"Your correspondent has a 1TB iMac with a 1TB external disk drive and is now feeling severely undernourished capacity-wise"
Sell the Mac, buy a real machine, buy a handful of 1Tb SSDs with the price difference.
Toshiba has reached the 10TB disk drive capacity level without the help of helium, providing OEMs with an alternative to Seagate and WDC. Its MG06 is a straight 25 per cent capacity upgrade on the MG05 with its six platters. Both are classed as enterprise capacity drives. The speeds and feeds are: 7,200rpm spin speed 6Gbit …
"40MB? My first personal computer used a portable audio cassette recorder for mass storage."
My first PC was an IBM AT with 10Mb hard drive, on that I had Lotus 123, Lotus Freelance and a word processing package - even with those installed I still had space to save my data files.
Ahh - the days when a programmer could program a chess game on a Sinclair in 1/2 k .....
My Heath H11 had a 256k 8-inch single sided drive.
To be honest, I sprung for a second drive for the low, low price of $500 when purchased with the kit ... and later I had paper tape (PC11), cards (CR11), and later still removable media hard drive (RK05) and DECtape (TU56). The H11 is probably the single best tool I ever invested in ... at least in the computer world. She still runs.
DEC kit was, and remains, the single best teaching environment for learning the concepts of computing. Shame the franchise was squandered away.
Same questions as for enterprise drives.
1. Warranty.
2. Rebuild times. Halve the nominal throughout number and divide it into the capacity. For parity RAID, divide by three. You're looking at a worst case of 36 hours during which you may need to engage the services of a priest. Do not try your own voodoo!
3. Vibration. Tolerances seem to have widened over the years. Consumer drives are shakier and noisier than they once were. This is a bad thing.
... but what if you are a priest already? or is that a case of finding a high priest, or a pope? :)
In all seriousness, though, there's a balance that will need to be struck between a bunch of factors and what's acceptable for the end user.
Another thing to factor into the calculation is noise and power consumption; more spindles means the appliance will drink more power, generate more heat, and need fans to dissipate that heat.
"Lots of bays with smaller affordable drives or smaller NAS units with a few of these monsters thrown in?"
Pros and cons to each approach.
Lots of small disks:- Less space 'wasted' on parity data if using RAID, cheaper per TB of storage, 6 or more disk NASes tend to be much more expensive as they're more business orientated, more disks means higher probability of failure.
Few large disks:- More space 'wasted' on parity data if using RAID, more expensive per TB of storage, small capacity NASes tend to be much cheaper, fewer disks means lower probability of failure.
In other words, I'm afraid there's no right answer, it depends on how much space you need, how much money you want to spend, and how much pain it will cause if you lose all your data.
There is no warranty service in Europe.
I had a 3TB drive fail within the officially declared warranty. The supplier refused to take it back as it was over one year and it was supposed to be Tosh Europe to handle it. Toshiba Europe sent it back to me as they do not do ANY WARRANTY on their hard drives. At ALL. In fact, it is possible to guess that as entering serial numbers/models in their warranty service website for Europe answers with a doodle on a stick.
So, no thanks. While I may use it for stuff like laptop drives things are deemed fully depreciated in a couple of years they are on my blacklist for any NAS and server drive purchases. No thanks.
To be fair, it is legally the responsibility of the *seller* to apply the EU mandatory 2-year warranty on consumer goods, not the manufacturer. Toshiba is not selling drives to consumers, so does not deal with warranties directly.
"Free of charge, 2-year guarantee (legal guarantee) for all goods"
"If goods you bought anywhere in the EU turn out to be faulty or do not look or work as advertised, the seller must repair or replace them free of charge or give you a price reduction or a full refund."
http://europa.eu/youreurope/citizens/consumers/shopping/guarantees-returns/index_en.htm
Sorry to hear that your supplier lied to you...
Sorry to hear that your supplier lied to you...
Most suppliers allow you to get an RMA only via their automated forms using account info from registration and/or proof of purchase. They also happily ignore Eu directive.
