Backblaze hard drive review

I have learned Hard drives are a strange bunch and I dont think any review from anyone tells the whole picture.
I myself have had nothing but problems with WD drives and have for years. But seagate drives always last a long time for me without issue. I ran the consumer 20 1tb drives in my first NAS for over 3 years only one drive failure the entire time. My second build I did with 2tb drives I read lots of bad reviews on seagates so went with WD drives. Within the first year out of 14 drives I had 4 fail. Switched over to seagate again and failure rate is back to 1 every couple years.
Yet I then have friends that have nothing but problems with seagates and th WD run fine. I think a lot of it has to do with hardware/type of use/environment on what works and what doesnt.
 
This isn't a review, or a small sample size. This is data from 30,000 HDDs in a production environment. This is as good as it gets.
 
As I understand it, WD and HGST are completely separate despite the latter being owned by the former. All due to regulatory stuff.
 
WD sold the Chinese part of their Hitachi purchase to Toshiba. HGST 3.5" drives come from the Hitachi Thailand factory that WD got as part of the deal; that factory was retooled after the flooding, and what is coming out of there since then is the really reliable stuff we're seeing now.
 
evilsofa said:
Nice, I hadn't seen the 2014 year in review report until now! Looks like the HGST 4TB drives win this round.
Click to expand...

Soon it will be the 6 and 8 tb drives.

Everyone can speak from personal experiences but there is a reason Seagate is forced to sell at a lower price. They did have a short stint at really terrible drives (I know, I bought a few of those 640gb drives) that plagues still then but their reliability is still sub par.
 
This is very much in line with my own experience of those three brands in both laptops and desktop machines.
 
evilsofa said:
Nice, I hadn't seen the 2014 year in review report until now! Looks like the HGST 4TB drives win this round.
Click to expand...

Sweeeet, my backup drive is one.
Strange the price has hardly moved in a year.

Two of my old drives still in service are Hitachi 1TB with 44K hrs on the clock.
Not a murmur of a problem, 30C sustained idle.
 
I'm not sure this is a fair representation of Seagate.

Perhaps if Backblaze hadn't farmed and crowdsourced all of those 3TB Segates as external drives from which they removed them from the enclosures and stuck inside their servers.

Or perhaps if they hadn't stuck all those consumer desktop-class disks in vibration heavy custom-designed rack cases with 24/7 load conditions. All these 3TB Seagates were taken from externals and then put into Backblaze 2.0 pods.

They made big improvements to vibration reduction in the pod 3.0 and 4.0 of which all those Seagate 4TB disks are housed in so it seems like that perhaps is what is making the difference.

Either way, the data is not conclusive if the issue is that the disks are bad or if the issue is how the disk were acquired or misused.
 
It is too bad backblaze does not mix HDDs within each pod. That would go a long way to removing some of the possible biases in their data comparing HDDs makes and models.
 
Well I can't blame them for not mixing as ANY data storage system always recommends using the same disk model in an array for best performance and things.
 
Since their arrays are decidedly NOT built for performance, that should not be a big concern.
 
FWIW we've had lots of problems with 3TB drives and won't touch them. Know this applies to Seagate, but believe it also applies to WD/HGST. Not quite as bad as the old 750GB drives; but close.

That said, it looks like BB is using consumer level drives. The mind boggles.
 
fastgeek said:
That said, it looks like BB is using consumer level drives. The mind boggles.
Click to expand...

They explained why they use consumer drives instead of enterprise drives in a blog post last year: enterprise drives had a higher failure rate than the consumer drives, even though they're being treated as well.

Also, I see today on amazon that 4TB enterprise drives cost $250 compared to $125 for 4TB consumer drives.
 
evilsofa said:
They explained why they use consumer drives instead of enterprise drives in a blog post last year: enterprise drives had a higher failure rate than the consumer drives, even though they're being treated as well.

