WD's Monster 2TB Caviar Green Drive, Preview Test

MojoKid writes "Today Western Digital is announcing their WD20WEADS drive, otherwise known as the WD Caviar Green 2.0TB. With 32MB of onboard cache and special power management algorithms that balance spindle speed and transfer rates, the WD Caviar Green 2TB not only breaks the 2 terabyte barrier but also offers an extremely low-power profile in its standard 3.5" SATA footprint. Early testing shows it keeps pace with similar capacity drives from Seagate and Samsung."

Comments

Powers of 2

[Hyppy]

It's really only 1800 Gigs.

Re:Powers of 2

[wild_quinine]

It's really only 1800 Gigs.

Ah, the drivemaker's kilobyte...

Re:Powers of 2

[Anonymous Coward]

Its nice they have that 32Mb of cache. way to go mixing your references.

Re:Powers of 2

[digitalderbs]

The capacity hasn't changed -- only their definition of a terabyte.

Re:Powers of 2

[corsec67]

You mean 1800 Gibibytes?

Re:Powers of 2

[pushing-robot]

Yes. Curse those evil companies, trying to replace our God-given units—like Furlongs, Hogsheads, and Binary Thousands—with evil, communist SI units. The fiends will stop at nothing to pollute the American way of life!

Re:Powers of 2

[jedidiah]

Yeah... nevermind units that fit in with what's being measured
or are computationally convenient. What we really need are
arbitrary metrics created by beaurocrats on a power trip.

Re:Powers of 2

[Anonymous Coward]

Err, no.

The new standard replaces kibi, mebi and gibi with cabbage, mofo and giblet.

Or do you think that's silly? Well, so does everyone else. Overloading the SI prefixes worked well for years (pedants and idiots notwithstanding) and we don't see any reason to change.

Re:Powers of 2

[fprintf]

The funny thing is both of my home computers have drives smaller than the missing 200GB from this 2TB drive. I really need to upgrade soon...

Re:Powers of 2

[pushing-robot]

Congratulations. You have just stated that powers of 10 are "arbitrary metrics created by beaurocrats[sic] on a power trip."

Let me explain this in simple terms:

Powers are two are convenient for machines.

Powers of ten are convenient for humans.

It's bad form to present data in an inconvenient format for the user (which, presumably, is human) no matter how "computationally convenient" your algorithm may be. You can use binary calculations all you want behind the scenes, but convert it to a format designed for human comprehension before displaying it to a human.

Re:Powers of 2

[Anonymous Coward]

Or about 1.21 jigobytes...

Re:Powers of 2

[MightyYar]

Your argument would carry more weight if the manufacturers were doing this for the benefit of humans. In fact, they mix units - using the 1024-standard units for cache. Tell me mixing units is friendly! :)

Re:Powers of 2

[moderatorrater]

Let me explain this in simple terms:

Powers of two are convenient for machines and how most technically literate people think of the unit.

Technically illiterate people don't give a shit, you could measure hard drive sizes in cubits for all they know/care.

The difference between 1012 and 1000 are small enough that for the layman, conversions can be carried out as if it were a power of ten. For those who need to do more precise calculations, 1012 is the preferred format anyway. This is a clear cut case of trying to impose a standard by fiat where there's already and perfectly serviceable one in place.

Re:Powers of 2

[Lord Ender]

Personally, I'm glad the industry is moving toward using standard international units. This is 2GB and 1800 GiB.

The silly thing about the situation is that some people prefer to use the wrong units for no good reason other than tradition.

Re:Powers of 2

[sexconker]

Let me explain this is detailed, complicated-for-you, terms.

Exponents are useful for counting possible combinations. Computers are logically binary and quantum. When dealing with storage, we use 1024, because it represents the number of possible configurations of a 10-bit sequence.

2^10 was chosen for convenience - it was close to 1000, which people are used to working with, and it provided a good separation between major units.

1000 was chosen by SI for reasons just as arbitrary, namely providing a good spacing. We have scalar units of 10 (decimeter, decameter, for example), but no one ever says "Go down the road 1.2 deca kilometers".

SI units (major units based on a factor 1000, with shitty units based on 10 for completeness) are for measuring.

Computery units, (major based on a factor of 1024, with minor units based on 2 as the basis), are for counting.

This is why clock speeds use 1000, not 1024. Clock speeds are traditionally measured, and not counted, and they do not operate on a binary quantum system.

This is why data storage is SUPPOSED to be described using 1024, while data transfer is described using 1000.

If you want to get down to it, all SI units are retarded, since the universe is quantum (it is). All measurements are merely inaccurate tools of convenience, and everything should be counted in universal quantums of space (Planck Length? I doubt it, unless it really is tortoises all the way down), time, etc.

To sum it up - SI is not right because it's "official". SI is WRONG for computer science. And if the universe is quantum, SI is technically wrong for everything. Calculus, too.

Re:Powers of 2

[JesseMcDonald]

No, I'm quite sure the OP meant 1800 gigabytes, or about 15.46 terabits.

Established convention is that bytes are measured in binary (powers of 1024), and bits in decimal (powers of 1000). There's no need to introduce ridiculous-sounding terms like "gibibytes".

(Incidentally, I suspect there would be a lot less resistance to these newfangled units if they'd had the sense to pick names people could be expected to say with a straight face...)

Re:Powers of 2

[larry bagina]

I prefer 1024, but to each his own.

Re:Powers of 2

[mmontour]

Your argument would carry more weight if the manufacturers were doing this for the benefit of humans. In fact, they mix units - using the 1024-standard units for cache.

RAM specifications use the 2^x numbering because the device is physically constructed as a square grid of cells with power-of-two numbers of rows and columns. There's a direct mapping between bits on the address bus and the cell that is selected.

In the early days it was convenient to say that 1024 was close enough to 1000, so RAM sizes were quoted in "KB". However, the error in this increases with each step up in size. By the time you get to the TB scale it's no longer a reasonable approximation.

Magnetic storage does not have this constraint. The sector size is (arbitrarily) set at 512 bytes and hard drives usually have an even number of read/write heads, but apart from that there are no powers of two. The number of cylinders on the drive, and the number of sectors per cylinder, are arbitrary.

Communication speeds (e.g. "9600 baud", "100 Mb/s") are also not specified in power-of-two sizes, because they are natively dealing with individual bits and not power-of-two-sized chunks.

Therefore, there is nothing wrong with saying that a drive is "1 TB with 32 MiB of cache". As long as the manufacturer uses the SI and kilobinary notation correctly, users should not complain. Save your anger for the marketroids at WD who come up with features like "IntelliSpeed" in order to sell you a 5400-RPM drive and make you think you're buying a 7200-RPM one.

tl;dr version: Just use the damn GiB/GB notation consistently and get over it.

Re:Powers of 2

[morgan_greywolf]

Powers of ten are convenient for humans.

Hey! I have 16 fingers, you insensitive clod!

Re:Powers of 2

[the_one(2)]

Powers of ten are not naturally more convenient for humans. It's just that we are used to it. We should have used power of 16:S. That would be convenient for both humans and machines!

Re:Powers of 2

[erroneus]

Okay, I suppose you are correct in your view. From now on, a foot is ten inches because a unit of measure not based on tens is inconvenient. After all, the metric system works exactly this way. Also, there will be ten hours per day, ten days in a week and ten weeks in a month... and of course ten months in a year which, for me, makes "December" far less confusing since the Dec in December means ten in the first place.

Yes, that is sarcasm. To me, a unit of measure, no matter how inconvenient, is a unit of measure. You do NOT change the unit of measure. You create a NEW unit of measure.

So to that end, I propose a new unit of measure for hard drives: LTC-KB

This would make the difference between KB and LTC-KB much more obvious. A KB is 1024 bytes. (A Kb is 1024 bits just so you know.) A MB is 1024 KB, or 1024 x 1024 bytes, and a TB is 1024 MB, or 1024 x 1024 x 1024 bytes. A LTC-KB is 1000 bytes, a LTC-MB is 1000 LTC-KB and a LTC-TB is 1000 LTC-MB.

Oh, forgive me for not mentioning earlier, but LTC means "Lie To Consumers."

Re:Powers of 2

[Anonymous Coward]

If you want to use powers of 2 to measure then please create a new unit. Do not hijack units a well-defined meaning, such as kilo, for something else, it's just insan

Re:Powers of 2

[Anonymous Coward]

As a general rule, metrics are arbitrary. Why not measure hard drive size for the purpose of sale the same way they are used? Do you have a good answer for this, or just that it would not be fair if some "bureaucrats" made companies describe their product accurately for their intended use? Seriously, just because government can be bad in some things does not mean government is bad in all things. The childishness of the latter attitude is a very destructive force in modern politics.

Re:Powers of 2

[LotsOfPhil]

Hey! I have 16 fingers, you insensitive clod!

At first I thought you did not use base ten. Unfortunately, no base can keep you from being a mutant.

Re:Powers of 2

[Al Dimond]

SI might be wrong for computer science, but SI prefixes have standard meanings. If we want prefixes that work better for computing (which we may well), then making new ones, just to be clear, is a good idea. Then if SI is wrong you don't have to use it, and you don't confuse everyone by using its terminology to mean something slightly different (which is much worse than using it to mean something very different).

Anyway, the power-of-two units make some calculations easier and many harder. Just because an N-bit MUX has 2^N inputs doesn't mean they'll all be connected to something. You have 4 384-byte memory modules, quick, how many kB? Um, what's 384/1024? 3/8 maybe? Having to mess with mutliplying/dividing by 1024 in the middle of back-of-the-napkin calculations where not every number is a simple power of 2 (even if many of them have lots of 0s at the end in binary, like 384 does) actually does suck unless you just give in and learn your multiplication tables in hex (if I was still doing driver programming I probably would have done just that).

Re:Powers of 2

[0xABADC0DA]

Let me explain this in simple terms:

Powers are two are convenient for machines.

Powers of ten are convenient for humans.

Powers of N are convenient for humans. Other societies have counted in base 20 (mayan,irish), 60 (sumerian), etc. People can even count in number systems with more than one base (5 and 20 for instance).. In contrast, binary computers are always going to use base 2 since it would be absurd not to.

So you have one side that's flexible and one that is not. The 'proper' solution then isn't to create two different units of measure, but for people to use a different base. Octal for instance... make thumb and big toe not count as fingers and you're all set (they could be use to mark powers instead).

Or wait, you weren't actually interested in solving the problem were you?

Re:Powers of 2

[Anonymous Coward]

And if the universe is quantum, SI is technically wrong for everything. Calculus, too.

Assuming the universe is binary. Check your premises -

Re:Powers of 2

[HTH NE1]

RAM specifications use the 2^x numbering because the device is physically constructed as a square grid of cells with power-of-two numbers of rows and columns. There's a direct mapping between bits on the address bus and the cell that is selected.

Magnetic storage does not have this constraint. The sector size is (arbitrarily) set at 512 bytes and hard drives usually have an even number of read/write heads, but apart from that there are no powers of two. The number of cylinders on the drive, and the number of sectors per cylinder, are arbitrary.

Now explain flash/solid state memory sizes and the "formatted capacity" of memory.

Re:Powers of 2

[Leafheart]

To sum it up - SI is not right because it's "official". SI is WRONG for computer science. And if the universe is quantum, SI is technically wrong for everything. Calculus, too.

And your rant has nothing to do with it. kilo means a factor of 1000 applied. No matter what the hell you are talking about. That is the beauty of SI, and that is how we all deal with it and where the confusion comes from. Now if you want to give another prefix for a power of 1024, go for it, just don't use what we ALL* know to be 1000.

Too say that when we move to computer kilo will stop meaning 1000 and suddenly become 1024, is the same as Microsoft implementations of a lot of things. Say it is one thing, uses the name we defined for that, and do another one completely different.

*by all I mean everywhere on the world that has heard or uses SI units.

Re:Powers of 2

[danwat1234]

NO, it IS 2,000 Gigabytes, and it is ~1800GiBiBites! The drive manufacture isn't lying to the people.

Re:Powers of 2

[Just Some Guy]

You mean 1800 Gibibytes?

I will never, ever, in my entire life, even once mean "gibibytes".

Re:Powers of 2

[corsec67]

Established convention is that bytes are measured in binary (powers of 1024), and bits in decimal (powers of 1000). There's no need to introduce ridiculous-sounding terms like "gibibytes".

So there are 8.5899 gigabits in a gigabyte when there are 8 bits in a byte?

Yeah, that is much more logical than Giga = 10^9 and Gibi = 2^30. Especially since Giga = 10^9 everywhere else, including, as you say, in gigabits.

Re:Powers of 2

[Larryish]

The thought of keeping 1800 gigs or 2 terabytes of information on one device makes me slightly queasy.

Re:Powers of 2

[pushing-robot]

Since when is "the number of possible configurations of a 10-bit sequence" relevant to anything? 1024 is It was picked as a rough approximation of 1000 in the early days of binary computers, because users couldn't make sense of the raw binary data and an accurate decimal conversion was too computationally intensive. It was a kludge from day 1.

And even in the "computer world", measurements don't magically work out to be powers of two. Clock speeds, network speeds, platter density—all of these are limited by the manufacturing technology and the laws of physics, not binary math, so calculations aren't fundamentally any "cleaner" in binary than they are in decimal. The only reason why binary is used in the computer world is that computers do math in binary (or, to be pedantic, hexadecimal).

It makes sense for computers to do internal calculations in their own native base. But by that same admission, it makes sense for humans to do math in their own "native" base, which is almost universally base 10. So when a computer presents data to a human, it should present it in the most human-readable format possible.

I agree with you on one thing: SI prefixes aren't right because they're "official". They're right because they let you do math in your head.

Re:Powers of 2

[Cowmonaut]

Except in computers. KB does mean Kilobyte in both SI and computer science, but it is NOT SI IN COMPUTER SCIENCE. A side affect of counting in binary and everyone is going to just have to live with it.

Re:Powers of 2

[HTH NE1]

Incidentally, I suspect there would be a lot less resistance to these newfangled units if they'd had the sense to pick names people could be expected to say with a straight face...

Like giga-binary-bytes? That is the long form of gibibytes.

It would be nice if we could say GB₂ instead of GiB, but we can't do that here since subscript tags were considered presentational and thus deprecated. Not that this page is under a strict DTD. Can't we at least have span.sub { vertical-align: sub; } in the stylesheet and allow GB<span class=sub> 2</span> ?

Re:Powers of 2

[Anonymous Coward]

You are correct in that computery units are based on powers of two, and that includes DASD. Anyone that claims DASD capacities are correctly expressed in powers of 10 should be modded "marketing shill".

Data transfer rates are traditionally expressed in powers of 10 because "modern" protocols, ( 1 start bit, 8 data bits, 1 stop bit), require 10 bits to transmit 1 byte, (or character) of data. So equating 1 baud to 1 bit, (and there are some liberties taken in that), a 300 baud dial up, acoustic coupler modem, could transfer around 30 characters per second as a max.

Re:Powers of 2

[AdmiralWeirdbeard]

i really dont mean this as a flame or troll, but is it really necessary for this post to be attached to every single article about a drive? yes. we know its not actually 2 gigs. Hard drives are all advertised thusly, anyone surprised by this... well, i dont know what to say to you.

Re:Powers of 2

[petermgreen]

Well binary units make sense for the size of solid state memories because they were almost always produced in power of two sizes (for good reasons, if they weren't the address decode logic would be a LOT more complex). Binary kilobytes also made sense for floppies since they were usually an integer number of power of two sized sectors.

And if your OS is already using a measuring system for memory and floppies doesn't it make sense to also use it for hard drives? MS clearly thought so. Unfortunately the hard drive vendors throught differently (whether for technical reasons or because it allowed them to advertise higher capacities is unclear) and so we ended up with two different systems in wide use.

Worse with each unit we go up the discrepancy gets worse. At kilo it was only 2.4% , at terra it's nearly 10%.

Re:Powers of 2

[Hyppy]

Once you get to the terabyte range, which I assume you were meaning to refer to, the difference is nearly 10%. That becomes a big deal.

Re:Powers of 2

[AdmiralWeirdbeard]

i just dont get where the surprise part comes in is all, i guess.

Re:Powers of 2

[Anonymous Coward]

Actually, it is 2TB but only a 1800 Gibibyte... You must be a windows user who does not get these correct terms shown on files?

Re:Powers of 2

[mdielmann]

Um, yeah, except that just about everything is stored on powers of two. This is as absurd as if hot dogs were sold in packages of 8 (or 1024) and buns were sold in packages of 10 (or 1000). AND they used the same term to describe both, until it got to the point where they could sell significantly less than was expected while using VERY small print to notify us of this change in wording.
There is absolutely NO reason to use base 10 numbering for computer memory of any kind, except that it allows manufacturers to use bigger numbers while selling less. The only mitigating factor is that now that they all do it, at least we're back to comparing apples to apples.

Re:Powers of 2

[mmontour]

Now explain flash/solid state memory sizes and the "formatted capacity" of memory.

At a low level flash chips have a row/column nature similar to DRAM, but there is additional complexity because some operations target a larger "erase block" rather than an individual byte.

Some embedded platforms present a raw interface to the flash memory and require the host operating system to provide support for bad blocks and wear-levelling. These would be specified in "MiB" along with a certain allowable percentage of bad blocks.

However, the more familiar approach (e.g. in a SD card) is to include an embedded microcontroller that presents a logical block interface to the host. This controller skips over the bad blocks, and also needs to use some of the good blocks to keep track of the logical-to-physical block mapping. Here it makes more sense to use SI notation.

I just checked an "8 GB" microSD card and found that it presents a capacity of 7969177600 bytes to the OS (before partitioning). So 0.4% of the rated capacity is not available to me. This is consistent with the typical fine print on the package, e.g. "Some of the listed capacity is used for formatting and other functions, and thus is not available for data storage" (from Sandisk's website). If the manufacturer had sold it as an "8 GiB" card I would be more upset because that would represent a 7.3% capacity loss.

Re:Powers of 2

[Anonymous Coward]

[...], since the universe is quantum (it is)

But maybe it's not at the same time. Did you observe it?

