Consumer NAS drives - are Helium drives more reliable?

[Hornet]

I'm planning to get some WD Red drivers for a new NAS, and I'm wondering if the helium drives such as 8TB could offer better reliability than the conventional drives?

My understanding is that at least in the enterprise segment, they are marketed to be more reliable. WD rates their Gold helium drives at 2.5m MTBF as compared to 2.0m MTBF for the non helium ones.

Their Red drives aren't rated differently though, so I'm not sure if being helium drives generally contributes to additional reliability, or it makes no difference for consumer ones.

Would appreciate any opinions. Thanks

 

[Luke M]

I don't really know, but I think it would be hard for a helium drive to be more reliable on a per-drive basis, simply because it has more parts - more platters and heads - than lower capacity drives.

 

[sinisterDei]

When you're only dealing with single digit numbers of drives, unless there turns out to be some model-wide manufacturing defect (like with the 3TB seagate drives a while back) you're basically relying upon luck as to whether or not you will see drive failures.

Manufacturers have drive failure rates very very low, and if you own ten drives and one of them fails, that's a 10% failure rate - far higher than the average failure rate for the drives as a whole. You can take a look over at Backblaze's drive reliability reports and see the types of reliability they encounter 'at scale' and realize it doesn't really scale downward in any meaningful way. Their highest annual failure rate on a widely deployed drive is ~3% on a few-years-old model Seagate 4TB drive. Everything else is sub 2%, most of it sub 1%.

What I mean by this is that with sub 1% drive failure rates, you really shouldn't have a drive failure. But you still might, because you might just be unlucky. That won't have anything to do with the particular brand or model of drive you have though - they're *all* pretty reliable. That would just be bad luck rearing its head.

Backblaze does talk about Helium filled drives - here and here - but they only comment thus far that they are a bit faster and use less power.

TLDR - it shouldn't matter in a way that is statistically relevant to anyone not buying drives in bulk (thousands)

 

[SamirD]

All drives fail--it's only a matter of when. So having a failure plan and and working larger drives into that failure plan will give you the answer. Also, look at drive warranties--if a manufacturer is warranteeing one drive for 5 years and another for 2, that tells me something about longevity.

 

[Valnar]

Don't know about more reliable, but they're generally regarded as weighing less. The more you know...

 

[SamirD]

Don't know about more reliable, but they're generally regarded as weighing less. The more you know...

I wonder how much of a factor the helium has in the weight.

 

[Valnar]

 

[zrav]

I wonder how much of a factor the helium has in the weight.

Helium filled drives can have thinner platters, so it does have an indirect influence on weight. Of course, they then add more platters...

MTBF isn't a good mechanism for predicting drive reliability even in theory. In practice the stated values have little correlation with actual drive failure rates, the peculiarities of individual models or batches have a far greater impact than some abstract estimate. Statistics like the ones gathered by BackBlaze are far more relevant. Particularly interesting is that they found no increase in reliability of enterprise drives over desktop drives (*when operated in the same environment <- that's the key point).