Example - More Computers and Amazon in the UK. More is a year for everything and the directive be damned. Amazon is even less - it is usually a month. After a month you are supposed to deal with manufacturer. However if the manufacturer does not want to deal with you - f.e. Toshiba or Huawei you have a dud on your hands and no recourse whatsoever.
Price is a bit steep for individuals at the moment. Actually, in 2015 I bought some 3TB HDDs for €130, so the price has eked up a bit, but is still reasonable.
Oh well, with these new 10TB beauties coming on the market, it means that I will be able to buy a few 8TBs in a few years when I need them for less than the €370 they are currently going for.
Finite, certainly, but how much? New discoveries keep being made, for instance: https://www.washingtonpost.com/news/energy-environment/wp/2016/06/28/the-world-is-running-dangerously-low-on-helium-this-discovery-reinflates-our-supply/ .
The world was supposed to reach the Hubbert Peak for oil in the 1970s, or so the thinking ran then. We aren't out of it yet, sadly.
When you take into account the full life time of both Oil and Electric cars, it becomes clear.
You can keep the electric cars until you have better batteries in mind.
Carbon footprint of E-Cars beat the heck out of oil cars, mostly because of the crap batteries mind you.
So there you have it, we're being led by the nose again to something our landfills will regret accepting, unless we all go into a serious recycling age, which we all know Corporate will decline.
"Carbon footprint of E-Cars beat the heck out of oil cars, mostly because of the crap batteries mind you."
Until you factor in the amount of electricity used in the refining process. It's about as much per tankful as to get the same range out of the electric car.
US 2014 usage of helium (from wikipedia):
Estimated 2014 U.S. fractional helium use by category. Total use is 34 million cubic meters.[122]
Cryogenics (32%)
Pressurizing and purging (18%)
Welding (13%)
Controlled atmospheres (18%)
Leak detection (4%)
Breathing mixtures (2%)
Other (13%)
So, balloons make up a part of that 13% other. Which means a small percentage of helium is used in balloons.
I'll keep using helium balloons to make people, especially kids, happy, thank you.
Just about every gas well produces significant quantities of helium, but most of it is vented rather than collected. It's still relatively rare and expensive though even if drillers don't think it's valuable enough to retain.
I'd still like to see industrial quantities on tap with MSRs.
Regarding the Hubbert peak: That was passed years ago. Don't look at the price of oil, look at the energy expended to extract it and you'll realise we're into the "tight oil" that wasn't worth extracting 50 years ago.
At the start of the 20th century, on average you'd expend about one barrel of oil to extract 100-150 barrels, By the end of the 20th century that barrel expended got you 25-30. Most new discoveries and fields are down below 10 barrels per barrel and things like alaskan tar sands are under 5 barrels per barrel expended (although paradoxically, nuclear generated heat would dramatically raise the rate of return and may end up being used when oil is a valuable industrial chemical feedstock instead of a relatively cheap primary fuel and feedstock.
got... Helium? [we can make it by building fusion reactors, yeah!]
Seriously, regarding Helium: it can leak through just about anything, so eventually that helium will leak out. The question is whether or not it will be in any significant amount within a few years' time.
I understand that the Helium atom is small enough to pass through metals. I don't know about glass, though. Plastic helium balloons deflate within a few days, whereas 'plain air' balloons stay inflated almost indefinitely.
So I have to wonder if a 5 year old helium-filled hard drive will still work properly. I mean, how often do you recycle drives on a computer?
"so eventually that helium will leak out"
Eventually. Long after the drives expire.
There's a difference between a thin permeable rubber membrane under slight pressure and a couple mm of aluminium with almost no pressure difference, in terms of the escape rate.
_IF_ we had MSR nuclear power technology (which would have been commercialised 30 years ago if not for one Richard Milhous Nixon), then the amount of available new helium each year would be _more_ than the volume of the US strategic reserve.
In that environment, MRIs would be cheap and airships probably revived.
I love the MTTF figures that all the disk vendors come up with.
2,500,000 hours Mean Time To Failure - that's 104,166 days, or 285 years. Disk drives haven't even been around 85 years, never mind 285.
