Request for Anand regarding Sandforce SSDs

[[ IronMan.Lok ]]

I know the units based on SF-2XXX get all the attention for now, but there a strange issue going on with units based on SF-1XXX, where units that have been rock solid for a long time simply start presenting problems. A 30 seconds freeze happens which may or maynot be followed by a 0xF4 BSOD if it was booted from SSD, or the SSD disappear from system if its a secondary unit... and the problem get worse gradually as the time passes... There's a more detailed description of the problem on this thread:
http://www.ocztechnologyforum.com/f...BSOD-(0xF4)-(not-related-to-sleep-hybernation)

What really bothers me, is that I've seen a lot of ppl with multiple RMAs complaining that none of their units lasted for more than X months, and after replacing it, the unit received lasted only for another X months...

So, what I'd like to ask is if you could host some public poll to see how long Sandforce 1XXX SSDs usually last, how many ppl out there have a Sandforce based unit that has lasted for more than a year, how many times did they have to RMA their unit, etc...

I mean, in the first months the unit is really strong and solid, it lasts long enough for ppl to claim they're good, but hardly a reviewer keep testing the same unit for a long period like 6 months or a year, so it's hard to notice if a product have a bug that manifest only after 6 months or 1 year...

And given the fact that certain SSD lines completed 1 year in the market recently, which coincides with the recent increase in faulty units reported, I dont know, Sandforce may or may not has a reliability problem over time, but that poll could greatly help in giving us a clue...

Whats your thoughts on the those issues? Could you help us SSD users please?

 

[Idontcare]

You can construct your own poll right here in the forums if you like.

 

[Zap]

Hello [ IronMan.Lok ], and welcome to AnandTech Forums.

You know that you can post your own poll here, right?

 

[[ IronMan.Lok ]]

oh I meant a public poll visible on the main site to attract most sandforce ssd owners possible... I mean, if I did this poll on OCZ supports forums, results would be skewed as only ppl with problems post there... Im not sure what kind of public would frequent this forum, specially because I'm pretty new to these forums myself, so I dont know if making the poll here would skew the results or not...

what do u guys think?

 

[chusteczka]

I think you would do well to poll Anandtech users here in the forums. You can easily do it yourself and your results would include mostly knowledgeable individuals while effectively removing those experiencing problems due to their own misunderstandings.

 

[Old Hippie]

oh I meant a public poll visible on the main site to attract most sandforce ssd owners possible

Let me get this straight....you expect the owner of this site to post a survey on the front page focused on a single companies products for complaints?

LOL!

 

[groberts101]

While I agree that the poll would be skewed if set up at the mfgrs support forum(for the very reason mentioned).. it may end up skewed here as well due to the simple fact that there are an enormous amount of anti-OCZ members on this forum.

I do believe it could be a good thread if the mod's stayed busy cleaning all the wise-cracking and sh**-talking along the way though. Especially since much of it comes from people that don't even own these drives. lol

As for my experience so far?.. I have 7 Sandforce drives(6 of them in R0 for almost 1.5 years) along with 3 other family/friends running them varying between more than 1 year and as little as 2 months without so much as a hiccup.

I personally set the drives/systems up and NEVER allow them to sleep the systems due to many/most of these so called "deaths" being caused by sleep transition issues(especially with many system defaults using S3 mode whick kills power to the drive and causes it to panic lock/brick. A firmware fault?.. most likely. A hardware fault?.. highly doubtful.

I also know for a fact that there is MASSIVE public assumption being made out there and there are VERY FEW actual damaged/dead drives and almost every single one sent back can be destructive flashed right back to life. For that reason alone I would love to get my hands on a case or 2 of those RMA'd drives along with the engineer's flash needed to revive them.

PS. just thought I'd mention the fact that the newer SF2000 series drives don't panic lock/"die" nearly as easily as the previous gen SF1000 series do.

 

[Old Hippie]

While I agree that the poll would be skewed if set up at the mfgrs support forum(for the very reason mentioned)..

