Question Refurbished HDDs

[Igo69]

I want to get refurbished Seagate Exos X12 12TB for storage and backup. Just want to know if you have any experience with refurbished Hdds.
thanks

 

[damian101]

Almost all my HDDs are refurbished enterprise HDDs, they seem to be very reliable, and are faster than consumer HDDs because of shorter access times, but are also a bit more noisy because of that.
Just make sure you get the SATA and not the SAS version, unless you also plan to get a SAS HBA.

 

[aigomorla]

nonono seagate is like a the 3 gorges dam waiting to collapse.

Get HGST He series or WD Gold. (basically the same drive) Especially if your looking into refurbs.

I don't care if its even enterprise for seagate, if you look at backblaze seagate has the worst numbers.

The only way i would ever trust seagate is if i had at least a Raid1/10 or a Raid-Z2 array.... even then i would also make sure i had a spare as backup incase one died.

 

[Jimminy]

I've wondered what "refurbished" actually means. Do they actually dismantle and replace parts, or is it just a matter of dusting the old drive off, formatting it, and checking the s.m.a.r.t data? And of course, a nice new plastic bag.

 

[damian101]

I've wondered what "refurbished" actually means. Do they actually dismantle and replace parts, or is it just a matter of dusting the old drive off, formatting it, and checking the s.m.a.r.t data? And of course, a nice new plastic bag.

I'm pretty sure it's the latter. At least for the vast majority of refurbished drives.

 

[aigomorla]

I've wondered what "refurbished" actually means. Do they actually dismantle and replace parts, or is it just a matter of dusting the old drive off, formatting it, and checking the s.m.a.r.t data? And of course, a nice new plastic bag.

depends on the refurbisher...
they check smart.
Some vendors will leave smart data, others will wipe it.

They do not take apart the drive, as taking apart you need a clean room, and it can open a new can of issues.
But refurbished is used drives, used parts, nothing is changed, unless its refurbished at the manufacture level.

Also i do not recommend a refurbished drive unless its data not sensitive to loss, or you are running some form of redundancy.
Lucky tho is redundancy on refurbished is still sometimes cheaper then new, so it works out.
IMO id honestly rather run redundancy on refurbished then no redundancy on new.

 

[mikeford]

With crazy deals on externals, many people who don't need a lot of drives buy the ext and "shuck" the drive, open the case and pull out the drive.

Refurb doesn't have any legal meaning, most cases the seller does a quick format, checks the smart, and wipes smart data, then puts it in a bag. Good chance many smaller sellers are just taking orders and having the drives drop shipped from a larger seller, never touching the drives themselves.

Given the low cost of drives, and the high true cost of data, EVERYTHING should be backed up to second layer of storage.

 

[SimplyComplex]

Refurbs and "used" are the exact same thing. I've purchased a LOT of used drives, but I always verify the manufacture year. I do not buy used drives > 4 years old.
People can tell me spin time is all that matters until they're blue in the face - it isn't. I used to do pc repair, and I saw that chronological age was a bigger factor in hdd failure than hours of operation.
I saw 10 year old drives dying with a few thousand hours of operation and 3 year old drives were generally fine with 20,000 hours.

The only two hdds I've owned that failed, were both backup drives more than 10 years old, and had operating times of < 2,000 hours.

 

[mikeford]

Enterprise grade drives fail at very different rates than entry level consumer drives. Its rare that servers full of good working drives get replaced and sold before the 5 year warranty runs out. As far as I can tell these are great choices for smaller NAS users, at least that is what I think, and what I am buying.

It pays to look at reports of actual drive failures, I think I've seen something called Blaze, with good info on which drives to look for or void.

 

[Jimminy]

Blaze reports are interesting, but it is a report of past history. Seagate came in with lots of failures, but does it mean that their more recent drive are crap? I don't know, but if it were my company, I would have been working hard to improve my products.

In the past, I had some very disappointing GM cars, but I honestly don't know if their later modes are still as bad. Also had a very crappy hyundai, but a much later hyundai was as good as any car we've had, including toyotas.

 

[mikeford]

Nice thing about buying refurbed enterprise grade drives, the exact model I am buying is what is in the current reports showing history over the last 5 years.

