How to verify Seagate hard drives running hours after used-sold-as-new scandal
A report from the German technology company Heise suggests that some Seagate customers have purchased new hard drives that were used for tens of thousands of hours already.
The details:
- At least fifty Heise readers have reportedly purchased used Seagate hard drives instead of new hard drives in the past couple of weeks.
- Drives were bought at different retailers, some of which official Seagate partners.
- Seagate Exos hard drives appear affected the most.
- Some drives appear to have been in use for at least three years and with tens of thousand of running hours.
The problem
Some readers who experienced the issue noticed signs of wear on the devices. A check of the drive's SMART values however returned nothing out of the ordinary according to reports. Usage was close enough to that of newly bought hard drives.
Heise reports that the checking of SMART values won't reveal the used nature of the drives. Users need to look at Seagate's FARM values instead. FARM stands for Field Accessible Reliability Metrics.
It is unclear how widespread the issue is and whether it is limited to Germany or other countries as well.
The following guide walks users through the steps of checking Seagate drive FARM values to determine the actual runtime of the hard drives.
Checking FARM values of Seagate drives
Only a few public tools are capable of looking up advanced drive metrics. One of the tools is Smartmontools. It is available for Windows, macOS, Linux, Unix and BSD.
The latest version of the tool can be downloaded from Heise's download server. Note that you need at least version 7.4, which is the latest at the time of writing. It is a command line utility. The installation on Windows adds several command files, which Windows users may execute from the Start menu.
It is furthermore possible to run commands directly. Here are the required commands:
- smartctl --scan-open: the command returns the hard drives.
- smartctl -l farm DRIVE: the command can only be run on Seagate hard drives. It collects FARM data.
Check the value of Power on Hours in the log file that gets created.
The second program that you may use is Seagate's own SeaTools application. You can download the latest version from Seagate. Again, it only runs on Seagate hard drives, so take that into account.
Closing Words
If you have bought Seagate hard drives recently, especially of the Exos family, then you may want to run the check to make sure that you got a pristine drive and not a drive that has been used for years in the worst case.
Now it is your turn. Have you bought Seagate hard drives in the past or do you prefer other hard drive manufacturers? Feel free to leave a comment down below.
Thanks for the article. For me, checking my 2 drives turned out to be easy and quick, and it was a huge relief when I checked the reported 2.6 years uptime for the oldest one against my purchase records and verified that the drives were indeed unused when I bought them. They were self-encrypting drives I picked up on eBay being sold as unused surplus, and they were both at virtually the same price as non-SEDs which is highly unusual, so I was wary enough to check SMART data and serial numbers when they arrived (separate purchases, months apart). When I saw this article, I was afraid i was hearing the sound of the other shoe dropping. Whew!!!
@negatron
those lower RPM drives make a lot of sense when you want to have lower power usage, since most of the power drain from a spinning drive comes from the need to keep them spinning.
Higher RPM = more drag and turbulence inside the drive (more heat & noise) = more power needed to keep the RPM constant (even more heat).
Even in a datacenter. Stepping down from 10k RPM drives to 5400RPM, if the increase in access times are tolerable.
At scale that means measurable savings, specially as cooling is far from free in a datacenter and has its own powercost. Easy to rack up (sorry about the pun) hundreds of watts on that 24/7. Kilowatts in larger setups. Just from saving a few watts here and there in each rackunit.
Just a bit of basic math: 5400RPM is still 90Hz ~ 11.1ms per revolution
10RPM is 166Hz ~ 6ms per revolution.
If those extra 5ms (worstcase) is not a big issue (raid does great bulk transfers, but with latency anyway), then the spinmotor can run at half the power cost (if not less) and you get drives that last longer.
On those high RPM drive you will often find that in practice the track to track latency can be worse than rotational latency, because moving that arm into position accurately for long swings is not trivial (specially since 10k RPM is a lot more “windy”), so you are “latency bottlenecked” on arm movement making further RPM gains minimal, on top of that already being on the diminishing 1/x curve.
Not even sure anyone makes 12/15k RPM drives anymore for the same reasons, it was a thing from a period before SSD came in practical sizes. Not worth it for the tiny gains when you can go SSD if you really need that extra speed.
