Used Seagate drives sold as new traced back to crypto mining farms

Julio Franco

Posts: 9,174   +2,102
Staff member
A hot potato: A widespread scandal involving used Seagate hard drives fraudulently sold as new has continued to escalate, with new evidence suggesting that the drives originated from Chinese cryptocurrency mining farms. The drives, many of which had logged 15,000 to 50,000 hours of prior use, were reportedly altered to appear unused before re-entering the retail supply chain.

The first reports of affected drives surfaced in January, when consumers noticed inconsistencies in supposedly new Seagate Exos disk drives used in data centers. The issue has since expanded globally, with over 200 confirmed cases across Europe, Australia, Thailand, and Japan. While Seagate maintains that these products did not come from its official distribution channels, the scandal raises serious concerns about unauthorized resellers and supply chain security.

An investigation by Heise suggests that the fraudulent HDDs were sourced from mining operations in China to mine Chia, a cryptocurrency that caused a surge in HDD demand before becoming economically unviable.

During the Chia mining boom, demand for high-capacity hard drives skyrocketed, leading to shortages and price surges. However, as the profitability of Chia mining declined, many operations shut down, flooding the market with used hardware. It now appears that some of these used drives were relabeled and resold as new, deceiving both retailers and consumers.

Although standard SMART parameters track HDD usage, these values were reset to obscure the actual wear and tear. However, a more in-depth check using FARM (Field-Accessible Reliability Metrics) values can reveal a drive's true operational history. Consumers concerned about their purchases can verify their drives using Smartmontools version 7.4+ (use command: smartctl -l farm /dev/sda) or Seagate's SeaTools software.

Retailers impacted by the scandal have taken different approaches to address customer complaints. Some have acknowledged that they unknowingly sold these manipulated drives and are offering refunds or exchanges.

Others have set up dedicated customer service portals to handle the issue, while a few insist on verifying the affected drives before providing compensation. Many retailers stress that they purchased these HDDs from suppliers they trusted, and they were unaware of any tampering before selling them to consumers. The Heise investigation cites several German retailers (Alternate.de, Wortmann, Proshop, Kosatec, Galaxus, among others), however they have also tracked similar used drives being sold in other European countries and regions.

Seagate has distanced itself from the fraudulent sales, saying that it did not distribute the affected drives. It urges affected buyers to report their cases via fraud@seagate.com. The company has launched an internal investigation and is collaborating with retailers and law enforcement to track the fraudulent resellers.

Permalink to story:

 
Though a terrible thing to happen.

15,000-50,000 hours usage out of a mtbf of 2,500,000 would seem to be a bargain of sold at used prices.
They were not sold as used drive. The title of the article specifically says they were sold as new (and at full price I assume). The MTBF isn't an accurate measure of how long a drive will actually last. Using that number would suggest a drive will last 285 years in operation. Obviously that's not going to happen. At any rate, these drives were most likely running for 1.7 to 5.7 years non-stop and then sold at full price.

PS. - Amazon does sell used Exos drives labeled as "refreshed" or "renewed" usually at half the price of a new drive. They are drives that were used in data centers and do not have their SMART attributes altered. They usually have 3-5 years of power on hours logged.
 
Title: "Used Seagate drives sold as new traced back to"

Article: "suggests that the fraudulent HDDs were sourced from mining operations in China"

Techspot, make up your mind - is it a fact, or for now, a theory?
 
Title: "Used Seagate drives sold as new traced back to"

Article: "suggests that the fraudulent HDDs were sourced from mining operations in China"

Techspot, make up your mind - is it a fact, or for now, a theory?

TS gave you a link , feel free to follow it- may need to translate if your german is not up to scratch

TS is right to be circumspect if not 100% confirm - may look like a duck, talk like a duck, but maybe framed like a chicken
 
TS gave you a link , feel free to follow it- may need to translate if your german is not up to scratch

TS is right to be circumspect if not 100% confirm - may look like a duck, talk like a duck, but maybe framed like a chicken
I'm concerned about the way TS expressed this to me, not what the link may contain. :)
 
Crypto mining farms? Hm, maybe try squeezing cryptowallets before sending it back?

Not that kind... Chia used "proof of space" method, HD space. It "plotted" files that you need to keep on your online Chia rig. HD space is for Chia what GPU power was for Ethereum while it was still in "proof of work".
 
15,000-50,000 hours usage out of a mtbf of 2,500,000 would seem to be a bargain if sold at used prices.

That isn't how MTBF works.

If you have a product whose MTBF is 100,000 hours, and you buy 100 of that product, you can expect a higher failure rate at the start of their lives (the beginning of the bathtub curve) and then 1 failure every 1,000 hours or 6 weeks (the flat portion of the curve), followed by a higher failure rate toward the end of their service life.

This would be squarely in the middle to latter portion.
 
That isn't how MTBF works.

This would be squarely in the middle to latter portion.

Well open to understanding something different, but IBM state MTBF aa exactly what I have interpreted and better yet, the drives have already gone past the early peek failure time. (If they have been abused is another thing outside MTBF)

Throwing less reliability and more drives as an example seems to be a poor example. Obviously the more drives you have the bigger a change of one of them failing
 
the drives have already gone past the early peek (sic) failure time

That was exactly my point.

Throwing less reliability and more drives as an example seems to be a poor example.

It was to easily illustrate the math involved. I will be more specific below.

Obviously the more drives you have the bigger a change (sic) of one of them failing

Which was my point. MTBF stands for Mean Time Between Failure; mean here being the arithmetic mean, or what most people call an average.

If you have only 1 of the product, it would follow that unless it suffers an early lifetime failure, it would fail (on average) after its MTBF. However, this averaging logic cannot actually be applied to a sample size of 1. Otherwise you'd be saying that on average these drives will survive for over 290 years. This is clearly not possible. MTBF is given for you to evaluate how often they will fail, taking into account how many of them you have. It has nothing to do with how long you have used a product for.

Once you are past the initial high failure rate early in a product's life cycle (which is what I originally pointed out -- these would be), they fail at a fairly consistent rate. If you have more than one of the product, it follows that (usually) they fail more or less at a consistent time apart from each other. Thus, you divide the MTBF by how many you have, and that tells you how often you can expect any one of them to fail.

However, this is also assuming that there are no interacting effects; for example, that one of the drives failing could never cause another drive to fail before it was going to. This is often not the case with these types of hard drives used in data storage applications, because they are usually set up in some kind of redundancy striping scheme. When one drive fails, it gets replaced (possibly automatically with a hot spare), and then the entire array gets subjected to an intense 100% read event (known as "resilvering") over every drive in order to rebuild the content that was on the drive that has failed, writing that content to the new drive. Thus, every other drive is now being subjected to more work, increasing their likelihood of failing, and the new drive is being subjected to a 100% write workload, increasing its likelihood of failing if it was operating as an idling hot spare the entire time.

For a concrete example, taking the 2.5M hours MTBF, and that you buy 200 of them, and assuming that one of them failing will not increase the likelihood of the other 199 failing, you could expect them to fail 12,500 hours apart (just under 18 months). This is far below the purported new warranty period -- a warranty that you will probably not be able to claim on, because you bought them from a dishonest seller.

However, these drives are nearing the end of their service lives, so they will fail more often than that. Furthermore, due to the situation described above, unless you are using them separately (e.g. one in each PC, or more than one in the same PC but not in the same storage group, you get the idea), one of them failing will increase the likelihood that the others fail, so they will fail even more often than that.

In short, even at used prices, if you're buying more than a handful of them, you got a raw deal. Which is what fraud is.
 
Last edited:
Back