Re:Powers of 2

[Anonymous Coward]

No, I'm quite sure the OP meant 1800 gigabytes, or about 15.46 terabits.

Established convention is that bytes are measured in binary (powers of 1024), and bits in decimal (powers of 1000). There's no need to introduce ridiculous-sounding terms like "gibibytes".

(Incidentally, I suspect there would be a lot less resistance to these newfangled units if they'd had the sense to pick names people could be expected to say with a straight face...)

Because terms like gigabyte or google/googleplex didn't sound strange or funny to the average person first heard of those, either.

Re:Powers of 2

[Anonymous Coward]

Sorry did you say ONLY 1800GB?

Re:Powers of 2

[lgw]

There's no good reason to use *any* system beyond tradition. There was a brief period of human history when humans were trying to compute values to many significant digits without the aid of computers, and during that period a "powers of ten" measurment system made sense, despite being inconvenient for anything *except* computation. That's was over decades ago, however. If you can't multiply by 12 or 1024 in your head, use a computer.

Re:Powers of 2

[MightyYar]

RAM specifications use the 2^x numbering because the device is physically constructed as a square grid of cells with power-of-two numbers of rows and columns.

No, they use 2^x numbering because the data is stored in binary. A hard drive also stores data in binary, so there is no reason for them to use different specs for the cache and platter... except marketing.

2^1 = 2
2^2 = 4
2^3 = 8
2^4 = 16
2^5 = 32 ...
2^10 = 1024

Sorry, that wasn't meant to patronizing... I just wanted to emphasize that binary data lends itself naturally to being represented by multiples of 1024, which is technically wrong from an SI standpoint but convenient and traditional in computers. Hell, even the term "byte" is 8 bits for no good reason other than history. Why do they embrace "byte" but not the other historical definition - except when showing cache size? The answer of course is marketing... they can sell as 1.819 TB drive as one that is almost 10% larger. And, of course, all the other guys are doing it. Lying about cache wouldn't gain them much - 32MB would only become 33.6MB... not even 5%.

As an aside, the person who came up with the prefix "kibi" never tried to pronounce it, and if he did, he certainly didn't do it in front of other people! :)

Re:Powers of 2

[KonoWatakushi]

In this case, powers of 2 are more convenient for both the machine and user, as that is how virtually all software and operating systems display disk usage.

It is bad form to present data in an inconvenient format for the user, and using the SI units does exactly that.

Re:Powers of 2

[tchuladdiass]

Another way of looking at it is:
RAM is manufactured in units based on powers of 2, since anything else would cause issues for programmers (at least ASM programmers, back in the early days). And it may cause issues with other hardware (CPU, etc) which can have optimizations for powers of 2 (i.e. bit shifts instead of multiply & divide). So rounding off the power-of-2-based numbers in RAM to the closest matching SI units is a convenience for people (who wants to say "536.87 MB ram" instead of "512 MB").

But hard drives are not manufactured in power-of-2 units. Instead, they are sector-cylinder-head units (at least originally). Only the sector size is a power-of-2 unit, everything else is arbitrary, which leads to a hard drive size not exactly fitting either power-of-2 or -10 units. Even the old 5-meg hard drives weren't exactly 5*2^20 or 5*10^6.

On the other hand, hard drives are used to store data that is held in RAM. So in a way, if the amount of RAM taken up is (for convenience) specified in base-2 units, then it would make sense to specify the hard drive space used in the same base. Which is why it makes sense for the file manager to report all sizes in base-2.

Re:Powers of 2

[jedidiah]

Your claims are assinine.

Powers of 10 are NOT convenient for humans when dealing with
binary based quanities. That's why the kilobyte was originally
standardized on 1024 to begin with.

The machine can deal with anything.

It's people that can't keep track of every significant figure
in the decimal representation of 1*1024^3.

How many sig figs does the traditional notation require?

How many does the new SI nonsense require?

Re:Powers of 2

[Anonymous Coward]

1000 was chosen by SI for reasons just as arbitrary, namely providing a good spacing. We have scalar units of 10 (decimeter, decameter, for example), but no one ever says "Go down the road 1.2 deca kilometers".

No because you're mixing two prefixes there!

You do get mixed up prefixes - I often see 75cl bottles of wine, but 750ml bottles of Pepsi.

Re:Powers of 2

[RzUpAnmsCwrds]

To sum it up - SI is not right because it's "official". SI is WRONG for computer science. And if the universe is quantum, SI is technically wrong for everything. Calculus, too.

No one gives a shit whether your measurement system is 'correct', as long as it's consistent. SI is VERY consistent for a system that spans so many fields. That's why it's better than the US customary system.

This is why data storage is SUPPOSED to be described using 1024, while data transfer is described using 1000.

There's nothing inherent to today's storage technologies that requires power-of-two capacities. We're not even using a fraction of the address space we already have, so sticking to a power-of-two size doesn't have any real benefits.

SI is WRONG for computer science.

Oh, so Computer Science is so important that we get to invent our own units, and use the same names as established SI units? Please. If you want to use binary units because they are convenient, go ahead and do so. But DON'T call them "tera", "giga", or "mega"; these terms already have SPECIFIC meanings and you can't just hijack them.

Re:Powers of 2

[WGAF]

Numbers all by their lonesome mean absolutely nothing. From a quantum perspective the ssd is the future and the mechanical disk is definitely the past. The best drives on the market (at least one of) are the WD300GB sata 3.0 raptors. Put those in raid or whatever you like. With todays speeds I doubt the 2TB model will keep up.

Re:Powers of 2

[borizz]

but no one ever says "Go down the road 1.2 deca kilometers".

That's because you already applied a scalar modifier (kilo) to the unit (meter).

Re:Powers of 2

[turgid]

I can count from -512 to 511 on my fingers, you insensitive clod!

Re:Powers of 2

[locketine]

Uhhhhhhhhhh..... Kilo = 1000 times the postfix. Period. Shall I reemphasize that? Kilo will always mean 1000 times whatever your base unit is. Computer science needs to follow this rule just like every other freaking discipline on the planet. Do you honestly think CS is the only place where physical constraints suggest using a non 10 based numbering system. Of course not! The SI system is designed to make engineers lives much much easier. How do you expect a mechanical engineer to know that a kilobyte isn't 1000 bytes when it would mean 1000 in everything they ever do that's not related to memory/storage? I had the SI system memorized freshmen year of college and yet a college diploma later I still have problems doing calculations involving transferring data across serial connections because of the misuse of the SI system in relation to storage space. If you want to keep your powers of two change the name of a kilobyte, megabyte, etc. to a non SI name.

Re:Powers of 2

[Sancho]

Only they don't use the notation consistently.

If the cache really is measured in MiB as you posit, the label on the drive in the picture uses an incorrect label (it uses MB.) The issue here is that they are using the labels inconsistently.

And it's inconsistent with regard to what the OS reports, too, so people who get their 2TB drives and then plug it in and find out that the OS thinks that it's considerably less than that get angry.

Re:Powers of 2

[Anonymous Coward]

That was painfully unfunny, just like everything from xkcd.

Re:Powers of 2

[WCguru42]

If you want to get down to it, all SI units are retarded, since the universe is quantum (it is). All measurements are merely inaccurate tools of convenience, and everything should be counted in universal quantums of space (Planck Length? I doubt it, unless it really is tortoises all the way down), time, etc.

To sum it up - SI is not right because it's "official". SI is WRONG for computer science. And if the universe is quantum, SI is technically wrong for everything. Calculus, too.

Okay, you go and do your designs with your special quantum system and some math other than calculus and I'll go and do my calculations with good old Newton and calculus. Let's see who gets those bridge/building/car/whatever designs done first, cheaper and just as effectively. SI units weren't chosen because they are the be all, end all of describing the universe. They were chosen for, oh God here it comes, the convenience that they provided to nearly everything in everyday life. Now if you work in a field where you should be measuring in Planks lengths then by all means, do that, but don't be trying to tell me that calculus and SI units and their prefixes have no use to everyone else.

And if you think SI is so wrong then why would you be using the same prefixes for something that is oh so right in it's binary essence. Why not just describe date as a 10-byte instead of kilobyte, or a 20-byte instead of megabyte. Ever think of that, just as easy to understand if you understand that in base two going from 10 to 20 is a factor of 1024.

Re:Powers of 2

[pushing-robot]

Define "computer science", though. Nobody is talking about designing CPUs with a base-10 architecture, or building 1000-bit-wide buses, or padding 32 bits per pixel out to a nice round 100 bits per pixel.

Even if the software industry adopts SI units and prefixes, Computers will still operate using their own native bases, and computer scientists will still be using the most sensible base for their work.

On the other hand, everybody else has a much easier time understanding base 10. If the SI prefixes are adopted, when computers display information to "everybody else" the computers would convert to proper base 10 (which is, I hear, the official base of "everybody else").

Right now we've got a kludge of a solution that's neither quite binary nor quite decimal. decimal_number x 2^10 (or 2^20, or 2^30...) isn't natural in any base, and was only chosen because it was an easier-to-compute alternative to a true decimal conversion. Nobody's doing themselves any favors sticking with it. Computer scientists would be better served with the native base, and normal users would be better served with base 10.

Re:Powers of 2

[Anonymous Coward]

Agreed. With ever increasing capacity of hard drives, the disparity between actual and advertised would make one think that eventually manufacturers are going to have to start throwing down more "accurate" or "real" figures (base 10.) ~200 gigs is a lot.

Re:Powers of 2

[x102output]

"Hell, even the term "byte" is 8 bits for no good reason other than history. "


I thought THAT was the reason for this whole mess? Not base 2 and base 10 conversion.

That's why you always divide by 8, if the drive is in Tera bits, and that gives you TeraBytes.

Re:Powers of 2

[Propaganda13]

I'd like to donate $1 million dollars* to you for your post.

*it's inconvenient for me to count to one million so I'll just round that down to zero.

Re:Powers of 2

[pushing-robot]

Argument 1:

Converting operating systems to report data in base 10 would require rewriting a few functions.

Converting humans to understand data in binary would take at least a lifetime, a concerted effort from all educators worldwide, rewriting every mathematical treatise and textbook ever written, and unquestioning cooperation from all human beings on the planet.

Argument 2:

We can date human civilization back around 7,000 years. We can date the decimal system to at least 5,000 years ago.

We've been using binary computers for roughly 70 years. We may have forgotten about binary computers in another 70 years. ....

I agree with you—base 10 isn't better than base 16, or 20, or 8, or 2, or 37. But it's more practical. Everybody is familiar with it.

That's the thing about the world: You usually have to reject the perfect-yet-utterly-impractical solution for the somewhat-flawed-yet-works-for-everybody solution. Or, in a word, compromise.

Re:Powers of 2

[DoubleReed]

computers do math in binary (or, to be pedantic, hexadecimal).

I'll see your pedantically discriminating between binary and hexidecimal, and raise you pedantically pointing out that hexidecimal is just a display format for binary data.
We may as well say this comment is written in English*
*: actually Times New Roman

Re:Powers of 2

[Samah]

To sum it up - SI is not right because it's "official". SI is WRONG for computer science. And if the universe is quantum, SI is technically wrong for everything. Calculus, too.

To sum it up, American in denial of metric's superiority to imperial. Multiples of 12? C'mon...
I agree with most of your argument though. :P

Re:Powers of 2

[EvolutionsPeak]

I will never, ever, in my entire life, even once mean "gibibytes".

Ack! What did you mean there?

*head explodes*

Re:Powers of 2

[sexconker]

Uh, no.
Calculus is wrong regardless, unless the universe is infinary.
SI is wrong because of the method and reasoning used. It could only be right coincidentally if the universe was decary, for all attributes.

Re:Powers of 2

[stmfreak]

I once worked for a company that built multi terabyte file systems. They tried to gloss over the difference between Mega and Mibibytes saying customers don't know or care. Then I had to explain that our 60TB system was actually only 54TiB and at $10k per TB, they might wonder where their extra $60K went.

Re:Powers of 2

[sexconker]

Uh, computers weren't developed with the plebes in mind. Neither were SI units.

You'll still get blank stares when you ask people the difference between a nanometer and a micrometer. People have had it pounded into them enough to know what a kilometer is, yet they still wouldn't understand someone who chose to say 3 megameters instead of 3,000 kilometers.

If retards hadn't bent to the will of marketers, more people would be familiar with the concept of 1024.

I never said kilo would stop meaning 1000. Fucking point out where I said that.
I'm merely stating that morans need to learn to deal with the duality of K meaning different things in different contexts. Lots of things have multiple meanings. Deal with it.

Re:Powers of 2

[sexconker]

WRONG.
Kilobyte means 1024 bytes.

Just because some assholes in suits say so doesn't make it true. Standards are nice, but things making sense is much nicer.

Engineers have no say in science. They apply the science done by the scientists, after plugging in their numbers into the equations figured out by the mathematicians.

I'd expect a mechanical engineer to deal with mechanical engineering problems. If they're dealing with computers in any way, they should know that 1 KB = 1024 bytes.

If you're dealing with oxen, you should know that a 12 oxen team has 10 oxen.

If you're baking, you should know what a baker's dozen is.

If you're sailing, you should know what a nautical mile is.

If you're broadcasting NTSC video, you should know that shit isn't really 30, nor is it 29.97 fps.

If you're drawing circles, you should know that pi isn't 3.14159.

It took you until college to memorize the SI system? You got a degree, and you still can't remember to use 1024? There aren't enough Os in LOOOOOOOOOOOOOOOOOOOOOOL for you, sir.

Re:Powers of 2

[sexconker]

Everybody else will have a much easier time using 1024 when dealing with computers.

When is it ever a problem other than when people forget the difference? Do morons really need help calculating things? 1024 MB = 1 GB after all. Pro tip: Most computers can do calculations for you.

Memory is byte-addressable, after all.
Clusters for storage are obviously using powers of 2.
Characters are 1 byte. (2 for foreigny-type stuff.)

Asking scientists to dumb it down is retarded.
Asking them to lie to users is retarded as well.

It's only an issue when some moron messes up and needs something to blame their mistake on. What, did you think 0x23 meant twenty-three? Does 777 mean seven hundred seventy-seven? No one ever bitches about hex, octal, or other schemes. It's always decimal cry babies who fucked up a calculation because they were lazy / out of their realm.

You don't see any computer scientists bitching about how they divided by 1024 instead of 1000 and fucked up.

Re:Powers of 2

[sexconker]

Yeah, you're wrong.
Permutations/combinations is fucking essential in computer science, el oh el.
But hey - that 8 bit color probably gives you great graphics, I mean, that's like, a lot of colors, right?

I already mentioned that the physical measurements are done in decimal. All logical (a bit is a logical element, not a physical one) stuff in a computer is (at the lowest level) binary. Clock speed is, decimal sure. Network speed is decimal, sure, when you're talking about BAUD. Storage density and network speed (in bits) is MIXED, and you've just opened up a whole other can of worms.

Hex is a representation for binary.
All actual work on any binary cpu is done IN BINARY.

Completely, thoroughly, wrong.

Re:Powers of 2

[sexconker]

Why did you post as AC?
You know what a fucking baud is, and you could've posted this in reply to one of the morans who talked about network speed using classical measurements.

Re:Powers of 2

[sexconker]

I observed it twice.
It was 0, then it was 1.
Also, it looked pretty granular.

Re:Powers of 2

[sexconker]

So which is it - based on 10 or 1000? Or 100?

Or am I supposed to forget about deci/deca centi/heca?

No reason you can't mix things that are just scalar multipliers. It's called math.

Seems to me 1 KM * 1 KM = 1 SQUARE KM, and not 1 K SQUARE M.

Re:Powers of 2

[sexconker]

You want to use our computers? Then adapt.

You know nothing about computers if you can't understand why they operate on a binary scale, and why it is correct to refer to things as such, and downright retarded to use something that is of a different base.

Lots of fields have things that mean different things.

Re:Powers of 2

[sexconker]

Scalar * scalar = scalar.
Also, 1.2 is a scalar.

12 (scalar) kilo (scalar) meters (unit).

Or does basic math now fall prey to the "standards" ?

FUCK morons who don't understand basic math and science.

Re:Powers of 2

[sexconker]

Exactly.
So when I'm working with computers, you'll fucking excuse me for using base 2.

Early computer scientists picked what they wanted. I'm sure some guy out there picked out the term gram, and some guy picked out the term bit, and some guy picked out the term giga, and some guy picked out the term bushel, and some guy picked out the term second, and some guy picked out the term hogshead.

So what if computer scientists used something that was already used for something else? That's never happened before, right? No one has to know the difference between two terms or words that look or sound the same, right?

Re:Powers of 2

[sexconker]

The metric system is better than the imperial system, sure.

I'm still waiting on metric time myself.

Re:Powers of 2

[Anonymous Freak]

Funny, I've never used 'bits' to mean decimal powers, except in measuring bandwidth; in which case, it's always in decimal powers.

Likewise, hard drive manufacturers have standardized on decimal powers for all prefixes.

So when they say "2.0 Terabytes", they mean 2 trillion bytes. (Not 15.46 terabits; 2 trillion bytes. Which would be 16 trillion bits.)

However, most mainstream OSes use the powers in the binary sense, so "2.0 Terabytes" would be 2*2^40 bytes. (Yes, I know that's the same as 2^41, but a Terabyte in this sense is 2^40, so I write it as "two of that".)

Yes, I also disagree with the use of the "binary prefixes" mebi, gibi, tebi, etc. I'd like to do away with the *NEED* for two separate prefixes. Just have the whole damn computer industry standardize on one or the other, and get it over with. Either ALWAYS measure in binary exponent, or always measure in decimal exponent. (I'm all for doing away with decimal exponent, then getting rid of the "mebi, gibi" garbage.)

It just happens that when you convert from 'drive manufacturer speak' decimal to 'OS speak' binary exponents, you get the increasing divergence between 'claimed' and 'reported'.