Just the big sore toe. Durability of NAND flash SS, specially the multicell kind. Often enough for casual consumers (assuming 3 years between replacing), not good enough for busy servers, they can kill them in months.
Not sure about buying WD over seagate. At least outside the US. Heard some talk about different firmware for export (probably weakened crypto in the full disk encryption support) + various stories in the past about firmware overwrite attacks bricking drives, targetted at “WD passport” drives and probably other models too.
Seagate had some stories about NAS insecurites too, but that was due to the nas device, not the actual drives.
I’ve been scammed on a WD HDD as well. Sold as 5400rpm but seemed slow… I later ran CrystalDiskInfo (excellent, free, recommended utility) on it and found the drive was actually 4800rpm. I didn’t know anyone still bothered making 4800rpm drives, but I guess WD still has a prodution line somewhere…
No shocked apology when I contacted support to tell them about this, just a “so you want to return it?”
That said, I still buy WD over Seagate HDDs.
Hope they also did proper capacity testing.
If the SMART data is faked/wiped, then…
Just happy I do not have one of those drive in need of testing right now.
Barracuda 14TB, rated at 250MB/s sequential write, that is close to 16 hours to fill.
Then about the same to do the readback verification step. 31-32 hours test.
Going to be worse if it is a fake.
Now you’ve wasted both money AND time. I vote that we hang the crooks from streetlights, bridges, etc.
Maybe build a trebuchet and sling them against a tall concrete building at short range.
If you want a good Seagate drive, buy their drives for NAS or their drives for CCTV devices. You get what you pay for and there consumer line has never been fit for use.
Why would anyone use HDD when SSD is so much better in so many ways?
Storage, if you need lots of storage.
pHROZEN gHOST, HDD are good for storage capacity (and backup), yes-yes, not as overpriced as SSD either. SSD have a finite lifespan, SSD are not much use long-term for “heavy use” scenario either. They [SSD] die far too easy (wear-out), like weak struggling rabbit between claws: snap-snap. Mechanical HDD are also much more compatible.
Data leaks also can occur from SSD, so not very “trustworthy”, they need power-reminder or slowly forget like sieve, or old man-thing. Harder to recover data too from SSD if they malfunction, e.g. malfunction like those poor quality scribbling-writing from chatbot-powered agencies.
Albeit SSD are fast yes, quick-quick like hurry-scurrying of clan-courier with important message for me… as I am the most potent of potents. Good-ideal for boot-kicking of machine OS, less seeking.
Also, you can bludgeon an enemy with a HDD… so very handy indeed!
Martin, those Seagate models, were mostly-reused old-decay server HDD drives, yes-yes; not surprisin’ they have been refurbished by bad tinkerings. You only have two real options, yes-yes, for mechanical HDD: Seagate or Western Digital.
There’s likely some nefarious, dastardly schemings going on somewhere in the chain, yes-yes, and it looked like most were OEM drives anyway, not magnificent workings of manthings. Workmanship second-class tinker like scribble-scribbling by agencies instead of true good engineer fresh-new.
The last HDD drive I purchased this year was from Western Digital, most true, it was not fake-scrap used half-dead…
You will call-know me as master-king! Whatever I say-squeak, you do.
Knowingly selling used hard drives as ‘new’ is a criminal offence in most countries and could result in substantial fines, or even jail time.
I haven’t use a Seagate HDD in many many years, because though new, that one was garbage.
WD puts the manufacture date right on the drive though I suppose someone could forge the decal.
WD HDD and Samsung SSD for me too.
The WD’s are not lasting as long as they used to though. It seems like the faster they get at transferring data, the shorter the life span.
I’ve always had issues with Seagate HDDs. They are the most prone to errors when there is a power outage.
And even without a power outage, I get more Reallocation Even Counts than with any other drive with similar usage.
I look forward to Seagate providing a robust, honest response to these allegations, promptly reimbursing the customers they cheated, and naming and firing the staff responsible.
Ha! No, that was a joke. I fully expect Seagate to stonewall, issue a few wordsalad press releases, blame it on someone else, and hope it all goes away.
I can remember a time when Seagate was the biggest name in drives, but I haven’t bought one of theirs in decades. And after reading this, that’s unlikely to change. Currently I’m using mostly WD HDDs and Samsung SSDs.