Now I understand there's some complex predictive statistics going on, but given my (and I'm guessing your) experience of many drives failing in under 5 years (and a substantial number do), then if the mean is 285 years then they are predicting more than half the drives will still be working after 285 years. BOLLOCKS!
+ MANY.
RAID is ***NOT*** a backup method. (Not even raid 1 or raid 10!) All RAID does is provide for a level of data integrity and performance.
Sure, you can have your primary storage be a raid 1, 5 or 10 array- as long as there's a drive that you are backing that data up to on a regular basis, you'll regret it when a drive dies (or the controller! I've seen that happen too!) or the OS/controller/NAS appliance decides to eat the array, or you have a virus put it's drooling mouth on it and trash everything.
If you seal a drive tight enough to keep in the helium (the only thing thinner being hydrogen), why not just vacuum out all the gas?
I heard something about the read/write heads floating above the plates, and apparently the gas acts as a lubricant or something?
The size and technical difficulties of the read/write heads has been a limiting factor for spinning rust for some time now. Maybe we should focus on getting those heads more efficient?
"I heard something about the read/write heads floating above the plates, and apparently the gas acts as a lubricant or something?"
The laminar boundary layer would prevent the heads contacting the disk (in normal operation).
When you have fluid flow, the molecules right on the surface of the pipe/container/whatever aren't moving. Molecules NEXT to those move, but not as fast as the total air flow. If you plot the fluid velocity vs distance from the surface, you get an exponential curve. The characteristics are based on velocity and the viscosity/pressure/etc. of the fluid. And, it forms what's known as a laminar boundary layer. Outside of that layer you get 'turbulent flow' which generally moves at the speed of the fluid. With spinning disks, this 'flow rate' is actually relative to the disk speed.
Anyway,
https://en.wikipedia.org/wiki/Laminar_flow
https://en.wikipedia.org/wiki/Boundary_layer
I expect Helium to have different laminar boundary layer characteristics than "plain air" and this is probably why the drive spins faster. It may also be physically smaller using Helium [which might have a thinner boundary layer] because the heads float closer to the disk [my speculation].
It's good to see non-He drives improving, but what these articles neglect to mention is the equivalent density 2.5" drives. I may be mistaken, but I thought 2.5" drives were favoured as server drives nowadays. Or would that be compute servers, but not data farms?
You can now buy 5Tb 2.5" portable drives for under $140 on offer * (are they He drives?). The equivalent 3.5" capacity (by volume) should be 15-20Tb for around $420-560. Does this mean 3.5" drive tech is behind 2.5" drives? The price of 3.5" drives can be dearer than 2.5" drives of the same capacity, despite using lower density platters.
* https://www.amazon.com/Seagate-Backup-Portable-External-STDR5000100/dp/B01LZP2B23/
" The price of 3.5" drives can be dearer than 2.5" drives of the same capacity"
1: Demand
2: Because they can.
3: The bigger drives have better seek times and faster transfer rates (not enough to justify the price difference)
Hard drive pricing is STILL higher now than it was before the 2011 Thai floods
Drive makers used to compete with each other, which resulted in prices staying low. With a comfortable duopoly in the market (Toshiba is a bit player that got tossed some Hitachi bones to keep chinese competition regulators happy), there's no real incentive to keep driving prices down until SSDs get within danger distance (about twice as much as rust seems to be the knee point).
There's no new technology coming down the pipeline. Seagate and WD both closed their research divisions 5-6 years ago. HAMR is an engineering problem that still hasn't reached marketability, so in order to increase capacity you're going to see more platters and helium until it does. By which time SSDs might well have undercut them.
8TB SSDs are £2.5-3k at the moment. That's about 25-40% more than what I was paying for 2TB SSDs a year ago (£1920 for the last batch of SM863s with 4TB ones being quoted at £5800 and a 3 month wait). Today a 4TB Sammy 850EVO will set you back £1350 and blow the pants off any spinner on just about every metric.
When that EVO gets to aroound £500, Seagate and WD will be worried - and when the SM863 version (now £2000) gets down that low, they'll start panicing) (8TB tosh drive is £200, 8TB WD or SG enterprise drives are about £290). At that point you'll get to see how much they've been scalping recently.