Re:Powers of 2

[conureman]

1800 Gigs should be enough for anybody's porn.

Re:Powers of 2

[root_42]

In the early days it was convenient to say that 1024 was close enough to 1000, so RAM sizes were quoted in "KB". However, the error in this increases with each step up in size. By the time you get to the TB scale it's no longer a reasonable approximation.

I don't get where this is coming from... When I was young(tm), which was way back in 1990, we used KB to denote Kilobytes, which meant 1024 Bytes. And that's that. The HD-makers screw us up, and some other idiot comes running up and introduces this KiB/KB crap. WTF? Dammit, we DEFINED the KB to be 1024 Bytes, and there was no SI standard that anyone ever cared about. And that was in friggin metric Europe! So, get off my lawn and take your stupid KiB with you! :)

Re:Powers of 2

[Samah]

I'm still waiting on metric time myself.

I think you mean this. ;)

Re:Powers of 2

[setagllib]

If people are to use advanced means of counting on their fingers, they may as well just use binary directly, for 1024 discrete values assuming 2 states per finger.

Re:Powers of 2

[Ilgaz]

After buying a 2 Gig SD memory card for my phone finding out it is 1800 MB , I call the situation hopeless. 200 MB really matter on that level.

Even Apple adopted their fake GB to disk utility not to confuse users, it says 1800MB (2 TB) I bet.

Re:Powers of 2

[locketine]

Cross discipline calculations happen all the time. The SI system is there to make such calculations easier. ME's are involved in creating those disk platters in HD's and they do need to know what a KB is. The problem isn't in knowing what a KB is but rather in knowing that dividing 1 megabyte per second into 1 megabyte doesn't give you the time it takes to transfer a megabyte in seconds; you want to know why? Because transmission rates are written using the correct interpretation of the SI prefix. They use kibi's to say what you think a kilo is. Use a different term if you don't want to follow the rules.

Notice how all of those examples you just stated don't use an SI prefix yet most of those disciplines use SI units like crazy. Wouldn't it be confusing to only use SI prefix's correctly in certain instances? What if instead of a bakers dozen they called it a decabagel. That would literally mean ten bagels but because the bakers decided to make it mean twelve bagels everyone would assume they were buying only ten bagels. Do you see a problem with this? Yeah, I know, people should be happy they're getting two extra bagels.

Tell me how many bytes are in a megabyte without using a calculator right now. Now tell me how many bytes are transmitted in a second on a 1 MBps connection. Notice the simplicity of the second example. Even though that's a more complicated question the simplicity in units makes it trivial.

Re:Powers of 2

[icebraining]

Every person I know who owns a computer on a daily basis understands that 1MB = 1024KB. Most people I know never heard of megameters, or kiloliters. In fact, I rarely see those scalars outside computing context. It seems to me that it would be much more difficult to change people perception from 1024 to 1000 than the opposite.

Re:Powers of 2

[dangitman]

Yeah... nevermind units that fit in with what's being measured

Are you really that stupid? Hard drives aren't built on a binary tree (like RAM). Decimal measures do fit what's being measured. That's why hard drives have always been measured in base 10. i think the engineers who designed them knew what they were doing. It had nothing to do with bureaucracy, it was about standards and logic.

Re:Powers of 2

[dangitman]

This is why data storage is SUPPOSED to be described using 1024, while data transfer is described using 1000.

So, use a different prefix when you use binary units. Don't use a prefix that is used everywhere else for decimal units. Is that so fucking hard? There are already defined prefixes for this.

Re:Powers of 2

[tylernt]

To sum it up, American in denial of metric's superiority to imperial. Multiples of 12? C'mon...

Well, I'm an American that thinks imperial is silly and we should switch to metric, but to be fair:

Multiples of 12 does have the nice feature of always being evenly divisible by 3 -- something many multiples of 10 can't do without ugliness like .333333 etc.

Re:Powers of 2

[MightyYar]

No a terabit would give you 125 gigabytes.

Re:Powers of 2

[shutdown -p now]

Byte is not a SI unit (and neither, obviously, are any derivatives, even those using SI prefixes).

Re:Powers of 2

[Anonymous Coward]

Converting humans to understand data in binary would take at least a lifetime, a concerted effort from all educators worldwide, rewriting every mathematical treatise and textbook ever written, and unquestioning cooperation from all human beings on the planet.

Kind like... learning the metric system? Or like... teaching people to say 'mibibyte' without sounding like an idiot?

The reason why the Imperial units are so great is because the units give nice round numbers for what people actually use. 1/4th of a cup vs 5.91 centiliters. How many gigabytes do you get with 4 one gigabyte sims? It's 4.39. How is that convenient?

Ironically, your same reasoning applies to Imperial units. It takes just a few operating system routines to convert and display them to the user even though the actual units the program uses might be different.

Re:Powers of 2

[Lachlan Hunt]

It's unfortunate that operating systems still report base 2 sizes using the base 10 notation. Data transmission and storage capacities and measured using base 10 units.

One solution would be for OS vendors to report all file sizes in base 10. But the change in numbers would be more confusing for users than simply changing the unit to base 2 (i.e. KiB, MiB, GiB, TiB, etc.). Hard drive manufacturers should then just list both sizes, as in: 2TB (1.81 TiB)

Re:Powers of 2

[Splintax]

This is what causes confusion when comparing download speeds, which are usually represented in KB/s or MB/s (bytes), and line sync speeds, which are represented in Kbps or Mbps (bits).

Re:Powers of 2

[Splintax]

And if the universe is quantum, SI is technically wrong for everything.

This doesn't have anything to do with SI, this has to do with the inaccuracy inherent in any form of measurement. In scientific contexts you usually provide an uncertainty for your measurements (or that uncertainty is implied by the number of significant figures), so when you say a metal rod is 1.00m long you don't actually mean to say it's precisely 1 meter long, but that the length is closer to 1.00m than 0.99m or 1.01m.

It's not really fair to say that SI is 'wrong' because we understand that measurement is a way of approximating the value of something when we can't count out its exact value. That's the context in which SI units are used, and they're very useful in that context.

Re:Powers of 2

[Tatsh]

Good one. Lol

LTC-KB = Kilobyte but it means 1000 bytes. 1024 bytes is a Kibibyte (kilo binary byte) according to Wikipedia according to the International Electrotechnical Commission standards defined in 1998-2000. http://en.wikipedia.org/wiki/Kibibyte

Re:Powers of 2

[Tatsh]

If you buy a set of DVDRs, you quickly find out they are measuring totally wrong on the package. It is definitely NOT 4.7 GB. This is a wild exaggeration if they would stop using SI GBs to measure (GB = 1000 MB). So if a DVD holds 4.7 SI GBs, then it holds ~4.34 GiB (Gibibytes, and I HATE that word).

Legality: the manufacturer LABELS the item stating that they measure a GB to be 1000 MB and not 1024.

Hard drive manufacturers do the exact same thing. A few years ago they were sued for not doing so, and won because they agreed to label their packaging about 'their' measurement of sizes.

Re:Powers of 2

[Splintax]

Yes, 'kilobyte' means 1024 bytes, but the point is that this is abusing the SI prefix 'kilo', which means 1000. We all understand that in CS sometimes 'kilo' means 1000 and sometimes it means 1024, but it's unnecessarily confusing.

Yes, we should expect somebody working with computers to understand when 'kilo' means 1024 and when it means 1000, but it would be nice if we didn't have to.

Standards are nice, but things making sense is much nicer.

It makes sense to have a unit that means '1024 bytes', but it doesn't make sense to call it a kilobyte, because anybody who's familiar with SI units will assume that 1 kilobyte = 1000 bytes.

Re:Powers of 2

[Splintax]

You seem to be getting awfully upset about this. Perhaps you'd have better luck convincing us if you tried to calmly explain why you disagree, rather than calling us 'morans' (ha).

Re:Powers of 2

[Splintax]

SI is based on 10, but unless there's a good reason to we normally use the units based on 1000 (kilo, mega, giga, tera) because changing units every time you go up an order of magnitude is tedious and unnecessary.

There's technically no reason you can't mix prefixes, but the GP's point was that it's odd to use different units (centiliters and milliliters) to refer to the same volume (0.75 L). It's not wrong, just unusual and confusing for those who aren't familiar with the SI system.

Seems to me 1 KM * 1 KM = 1 SQUARE KM, and not 1 K SQUARE M.

Your math's off - if you're squaring 1km and want to express the answer in square metres, then you need to square the 'kilo' prefix as well, giving you 1 'mega-square metre'. Although it's more conventionally expressed as one square kilometre, like you said.

Re:Powers of 2

[erroneus]

That is absolutely lovely, but that actually proves my point. The unit of measure "kilobyte" was defined LOOOOOONG before 1998; LOOONG before 1988; LONG before 1978 and I would venture to guess long before 1968 even though a kilobyte in those days was ridiculously expensive and hard to imagine. But I'll see your one link:

http://www.tekmom.com/buzzwords/zdkilo.html

And raise ya three:

http://searchstorage.techtarget.com/sDefinition/0,,sid5_gci212444,00.html
http://www.answers.com/topic/kilobyte
http://www.pcmag.com/encyclopedia_term/0,2542,t=kilobyte&i=45822,00.asp

And there are a LOT more places that state what I learned over 30 years ago. Best not to redefine units of measure. It confuses things especially when referencing older works using those measures. It also confuses and sometimes even angers the consumer. But let's look at it another way -- one day, you just might end up with less than a gallon or litre of gasoline one day because someone decided it would be okay to change a unit of measure to something they found more convenient. But that shouldn't bother you should it?

Re:Powers of 2

[Splintax]

Yep, overloading the SI prefixes worked well for years. Maybe you and I understand the overloading, but many people don't, and it's unnecessarily confusing. I think the fact that a class action suit was brought against Seagate as a result of this confusion indicates that overloading the SI prefixes hardly 'works well'.

Re:Powers of 2

[lpq]

I try to be careful and use GB for 1024 and GdB. Has to be the right audience I suppose... :-)

Re:Powers of 2

[this great guy]

There is absolutely NO reason to use base 10 numbering for computer memory of any kind.

Uh, yes. It's called standardization. Don't modify the meaning of the prefixes to fit whatever you are measuring. If a vendor start selling storage devices internally using base 3, I don't want him to invent a new way of using prefixes ! However it's perfectly ok to create new prefixes, like Mi, Gi, etc.

Re:Powers of 2

[Tatsh]

Well, I completely agree. I hate the terms Kibi, Mebi-, etc. Stupid and random. I for one would be embarrassed to use those words anyway.

Re:Powers of 2

[Anonymous Coward]

But kilo, mega, giga, terra are SI prefixes... I can still use those, whether I'm counting seconds or apples or bytes.

Re:Powers of 2

[mdielmann]

What you call something and what you measure are two different things. For years, storage was sold in what we now KiB or MiB or GiB, and listed as KB, MB, and GB. When they changed the standard, they didn't change the wording, but the definition they used. Which is fine, if you tell consumers that's what you're doing. It would be no different than switching from the Imperial Gallon to the US Gallon and not telling you that I had. Especially if no one had used a US Gallon before this happened.
As for base 10, standardization doesn't require its use. Standardization would require that the notation be in base 10, not the measurements we used for that particular item, bytes in this instance.

Re:Powers of 2

[Anonymous Coward]

Hm, what if disks were sold as a decimal count (with SI prefixes) of disk blocks, each of which are 4096 bits (or 512 8-bit bytes (or "octets" to use ISO terminology))?

Your 6.00e10 octet RAID becomes a 1.17e08 block RAID.

This makes sense since generally speaking drives are block-addressable, rather than byte-addressable.

Re:Powers of 2

[sexconker]

There are no defined prefixes for it other than K, M, G, T.

Ki, Mi, etc. are bullshit that was made up because the morons in the "standards" bodies couldn't stand up to the fucking storage manufacturers. These "standard" prefixes will NEVER be adopted outside of the pretentious fucks on wikipedia. No computer scientist worth his silicon will ever adopt them.

Lots of things have two meanings. Lots of things require at least some semblance of knowledge. Deal with it.

Re:Powers of 2

[sexconker]

Exactly why I said
1 KM * 1 KM = 1 SQUARE KM
----AND NOT----
1 K SQUARE M

You seem to have missed the point - scalars are just numbers, and standard arithmetic math applies to them. They are not special, and there is nothing wrong with mixing them.

It's weird, sure, but no weirder than saying 3000 KM instead of 3 MM, or saying 1 tera thing = 1 million, million Watts, as the news so often does when explaining large numbers to the plebes.

Re:Powers of 2

[sexconker]

I've explained it thousands of times by now.
You seem to be awfully wrong about this.
Perhaps you'd have better luck understanding it if you took a computer science course.

Also, google "morans" if you think I've spelled it incorrectly. Image search would be best.

Re:Powers of 2

[sexconker]

Doesn't make sense to call a lot of things a lot of things.
Get back to me when you've sorted out flammable and inflammable, all homonyms, and why restaurants give you a bill and call it a cheque.

Re:Powers of 2

[sexconker]

It's wrong in the sense that they are using 10 and 1000 for factors, and that (if the universe is quantum) counting would be more accurate and more useful than measuring.

The universe may not be binary, but it's more efficient to use binary than 10 when representing it (assuming the universe doesn't happen to be decary).

Of course, this only matters on a quantum scale. Shit will still be x meters tall. If you want to get pedantic about it (and who doesn't when 1000/1024 comes up) counting states/quantum attributes is how you "measure" any quantum system.

Re:Powers of 2

[sexconker]

Transmission rates - in bits/second - are calculated using the CORRECT 1024 base.

When you are counting logical units you use 1024.
When you are measuring physical things you use 1000.

1048576
1048576 (max theoretical raw transfer rate)

I didn't use a calculator.
How hard is it to multiply things by 1024? Do it in steps of 2 if you have to.

Let's see how far I can get without a calculator or long pauses:

2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 65536 131072 262144 524288 1048576 2097152 4194304 8388608 1677216 33554432 67108864 134217728 268435456 536870912 1073741824

That's 2^30.
Most people can't do that, sure. Most people aren't engineers or scientists. Most people can't tell you the difference between a nanometer and a micrometer.
Either way - are you telling me engineers can't do math? Are you telling me they don't have calculators? Are you telling me their back-of-the-envelope calculations have to be that precise?

Re:Powers of 2

[locketine]

http://en.wikipedia.org/wiki/Kbit/s#kilobit_per_second
I'm sorry but you're wrong. There is a huge disconnect here in CS and one of my CS profs spent a half hour of class time making sure everyone realized it.

That's mighty impressive. I can only get to 65536 before having to pause. But how long did it take you to get to 1048576? Was it instantaneous? I doubt it.

You're right, most people don't know the extreme ends of the SI system but just about everyone knows what Mega, Kilo and micro means thanks to updated chemistry classes in high school.

Why is it important to make it difficult to do these calculations? What is it achieving? Why must you insist on misusing the SI prefix when you could do what the telecom industry did and invent the kibi,mibi,tibi prefix's for your beloved binary counting system and leave the SI system alone.

Don't assume what has been done in the past is the correct way to do it.

Re:Powers of 2

[Anonymous Coward]

I think you need some air. It's kinda close nested this far down.

Re:Powers of 2

[Anonymous Coward]

Actually, it's 11 bits. You have to include the wait bit after the stop bit. But 10% is close enough for most people.

Re:Powers of 2

[Samah]

Multiples of 12 does have the nice feature of always being evenly divisible by 3 -- something many multiples of 10 can't do without ugliness like .333333 etc.

True, but I like the fact I can increase the number by a magnitude by simply throwing a 0 on the end (or moving the decimal place). It's not as easy to do that with 12s.
Maybe we should use multiples of 60 for everything! Divisible by 2, 3, 4, 5, 6, 10, 12, 15, 20, 30...

Re:Powers of 2

[sexconker]

I actually have them memorized up to 16777216 just because I've used them so much.

Good ol' sixteen triple seven two sixteen.

And wikipedia is wrong again. They're just espousing the ibi bullshit in that article.

K, M, etc. can be either upper case or lower case, and it means 1024, when talking about bits.

In order of what you're most likely to see:

Kbps means kilobits per second.
KBps means kilobytes per second.
kbps means kilobits per second.
K alone implies B, unless otherwise stated. (Though this one is so horribly abused by marketing, I would just say assume it implies b.)
k alone implies b, unless otherwise stated.
kBps means kilobytes per seond.

ibi was made up by some group with no authority to do so. The field laughed and rejected it. It has only added to the confusion.

What was done is correct because ANY assignment of prefixes is for convenience, and the decisions are at least somewhat arbitrary. Since the original usage of 1024 along with the K, M, etc notation is functionally complete, it is correct. Since it is the original system widely adopted, it should be preferred over all others to maintain compatibility and reduce confusion.

Just because it's stepping on the toes of another field doesn't mean it's wrong. Just look at the mess we have in math and physics. Look at all the horrible attempts to resolve it using the greek alphabet, stylized characters, little squiggly things, odd notation, subscripting, superscripting, etc.

Re:Powers of 2

[dangitman]

Ki, Mi, etc. are bullshit that was made up because the morons in the "standards" bodies couldn't stand up to the fucking storage manufacturers

Why are they bullshit? They solve an important problem, therefore are useful. And what does it have to do with storage manufacturers? They had nothing to do with them becoming SI or IEC units.

These "standard" prefixes will NEVER be adopted outside of the pretentious fucks on wikipedia. No computer scientist worth his silicon will ever adopt them.

Nonsense. Serious computer scientists will. It's only whiny retards who live in their mother's basement who are so emotionally attached to the messy kludge from the past.

Lots of things have two meanings. Lots of things require at least some semblance of knowledge.

Yeah, so what? How does that make ambiguity in science and engineering units a good thing? It's completely stupid - not something that "requires knowledge" - it's something that requires being illogical and inconsistent. Two characteristics that should have no place in computer science.

Re:Powers of 2

[Anonymous Coward]

If I ever hear someone mention "gibibytes" again I'll strangle a beaver in retribution.

And possibly a squirrel.