I wasn't talking about any type of support but you can't expect a neutral review/info site to single out any manfg for different treatment.

Or doesn't that even cross your mind?

 

[[ IronMan.Lok ]]

First, as far as I can tell, this is not a single manufacturer problem, all sandforce SSDs are having problem, so it would rule out "anti-ocz" fanboys

Second, old hippie, I don't understand your point, I came to anandtech because he already called "bs" over the very same company (sandforce) on an article about its SSD reliability (but it was limited to newer drives based on SF-2XXX)... He also called out on problems from Intel and Micron in the past... and actually, that's exactly what I call neutral stance, as I expect tech journalists to confront companies that put defective stuff on the market, and not just "parrot" whatever companies tell them to, or simply hide the ugly stuff from public...

All I'm asking is to put SF-1XXX on focus too, because there's a good chance they're presenting a latent bug given the high amount of reports lately... on Anandtech's SF-1XXX reviews, it was never mentioned, obviously because this problem usually happens after months of use, so yeah, I think it would be a good use of a "neutral site" to update the situation and say "hey guys, remember that product I reviewed a year ago? lots of them are presenting problems"

and please, don't confuse the issue I'm talking about with the sleep problem, this is not related at all, Im talking about a lot of ppl that had very stable units (without sleep enabled) for months, and suddenly the unit started to degrade, freezing and BSODing with certain specific tasks...

the amount of ppl with this problem is too high to be ignored, and most ppl either confuse this issue with sleep problems, or use the argument "I THINK theres only a small percentage with this problem"... well, the poll would put an end to the discussion of the second point

 

[Idontcare]

the amount of ppl with this problem is too high to be ignored, and most ppl either confuse this issue with sleep problems, or use the argument "I THINK theres only a small percentage with this problem"... well, the poll would put an end to the discussion of the second point

Surely you can recognize that no poll would actually put the discussion to rest because the poll itself will be questioned.

Who takes the time to vote in a poll? The verbally outspoken.

Who's more likely to be verbally outspoken? Those who are irritated, annoyed, or otherwise jaded.

I can already tell you in advance how your post-poll conversations are going to go - if the poll shows a large percentage of issues with Sandforce drives, the arguments will be "but that's only because the small minority of people with sandforce drive issues are likely to vote, you haven't really sampled a cross-section of happy sandforce owners yet".

Or, if the poll shows a small percentage of issues with Sandforce drives, the arguments will be "but that's because you held the poll on a tech website that people with SSD issues wouldn't have the luxury (because they ARE dealing with SSD issues at that time) to visit and vote in a poll, they are busy over on OCZ's support forum trying to get their SSD's functional, duh".

Polls settle nothing because the ensuing arguments then become about the poll itself, what it actually sampled in the demographics and what it failed to capture.

In short, its a fool's errand.

 

[Coup27]

Idontcare you've been modding forums for too long.

I used to be the main admin on a music forum which had ~100,000 members.

Had to give up in the end. Now I just casually browse various forums. Much less stress

 

[Idontcare]

Idontcare you've been modding forums for too long.

I used to be the main admin on a music forum which had ~100,000 members.

Had to give up in the end. Now I just casually browse various forums. Much less stress

LOL, so true. But tell me I'm wrong about the net result of polls Surely you've seen it to in your time, over and over again.

 

[Coup27]

LOL, so true. But tell me I'm wrong about the net result of polls Surely you've seen it to in your time, over and over again.

I think I've seen it all on forums. Although polls on here would be different than polls on music. Music polls are generally personal preference.

What I did discover during my sentence is they're probably 4 years I'll never get back. Looking back, I cannot believe how much time I sunk into forums. When one of the mods went AWOL I ended up covering the Florida forum for 6 months as well. It became my life lol.

 

[blackened23]

While I'm using a sandforce 12xx SSD with no problems, the reports about the 2281 series drives are extremely worrisome. I've never seen a product with so many reports of issues in years.....someone needs to step it up with the QA among all the manufacturers.

I don't understand why this is going on months and months past release - if someone like AMD had released say, a videocard with this many issues they would get killed by the press.

