To the RecoveryForce CEO: What makes the ST3000DM001 fail so much?

[eton975]

From what I've read on your forum, a lot of the ST3000DM001s that come into your lab have crushed heads, which cause the failure. You mentioned a low head flying height somewhere, but doesn't that apply to most other Barracuda/Desktop HDD drives which don't experience such high failure rates (e.g. ST4000DM000)? What makes these drives particularly vulnerable to this kind of failure?

How does the design and operation of the ST3000DM001 differ from the WD/HGST/Samsung/Toshiba drives? Are there any particular 'sub-models' of the ST3000DM001 (e.g. particular type of firmware/number of platters/case design/9YN166 vs. 1CH166 vs. 1ER166) that fail more than the others?

 

[Soulkeeper]

I don't think anyone has really done a study into this in detail, it's mostly just speculation and counting failures. Seagate hasn't stated anything on the matter that i'm aware of, but it appears their newer/larger drives are fairing better. One thing worth considering is the DM drives were essentially the first 1TB/platter drives on the market, also this is when they started ramping up the chinese manufacturing. Not to mention the low price and reduced warranties. So essentially the newest tech in the cheapest combination, not surprising in retrospect.

 

[Spungo]

Seagate's problem is that they put zero thought into quality control. They rush crap to market as quickly as possible, as cheap as possible. Look at 8TB drives and see if you notice anything unusual:
http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&IsNodeId=1&N=100167523 600376735
Seagate: $258
HGST: $539

Is it safe to assume there's a reason one is half as expensive? Was it tested, or did they push it out as quickly as possible? The Xbox 360 was rushed to market with little or no testing just so it would hit the market before the PS3, and we saw how well that turned out:

That's exactly what Seagate does. They want to be the first to have a certain hard drive on the market. The only way to do this, aside from being a generation ahead in terms of technology, is to cut out all of that expensive and time consume quality control testing.

 

[Elixer]

Not to mention the low price and reduced warranties. So essentially the newest tech in the cheapest combination, not surprising in retrospect.

This is the biggest problem, their warranties went from 5 years, all the way down to 1.
What is next, 30 days?

While warranties don't tell the full story, it does start to paint a picture on how reliable a company thinks their products are.

 

[dave_the_nerd]

Seagate's problem is that they put zero thought into quality control. They rush crap to market as quickly as possible, as cheap as possible. Look at 8TB drives and see if you notice anything unusual:
http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&IsNodeId=1&N=100167523%20600376735
Seagate: $258
HGST: $539

Is it safe to assume there's a reason one is half as expensive? Was it tested, or did they push it out as quickly as possible? The Xbox 360 was rushed to market with little or no testing just so it would hit the market before the PS3, and we saw how well that turned out:

That's exactly what Seagate does. They want to be the first to have a certain hard drive on the market. The only way to do this, aside from being a generation ahead in terms of technology, is to cut out all of that expensive and time consume quality control testing.

The Hitachi is available for about $100 less elsewhere. At least as far as I can google in five minutes, it looks like:

1) It uses Helium, the Seagate doesn't. (More expensive to make.)

2) It doesn't use SMR, the Seagate does. (Faster than the Seagate.)

I think it might use more platters than the Seagate.

A faster, more complicated piece of hardware should command a bit of a price premium, imho.

 

[RecoveryForce]

From what I've read on your forum, a lot of the ST3000DM001s that come into your lab have crushed heads, which cause the failure. You mentioned a low head flying height somewhere, but doesn't that apply to most other Barracuda/Desktop HDD drives which don't experience such high failure rates (e.g. ST4000DM000)? What makes these drives particularly vulnerable to this kind of failure?

Just as Western Digital had significant issues with PCBs when they transitioned from a marvel chipset to the currently used ROYL chipset, there are always learning curves. Over the past year, I have found that not all Seagate DM issues are as bad as they look, but can very quickly become a reality of not handled correctly. So, this is how we currently assess Seagates:

- Does it spin up and stay spinning?
- If nothing, PCB
- If sounds like trying, stuck spindle or weak PCB
- If spins, clicks and spins down, heads and/or media damage
- If spins, clicks and stays spinning, possible weak head and/or firmware issue

- What is the terminal output showing?
- damaged media cache?
- damaged translator?
- weak head(s)?

As of right now, we are seeing less than 10% DOA fatal head crashes coming in with these drives (most of which are the 3TB capacity) and probably another 10-20% with weak heads that require head changes. The rest are usually recoverable with some firmware tweaks using PC3000.

How does the design and operation of the ST3000DM001 differ from the WD/HGST/Samsung/Toshiba drives? Are there any particular 'sub-models' of the ST3000DM001 (e.g. particular type of firmware/number of platters/case design/9YN166 vs. 1CH166 vs. 1ER166) that fail more than the others?

We see all brands and models of drives and they all have their issues. I'd focus more on backups than to figure out which is the one less likely to crash.

Within the Seagate series 9YN, 1CH and 1ER series, the 1ER series has a locked down firmware that my lab has yet to find a work around. So, on those drives, even with what we are sure is a firmware issue, we cannot resolve just because Seagate has locked us out from fixing it. I haven't noticed a huge difference between the 9YN and the 1CH.

I started to log Seagate DM series drives, just to see if I could spot a pattern. Being so busy, I failed to keep adding and haven't had time to try to compare the issues with each series drive.

I'm not sure that this really answers your questions fully. But hopefully it helps a little.