Re:Powers of 2

[locketine]

How about IEEE? Are they official enough?
http://standards.ieee.org/reading/ieee/SB/Jan97/bearer_jan97.pdf#page=5
read the article on page 5. I'd find you the standards manual regarding this but that costs money.

Here's some more stuff:
http://physics.nist.gov/cuu/Units/binary.html (definition of correct system)
http://members.optus.net/alexey/prefBin.xhtml (history of the change)

NIST is also a very respectable source.

Re:Powers of 2

[sexconker]

Storage manufacturers created the confusion.

To anyone who card about 1024/1000 in calculations, remembering wasn't an issue.

When storage manufacturers decided to lie, there were numerous lawsuits because people were getting less storage than they expected.

ibi was introduced afterwards.

Logical and inconsistent? Using 1024 is perfectly logical (more so than 1000). Computer science has been completely consistent about it.

Your beef is that your precious letters were reused. Too bad. Happens all the time. Nobody owns them - not you, not the engineers, not the standards bodies.

Re:Powers of 2

[sexconker]

The IEEE is a joke.

No standards body has the right to decree random standards and expect everyone to follow it.

Re:Powers of 2

[locketine]

Being a computer engineer I'm required to follow their decisions just like electrical engineers. You'll have to use our drivers and hardware so be prepared for more confusion down the road... These aren't random in the slightest either. Good luck existing in the real world where people have to work together to accomplish things. I'm done.

Re:Powers of 2

[dangitman]

Storage manufacturers created the confusion.

No, they did not. The confusion was created when some geniuses decided to use "kilo" to mean 1024, when everywhere else it means 1000.

To anyone who card about 1024/1000 in calculations, remembering wasn't an issue.

Remembering is not the issue. It's when you get something written down, and it's not clear which unit is being used. Do you just divine which unit is being referred to?

When storage manufacturers decided to lie, there were numerous lawsuits because people were getting less storage than they expected.

Again, such bullshit. Storage manufacturers did not "decide to lie". When IBM made the world's first hard drive, it was measured in decimal units. Did IBM lie to themselves? It's not as if that drive was a consumer product. You don't have any evidence of an intention to deceive, because there is none. And why is it only the fault of the storage manufacturers? Many other measures in computer science are decimal, such as network speed. Is that some sort of plot to deceive you?

When storage manufacturers decided to lie, there were numerous lawsuits because people were getting less storage than they expected.

Any idiot can file a lawsuit, it's not proof of anything. If I sue you for shooting my cat, is that proof of you shooting my cat? None of those cases had a guilty verdict, or any findings of deception.

ibi was introduced afterwards.

That's called progress. The metric system was introduced after the "hogshead". Does that make it invalid or inferior?

Logical and inconsistent? Using 1024 is perfectly logical (more so than 1000). Computer science has been completely consistent about it.

No, it hasn't. As noted above, decimal measures are used in many areas. Are you just pulling stuff from your ass? You appear to have little grasp of the facts, and are basing your stance on nothing but old wives' tales and nostalgia.

Your beef is that your precious letters were reused. Too bad. Happens all the time. Nobody owns them - not you, not the engineers, not the standards bodies.

Having multiple arbitrary definitions for a unit is not a good idea. You can measure things in hogsheads to your heart's content. but engineers and scientists like having clear and well-defined units. I haven't seen any serious and rational objection to this - the only ones are "but we're used to this way" and "those mebi things sound funny." Nothing of substance, just the whining of people opposed to progress.

Re:Powers of 2

[sexconker]

You're not required to follow their decisions or use their standards.

If everyone listened to the IEEE, IETF, etc., we'd never get anything done.

Re:Powers of 2

[Repossessed]

There's nothing inherent to today's storage technologies that requires power-of-two capacities. We're not even using a fraction of the address space we already have, so sticking to a power-of-two size doesn't have any real benefits.

The hardrive may not need to be based on powers of two, but the filesystem does.

You are, however, somewhat correct about the hijacking of the powers of ten nomenclature.

That was quick.

[jcr]

Wasn't it only about a year ago that 1TB drives hit the market?

-jcr

Re:That was quick.

[Anonymous Coward]

Actually it has been two years. I believe it was a Hitachi in early January 2007.

Re:That was quick.

[petermgreen]

looks like it was arround january 2007 that they were announced ( http://www.macnn.com/articles/07/01/05/1tb.hard.drive.unveiled/ ). So that would put about two years between the announcement of the first one TB drive and the announcement of the first 2TB drive. I would expect a similar gap between the shipping dates.

Re:That was quick.

[Anonymous Coward]

At that rate, we're only one year away from the next size limit: The MBR partition table is limited to 2TB partitions (with 512 byte sectors) and can address at most 4TB (maximum start offset at 2TB, maximum partition size 2TB). Is your OS ready for GUID partition tables?

Re:That was quick.

[rolfwind]

No, 1TB drives hit the market about 2 years ago. In early 2007:
http://www.pcworld.com/article/128400/hitachi_introduces_1terabyte_hard_drive.html
http://en.wikipedia.org/wiki/History_of_hard_disk_drives#1980s_to_present_day

Re:That was quick.

[aqk]

Fukit, man.
I'm going back to punchcards.
80 bytes (or characters) - THAT I can still understand.
Now where's my green card...

You punks wouldn't know what I'm even talkin' about...
.

just what I need

[Coraon]

I was worried I would have to start deleting from my *cough* adult movie collection *cough* to make more room

Re:just what I need

[geekoid]

I just leave mine on the internet.

backups

[Reality Master 201]

What the hell do you do to back up your 2TB drive?

That much storage in a single unit seems kind of dangerous.

Re:backups

[Moridineas]

What the hell do you do to back up your 2TB drive?

2 other 2TB drives?

Re:backups

[Greyfox]

Another 2TB drive. They're cheap enough to keep a spare around. If you're paranoid about leaving it in the case you can pick up an external SATA or USB rig and only plug it on when you're running backups.

Re:backups

[Reality Master 201]

Probably from the same manufacturer, with the same failure rates.

Re:backups

[JCSoRocks]

I've got two 1 TB drives in my home server. I use software RAID to keep my data safe and happy. You pretty much have to buy these things in pairs if you care even a little bit about the data. Unless, of course, you're buying this to use as a back up.

Re:backups

[Xemu]

Another 2TB drive. They're cheap enough to keep a spare around.

The problem with that is of course that mirroring simply mirrors problems on the primary drive to the secondary drive. So you end up with two working drives with broken data on it.

It's the data you want backed up, not the drive.

Re:backups

[timeOday]

That much storage in a single unit seems kind of dangerous.

I never understood this argument. Say you have N drives each with capacity C/N (e.g. C=2TB, N=1 for this new drive, or C=500GB, N=4 as you prefer) and probability P of each drive failing in a given time interval. Your expected data loss is N*P*C/N, which is independent of N. So what's the gain from more drives?

Heck, assume you don't want the hassle of multiple partitions so you use logical volume management to concatenate the drives (simulating the larger disk). Since any failure kills the whole thing, it's even worse - N*P*C.

I guess maybe your are thinking of RAID5? But is this an enterprise-class hard drive? I'm not buying (or buying electricity for) 3x 1TB drives instead of 1x 2TB drive just to protect my PVR recordings. And since RAID (regardless of level) is not a backup, if the data is any more important than PVR recordings, you still need backups with or without RAID. So all RAID5 gives you is decreased time to recover from a broken drive, by making you buy a spare up front. Obviously decreased downtime is critical for an important server, but not for the vast majority of home PCs.

Re:backups

[corsec67]

Except RAID isn't a backup, so your data isn't that "safe and happy".

Re:backups

[Wdomburg]

A second drive doesn't necessarily mean mirroring.

Re:backups

[ChienAndalu]

Bittorrent?

Re:backups

[Anonymous Coward]

Not dangerous if its part of a RAID array...of like 5x 2TB drives...ooohhh yeah.....excuse me, I have to go change.

Re:backups

[Anonymous Coward]

Meh. There are at least three or four different manufacturers that sell big, cheap drives. So what's the problem?

Re:backups

[MightyYar]

I'll probably buy 2 when the price comes down a bit. I'll keep data that doesn't really need to be backed up on it, and then periodically clone it to the other drive. Things like my movie collection (have the originals), iMovie projects (have the original tapes), virtual drives. On 2TB, I could fit all of my 13-GB-each iMovie projects and all of my DVDs in whole. I currently archive the iMovie projects to DVD and use a shelf to hold the movies.

My Mozy backup for everything else is just under 300GB and I'm on DSL. If I were on FIOS, I'd have no qualms about backing up the full 2TB.

Re:backups

[dlcarrol]

2 other drives: one in the removable tray, one in the firesafe*. Swap monthly.

*firesafe not included

Re:backups

[Anonymous Coward]

It's amazing how this comment, in some form or another, is inevitable when RAID enters into any discussion.

Re:backups

[Anonymous Coward]

Stop deleting stuff. Forever.

Buying a game? Watching a movie? Sure, but first I make a copy of it on my drive, so I will never have to meddle with the original DVD.

Re:backups

[Anonymous Coward]

that's because same troll is trollin'. "Doo doo doo look at me, i'm an expert, i'm going to argue semantics with you hurr derrr durrrr."

Re:backups

[Domint]

I think it's worth pointing out that what you suggest as a solution implies RAID 5, which is not a backup solution. There was an article on slashdot in the recent past about someone making that same mistake, and losing their entire hosted blog site (if someone could link the article that would be great - I'm having a heck of a time finding it).

Re:backups

[Reality Master 101]

Just in case anyone is confused that this imitator is me, this is NOT me making this stupid comment. :)

[it's stupid because having the same failure *rate* does not mean they fail at the same time. Sheesh.]

Re:backups

[Anonymous Coward]

Well, its a relatively safe comment to make right now.. especially since a blogging site thought mirroring was enough... until their drives were blanked and the blank was copied to the mirror (It was on /., just can't find the link anymore). The thing with raid is that it really isn't a backup solution unless you are running something like raid 5+1 or similar.

Re:backups

[AntEater]

Well, based on my experiences, I would strongly suggest that you purchase some non-Western Digital hard drives for your backups. That is, unless you like being on a first name basis with their technical support agents.

Re:backups

[urbanriot]

What a thunderous grasp of the obvious, Captain obvious - I assume that's why it's not called BAID. I do, however, contest the second part of your statement as my data on my personal RAID-6 array has been very "safe and happy" for a couple years now and survived a few catastrophic drive failures without the use of an expensive backup solution.

Re:backups

[binarylarry]

hmm, nope.

No one was confused. Nor would they care if they were.

Re:backups

[ColdWetDog]

Actually a multi terrabyte RAID 5 drive is a nice bit of the backup solution. No, it's not the be all and end all of backups. You still need separate completely off line and off site backups. But since modern RAID boxes can tell if a drive is bad, you get to look at the blinken light, go "oops", pull the drive and plug another one in. Wander off to Slashdot for a few hours and poof. Your data is back. No muss, no fuss. I like that part.

You can never be too rich, too thin or have too many backups. At least I get the chance to do one out of three....

Re:backups

[Domint]

If I had mod points, they'd be yours. It boggles the mind how many people equate RAID to backup - RAID only ensures data availability, not integrity. If the bits get fubar'ed on one drive, that will propagate across the rest - leaving you with no data.

Re:backups

[Anonymous Coward]

Lots of tapes.

As for dangerous, dropping six 320 GB drives on your foot would do lots more damage than dropping a single 2TB drive. Perhaps they could use this as a selling feature!

Re:backups

[dougmc]

No, he didn't imply raid 5. (You did seem to infer that, however.)

Raid 1 would only require two drives. But having three lets you do raid 1, and then periodically remove one of the drives and store it safely offsite and then replace it with the other drive and let the raid rebuild itself. That *is* a proper backup -- though personally I'd just periodically update one backup drive with rsync or something similar with the other backup drive being kept offsite.

Re:backups

[Domint]

Then you are just biding your time until something really subtle happens that screws the whole thing up. For example. what happens if the filesystem becomes corrupt? That corruption is propagated across your entire RAID array, leaving you with 4+ fully functional yet useless drives and you'll be left kicking yourself for not performing routine backups.

Re:backups

[dougmc]

And it's amazing how many people don't understand why.

Re:backups

[berend botje]

If N=16, you lose one-sixteenth of your data. If N=1 you lose all of your data.

You are right the first scenario is sixteen times more likely to happen, so mathematically there is no difference.

For practical purposes I'd take the small dataloss...

Re:backups

[DaHat]

Use a Windows Home Server?

Re:backups

[Moridineas]

If you were replying to me, no, I don't agree with what you said. It does NOT imply or require RAID5 at all.

For instance, maybe those two drives are in a mirroring configuration. Maybe they are in two separate computers. etc.

That's how I run our small business backup. It's by no means perfect, but it fairly well guarantees minimal data loss in case of a catastrophic main server failure, main office burning down etc. It's also not prohibitively expensive...

Main office has a server.

Owner's house has a 24/7 backup server with two drivers in a mirror configuration rsyncing from the main office every night.

Office warehouse has another backup server (no mirror in this one) that also rsyncs every night.

The weakness of this is that if somebody accidentally deletes a file, and wants it back the next day, tough luck. We DO supplement this system with most workers having portable USB drives they take home. Obviously this is not possible or desirable for many offices...

Re:backups

[berend botje]

Or, the other extreme: delete what you don't need anymore.

Do you really need DownloadedSoftware v0.7 when the current version is 4.2? Do you even need that version stored locally?

Are you really going to see that movie again? Is ripping it so harddisk even worth the time? Deletion is your friend. No worries for data you don't have. Easier (smaller, faster) backups, too!

Re:backups

[Moridineas]

I've had worse luck with Seagate...

But given that I'm dealing with small numbers of drives, I don't think there's any statistical significance to my experiences.

Re:backups

[Lord Jester]

Depends on the level of RAID. I have yet to see this in any of my RAID 5 arrays.

Not arguing about the declaration that RAID is not a backup, but it is, by definition, redundant.

If I have critical data, it is stored in two (or more) locations (both electronic and physical).

Re:backups

[Ephemeriis]

I've got two 1 TB drives in my home server. I use software RAID to keep my data safe and happy. You pretty much have to buy these things in pairs if you care even a little bit about the data. Unless, of course, you're buying this to use as a back up.

RAID is not a backup!

RAID protects you from a drive failure or two, sure... But what if you delete something you shouldn't have? Or something important gets corrupted? Or you get a nasty virus? Or your house burns down? Or you get some kind of nasty lightning strike that takes out your entire RAID? Or you run one of Seagate's firmware updates?

A backup is a periodic snapshot, preferably stored off-site, that you can go back to if something horrible happens to your data.

You can certainly use HDDs as backup... Or even a RAID... But you need to move your data to those drives and then take them offline and put them someplace safe. Just simply RAIDing your working set doesn't give you a backup.

Re:backups

[Domint]

If you don't have the data replicated periodically outside the RAID, it's not a backup solution at all - I could, within the OS & filesystem, delete files arbitrarily and the RAID will happily remove those bits from all the drives. Thus, it is not a backup solution. If those bits aren't anywhere else, they are gone for good.

Re:backups

[Anonymous Coward]

Its not only smaller it is also more frequent, you lose less data 16 times instead of all the data once. 1 instance of downtime is a lot better than 16.

Re:backups

[Domint]

Ok, maybe I did leap to my nearest conclusion, but I've seen it happen enough where people assume that because they have RAID, they don't need to think about anything else for backup - your solution is a valid one though, since you are periodically ensuring that the data is stored outside the live array.

Re:backups

[corsec67]

Depends on the level of RAID. I have yet to see this in any of my RAID 5 arrays.

sudo rm -rf /

Doesn't matter how many levels of RAID you have, in that case your data is gone pretty quickly.

I have had something similar happen to me on a RAID 1 array. The ReiserFS partition got really messed up, luckily I had just backed up everything in that array, but the RAID didn't do anything to save my data.

Re:backups

[Anonymous Coward]

Wasn't there something on TheDailyWTF.com that this method killed the main drive real quick?

Re:backups

[Domint]

If a file gets deleted and is not recoverable, then it is not backed up. Period. If you are just replicating data to another drive, you replicate errors and corruption as well. I don't know about where you work, but if one of my coworkers came in and said "hey, I seem to have lost file such-and-such, and need it back" and I told them it's gone for good since they didn't come to me within 24 hours, I'd lose my job.

Re:backups

[larry bagina]

I've always had good luck with WD, and found Maxtor to be unreliable. I've also seen /. posters claim that Maxtor is the shit and WD is just shit.

Re:backups

[rthille]

That's a horrible idea. If you do that, there's a window where you have no RAID protection of changes to the filesystem. If you really want to safeguard that data, you need 4 drives. 2 in a RAID mirror, and then 2 more so you can be sure that all your data isn't at the same location, and you can keep your mirror of new data.

Re:backups

[nizo]

Though if your primary drive is spitting out bad data, it doesn't matter what backup media you are using.

Re:backups

[theJML]

I'm with you. I've had nothing but the best of luck with WD. Hell, I still have a 15GB drive that until a year ago was basically on 24/7 since I got it in '98 and haven't had any issues with it.

Even at the worst of times with WD, I'd have to say it's no worse than Seagate or Maxtor. Though recently I've moved to Seagate due to their SAS (and sorta SAS ES2 models) Superiority. I'd still put 'em in Desktops.

Re:backups

[ceoyoyo]

Set up an FTP server with anonymous access to your 2 TB drive.

Let's be honest: you don't have 2 TB worth of your own stuff. You've got 2 TB worth of porn and Bittorrented movies.

Re:backups

[chrysrobyn]

What the hell do you do to back up your 2TB drive?

Can I use it as my backup server?

I have an iMac with only a 500GB hard drive, so I figure a 2TB drive would store a lot of Time Machine differential backups. Of course, somewhere in there it becomes an archival server instead of just a backup server...

Re:backups

[Anonymous Coward]

They won't have the same failure rate if one is sitting unpowered and is only plugged in when you decide to make backups.