 

[[ IronMan.Lok ]]

Surely you can recognize that no poll would actually put the discussion to rest because the poll itself will be questioned.

Who takes the time to vote in a poll? The verbally outspoken.

Who's more likely to be verbally outspoken? Those who are irritated, annoyed, or otherwise jaded.

I can already tell you in advance how your post-poll conversations are going to go - if the poll shows a large percentage of issues with Sandforce drives, the arguments will be "but that's only because the small minority of people with sandforce drive issues are likely to vote, you haven't really sampled a cross-section of happy sandforce owners yet".

Or, if the poll shows a small percentage of issues with Sandforce drives, the arguments will be "but that's because you held the poll on a tech website that people with SSD issues wouldn't have the luxury (because they ARE dealing with SSD issues at that time) to visit and vote in a poll, they are busy over on OCZ's support forum trying to get their SSD's functional, duh".

Polls settle nothing because the ensuing arguments then become about the poll itself, what it actually sampled in the demographics and what it failed to capture.

In short, its a fool's errand.

[/quote]

imo, the poll can be useful depending on how you gonna interpret its results and what kind of info you are gonna extract from it, independent on what sample it covered...

ok, for example, if 1000 ppl answer this poll, and this 1000 ppl are angry with sandforce products, even if its not the best sample of the market, it means one thing, there were 1000 ppl that were annoyed enough to go on internet and complain, because something is wrong with the product... specially if you consider SSDs enthusiast low volume products, they dont sell by the millions like HDDs

what I mean, it doesn't matter who "win" the poll, if there's enough ppl (in absolute numbers) with units not lasting more than a year, the its becomes clear that something went wrong somewhere

about blackened23's post, I think press ain't going after sandforce's head yet because SSDs aren't widespread enough in market...

 

[Idontcare]

Yeah I still don't see it happening. I think we won't know the statistics here until a large independent 3rd-party study is undertaken the likes of Google's massive spindle-drive study.

Consider how the original Jmicron-based SSD"s were finally caught out, it wasn't a poll or collective inputs from dissatisfied users...it was from tenacious and persistent assessment and analysis by a handful of highly motivated and well equipped/resourced individuals.

Anand himself might be the person for the job, again, with these Sandforce issues...but is he really interested in doggedly pursuing to crack that nut, again?

But a poll, I just don't see it telling you anything you don't already suspect or know, for sure there are thousands of end-users impacted, just look at the OCZ forums.

* I forgot to add the obvious "conspiracy" related outcomes to the poll when I referenced it above...if Sandforce drives are voted as overly problematic then there will be some folks who claim "its a conspiracy, Intel's marketing division paid Chinese workers WoW Gold to vote en masse against SandForce!"...and likewise if the results are that Sandforce is voted overwhelmingly good "its a conspiracy, OCZ's marketing division paid Chinese workers WoW Gold to vote en masse for SandForce!".

End result -> fool's errand to collect the poll stats as the endless bickering over the legitimacy of the poll stats will forever ensue.

 

[Old Hippie]

Second, old hippie, I don't understand your point, I came to anandtech because he already called "bs" over the very same company (sandforce) on an article about its SSD reliability (but it was limited to newer drives based on SF-2XXX)... He also called out on problems from Intel and Micron in the past... and actually, that's exactly what I call neutral stance, as I expect tech journalists to confront companies that put defective stuff on the market, and not just "parrot" whatever companies tell them to, or simply hide the ugly stuff from public...

Think what you want.

I doubt it makes any business sense to focus a poll on OCZ's shortcomings unless you add every company to the poll to avoid bias claims.

I really don't care but it just doesn't sound like "good business" to me for a site to sponsor a poll like this.

 

[[ IronMan.Lok ]]

@old hippie

why do you insist saying "OCZ" when I'm saying "Sandforce" all the time?