 

[Igo69]

Updating this thread with new data from Guru3d & Backblaze:

Backblaze Outs 2020 Hard Drive Stats for HDDs - Reliability Increased

In 2020, Backblaze added 39,792 hard drives and as of December 31, 2020 we had 165,530 drives under management. Of that number, there were 3,000 boot drives and 162,530 data drives....

www.guru3d.com

 

[mikeford]

6 months ago I decided to go with refurb drives, its still my choice, but I am now a LOT more fussy about who I buy from and their warranty and return policy. Next batch I buy I suspect I will be even more fussy.

 

[nosurprises]

Blaze reports are interesting, but it is a report of past history. Seagate came in with lots of failures, but does it mean that their more recent drive are crap? I don't know, but if it were my company, I would have been working hard to improve my products.

In the past, I had some very disappointing GM cars, but I honestly don't know if their later modes are still as bad. Also had a very crappy hyundai, but a much later hyundai was as good as any car we've had, including toyotas.

I think Seagate drives are fine for the past few years (they were pretty poor). At around 1%, it's not bad -- though it's not as good as HGST's 0.5%; it's just depends on how much you want to pay for that or how much does 0.5% difference makes in your usage.

 

[mindless1]

Nice thing about buying refurbed enterprise grade drives, the exact model I am buying is what is in the current reports showing history over the last 5 years.

But how many hours are on them? I'm wondering where these drives are coming from, seems likely they could've been used 24/7 for X # of years then retired because the risk to keep using them was going up.

 

[mikeford]

But how many hours are on them? I'm wondering where these drives are coming from, seems likely they could've been used 24/7 for X # of years then retired because the risk to keep using them was going up.

Start with 2.5Mhrs MTBF, using the first 1% for a 90% reduction in cost isn't a bad deal. Nothing suggests its the failure rate causing replacement vs moving to the next gen in capacity.

 

[mindless1]

^ I was asking more what the specifics drives you received, already had on them for running hours.

MTBF isn't significant in HDD lifespan expectations due to how it's calculated. It does tell a bit about quality control, as might the warranty (if the new purchase price were no higher), but really useless for any real world purpose. Have you ever owned ANY HDD that lasted 2.5M hrs running 24/7? Of course not. That's 285 years.

It would shed some light on infant mortality rates but you'd typically already be well past that period with a used drive.

Suppose it is only 1% of a 2.5M hr. MTBF, that puts it at almost 3 years old. If that's the case, the reality is that it is probably closer to 40% through its (average) lifespan, which may still be a good deal depending on how you look at it, but at that point, it's going to have a shorter remaining lifespan than the average new consumer grade HDD. It is a bit ironic but I'd rather pay a little more for a consumer grade HDD that has a longer expected remaining lifespan than a used enterprise drive.

Keep in mind that I'm stating this within the context of this topic, that the desire is to get a used 12TB HDD. That is a significant capacity still, and a lot of platters spinning for (years?) while data tends to show that higher capacity HDDs with more platters have higher failure rates. Personally, I wouldn't do it. The last thing I would do is buy a few years old drive with a lot of platters, enterprise or not.

Here's a bit dated article from Backblaze, but still interesting reading:

Enterprise Drives: Fact or Fiction?

Last month I dug into drive failure rates based on the 25,000+ consumer drives we have and found that consumer drives actually performed quite well. Over

www.backblaze.com

 

[mikeford]

12TB drives might as well be moon rocks for my use, as I would need 5 or 6 of them to get what I get now with 8 smaller drives in two NAS units.

Capacity is key for my choices, what size do I need to accomodate my planned storage needs in a minimum 4 drive Raid 5 NAS. First batch was 4TB drives, second batch 6TB drives, and maybe at some point migrate the 4TB drive system to 8TB drives. What are the chances that 4TB will have any practical use in a Raid 5 min array in 3 to 5 years, close to zero I think.

Enterprise drive users have thousands of drives to cycle every year, value at time of sale could play as much a part as any other factor. 6 months ago I paid $53 per 4TB drive, when 8TB drives start to cycle out, the price in NAS user quantity will drop.