1

When I say best I mean most reliable. I have a Drobo FS with 5 drives in it. I have been using older Seagate Barracuda 3TB drives but they seem to be failing a little too often for my taste. It could be the older drives are less reliable and the newer ones are better. I’d prefer not to have a “green” drive spinning at 5400RPM but I don’t need anything like a 10K drive. What would you put in your home NAS in the 3TB – 4TB size range?

@Backblaze

9 Spice ups
2

I use the same things we put in our backup target NAS at our office, datacenter, and at most client sites: HGST DeskStar NAS 7200RPM drives. They seem to provide better performance and fail less often than WD Red drives.

My Synology at home has 5x 4TB HGST DeskStar NAS drives in an SHR-2 array.

3 Spice ups
3

+1 for HGST DeskStar NAS drives (large capacity, not 1TB)

4

Hard drives can be a fickle thing. I have used pretty much every brand in every device, but the best life and least fail that I’ve heard about or seen reviews on the last two years is fujitsu.

5

WD Reds are pretty solid.

2 Spice ups
6

I’ve heard pretty good things about the WD Red NAS drives. Might give those a shot.

1 Spice up
7

Seagate all the way!

1 Spice up
8

Depends on the RAID level

In any parity RAID like 6 then the WD RED isn’t going to be reliable. In mirror or RAID 10 they will be much better

Whatever way you go the new WD RED Pro (which is their old SE range renamed) is an excellent drive. Its 7200rpm and has a URE rating of 10^15 so will perform well in parity RAID (if you have to). Put 4 of those in RAID 10 and you have an excellent combination

4 Spice ups
9

Hey there, sorry to hear about any issues with your older Seagate drives, but I wanted to let you know that we have since made some great improvements and updates to our products. Take a look at our current items and let me know if you have questions!

3 Spice ups
10

Keep in mind that unless someone is commenting about the hundreds or thousands of drives that they maintain and keep reliability statistics on, their comments are just anecdotes.

Google did a study in 2007 that showed that consumer drives and enterprise drives had basically the same reliability.

Backblaze did a similar study in late 2014 that showed the same result.

As for reliable brands and sizes, check out Hard Drive Reliability Update: September 2014

Their results show HGST drives to be the most reliable, followed by WD and then Seagate.

1 Spice up
11

WD Red Pro.

1 Spice up
12

WD RED

13

I use WD red they are a solid drive and great price

14

People keep quoting this “study” but in reality it doesn’t really apply to any real use scenario.

Backblaze shucks external drives (no one in their right mind would do this for any valuable data), gets consumer desktop drives (for them that’s fine, because of how their business operates and how it’s cheaper for them getting these drives instead of NAS or enterprise drives) and throw them in a huge array where the drives are prone to heat and vibrations. This is nothing like what the drives would normally experience in a normal use scenario, such as OP’s current situation.

You need a study with normal consumer drives used in consumer equipment before you can claim any sort drive is more reliable than another in normal use scenarios.

1 Spice up
15

Sorry Gareth, but I don’t agree with a few of your points.

Backblaze gets some of their drives by shucking external drives. I don’t see the problem here. The drives inside external USB shells are the exact same models that are available for individual purchase. (No, I don’t truly understand why it is cheaper to buy a drive with a case, power supply and SATA/USB circuitry wrapped up in a nice retail-friendly box than it is to buy the raw drive, but hey)

Backblaze also have hundreds of NAS and enterprise class drives in the study (eg WD Red, Seagate Barracuda XT)

You are right that this study is not in a normal consumer environment. But it is similar enough for the conclusions to remain valid. (Yes, there will be heat and vibration in the BB environment, but the servers are situated in a dust-free environment-controlled data centre, as opposed to a home NAS sitting on a shelf covered in dust…)

Saying that this study is not valid because the conditions are different is like saying a car reliability survey based on taxi cabs is not valid because a you don’t drive your car the same way as a taxi driver drives their cab.

The important thing here is the sheer number of drives included in the study. No one else has published the drive comparison results of a study of this magnitude. (Google deliberately left brand info out of their 2007 published results.)

16

Shucking is bad because it removes any warranty you had on that drive. No one wants a drive without warranty.

I disagree that the study is similar due to the fact that it’s just not. Having something in a server rack in a controlled environment vs having it in a home office is completely different. Yes the study is valid if you’re going into the same business as Backblaze, but it’s not the same for home users.

The analogy you gave isn’t a good representation of the situation. It’s more like comparing car reliability when you’ve got 2 cars, one driving on a highway for it’s whole life, and one driving on rural roads for it’s whole life. The rural car would be put under a lot more stress due to vibrations and a lot of stop/starting and would be expected to fail sooner than the highway car.

They’re used in different situations and as you can see in the results, some fail when in this particular situation. Until there’s a more in-depth study showing a lot more data (average lifetime of the drives would be fairly relevant I’d imagine) I’d say these results are inconclusive for OP’s use case.

17

WD Red or Purple, depending on your write vs read needs. I have Reds in both NAS units, and Purples in my Media Server.

18

I have a pair of Seagates in my QNap at home, no issues at 6 months or so. The power of a sale at Newegg! Really, for a low stress job like that, whatever is on sale is first on my list. I don’t buy beer that way, though!

At the shop I’ve got two stacks of WDC Red’s in external RAID5 arrays, bulk storage for VM’s to back themselves up to just for belt and suspenders. The new 30TB Segate NAS I just picked up is taking over that job, crazy pricing made me NOT roll my own. So, I like both WDC and Segate NAS drives, haven’t tried HGST, maybe my next NVR.

19

I swear by Seagate hard drives, but will not use their 3TB drives. Every single one failed, or started to fail within 6 months. Moved down to 2TB, and have had them running 24/7 for about 2 years now, not one issue. Never use ‘odd’ numbered anything - 768MB RAM, 3-core CPU, etc… you get the point. It’s bad luck (doesn’t apply to 1TB, 1GB, Single-Core, etc)

20

Definitely going to agree with the WD Red suggestions. A lot of the reviews online state how reliable it is, so I would definitely make it an option.