@Idontcare

while I agree ppl will question that, there won't be a "single and valid" interpretation of its results... I know fanboys will raise tin foil hat conspiracies, but, every single person will have its own interpretation of the results, and ppl willing to solve the problem, will have another tool to tackle the problem...

also, take into consideration that what I wanted to accomplish with the poll is not to get the satisfaction level from customers, but the following:

- to discover how many ppl does have stable units after a year of constant use and how many doesnt...
- for the ppl that does have their units working fine for more than a year, what's their combination? FW, chipset, raid'd or not, how frequent does secure erase, etc...
- for the ppl that doesn't, how many times they RMA'd their units, and if the replacement solved the problem

also, as a side effect, it would "force" companies to break the silence... having a defective products happens to a lot of companies, even the mighty intel, but acting like nothing is happening, not coming clean about it is terribly wrong... for now, they're acting like nothing is wrong with SF-1XXX, and they would be acting like that for SF-2XXX too if Anand didn't call them out

 

[MobiusPizza]

I recently bought a sandforce SF-1222 mSATA drive for my Lenovo X220. The brand is Renice. I used the drive as Windows 7 drive for just over a month, and problems begin to surface.

As op mentioned, the machine will freeze randomly, or quite certainly when high I/O utilisation. It sometimes produce a hard freeze, or a softfreeze, or a blue screen with stop 0x00000F4 as op mentioned.
I can boot into windows fine. It runs chkdsk c: /f fine. It fails chkdsk c: /r if and only if some bad flash blocks contain data file. If the bad block resides in free memory, chkdsk c: /r also runs fine (This I have tested after reinstalling Windows).

Tried AHCI drivers from Microsoft and Intel. Same problem. Tried using Acrnois to back up data. Always fail at same place with CRC error on a block. So my guess is somehow SF controller is very bad at handling blocks with CRC error, and will slow I/O performance to a freeze.

When there is a soft-freeze, the HDD light is always on and the disk performance becomes abysmal. Resource monitor shows this:
http://s233.photobucket.com/albums/ee209/AnnihilatorX/?action=view&current=ReniceX3.png

Then I reinstall Windows 7 completely. A clean installation, but did quick format only. Guess what, problem still persisted.

I'd advise anyone to refrain from buying SandForce drives, like me who was sold by the benchmark figures. Should have got an Intel 310.

Someone did mention on a forum he had OCZ agility drives and fixed these issues through Secure Erase. But to me that's not an option as I can never seem to get HDErase to work with my controller.

 

[Old Hippie]

I'd be returning that unit.

 

[MobiusPizza]

Well done that. But I doubt the replacement will last until the root of the problem is solved by none other than SandForce themselves.

 

[Coup27]

Well done that. But I doubt the replacement will last until the root of the problem is solved by none other than SandForce themselves.

To be honest, I get the feeling SF don't actually know what is wrong with their drives. 1st or 2nd gen.

 

[MobiusPizza]

To be honest, I get the feeling SF don't actually know what is wrong with their drives. 1st or 2nd gen.

It doesn't help when they hides behind OEMs and says in their web page any support issues should be forwarded to OEMs, not them.

I've actually offered to send back my SSD to Renice the OEM directly, rather than back to my point of purchase, so that someone may get first hand on the problem.

I now convince I should have sent it directly to SandForce headquarters with a note From Russia with Love.

 

[groberts101]

I love my Sandforce drive's(mainly because they work, lol)... BUT.. they should have had their asses hualed into a massive mfgr wide class action lawsuit 6 months after release. Here we are 1.5 years later revisiting all the same bull*hit.

And all the one's happily running one and sleeping their system's out there?.. they won't even know what hit em' till it's too late. Pretty sad stuff, really.

 

[Gigantopithecus]

I have SF 1xxx and 2xxx SSDs deployed in various systems. None of them are 'mission-critical' drives, all of them can be replaced and the systems can all be back up and running again in under an hour with minimal effort.

IMHO SSDs are still luxury items, at least to me. They're really nice to have, but I simply don't trust anything other than my Intel G2s.

Of course, not having spent any of my own money on the SF SSDs makes me pretty charitable towards them.