out of spares and the process will fail sometime in the future; and, (b) a lot of Bad Sectors have been found, which probably indicates that more will be found, perhaps even more frequently than in the past. Net result: replace this HDD NOW before you find it impossible to get good data from failing Sectors.

So, if that really is what has happened, the Reallocated Sectors Count from the HDD's SMART data may be enough to satisfy Seagate that the unit is failing and needs replacement. It depends on what the alarm limit for this parameter is for your HDD unit.

Hey! Thanks for the answer

I see. Yeah, by "Low Level Format" (I don't think that's the actual name for it, is it?) I meant filling the drive full of zeroes, using a third-party tool. Took a while, though, and it didn't report any bad sector for some reason.
I think I understood the "repairing" process after your answer. I believe there's no way to know how many spares does my HDD has, right? Because if it's using 88 sectors out of 100 spares, that's too much for a three months old HDD, but if there's 1000 spares, well it ain't that bad. By any way, it reports itself as failing, so I guess it's really running out.

I've tried to give this drive one more go as an normal drive, since it doesn't give out any bad block erros (and because I can only send it away monday), I've installed it on my main notebook and installed Windows 10 on it. Well, after installing, it wanted to update do Build 10049.. it's 8:13PM and it's stuck at 30% since 11AM so... I guess this drive is really done for.
In comparison, my PC did the very same process at the same time, and it was finished less than half an hour later.

The reasons that made me create this thread was:
1 - If a "low level format" could actually FIX a defective HDD
-- Now I know it doesn't actually FIX, but pushes the defect to another place so I can't see it. But it can only push it away so much.

2 - If Seagate would accept this drive, even though their own software doesn't find anything bad anymore
-- I was worried because SeaTools doesn't report it as a bad drive anymore, with any test (even SMART test), and from what I understood from the warranty, they'll test it with SeaTools, and if doesn't complain about anything, it means that the drive is fine (which is obviously not).
If they take a look at the SMART log, they'll see there are reports of uncorrectable error, command timeouts and high reallocated sector count, and I hope that's enough to make them replace, but it's better to be on the safe side and call them before sending it over.

I've never done a RMA before (although I've returned a few products to sellers), so that's pretty weird for me. Hope it all turns out fine

--------------------------------------

So it finally finished installing the update. Phew!
It seems it still has the issue that made me aware of a defect on the drive: 100% Disk usage, 4000~8000ms latency, <10kb/s read and write.
SeaTools still reports it as a good drive by any means.

But hey, should I send it formatted (which seems to not trigger any errors) or send it as-is, with my files (not sensitive information, but steam with auto-login, configured Chrome browser, profile pic etc), if it ever reports an error?

----------------------------

http://knowledge.seagate.com/articles/en_US/FAQ/171691en?language=en_US

I've tested it through a Seagate external drive, internally in a PC and internally in a Notebook. They've all the same issues
Does anyone have experience with RMA and stuff? I believe I'll run through the "restart your computer" steps, and I'd like to directly go to advanced support