Re:backups

[NXIL]

http://hardware.slashdot.org/article.pl?sid=09/01/02/1546214

Re:backups

[blahbooboo]

My Mozy backup for everything else is just under 300GB and I'm on DSL. If I were on FIOS, I'd have no qualms about backing up the full 2TB.

Good luck with Mozy. Try to do a simulated catastrophic recovery some time, most people report they don't work properly or well. Their Mac client is especially bad. Do yourself a favor, get a different provider...

Re:backups

[Anonymous Coward]

mod parent -1 redundant.

Re:backups

[nxtw]

Well, based on my experiences, I would strongly suggest that you purchase some non-Western Digital hard drives for your backups. That is, unless you like being on a first name basis with their technical support agents.

I had a WD desktop drive show signs of failure a few months ago (SMART showed offline unreadable sectors.) I was able to get a replacement just by filling out a few forms on the web.
Seagate's the same way, but they charge a nonrefundable fee for advance replacement (where they send you a replacement drive first.)

Re:backups

[speedingant]

It isn't proper RAID.

Re:backups

[ACMENEWSLLC]

Mozy.com or Carbonite.com will back that up. Assuming just because you have a 2TB drive, you don't start out with 2TB of new data.

Restores are another matter.

Re:backups

[chrysrobyn]

I do, however, contest the second part of your statement as my data on my personal RAID-6 array has been very "safe and happy" for a couple years now and survived a few catastrophic drive failures without the use of an expensive backup solution.

I'd like to propose an alternate view of "safe". RAID-6 promotes catastrophic failures on the software side quickly -- like a virus or trojan (you're only as safe as your dumbest user at his drunkest and tiredest) or a 15 year old cousin's "rm -rf /" / "deltree /y c:\". RAID-5/6/0 is "safe enough" for a lot of things on a desktop, but no amount of redundancy makes it unequivocally "safe" when it's updated live.

Simply adding a NAS drive with automated hourly differential backups would increase the amount of "safeness" your data enjoys, in addition to providing some archival support. My wife had her hard drive die in her computer last year and within an hour of getting the replacement, the image was pulled off the NAS and she hadn't lost anything but a little browser history. Certainly your desktop is doing better than hers because of her single point of failure, but could you recover from a catastrophic failure as easily? Or are you relying on that catastrophe never happening?

There's a fine line between "safe and happy" and "lucky and dumb". I don't know where RAID-6 is with respect to that line, but I'm certain it's not very far from the other side of it.

Re:backups

[Anonymous Coward]

Alternatively, you could stare out the window when the inevitable RAID argument touches off any time Slashdot posts anything at all about drives, or backups, or any sort of failure of anything related to computers.

You'd be staring because you used ZFS and everything was copy on write with no holes, and you could add, remove, replace disks without worrying about your data.

Re:backups

[danking]

You can never be too rich, too thin or have too many backups. At least I get the chance to do one out of three....

Say that again when your critical organs start failing from a lack fatty tissue in your body.

Re:backups

[Blackknight]

I'd use RAIDZ2 with periodic snapshots.

Re:backups

[rickb928]

Manufacturers quoted failure rates and MTBF are well-thought out, but meaningless, statistical derivatives.

The only failure rate that means anything is the one measured by the weight of your trash can.

An MTBF of 200,000 hrs. claims that there will be some power-on failure approximately every 22.8 years. Funny.

Actually, every drive manufacturer can claim a failure rate of 100%. They all will die. What's the out-of-box, 1,2,3,5,7 year failure rates?

I left some 7-year old Fujitsu drives running back in 2005. Stern instructions to not let them get turned off and go cold. I didn't expect them to start again. Who knew the UPS would be so critical...

Re:backups

[Anarke_Incarnate]

RAID 5+1 is not a backup solution either. You have created a fault tolerant scenario. Backups are there for after the rebuild when you find out that there are additional issues. The issue is not just the part where a drive can be unusable. People have already given scenarios about what can happen to the data. Any "Backup" solution should keep the data in at least 3 places, one of which, should be portable. Tapes, DVDs, removable disks, etc that can be away from the site.

What good is your RAID if you just had a power spike take out the disks along with their motherboard? Small flood in the room took out your tapes and system? Data's all gone now.

Re:backups

[drgruney]

I remember people saying that about 15GB drives. And yes... I'm sure other slashdotters recall the same issue at smaller drive sizes.

Re:backups

[Joce640k]

"What the hell do you do to back up your 2TB drive?

That much storage in a single unit seems kind of dangerous."

I'll use one to back up my 1Tb drives.

Then, when it's proven itself and 4Gb drives appear, it might go into my desktop machine and be backed up onto the 4Gb drive.

Re:backups

[Grapes4Buddha]

That's why I use:

Backup Assisting Networked Disks - Archiving Immense Data

That's right, a BAND-AID

Re:backups

[TechHSV]

The problem is you need off-site storage. RAID doesn't matter if your house burns down. To do this well, you'll need to buy a total of three drives. One to use for your system, one to be stored in a safe deposit box, and one to be used in transit.

Re:backups

[nxtw]

Actually a multi terrabyte RAID 5 drive is a nice bit of the backup solution.

RAID 5 gets increasingly dangerous as drive size increases. Going strictly by published unrecoverable read error rates (1 per 10^14 bits read on recent Seagate desktop drives), the chance of data loss during a rebuild can be very high - 48%, assuming a five-drive array of 1.5 TB drives with this failure rate.
Of course, these figures don't mean that 10^-14 of all bits read will result in a failure. It also doesn't mean that an error will manifest as a flipped bit - instead, one or more sectors will be unreadable (512 bytes each).

The risks of large capacity RAID5 arrays can be mitigated by using more reliable drives or a system that can handle more failures. WD's desktop drives have a failure rate of 1 in 10^15; enterprise drives from all manufacturers usually have a failure rate of 1 in 10^16. And RAID6 can sustain two failures of any drive. Using six 1.5 TB drives with a 10^-15 bit error rate and RAID6 has a failure rate during rebuild under 1%.

Re:backups

[MadnessASAP]

Most RAID controllers already support that, it's called a hot standby. It's just a spun down drive attached to the controller that is activated and added to the array when a drive fails.

Re:backups

[bendodge]

There's another problem. Take a look at this excellent article:
http://blogs.zdnet.com/storage/?p=162

SATA drives are commonly specified with an unrecoverable read error rate (URE) of 10^14. Which means that once every 100,000,000,000,000 bits, the disk will very politely tell you that, so sorry, but I really, truly can't read that sector back to you.

...

Disk drive capacities double every 18-24 months. We have 1 TB drives now, and in 2009 we'll have 2 TB drives.

With a 7 drive RAID 5 disk failure, you'll have 6 remaining 2 TB drives. As the RAID controller is busily reading through those 6 disks to reconstruct the data from the failed drive, it is almost certain it will see an URE. So the read fails. And when that happens, you are one unhappy camper. The message "we can't read this RAID volume" travels up the chain of command until an error message is presented on the screen. 12 TB of your carefully protected - you thought! - data is gone. Oh, you didn't back it up to tape? Bummer!

Re:backups

[Anonymous Coward]

Wonder how long it would take D.B.A.N. to do a 3 pass on a 2T Drive?

Re:backups

[MobileTatsu-NJG]

What the hell do you do to back up your 2TB drive?

Eh that depends on what you do with the machine. In the last few years, I've found that maybe 1/4th of my HD contained really-necessary-to-back-up data. The rest was stuff that I didn't want to lose, but if I lost it I wouldn't be terribly upset.

That's just me, though. I mean, the content came from somewhere, it didn't just magically breed on my machine.

Re:backups

[BobMcD]

If you don't have the data replicated periodically outside the RAID, it's not a backup solution at all

This is merely a disagreement over the terms RAID and backup. Many tape backup systems would likely also fail to meet your definition by this standard.

Quoth the poster:

No, it's not the be all and end all of backups

These are the key words, because in all reality, NOTHING is the end of backups.

Do you have backups of your backups? Backups of your backups of your backups? Well, what if someone deleted something off of the backup of the backup of the backup's backup? Would you know?

The chain has to stop at some point, or the backup system itself becomes heavier than the value of the data.

The point is, RAID lets you survive a drive failure. This is, typically, the most common reason one would have a backup (in the sense of the word you're using) and it meets this need quite well. Not perfectly, but again, nothing ever will be absolutely perfect, so you may as well draw the line based on the value of the data vs your means to provide protection.

Re:backups

[Therefore I am]

I stopped chasing big drives after 500GB. The sense of impermanence gets a little stronger with each larger offering. I don't even want to think about the anguish of grinding noises from a fully loaded 2TB. This will only make sense if the data store is permanent. Until then, such drives can stay on the shelf.

Re:backups

[DaHat]

And? In many respects it's better than RAID!

Re:backups

[Anonymous Coward]

Now I have to get a drobo firmware upgrade... /lesigh...

Re:backups

[Just Some Guy]

Depends on the level of RAID.

Nope. "RAID" has absolutely zero relationship to "backup".

Re:backups

[Anonymous Coward]

They'd probably notice that I'm not the scat munching freak you are.

You mean that you're a *different* scat munching freak, then? Thanks for clarifying that!

Re:backups

[the_B0fh]

Wrong. As pointed out above, you have a hole when you break the raid to remove the drive, and need to resync it to the new drive.

What you want is something that allows you to create a snapshot. Something like this: http://blogs.sun.com/mmusante/entry/rolling_snapshots_made_easy

Re:backups

[Anonymous Coward]

What the hell do you do to back up your 2TB drive?

Huh? I don't see an issue. This would fit on well under 2 million floppies.

Re:backups

[Gilmoure]

I have 750 GB of backed up DVD movies for my Apple TV. Sure, another 1 TB drive will work for backup for now, but I'll hit the limits of a single drive for my collection in the next few months (just got both seasons of ROME and all of Deadwood for Xmas).

Re:backups

[kallisti5]

Seagate proved this true.. I has a raid 5 with 10 2TB drives!! Super reliable! *updates firmware* OMG MY DATAS!

Re:backups

[Moridineas]

That's a kind of silly response, especially after your first post showed a total misunderstandings of the technology involved...

As I said, individual workers have their own portable disks they take home. The network share that is rsynced is almost ENTIRELY backups from people's personal workstations. Additionally, I take daily snapshots of all databases (mysql,postgresql) and a few other directories and keep historical backups of those things. Our workflow does not need historical backups of everything... If you wanted to modify this scheme to keep more historical backups there are definitely ways of doing that with snapshots, or zfs, or as another helpful poster mentioned--rdiff-delete.

As I said in my post, the point of our backup scheme is not to have a historical backup of every file that ever existed. We don't need that...

TIMTOWTDI, the right tool for the job, and all that good stuff...

Re:backups

[Killjoy_NL]

Heh, I'm on the other side of the fence, I had 6 WD drives and they all failed within 2 years.
Haven't had Maxtor or Seagate fail yet and they've been chugging along on heavy load for years by now :)

YMMV

Re:backups

[darrylo]

Also, possible power failures may be an issue for some. If the rebuild was initiated because of a power failure, and another power failure comes along during the rebuild .... Kablooey.

A UPS is a virtual necessity. However, with larger arrays, the rebuild time is likely longer than the amount of time that a UPS can supply power (for home-based solutions -- I'd hope commercial sites wouldn't have this problem).

Also, does anyone know if ZFS can fix these sector errors without rebuilding? I know that ZFS's disk scrubbing can correct readable corrupt data, but I don't know about unreadable sectors.

Re:backups

[Xoron101]

What about:
del c:\*.* ./s ??

Don't forget that stupid users (running powerful commands) are O/S agnostic.

Re:backups

[ninjackn]

But wouldn't each drive have it's own independent probability of failure?

The probability of losing all your data for n drives is:
P_total = P_N * P_N-1 * ... * P_1

If we assume all drives have the same probability of failure (P) then
The probability of a single 2TB drive failing is:
P_2TB = P
And the probability of two 1TB drive failing is:
P_2x1TB = P * P

Since P < 1 we know P*P < P thus the probability of loosing all your data is greater with one drive.

Re:backups

[Reality Master 301]

Who, me?

Re:backups

[Rockoon]

So at most 2 prior images?

The real problem with backing these suckers up is how long it takes. HotHardware has this thing listed at an average 88MB/sec transfer rate in benchmarks (noting that the firmware may not be final.) They do not have write speed listed but we can be generous and allow write speed to match read speed.

Math is fun. Lets go crunching!

It will take over 6.31 hours to read/write the entire contents of one of these, and another 6.31 hour to verify the copy. Thats 12.62 hours of constant spinning and head moving.

Want a daily backup? Then 52.6% of every single day just backing it up, and a lifespan of probably 6 months.

Maybe you want to back it up over a network connection to some place offsite. Lets say you got a fairly phat pipe, like a T3 which in fantasy ideal land peaks out at 44.736 Mbps. It would take 99 hours to transfer one of these drives over such a pipe. If this were a weekly backup process, 59.2% of the week would be taken up making the copy.

It would take 426 single layer DVD-R's if you wanted to burn it. I'll leave the calculation of how long THAT would take to someone else. I know that if I was making such a copy, it would never get completed.


The proper use for these mega-drives is for daily backups of smaller drives, or in raid5 with your fingers crossed.

Re:backups

[Mordstrom]

Dear god why can I not have mod points when I know what the hell is being discussed! The article quoted in the previous post is SO right and is compounded by the proclivity of manufacturers to hook directly into the firmware on the drive's controller to get even faster performance times. You will _almost_ never be able to recover a failed SATA drive in a RAID 5 configuration where the manufacturers have played with the SATA drive's firmware for performance because of this very issue. SATA drives != fault tolerant enterprise solution! I have personal experience with this phenomena having been sold down the river by a nameless company which no longer does business in the US; I suspect because of this very issue.

Re:backups

[codeoblast]

That was "Why Mirroring is Not a Backup Solution" the sad tale of journalspace.com: http://hardware.slashdot.org/hardware/09/01/02/1546214.shtml

Re:backups

[MightyYar]

Try to do a simulated catastrophic recovery some time

So if I call them and ask for the recovery disks they won't work?

I've never tried that - only used the service to recover files on a onesie-twosie basis, but I've never had a problem.

Their Mac client is especially bad.

Yeah, but at the end of the day it can be forced to do a backup.

Do yourself a favor, get a different provider...

What alternate provider? I don't see anyone competing with them on Mac.

In any event, I don't rely on Mozy exclusively... I also backup to an external drive. I use Mozy because I have had a lightning strike in the past that wiped out both master and backup. And then there was the issue where one hard drive fell onto the one below it, which happened to be the backup :( So now, offsite backups...

Re:backups

[bigstrat2003]

You still need separate completely off line and off site backups.

If you don't have the data replicated periodically outside the RAID, it's not a backup solution at all...

Wow, I think this is a new low for /. I understand not RTFA, and I even sort of understand not RTFS, but not RTFP is pathetic, man.

Re:backups

[Killjoy_NL]

Sometimes older versions of software work better with some of the programs I use at work.
In this case an older java version which is needed for accessing building maintainance server.

Re:backups

[Lord Jester]

If I do sudo rm -rf / I get what I deserve.

However, like I said, if it is critical, it is backed up in multiple places.

Re:backups

[Lord Jester]

As I said. "Not arguing about the declaration that RAID is not a backup, but it is, by definition, redundant.".

I was refering to the statement that Domint made, "If the bits get fubar'ed on one drive, that will propagate across the rest - leaving you with no data" which only applies to mirroring.

Re:backups

[Hyppy]

I would grab two of these, put them in an OpenFiler set, and use them as disk backups of evey OTHER drive I have in the home.

Primary storage? Ehh...

Re:backups

[blahbooboo]

Check out Jungle Disk. Not expensive, great Mac client, and based on the Amazon S3. The premium/pro software version is really great as it has bit level file updating.

http://www.jungledisk.com/

Re:backups

[Anonymous Coward]

How do you back up a 1 TB drive?
How do you back up a 750 GB drive?
How do you back up a 500 GB drive?

Yes, it's dangerous - you've got a point there. But this has been true for any and every hard drive released in the past couple of years.

Re:backups

[Reality Master 201]

Actually, for native speakers of English, your interpretation isn't the default one; if I'd said "I'm not the same scat munching freak you are," then you'd have something.

One could probably make a good case for a decent paraphrase being "I'm not as much of a scat muncher as you are," being the default or a common alternative.

Re:backups

[TheThiefMaster]

MTBF is only useful for larger samples.

If you have 1000 drives with MTBF 200,000 hours, then on average you'll have one fail every 200 hours (200,000 / 1,000). In other words it helps you predict the parts cost of maintaining a datacenter.

With a small number of drives the MTBF is obviously longer than the drive's lifetime, but the MTBF is not meant to help you predict the life of a single drive. That's what the length of the warranty is for.

Re:backups

[Trixter]

Simple: Buy two, and use one in an eSATA hard drive cradle/dock. Once a night, turn on the dock, back up the data, then turn it off.

Re:backups

[lgw]

This is merely a disagreement over the terms RAID and backup.

Not, not disagreement, merely someone being wrong. It's just not a "backup" if it doesn't let you recover from accidental deletion. Technical terms often have such specific meanings. Some of us do storag for a living, and have to actually care about the difference.

Well, what if someone deleted something off of the backup of the backup of the backup's backup? Would you know?

If it really matters, you use WORM storage. For everyone else, merely having the data both offsite and offline is adequate. Also, you never want a backup of a backup (well, except in some Microsoft solutions, but no one thinks that's smart), you simply retain old backups for some length of time.

The point is, RAID lets you survive a drive failure. This is, typically, the most common reason one would have a backup

False. About 85% of restore requests are for "oops": data that was accidentally deleted, inappropriately saved over, or the like. Only 15% or so or restore requests come from hardware failure.

Re:backups

[TheThiefMaster]

If the bits get fubar'ed on one drive, that will propagate across the rest - leaving you with no data.

Nope, if the bits get foobar'ed the sector checksum fails and the data from the "OK" drive is used.

The disks themselves can protect against minor errors.
Raid protects against small to medium amounts of hardware failure.
Backups protect against large amounts of hardware failure, and limited amounts of stupidity.

