SSDs

[VGER]

Hi,

I had an incident concerning SSDs a while ago that set my blood boiling. Maybe you all have some further ideas about it that you could share with me. (If you make it through the longish report - sorry, I'm in my talkative mood today... )

Around the turn of the year 2011/12 I bought a Intel 320 SSD for my laptop (Dell Vostro 1520). I made sure to always keep the drive firmware current, since there were some issues around regarding the "8 MB bug".

About six months later there was another firmware update which I promptly installed. One or two weeks later I was happily working away on the laptop when suddenly Windows crashed with a blue screen. There was considerable disk access at the time since I was just copying a SD-card full of camera images to the SSD. When I rebooted the laptop, no OS was found. I then called up the BIOS and found the drive being detected at 8 MB - the exact failure mode the constant firmware updates should have prevented.

After some discussion with data recovery experts, Intel and the vendor I'd purchased the drive from I sent it back for replacement. (Intel was no help at all, the recovery guys didn't give much hope anyway, and the vendor naturally wanted the drive unopened to be elegible for replacement.) I got a new SSD of the same make a few weeks later. But I couldn't wait that long.

So I purchased another SSD, this time a "Verbatim Black".

About a month later, the laptop was taking more and more "breaks" - short freezes during normal work. I checked the disk and found numerous bad sectors and damaged files. I tried to run a backup, but portions of the data could no longer be accessed. Whatever, I didn't loose anything this time. (The first time I was in the field working with data that had just been newly generated and for which there wasn't any backup yet. Losses were heavy.)

Now the curious part: To send the Verbatim SSD back, I ran a security erase on it (using a DOS tool.) After this, I did a read-verify of the surface just for fun and was surprised that I could find no trace of any error. I formated the disk and tried again, but still no defects at all.

HD-Sentinel still displayed a drastically lowered "health" value, but when you looked closely, you could see that the SMART data was showing multiple "controller failures/data transfer errors", not surface errors ur "uncorrectable sectors". (I don't have the drive installed at this time, but if aynone is interested in a HD-Sentinel report I can certainly create one.)

At about that time, the replacement for the Intel SSD arrived and I installed this drive in the laptop, which I'm using since then without further issues.

But the incident still puzzles me:

- The Intel death could be a simple random defect - apparently when the controller dies, it always shows 8 MB. It dosen't have to be the firmware problem.

- But: The Intel firmware bug was rumoured to be caused by unexpected drive shutdowns, such as power loss without prior shutdown procedure. Which might exactly be what happened when the Verbatim drive recorded its "controller failures". Only the Verbatim didn't kill the hardware over it but silently ignored them as well as it could.

Could it be that the SSDs are not even at fault, but the laptop? A flakey data connection or something like that, which the drives just reacted differently to?


As described, I'm now back to the new Intel 320, so far without problems. But there is still that uneasy feeling... And what am I supposed to do with the Verbatim? I can't very well return it, now that it is working again. But trusting it with real data doesn't seem like a good idea either.


Has anyone ever seen something similar?


Regards

[hdsentinel]

The operating environment can cause serious problems for both hard disks and SSDs.
Flaky connections, data transfers and unstable power source may both cause these kind of problems.

Personally I agree that the problem with the Intel SSD is related to the improper shutdown. But this may happen any time (even if it should not) and personally I'd worry if this happened even with the latest firmware version which should prevent this. This way such Intel SSDs are like time-bombs which can any time (without prior notice) fail and prevents any kind of data recovery as you can see.

But the Verbatim drive may really experienced some bad sectors which may not be related to the current system.

Yes, I can confirm SSDs can have bad sectors as well, exactly as hard disks. After write to a specific sector, the SSD verifies the contents of the sector and
if the CRC check shows that the written information is damages, the SSD automatically writes the data again to a spare sector and re-directs all further reads and writes to that sector.

This is exactly what hard disks do - just on SSDs the whole process is much quicker and there are no noticeable signs, for example no "clicking" or other weird noises, no retries / substantial delays and so.

I suspect the secure erase function forced the SSD to correct problems this way - and then all further reads/writes/tests use the remaining and spare area. This is why no propblems reported after that.

I'd recommend some testing to verify if the SSD has further problem(s) or if the status is now stable and the device could be used without problems.

For more information about these bad sectors and further steps about how to examine and improve the situation, please visit
http://www.hdsentinel.com/faq.php#health and
http://www.hdsentinel.com/faq.php#tests
http://www.hdsentinel.com/faq_repair_ha ... _drive.php

Personally I'd be curious to see report(s) about the Verbatim drive, now and/or after some testing. This way it is posible to check how the status changes.

[stinklyonion]

I don't have any problems with my SSD yet. As far as I know, everything works great on my end for both HDD and SSD.

[JTunn]

I've had my samsung EVO SSD (http://www.shoppingexpress.com.au/laptop-accessories) for quite some time, and haven't experienced any problems whatsoever (am running diagnostics very frequently). I'd say 2 bad SSDs may mean that the problem lies elswehere, probably in your other hardware as previously stated. :/