Also, I see today on amazon that 4TB enterprise drives cost $250 compared to $125 for 4TB consumer drives.
Click to expand...

Keep in mind the cooling in those backblaze pods are terrible. They have also made the argument that cooling doesn't matter. On consumer drives, it probably doesn't because they are designed for low I/o high temp. Enterprise drives are designed for high I/o low temp.

In the little blog post about temperature they don't even mention any real enterprise drives, makes me laugh a little.

Even their drive reliability ratings reflect that the Seagate low power drives last longer than the regular Seagate drives, which is a contradiction of what everyone else will argue. If the drives afr is calculated at 25c, the drive should be kept as close to that as possible to achieve the measured afr. I would be willing to bet that if the 7200.11 drives were kept at 30c they would last longer.

I can guarantee the drives in that pod reach 40c-50c. They are using Dynatron 120x120x25mm 2600RPM Case Fan. Crappy Generic fans that push 96cfm. 3x120mm pushers and 3x120mm pullers isn't enough for the amount of drives they have in the enclosure. In other words they have 300CFM cooling the entire system, not enough for a motherboard, CPU, PSU, SATA card and 45 drives. They should be using fans push 250CFM or so each so they have 750CFM moving through the case.
 
Last edited:
backblaze report is completely flawed in their testing in so many ways, i would ignore it.
 
bds1904 said:
I can guarantee the drives in that pod reach 40c-50c.
Click to expand...

Even if true, that may not hurt the longevity of the drives. The famous 2007 google study of HDDs found that the lowest AFR was for HDDs in the 35C - 45C temperature range.
 
SirMaster said:
I'm not sure this is a fair representation of Seagate.

Perhaps if Backblaze hadn't farmed and crowdsourced all of those 3TB Segates as external drives from which they removed them from the enclosures and stuck inside their servers.
Click to expand...

That was happening only for few months after flooding. It represents ~6% of their storage, as explained in the current article.

SirMaster said:
Or perhaps if they hadn't stuck all those consumer desktop-class disks in vibration heavy custom-designed rack cases with 24/7 load conditions. All these 3TB Seagates were taken from externals and then put into Backblaze 2.0 pods.
Click to expand...

Again, no, they are not external drives. Or do you think http://www.newegg.com/Product/Product.aspx?Item=N82E16822148844 is an external drive ?

We did have a period where we used externals (https://www.backblaze.com/blog... ) but it was rather brief and it accounts for less than 6% of our entire fleet, and the stats from those mirror the internals of the same style.
Click to expand...
 
MrGuvernment said:
backblaze report is completely flawed in their testing in so many ways, i would ignore it.
Click to expand...
Care to elaborate? What I've read isn't testing; the link here goes to quantitative measurements from a medium-sized cloud deployment. What is flawed about that?
 
faugusztin said:
That was happening only for few months after flooding. It represents ~6% of their storage, as explained in the current article.
Click to expand...

6% of their entire storage fleet yes. They have 41,213 disks so 6% is 2,472 disks that are from externals. They only have 1,163 3TB Seagate ST3000DM001 disks, so they could ALL be from externals.

In fact, according to their data, they have had just 1,339 disk failures out of 41,213. For all we know, all of the failure could have been from external disks that they shelled. There are just too many unknowns to really get meaningful data out of this.

faugusztin said:
Again, no, they are not external drives. Or do you think http://www.newegg.com/Product/Product.aspx?Item=N82E16822148844 is an external drive ?
Click to expand...

Yes, they are. That is the disk model that came out of the 3TB Seagate external enclosure back in 2012 and still comes out of it today. Seagate has pretty much always used their desktop models in their externals.

http://www.newegg.com/Product/Product.aspx?Item=N82E16822178117

But it's well known that electronic companies bin their products. Lower binned HDDs end up in external enclosures which is also why they can generally be sold cheaper and are sold with a shorter warranty.
 
Last edited:
You must log in or register to reply here.
Back
Top