Re:backups

[diskis]

That MTBF actually makes sense, when you know what it's measuring. In addition to MTBF a harddrive is rated to last 3 or 5 or so years. If you replace this HDD every 3 or 5 years, you will have a failure every MTBF period or 22.8 years. MTBF doesn't make any sense for consumers, but comes in handy when maintaining bigger SANs.

Re:backups

[BobMcD]

Technical terms often have such specific meanings.

Only technicians force themselves to think in strict technical terms. You may well be correct that your definition of the term lines up better than the casual use I'm describing here. You'd also be 'wrong' in the sense that you're failing to communicate with anyone who uses the word 'backup' in a common-sense manner, rather than the strict technical definition.

It may shock you to discover this, but humanity did exist before technology. Many of the words used to describe technical things had established meanings well before the technical ones came along. A 'backup' firearm, for example, has nothing to do with tapes. And yet a 'backup' it certainly is. It is in this way RAID is a 'backup' as well - it represents a contingency.

For everyone else, merely having the data both offsite and offline is adequate.

I find it interesting how you are able to limit your 'technical' term in such ways, and no one else is.

You're also failing to see beyond your own environment.

About 85% of restore requests are for "oops": data that was accidentally deleted, inappropriately saved over, or the like.

This is in a multi-user, corporate environment. What about a single-user scenario? Do you really think you can drum up someone that will willingly admit that they are likely to delete their own critical files? Of course not. Only someone else would delete my files.

Re:backups

[diskis]

No. That is not a proper backup. What if the RAID controller partially fails and fills your "backup" disk with garbage? Then you move it to the second array and mirror the garbage.

Also, rebuilding drives causes puts extra strains on the disks, especially if they are under a heavy load normally. I've seen cascade failures which started on one disk, breaking more and more disks when the rebuild started. 7 of 32 disks in the SAN failed within a few hours.

Re:backups

[daybot]

No one was confused. Nor would they care if they were.

Nice. Have you thought about counselling as a career? Or maybe life coaching?

Re:backups

[MightyYar]

That won't work for me... my 300GB would cost $45/month just to host, plus upload costs. That's a lot more than my current $5/month.

Again, Mozy isn't my primary backup plan. I mirror my drive periodically. My photos are also backed up to DVD, as are my home videos. These are then distributed to my parents as well... sort of a de-facto offline backup. I use the very good DropBox for all of my current projects, which makes my data even more redundant as my machines sync with one another and the DropBox server (with versioning!). My friends and I all share our MP3s, so that's another de-facto backup. So if Mozy completely flakes and loses ALL of my data, then it would have to be coincident with some sort of regional catastrophe for me to lose everything :)

Re:backups

[wastedlife]

Can't tell if you are joking or serious. If joking, don't quit your day job. If serious, I think you need to go back to school and take a statistics class.

Re:backups

[TheThiefMaster]

I did say "average" time between failures. It could of course vary between multiple failures at once (0 seconds) and no failures ever (though proving that one could take a while).

The length of the warranty quip was semi-serious, in that the length of the warranty is set at a point where the manufacturer would still make a profit, so odds are in favour of the drive not failing within warranty.

Re:backups

[Domint]

Do you really think you can drum up someone that will willingly admit that they are likely to delete their own critical files? Of course not. Only someone else would delete my files.

Must be nice to be perfect, eh? Back here in the really real world people can and do make mistakes. I myself have inadvertently typed rm some_file instead of mv some_file - stupid little things like this happen all the time.

Re:backups

[Greyfox]

Well ideally you have a backup process that stores multiple iterations of changes on your backup media and lets you roll back to any arbitrary date. I've worked with several pieces of software that could do this over the years (I believe that's how Apple's Time Machine works) but can't think of any OSS alternatives off the top of my head. All the files I don't want to lose will easily fit on a DvD. Sure I might have an astoundingly large goat pr0n collection, but it wouldn't be too difficult to reassemble that if I needed to. Applications chew up the majority of the space on my hard drive and they're not hard to reinstall, either.

Re:backups

[BobMcD]

Obviously, but few actual users consider such things in their planning.

Re:backups

[Sancho]

Failure to use words correctly leads to circumstances where people think that, for example, that the Theory of Evolution is comparable to the theory of Intelligent Design. Hey, they're both theories, right?

Put in the context of storage technology, if someone was providing a backup service in their data center, but really they just meant that the drives that they provide to you are RAIDed, you'd probably be pretty pissed off when they were unable to restore from backup.

Words have meaning for a reason.

Re:backups

[WuphonsReach]

RAID5 is absolutely a bad idea once drive sizes went past the size where rebuild times take more then an hour or two. Too great of a chance that a second drive will fail during the rebuild window due to the stress.

RAID6 with one hot-spare should be considered the minimum safe level. For 15-drive arrays, I prefer RAID6 over 13 spindles with 2 hot-spares.

Re:backups

[afidel]

Or 3 drives, RAID1+removable backup drive doing occasional sync's. Personally I like the setup I did for my dad's small business. RAID1 on the server which does VSS snapshots twice daily then the app dumps its DB nightly to his workstation which then copies the backup to his business partner via VPN and the workstation is also backed up to a Mozy Pro account. The server cost about $1K and the VPN routers were about $150 a piece, the Mozy account costs like $10/month, all in all really cheap for data protection better than many large companies =)

Re:backups

[LurkerXXX]

RAID is not a backup.

RAID is not a backup.

RAID is not a backup.

How do you restore an overwritten file from your RAID?

How do you restore it if a bugler steals the computer?

If a lightning strike takes out the power supply and attached drives?

RAID is not only 'not the be all and end all of backups", it's not a backup at all. It's a way to reduce down time.

Re:backups

[Anonymous Coward]

"having the same failure *rate* does not mean they fail at the same time"

In my experience, murphy will make sure that, however unlikely it seems, it does mean that they will fail almost at the same time...

I would not be surprised that the actual probability of a double failure in a short timespan is much higher than most people think. The same design == the same weaknesses...

Re:backups

[Anonymous Coward]

Your argument fails completely because you cannot just stuff probability of failure into "P" and use it as a constant. As previous responses pointed out, the probability of failure depends on the brand and model - all drives are not created equal.

"single-point-of-failure" vs "redundancy" arguments are not something you can just ignore.

Re:backups

[Reality Master 101]

I would not be surprised that the actual probability of a double failure in a short timespan is much higher than most people think.

It's far more common to use exactly the same model of drive purchased at exactly the same time in multi-drive configurations than not. The fact that double failures are pretty much unheard of would pretty much doom your theory.

Re:backups

[Reality Master 101]

I'll bet you're a laugh riot at parties.

Re:backups

[Reality Master 201]

Quite so. I have lovely singing voice, as well.

Re:backups

[TheRaven64]

That's what filesystem snapshots are for. Set a cron job to take hourly or daily snapshots and then keep the device mirrored on different sites. With ZFS you can create snapshots as easily (and cheaply) as creating directories, and zfs send / receive lets you keep two remote storage pools in sync.

Re:backups

[TheRaven64]

You only need that much throughput if you're writing that many files. You can use something like rsync or zfs send / receive to only transfer the changed files or blocks, which means your network only needs to be fast enough to stream the changes. My laptop has been on for 11 days and has written 57GB to disk (most of which is swapping). If I wanted to send this over the network it averages out to 63KB/s, which is about the upstream bandwidth I get from my cheap consumer-grade ISP and a tiny fraction of my local network bandwidth. If I ignored writes to swap then it would be a lot less. I've also been running pdflatex a lot recently generating a 5MB file every few minutes, which probably accounts for a significant fraction of the remainder - I could easily write this to a temporary location which wouldn't be backed up, since I can regenerate it from other data easily.

Re:backups

[TheRaven64]

sudo rm -rf /

Would complain first that / is mounted read-only. If it got to /usr or /home where there is important data then it would need you to clone the last good snapshot, so the data loss depends on your cron settings.

Re:backups

[setagllib]

ZFS will just remap the bad sector. If a redundant copy of the data is still available, the remapped version will contain that data. If anything really is lost for good, you'll be told exactly what file it was.

ZFS is a bit too good to be true. Just hope btrfs catches up soon.

Re:backups

[setagllib]

So is Linux with LVM on top of MD software RAID. Home Server just gives you a GUI, the functionality given once built is no better.

And ZFS is even better than LVM on MD. btrfs may some day be even better than ZFS.

Re:backups

[Shadow-isoHunt]

Raid *is* a backup if you're doing raid 1, and you intentionally remove the "B" mirror to replace it with new disks. Rebuild the array. You've now got fresh drives as the B mirror, old drives as the A mirror, and a third set of drives to haul offsite. Yank the A mirror next time around, and you keep fresh drives in production with snapshots stored somewhere safe. It's all in how you use the technology.

Re:backups

[ZosX]

Well it depends. I can't possibly bear to delete all 60gigs of star trek TNG or my Northern Exposure collection or Twin Peaks or Deadwood or etc etc etc. Do I delete the HD rip of Blade Runner? Do I delete all those MP3s I've never listened to (yet)? Or do I just go out and buy another drive and start filling it? What about the hundred or so PSX isos I have on my drive? The 30,000+ MP3s? Every issue of batman? Sandman? The Invisibles? How could I possibly delete any of it?! I'm starting to feel like my drive is a treasure trove of nostalgia and culture. Who knows. Maybe one day I'll die and my collection will become the next library of alexander to some future culture? That is once we have permanent holographic storage. Imagine 2PB in your pocket! It can't be that far away...1 gig drives were so insanely large 15 years ago.

Re:backups

[MBGMorden]

He said if the bits get fubared - not necessarily from a hardware failure. Software failures can cause data corruption in such a way that it will propagate the errors right on along to the redundant drives. Or, you could simply accidentally delete your data and the delete will propagate too.

The reality is RAID only protects you from hardware failure - often times there's a need not only to get to your data, but to get to back to your data the way it was the previous day. Backups help there.

The depressing thing is that good backup solutions are expensive. If you want to backup 2TB worth of data on a regular basis, then CD-R's or DVD-R's aren't really going to cut it. LTO3 or LTO4 tapes are a good solution and are what I use to backup my servers at work (where I have a budget to afford such tools). Just the drive to handle them will cost 10x as much as many people's computers though. RAID is attractive because it's relatively cheap.

In the end I've taken to keeping 2 1TB drives in a mirrored RAID on a NAS within my home network. Then, rather than backing up all that data, I essentially accept the fact that I COULD lose all of it in certain situations (house burns, both drives fail, software corruption), and instead backup only my most important, irreplaceable information onto a DVD-R periodically (about every 2 weeks), which I then keep in my desk at work (encrypted, just in case).

Re:backups

[speedingant]

Like when the OS becomes corrupt, you can't restore the RAID set again? Or because it's running a Windows OS you have to have virus protection or else you open yourself up to all sorts of vulnerabilities. In many many respects, it is not better than proper standardized RAID.

You're better off building something decent. I built my media server using UnRAID. It offers full expandability, one Parity disk with no parity striping, and no striping of data.

Re:backups

[geekoid]

hahaha...

Why don't you just post "I'm Old"~

They're a number of tape drives and NAS solutions for this.

In fact there is a device that turns all your USB external drives into a NAS.

Re:backups

[corsec67]

That could work, but you would be abusing RAID in a way that it isn't designed to do.

What if it mirrors the wrong way, replacing your new data with the old data?

You really want to suggest a "backup" scheme that is right out of TheDailyWTF

RAID != Backups.

Re:backups

[Shadow-isoHunt]

Rebuilding the array incorrectly is PEBKAC. Unless you're using Promise cards to do your RAID, at which point your array will figure out a way to confuse array members anyways.

Re:backups

[Pentium100]

Do you really need DownloadedSoftware v0.7 when the current version is 4.2?

v0.7 may work on, say Windows NT4, while v4.2 may not. Or v4.2 may be too bloated. Or some other reason. Also, deleting the setup for v0.7 may only save me 1MB on a 250GB drive, while taking me the same time to locate it and decide if I want to keep it or not.

Do you even need that version stored locally?

In case the program is no longer available.

Are you really going to see that movie again?

I don't know. I saw it one time, kinda liked it, so, maybe I'll want to see it again sometime. I cannot be certain that I'll be able to download that movie again (torrents die), while if it is on a DVD or a LTO-1 tape, I'll be able to watch it whenever I want.

The same reason why I record a lot of TV shows (that are not available for download). Yes, the broadcaster has a copy of the TV show, but it is useless to me if I can't get access to it.

Re:backups

[Shadow-isoHunt]

And I'll point out that in the case you linked to, again, it's PEBKAC. The result also would have been the same if he was doing backup tapes, so your point is moot.

Re:backups

[timeOday]

The point is, RAID lets you survive a drive failure. This is, typically, the most common reason one would have a backup (in the sense of the word you're using)

I disagree, I think accidental deletion / alteration is the most common need for a backup. Neither of us has any solid statistics, but it doesn't matter which is "most common" if we agree both are "pretty common."

Re:backups

[drsmithy]

Want a daily backup?

Your figures are assuming a complete copy of the drive is taken for every backup. Even in the days of drives measured in megabytes, that would be a pretty dumb way to do it.

The proper use for these mega-drives is for daily backups of smaller drives, or in raid5 with your fingers crossed.

Better to use RAID6 and avoid the cramp in your fingers. :)

Re:backups

[DaHat]

You realize that I provided a couple of links to what I was talking about, links that contain supporting information... you gave us... acronyms that all have meaning to... what % of /. readers?

Lucky for you... despite hating Linux as I do, I am familiar with what you describe.

The key difference between Windows Home Server and what you provide... is that I would have no fear dropping off a pre-packed device at my parents house running WHS and telling them they are on their own for configuration and support... it's just that easy to use. Can you say the same for any pre-packaged Linux based solution?

Re:backups

[CasaVacas]

Although insightful, you annoy me. What are you? Cylon? Stop writing in variabels. Robots are annoying!

Re:backups

[setagllib]

You mean like any of the many commercial NAS machines which run embedded Linux distributions? I've heard they do ok.

Re:backups

[noidentity]

That much storage in a single unit seems kind of dangerous.

I never understood this argument. Say you have N drives each with capacity C/N (e.g. C=2TB, N=1 for this new drive, or C=500GB, N=4 as you prefer) and probability P of each drive failing in a given time interval. Your expected data loss is N*P*C/N, which is independent of N. So what's the gain from more drives?

If everything is on a single drive and it completely fails, it's all lost. If it's on two drives (each half the capacity of the original), if one fails, only half is lost. If I have two drives of equal size and one is a mirror, complete failure of one means I lose nothing, and can immediately replace the drive and re-mirror. Thus, multiple drives is better.

Re:backups

[DaHat]

Dunno, how many of those Linux based NAS devices can have new hard drives added to their storage pool in a way that is transparent to the end user other than the couple of mouse clicks required? How about when it comes time to remove one of those drives (when looking to replace one of the itty bitty ones with something larger)?

Re:backups

[DaHat]

> Like when the OS becomes corrupt, you can't restore the RAID set again?

Again, Windows Home Server's Drive Extender != RAID.

And yes... if the OS becomes corrupt, even if the main hard drive becomes corrupt you can still recover your files (provided you have more than one HD in your system and the folders have duplication enabled (which is the default in multi-hd systems)). In fact unlike a traditional raid setup, in case of an emergency you can actually yank out a HD from a Windows Home Server and access it normally on a regular Windows PC in case you don't have time to go through the recovery process.

> Or because it's running a Windows OS you have to have virus protection or else you open yourself up to all sorts of vulnerabilities.

Assuming you are running random software on the box, otherwise it tends to be pretty well locked down and running in a home environment where it is far more secure than the average server.

> In many many respects, it is not better than proper standardized RAID.

Again, read the linked to post above to see just how wrong you are. And with this grand RAIDness you keep advocating... do you think the average home user is able to manage proper standardized RAID? I don't. Heck, it sounds like you yourself aren't!

> You're better off building something decent.

And yet you still have yet to quantify how Windows Home Server isn't 'decent' beyond the provably false assertions above.

> I built my media server using UnRAID. It offers full expandability, one Parity disk with no parity striping, and no striping of data.

That may work great for you... but how about for your mother? Mine would have no problem with a Windows Home Server and me NOT providing tech support.

Re:backups

[Splintax]

That's taking it a bit far. Sure, RAID isn't a total backup solution, but it does provide some protection against one form of data loss (hardware failure).

Re:backups

[cheetah_spottycat]

That's easy: On a second 2 TB drive.

Re:backups

[igb]

Of course, if you're using ZFS the checksums should make this less likely: as soon as you start getting back checksums off the disk it's marked bad, even if at a hardware level it purports to be OK. Other filesystems are available, of course.

Re:backups

[BobMcD]

As I said to another poster, I doubt any single user suspects they will accidentally delete files that are important to them. In a multi-user environment, sure. I was getting at the psychology behind selecting a contingency for the data on your shiny new 2TB drive.

Re:backups

[drinkypoo]

SATA drives are commonly specified with an unrecoverable read error rate (URE) of 10^14. Which means that once every 100,000,000,000,000 bits, the disk will very politely tell you that, so sorry, but I really, truly can't read that sector back to you.

I haven't even taken stat and I know that means no such thing.

Those specifications are set so that if a little data is lost (which happens) that the customer has no leg to stand on.

I've seen less and less errors on disks as their capacity has gone up. The drive is more and more likely to correct a marginal ECC and rewrite the sector automatically, marking the old one as bad, after it takes multiple reads to get a sector in the first place. So disk degradation has become less of an issue for users as capacity has increased, coincidentally, because the rest of the disk has been getting more sophisticated as well.

Re:backups

[petermgreen]

The problem with that is of course that mirroring simply mirrors problems on the primary drive to the secondary drive. So you end up with two working drives with broken data on it.
I think the answer to that is it depends on what you are storing.

If you are storing files that never change (e.g. completed projects, recordings of TV shows, porn, software installers etc) then you can just have your backup script copy only newly added files.

A similar thing can be done for subversion repositries, you can use svnsync to copy accross the new commits without risking copying corruption to older commits.

OTOH if you are storing large work in progress files or a large database backup becomes much tricker because your backup set can get much larger than your working set with a reasonable backup strategy.

Re:backups

[speedingant]

But.. but.. A standard user would buy an off the shelf netgear box and let that do the rest. Not build a PC, chuck in HDs, and then configure the hellish OS called Windows Home Server.

There are cheaper, more efficient, and better products out there. Just type in Windows Home Server Raid into Google, and see what horror stories come up. Corruption of data, complete loss of data due to proprietary Microsoft bullshit. Don't get sucked into the marketing.

Perfect

[timeOday]

Spindle-drives are inherently slow anyways, so I think the combination of a big, power-efficient drive (never mind the speed) for movies and an SSD drive for everything else is ideal.

Nice

[Anonymous Coward]

The cache on this drive is 8x larger than the capacity of my first hard drive.

Re:Nice

[binarylarry]

So? I'm sure the onboard controller is probably 8x faster than your first processor as well. ;)

Re:Nice

[Mattsson]

I'd think it would have be faster than 7.88MHz... =)

Re:Nice

[Anonymous Coward]

and it's 2000000 times as large as my first hard... ahem, floppy drive.
just wait until you are my age, you'll be even more impressed.

Really Enough

[Anonymous Coward]

Two Terabytes should be enough for anybody

WD20?

[argent]

It'll be so slick when the 4.0 TB WD40 comes out.

Re:WD20?

[LUH 3418]

Except they'll call it WD4X or something to avoid getting sued!

Re:WD20?

[ubercam]

Yeah, it's SO easy to confuse a water displacing lubricant with a hard drive. It happens to me all the time!

But seriously, for the makers of WD40 to sue based on a trademark violation like that, I'm pretty sure it has to be in the same ball park... i.e. no reasonable person is going to confuse WD40 with a hard drive. If Western Digital started manufacturing a lubricant called WD40 or even something close like WD50 or WD20, the makers of the real WD40 would have something to complain about.

Re:WD20?

[acmwallace]

oh man.... I can't believe you went there.

Re:WD20?

[CopaceticOpus]

Yeah, it's SO easy to confuse a water displacing lubricant with a hard drive. It happens to me all the time!

Followed by a rather embarrassing trip to the emergency room, I imagine.

Re:WD20?

[argent]

Yeah, it's SO easy to confuse a water displacing lubricant with a hard drive. It happens to me all the time!

I hope you keep good backups.

Re:WD20?

[argent]

Followed by a rather embarrassing trip to the emergency room, I imagine.

I'm pretty sure you're thinking of a different kind of lubricant. Or do people really use petroleum-based lubricants in a manner likely to make for embarrassing trips to the emergency room?

Re:WD20?

[Anonymous Coward]

weapons of mass destruction 40? oh yes, can't wait. i'm sure we'll bomb them.

Re:WD20?

[argent]

Steady on, you're not SUPPOSED to take LD50 of WD40 before posting to Slashdot.

Re:WD20?

[noidentity]

It'll be so slick when the 4.0 TB WD40 comes out.

But it's a solvent, not a lubricant.

Re:WD20?

[argent]

But it's a solvent, not a lubricant.

Well, you know what they say. If you're not part of the solution...

backups-Blowups.

[Ostracus]

Unless they're all the same model made in Thailand.

Re:backups-Blowups.

[Anonymous Coward]

That's why you diversify and buy from at least 2 different manufacturers. The probability of both failing within a small time period is negligible. Do this for your backup drives as well as your raid arrays.

But how reliable is it?

[jollyreaper]

Agent smith: What good is 2 terabytes of porn if you are unable to access it?
Keanu: (glances worriedly at his zipper)
agent smith: (palm to face, shakes head) The hard drive, you imbecile, the hard drive.

Re:But how reliable is it?

[Anonymous Coward]

Wow, someone needs a lesson in humor. That wasn't funny, just stupid and prosaic.

Re:But how reliable is it?

[4D6963]

Wow, someone needs a lesson in humor.

Yeah, someone does. "Who" is the question..

Re:That was quick, but normal

[bigsexyjoe]

Storage generally doubles every year. Exponential growth is a wonderful thing.

Green Caviar?

[auric_dude]

No thanks, looks and smells a bit fishy to me.

Re:Green Caviar?

[philspear]

I know, it will TOTALLY clash with the rest of my yellow PC.

Re:Green Caviar?

[larry bagina]

that's not tartar sauce!

Re:Green Caviar?

[Paracelcus]

It only smells fishy if you don't wash it.

Re:That was quick, but normal

[Anonymous Coward]

Isn't that only geometric growth?

Fools say it's DANGEROUS!

[TheMiddleRoad]

Every time a new, larger drive comes out, people say, "That much data in one drive is dangerous!"

So here's what you do. Go buy ten 200GB drives. RAID them together. Who do you think will lose data, you, with ten times the possible failure points, or me with only one?

Just back it up, biznatch!

Re:Fools say it's DANGEROUS!

[Ephemeriis]

Every time a new, larger drive comes out, people say, "That much data in one drive is dangerous!"

So here's what you do. Go buy ten 200GB drives. RAID them together. Who do you think will lose data, you, with ten times the possible failure points, or me with only one?

Just back it up, biznatch!

Well, of course backups are the solution. And anyone with half a clue and some important data has nobody to blame but themselves if they don't have a backup.

But if you've got 10x 200 GB HDDs, and one of them fails, you've only lost 200 GB. And in a RAID setup you might not even notice that single drive failure...which means you can easily slot in a new drive and never lose any data.

While if you're running 1x 2 TB HDD, and that one drive fails, you're pretty much hosed. In this situation you'll be rebuilding your working set from a backup, which might very well take a while. It's certainly more disruptive than slotting in a new drive while your RAID keeps everything up and running with no downtime at all.

The part that concerns me is that live storage seems to be outpacing backup capacity. Sure, LTO4 can hold about 1.5 TB with decent compression... But that isn't even the full capacity of a single one of these drives. RAID a few of them together and you've completely exceeded that tape's capacity. It's getting to the point where the only solution is going to be dumping data to more HDDs, or selling robotic tape libraries to everyone.

Re:Fools say it's DANGEROUS!

[4D6963]

Ten times the possible failure points??? What the hell are you talking about, a RAID 0 array?

Re:Fools say it's DANGEROUS!

[Archangel Michael]

Raid is for High Availability not a Backup Strategy.

Offline (Offsite) Storage is for Backup, not for High Availability.

These are two separate problems, that are often confused. Backup is way more expensive than RAID, because it requires more time and effort. It often requires some level of human intervention on a regular basis for media cycling and verification of procedures. Most people's data isn't worth the effort.

Re:Fools say it's DANGEROUS!

[mikael]

So here's what you do. Go buy ten 200GB drives. RAID them together.

Which may not be a good thing if your office/apartment is hit by a flood/fire/earthquake or other disaster.

If the information is that important that you need ten drives, you would want at least two locations.

Re:Fools say it's DANGEROUS!

[segedunum]

But if you've got 10x 200 GB HDDs, and one of them fails, you've only lost 200 GB.

The odds are that you're going to get more than one drive failure in that array when compared with larger drives and less of them, and with multiple drive failures you spend more time in array rebuilds and then comes the greater chance of a complete array failure as a result of those two factors - something you really don't want.

While if you're running 1x 2 TB HDD, and that one drive fails, you're pretty much hosed.

The bleeding obvious. Well yer, because you don't have any redundancy. You're not comparing like with like there.

Re:Fools say it's DANGEROUS!

[evilviper]

It's getting to the point where the only solution is going to be dumping data to more HDDs

What's wrong with that? Data storage is data storage...

Just because tape is the old man, that once upon a time was the end-all solution for backing up everything, doesn't mean there's really any benefit to it, today.

Tape backups are more of a legacy system that just kept on being used and minimally improved as technology progressed. The writing has been on the wall for well over a decade.

Re:Fools say it's DANGEROUS!

[Ephemeriis]

It's getting to the point where the only solution is going to be dumping data to more HDDs

What's wrong with that? Data storage is data storage...

Just because tape is the old man, that once upon a time was the end-all solution for backing up everything, doesn't mean there's really any benefit to it, today.

Tape backups are more of a legacy system that just kept on being used and minimally improved as technology progressed. The writing has been on the wall for well over a decade.

Tapes are relatively small, sturdy, and lightweight. An LTO4 holding 1 TB is quite a bit more portable than a 3.5" HDD with a similar capacity.

For years now we've been able to sell our clients a single-tape backup solution... Leave a tape in the drive when you go home for the night, the backup runs, you grab the tape and put it someplace safe in the morning. It's fairly easy to get people to throw a single tape in their purse/briefcase and take it home with them. And we don't have to worry too much about those tapes being jostled or dropped.

Laptop drives are nice and portable... And there are some great external USB drives... But I worry about reliability. I'm not sure how well a HDD would hold up to being jostled around in a purse or briefcase all day long.

Re:Fools say it's DANGEROUS!

[Anonymous Coward]

And in a RAID setup you might not even notice that single drive failure...which means you can easily slot in a new drive and never lose any data.

If you don't notice a drive failure:
1) You don't have it set up properly
2) Why would you decide all of a sudden to plug in a new drive?

Re:That was quick, but normal

[ChienAndalu]

Apperently they are the same. I was a little bit surprised, too.

More room... or backup?

[TheMiddleRoad]

The problem with a larger drive is I fill it quickly. Should I buy a 2TB drive and use it to backup my already full two 1TB drives, or should I just add storage? Oh, the agony!

Re:More room... or backup?

[Ngarrang]

You should buy (2) 15-disk arrays. One for active use, the second for backups. At 15 disks, how quickly could one fill 28TB of space?

Re:More room... or backup?

[TheMiddleRoad]

Figuring 200GB a month of downloads (darn Comcast!) plus ripping DVDs and HDDVDs, I'd say a couple years. And since my wife would leave me for buying all those disks plus a case to handle them, and that our power bill would skyrocket, and that the noise would be obnoxious, I'd have plenty of time to devote to ripping content! ;)

Re:More room... or backup?

[Anonymous Coward]

Maybe you should just pirate less.

Re:More room... or backup?

[tompaulco]

Good grief. I remember I used to have a problem with the available drive sizes filling up too quickly. I had a system with four drives and then I had an Iomega Jaz drive with four disks.
But nowadays, I seem to be getting along just fine. I bought a Dell with an 120 GB drive back when up to 240 were available. That lasted me fine up until about two years ago, when I put in an additional 60 GB drive from an even older machine that I had laying around.
Of course, now I am a hypocrite because I went and bought a 320 GB drive, copied everything off of the 60 onto that, and then disabled the 60. I'm currently using 63 GB of my 320 and 90 on my 120.

I'll lose less data

[Reality Master 201]

My RAID setup would use drives from different manufacturers and production lots, and contain hot spares.

Re:I'll lose less data

[PitaBred]

Still just as vulnerable to fire. RAID is not a backup system... it's simply a single-site reliability system. If you want your data safe, you copy it to another drive/array/whatever, and keep that safe and separate from the "original", also known as "backing up".

Re:I'll lose less data

[Anonymous Coward]

Would use? This is slashdot, the proper term would have been "uses".

Log off daddy's computer and go to bed now.

Re:I'll lose less data

[segedunum]

My RAID setup would use drives from different manufacturers and production lots

Stupid. You just end up with wildly divergent performance and increase the chances of failures.

Re:That was quick, but normal

[Anonymous Coward]

Hard drive capacity is no longer exponential. They have hit some limits that are pretty hard to overcome. They're still making progress but not nearly as fast as in years past. Additionally, drives larger than 640 GB or so seem to have some reliability problems. I just recently upgraded my RAID arrays and went with smaller 640 GB drives because they have proven more reliable even though it would have been cheaper for me to go with newer larger drives.

The OP was wrong about it being one year anyway.

I hate hard-drives. I wish SSD technology would improve. It's not just price, the current drives are unreliable as hell. I trust regular old mechanical spinning devices a lot more than the current SSD crap.

Good foil for Seagate

[jedidiah]

It will be nice to have someone besides Seagate in this space.
Perhaps they will be motivated to get their act together. If they
don't those of us that buy these kinds of drives at least have an
alternate vendor now.

Re:Good foil for Seagate

[muckdog]

Seagate does not own Western Digital. You are likely thinking of Maxtor.

Re:Good foil for Seagate

[Anonymous Coward]

Right, my bad.

Re:Good foil for Seagate

[Langfat]

Seagate does not own Western Digital. You are likely thinking of Maxtor.

Maxtor doesn't own Western Digital either ;)

Mine's Better!

[andersonEE]

I think it's better to have multiple smaller drives than a single big one. My 2 500 gigers were $65 each. I have everything important on both so when one goes, it won't be a major loss.

Re:Mine's Better!

[berend botje]

And do you keep them both in the same closet? :-(

Re:Mine's Better!

[Anonymous Coward]

I think it's better to have multiple smaller drives than a single big one. My 2 1TBers were $130 each. I have everything important on both so when one goes, it won't be a major loss.

Re:Mine's Better!

[Anonymous Coward]

I think it's better to have multiple smaller drives than a single big one. My (2) 1.5 TBers were $130 each. I have everything important on both so when one goes, it won't be a major loss.

Then my external array with (4) 1TB drives in a RAID5 config has it too. As does my laptop - in a truecrypt volume. And I leave some at work on an EMC DMX3 frame where it is backed up to DLT4 tapes nightly and replicated about 500 miles away to another data center.

If only my pron collection had this type of redundancy! ;)

Re:Mine's Better!

[Meumeu]

And do you keep them both in the same closet? :-(

In the closet? My drives are out and proud, you insensitive clod!

Re:Mine's Better!

[linj]

I think it's better to have multiple smaller drives than a single big one. My 2 500 gigers were $65 each. I have everything important on both so when one goes, it won't be a major loss.

Keep redundancy away from Creative, though: mein leben!

Note: I own many Creative products, including soundcards, and they have yet to fail on me. (:

2,000,000,000,000

[desmondmonster]

Seriously, at some point won't all this space be enough for anybody?

Re:2,000,000,000,000

[Shikaku]

http://www.microsoft.com/windows/windowsmedia/howto/articles/UnderstandingHDFormats.aspx#storageneedsforhdcapture

Short answer: hell no.

We will always find something more to store. See above for a concrete example, recording raw HD video. For a more abstract version, see: everyone's DNA, biometrics, IRS/similar, etc. etc. forever.

Re:2,000,000,000,000

[forkazoo]

http://www.microsoft.com/windows/windowsmedia/howto/articles/UnderstandingHDFormats.aspx#storageneedsforhdcapture

Short answer: hell no.

And, if you get into any visual effects for your hours of uncompressed video, you may eventually discover the joys of the multilayer exr format. It's currently becoming very popular for rendering multipass CG in the effects world, and basically it allows you to render one file which contains separate images for matte passes, diffuse shading, base color, specular lighting, reflection, etc., etc. The new workflows available make this technique suddenly much more popular. And, a lot of studios will render out 32 bit per channel instead of the 10 bit per channel listed in the table you linked. So, multiply the data rates at the top end of that chart by a factor of at least fifteen when talking about CG. Then double it because it'll be done progressive instead of interlaced. Then double it again because you wouldn't render to 4:2:0 - you'd render to 4:4:4. Then quadruple it because you want to work at 4k instead of 2k HD. Then, if you want to work at the high end in 8k, quadruple it again. (Though, not many people are currently working in 8k, and those who are do so at 24p, not 60p. So, the final quadrupling is probably unfair.)

Now, think about how many iterations you go through as the director says "Make this part faster. Make the wing flaps longer. Make Jar Jar die." Whatever. You wind up with umpteen version of a sequence that you want to keep around for review and comparison.

So, yeah. There are plenty of fields where 2 TB is a tiny joke, rather than being enough for a lifetime of data. I just happen to be involved with one of them. :) Some studios passed the 100 TB online mark years ago. Hollywood will take all the storage the engineers can give them. And big GPU's. :)

Re:2,000,000,000,000

[Benfea]

People who illegally download movies need a LOT of space. I'm not sure what other normal consumers would possibly need that much space, as there's only so many baby pictures you can take before the kid's off to college.

Re:2,000,000,000,000

[Anonymous Coward]

I don't illegally download movies...

But 1TB does not suffice for my storage needs.

Backing up BlueRay movies takes a lot of space.... I even remux the files to exclude additional languages and features... so far the average file size is about 19GB per movie.

Yes, these are legitimate backups. The discs are on the shelf... If you have kids you understand the value in keeping things not broken.

Great,..

[Anonymous Coward]

Now RAID-5 arrays will take 5 days to initialize and resync.

How long till they hit the stores?

[jedidiah]

The 1.5TB Seagate drives have only been available from a small
number of online vendors for a short while now and just now
became available from brick+mortar outlets like BestBuy.

Any idea when these drives will hit a shelf at Frys?

The new retail packaging is relatively tiny BTW. You could blink and miss the new boxes.

7 hours?

[mikeee]

It's worse than you think. Even if you have a place to back it up, the I/O rates on modern HDs aren't increasing nearly as fast as capacity. Reading at top speed, it would take almost 7 hours to pull all the data off this drive, even if you have someplace to put it. Similarly, if you're using it as part of a RAID set, it'll take that long to rebuild if you have a failure.

Pretty soon the MTBF on these drives will be a significant fraction of (capacity)/(read rate); that will make for fun all around.

Re:7 hours?

[craighansen]

It's worse than the MTBF implies. It used to be, drives would usually start failing with a few bad blocks here and there. The last rounds of failures I dealt with, the usual failure mode was loss of the entire set of data on the drive, probably due to head or controller failure. It's also made worse by the way controllers are no longer interchangable, and by placing controller firmware on the platter instead of solid-state storage. With a few bad blocks type of failure, you could always rebuild a RAID5 so long as the bad blocks on two drives don't occur at coincident locations. When the whole drive goes south at once, any failure in the other drives means data loss. What I'd like to see specified in addition to the MTBF is the MTBF weighted by the amount of data loss. 10^-14 multiplied by drive capacity of 10^12 means a reliability index of only 10^-2, while 10^-14 multiplied by a kB of bad blocks multiplies out to 10^-11. The type of failure makes for an enormous difference.

Re:7 hours?

[shutdown -p now]

It's worse than you think. Even if you have a place to back it up, the I/O rates on modern HDs aren't increasing nearly as fast as capacity.

Actually, that's something that I don't quite understand. Having more storage in the same form factor implies higher data density, doesn't it? Even with the same RPM, wouldn't it mean the proportional increase in sequential reading speed (since RPM essentially defines how fast it takes to read a whole single track, and there's "more bits" on the track with higher density)?

What am I missing here?

And, by the way (perhaps it's relevant) - why aren't there any 10,000 RPM HDDs of size >500Gb?

Re:7 hours?

[evilviper]

Reading at top speed, it would take almost 7 hours to pull all the data off this drive,

Yes, but that makes it increasingly likely that only a fraction of the data on the disk actually changes between backups. And with the increasing througput, incremental backups are quite fast, as is an rsync...

32MB On Disk Cache

[3p1ph4ny]

I was thinking about this the other day, but, does the 32MB on disk cache really matter?

Think of it this way: the Linux kernel does disk caching with my free RAM (which I generally have more than 32MB of) according to some reasonable locality scheme (LRU or something).

If the HDD does the same caching according to nearly the same principles, won't the data on the disk cache nearly always be a subset of the disk cached in RAM? Meaning: doesn't the disk cache have no effect whatsoever?

I'm genuinely interested in an answer to this question, even if it is a little OT. Please burn a little karma for me :)

Re:32MB On Disk Cache

[drsmithy]

If the HDD does the same caching according to nearly the same principles, won't the data on the disk cache nearly always be a subset of the disk cached in RAM? Meaning: doesn't the disk cache have no effect whatsoever?

No, because the OS does not know about the physical layout of sectors on the disk and the HDD controller does. Therefore, it can reorder requests appropriately to maximise performance.

Disabling the cache on a hard disk gives a massive performance hit, especially for writes. They become nearly an order of magnitude slower.

Re:32MB On Disk Cache

[Bill Dimm]

On a related note, I've always wondered if drives cache data from unrequested blocks that the head hovers over. If they do, that would be an example where the drive cache isn't made redundant by the OS caching requested blocks in system RAM.

To be more explicit, suppose a large file is written to disk, and the disk isn't fragmented, so the blocks are fairly contiguous when written. When reading that file back in, suppose our program reads a chunk of data, does some processing on it, then reads the next chunk, etc. While it is processing a chunk, the disk continues to spin, so the head is no longer positioned over the next chunk when it is requested. In the worst case the program must wait for the disk to do almost a full revolution before the next chunk can be read, unless the drive caches the blocks the head is hovering over in anticipation of them probably being requested sometime soon. Do drives do that?

If drives don't cache that way, it seems it would behoove programmers to read data into a large buffer before processing, rather than interleaving processing with small reads.

Re:32MB On Disk Cache

[phantomcircuit]

The disk cache is simply another layer of cache, the link between main memory and the CPU is faster than the link between the hdd and the CPU, so if it is cached in memory it will be faster than if cached on the disk but faster than if a read has to occur.

Re:32MB On Disk Cache

[Anonymous Coward]

It is there more to help the controller reorder writes and reads.

So you get 20 writes/reads all to sequential locations you can 'combine them up' when the platter finally rotates around and the head is in the correct spot.

Think of it as more of a way to buffer out things as the drive is mechanical and takes time to move its physical parts around. So you can read and reread spots that are semi close to each other and not have to hit the platter. Saving power and time.

Otherwise it would HAVE to go out to the disk.

Now yes in theory it should hit the OS cache first. But what if that info is flushed out or not in the cache at all? Also I think you can tag items to not live in the OS cache. Also what if your OS has not concept of 'os cache' such as a NAS or simple embedded system?

Think more about it this way. You should be asking why do I need an OS cache? Why is that cache on those drives MONSTER size in the gig ranges? Why am I wasting system memory on hard drive cache?

Re:32MB On Disk Cache

[Anonymous Coward]

The disk controller's cache is not entirely used for filesystem data itself. Much of the cache is used to keep track of disk internals so that the disk can perform far better than having to read on-disk indexes when it needs to service your request.

Re:32MB On Disk Cache

[n1ckml007]

Yeah the HDD is in LBA, and the OS speaks CHS.

Re:32MB On Disk Cache

[this great guy]

No, because the OS does not know about the physical layout of sectors on the disk and the HDD controller does. Therefore, it can reorder requests appropriately to maximise performance.

Yes the OS does know about the physical location of sectors ! Sectors with lower LBA addresses will be close to the outer cylinders, and sectors with higher addresses will be close to the inner ones. In the Linux kernel, the whole point of the I/O scheduler layer is to use that knowledge to optimize the movement of disk heads. For example the "elevator" scheduler attempts to move the heads from the outer cylinders, to the inner ones, and back, etc.

Re:32MB On Disk Cache

[Anonymous Coward]

Are you sure about this? I thought that modern disks hide where the sector is physically located on the platter(s).

Re:32MB On Disk Cache

[Anonymous Coward]

None of modern OS-es speaks in CHS.

Re:32MB On Disk Cache

[swilver]

The disk cache on most modern drives is probably more like a buffer. I always thought it works like this:

1) OS requests some specific sector.
2) Hard drive moves head to appropriate track.
3) As soon as head is positioned, transfer everything that passes below the head into the cache.
4) Once complete track is read (a full revolution of the spindle), look at the cache, find the "special marker" that indicates the start of the track and then from there find the sector requested.

That last part is because the data passing below the read heads will start at some random point (it won't magically start at the start of the track). It's efficient to store everything passing below the read head immediately so you can simply "shift" the data in memory once a revolution completes (instead of waiting for the start to pass below the read head, and then read everything from there, which would take upto 1.5 revolutions on average).

So, I always figured this "cache" memory is just to assist with reading full tracks. I don't know how much data a track these days contains but I assume a 32 MB cache could be used to store a couple of tracks atleast, but it could also just be one track.

Re:32MB On Disk Cache

[drsmithy]

Yes the OS does know about the physical location of sectors !

No, it does not. Exhibit A: bad sector remapping is completely transparent to the OS. Exhibit B: SAN LUNs that appear to the OS as a single block device may in fact be multiple levels of arrays disk different disk types and RAID levels stacked on top of each other (eg: Sun's 7000 Series that uses SSDs as an additional caching layer). Heck, even low-end SANs let you move a LUN from something like a 3-SATA-drive RAID5 to a 10-FC-drive RAID10 completely transparently to anything that might be accessing it (before, during, and after).

In the Linux kernel, the whole point of the I/O scheduler layer is to use that knowledge to optimize the movement of disk heads. For example the "elevator" scheduler attempts to move the heads from the outer cylinders, to the inner ones, and back, etc.

Linux (and other OSes) may assume that higher and lower LBA addresses equate to inner and outer areas of the "disk", but they have no way of knowing this. It has been a good 20 years since even normal consumer-level hardware started hiding the physical layout of sectors on drives from the OS (anything with an IDE or SCSI disk in it is doing this). This is why cache on the physical "disk" side is so important - because only it can know the optimal way to reorder physical disk operations.

The purpose of the IO schedulers is to account for a) different user requirements and b) different storage architectures. There is no Linux (2.6) IO scheduler called 'elevator', they are 'noop', 'deadline', 'anticipatory' and 'cfq'.

Re:32MB On Disk Cache

[TheRaven64]

You have two sectors, one at address n, one at address n+1. Will you need a seek to read the second after the first? The disk controller knows the answer to that, but the kernel doesn't. They might be on the same track, or they might be on adjacent ones. If they are on adjacent tracks then you need to move the disk head, which can take milliseconds. Even a one millisecond seek per read drops the maximum throughput for a hard disk to 500KB/s.

Re:32MB On Disk Cache

[toddestan]

I've always wondered why disk caches are always so small. Why hasn't some disk manufacturer gone and put a 1GB cache on the drive? We're talking something like $15 of memory, and it doesn't even need to be very fast. Even if the memory was useless, it's not like it stopped the graphics card manufacturers from putting 256MB of memory on my Radeon 9250 in order to try and stand out.

Re:32MB On Disk Cache

[this great guy]

True, sector remapping is transparent to the OS. But this is a pathological case for arguing whether the disk cache is worth it or not. Most sectors are not remapped.

You are of course correct that if a SAN exports a LUN to an OS, its I/O layer won't know the intricate details of that block device. But this is off-topic, as pretty much all SAN devices implement their own caching layer anyway, the latter is completely independent from the disk caches.

About higher and lower LBA addresses, OSes make the assumptions I was talking about because they are virtually always true. This is why the throughput graph of a hard drive benchmark always looks like a curve that plunges for LBA addresses increasing in value. Disk caches become less and less relevant nowadays especially with NCQ as many benchmarks show.

BTW "elevator" is the name of the I/O scheduler in 2.4 kernels. Yes, in 2.6 they have 4 different scheds with other names.

Re:32MB On Disk Cache

[n1ckml007]

Interesting: http://www.diydatarecovery.nl/dp_manual/vista.htm It looks like Vista doesn't speak CHS.

Breaks a Barrier?

[Slash.Poop]

the WD Caviar Green 2TB not only breaks the 2 terabyte barrier

In order to break a barrier don't you have to go through said barrier?
2.01 would have broken it. As it is, the WD Caviar Green 2TB stopped at the barrier's gates.

Performance looks surprisingly good!

[arugulatarsus]

This hdd seems to be competing with the spinpoint f1 and the latest of seagate's 7200 RPM hdds. The kicker is this is a "green" series drive. It uses variable RPM technology. It actually spins at 5400 RPM quite often.
I'm still not convinced going green on the HDD will save energy as it drops 10 watts on your total load. In an array of arrays, there may be savings though. Gamers, remember, your power supply/CPU/video card are the biggest culprits. Lower power will generally equate to lower hear and thus less breakdowns though.
I'll wait a few months to see if there are recalls. If there are none, this drive looks like a winner.

Re:Performance looks surprisingly good!

[swilver]

IMHO, slower spinning drives for mass storage are a win-win.

When all you want is mass storage, then speed is just secondary. Most of my large drives are maybe accessed a few times a day at most, and when they are accessed, it's usually at speeds far below the drives top speed.

The advantages of slower spinning drives are plentiful, and all are desirable if all you want is reliable mass storage:

1) Less heat production
2) Less noise
3) Less wear (more reliability)
4) Less power consumption (the "green" part)

I hated it when I had to go to 7200 RPM drives for my RAID arrays, just because nobody was making 5400 RPM drives anymore. I'm glad however that the manufactures now again see that designing drives for reliable mass storage is just as worthy a goal as designing drives for maximum speed. Since they'll be losing that last race really soon (if they haven't already) vs. SSD's, I expect to see more focus on the first.

Re:Performance looks surprisingly good!

[arugulatarsus]

That is true. IMO in most cases this hard drive looks great.

My only issue is that most examples of hard disk bottlenecks I've seen were due to concurrent file access. Raiding is great for bandwidth... not so much for access time. Let's say someone is watching a video while you're doing some background re-encoding. A very common task btw. The movie may stutter. Now let's throw in some other tasks all with auto-save enabled, like word and a driver / antivirus program. Your seek time will become the limiting factor very very fast. The seek time on 5400 rpm drives is not that of an equivalent 7200. NCQ helps, but ncq shines even more on a pure 7200.

As for your statements:

1) Less heat production

2) Less noise

3) Less wear (more reliability)

4) Less power consumption (the "green" part)

They are true while talking about similar technologies. IIRC seagate stated using a fluid bearing on their 7200 line that made them produce less heat, noise, wear and consequently lowered their power requirements. I am not convinced without seeing tested results that a fluidic bearing would actually decrease it's friction at lower temperatures. It is supposedly non-newtonian.

As for the drive manufacturer's goals of speed/reliability and cost (important and not mentioned I think) most companies have several lines of products. An enterprise or business line for reliability and a gamer / home series for cost. In all seriousness WD/Seagate (re / es respectively) have a line of drives made for raiding. Those are the ones that should be used, they are tested and have higher tolerances. Think of them as the center cpus in a wafer. Then again, isn't the premise of most non-zero raids to kinda expect a hard drive to die, so you add a layer of redundancy?

Re:Performance looks surprisingly good!

[drsmithy]

Raiding is great for bandwidth... not so much for access time. Let's say someone is watching a video while you're doing some background re-encoding.

RAID (generally speaking) improves both bandwidth and access time.

You need a pretty malicious corner-case to get a worse average access time out of a RAID array than a single disk (especially if you include downtime for failures in that average ;) ).

Re:Performance looks surprisingly good!

[mcvos]

This hdd seems to be competing with the spinpoint f1 and the latest of seagate's 7200 RPM hdds. The kicker is this is a "green" series drive. It uses variable RPM technology. It actually spins at 5400 RPM quite often.

"Quite often"? Do you mean it's really variable, intead of merely unspecified, as with WD's other Green Power drives? There they suggested variable speeds when it wasn't. The original WD GP drives rotated at 5400 rpm all the time. (And personally I don't think it's possible to vary the spindle speed of a drive during operation.)

They didn't dare to mention that because they were afraid people might see 5400 rpm as slow and old fashioned, when in this case it's actually modern, efficient, and almost just as fast as other manufacturers' 7200 rpm drives. An impressive feat that they should be proud of, instead of trying to disguise it with misleading specs.

Gamers, remember, your power supply/CPU/video card are the biggest culprits.

True, although modern CPUs (like Intel's E8x00 series) also don't consume all that much power anymore (3W idle, 30W during load, I read once), power supplies are getting more efficient, and even some modern video cards try not to use 100 W to draw the desktop.

Lower power will generally equate to lower hear and thus less breakdowns though.

And less noise. There's a lot of benefits here.

What's the effing power consumption?

[loshwomp]

I read both TFA and Western Digital's press release. There is not one actual number behind any of the claims of "low power". Guys, we do have ways of quantifying power consumption, you know.

Re:What's the effing power consumption?

[rthille]

Well, you could click twice more, once on this (in the linked PR) http://wdc.com/en/products/Products.asp?DriveID=576 and then on the "Specifications" tab (I hate web 2.0 shit like this where you can't properly link to content).

Power Dissipation
        Read/Write 7.4 Watts
        Idle 4.0 Watts
        Standby 0.97 Watts
        Sleep 0.97 Watts

For comparison, here are the number for the 1TB (32MB cache)
Current Requirements
        Power Dissipation
        Read/Write 5.4 Watts
        Idle 2.8 Watts
        Standby 0.40 Watts
        Sleep 0.40 Watts

I don't understand why Standby/Sleep power use more than doubled... As for the Active, I assume that's due to spinning 2x the platters and added processing power to be able to process the data coming off those platters 2x the speed.

Re:What's the effing power consumption?

[Spoke]

You missed it. From TFA:

We'll be looking at power as well but WD claims this drive drops in somewhere around 7 Watts under read/write load and 5 Watts at idle.

That's about 50% less power than 7200rpm drives that hold 1TB which typically idle around 7-8 watts and peak around 10-11 watts. It's a little bit more than the WD GP 1TB drive, I think about 20% more. Certainly in terms of storage/watt, it's the class leader.

Re:What's the effing power consumption?

[Anonymous Coward]

My guess would be some combination of the increased number of arm+heads, spin-up/spin-down depending where they account for that in their averages or increases from the control circuitry (which seems like the least likely of the three).

Re:What's the effing power consumption?

[Anonymous Coward]

It is very simple, while idle, the memory is still powered, the cache is double so the power consumption is also doubled.

A bigger issue is the WD marketing (and support) department lying about the spindle speed. They suggest that the spindle speed dynamically changes from 5400 to 7200 rpm while the drive is in use. This can be done with a CD-drive, but that is a total different technology.

With current technology it is impossible to build a drive head assembly that works on two different speeds. The type of head, magnetic flux generated and distance to the platter made by the lift caused by air is finely tuned to a specific speed. These are designed and build by external companies that do this for more than one manufacturer. That is why you only see 5400, 7200, 10.000, 15.000 rpm 3.5inch drives and never a speed between them.

All the current Green drives have one fixed (5400 rpm) speed. Maybe in the future they will offer models under the Green label that are 7200 rpm. But never a drive that varies the speed based on usage.

WD: just be honest and explain customers that because of the higher density and bigger cache, these drives perform like 7200rpm but are more energy efficient and quiet because they run 5400rpm.

Re:What's the effing power consumption?

[Anonymous Coward]

The increase in standby/sleep is to make sure the wormhole to the dimensional pocket stays open.
Not to mention that every 38 minutes, the wormhole has to close and reopen due to the laws of nature, so ACTUAL "standy" usage is quite a few orders of magnitude higher.

Re:What's the effing power consumption?

[rthille]

No, the cache is 32MB on both the drives for which I was quoting specs.

1012?

[rthille]

Huh, I thought Ki was 1024, you know 2^10... learn something new every day...

Re:1012?

[morgan_greywolf]

wB | Whooooosh Bytes | -1.937528 | Also calculated on a Pentium FPU.

rdiff-backup

[quarterbooty]

You should look into rdiff-backup instead of rsync for your nightly backup to the offsite location. rdiff-backup keeps a set of compressed reverse-diffs in each directory that is backed up so that you can restore a file that's lost.

Re:rdiff-backup

[legoburner]

I was going to recommend the exact same thing but you beat me to the punch. rdiff-backup is fantastic but would be truly indispensable if they would get rdiff-backupweb working properly!

If nobody else is gonna ask...

[Anonymous Coward]

Dude, how much porn do you really need? Jesus!

Western Digital past experience

[Tatsh]

I do not think I will be buying this one, or another WD. It is really hard to witness so many dead hard drives (including many DOA) and have your own experiences with their hard drives that just die so quickly. And another thing, why is every WD so damn big? They squeeze into every slot you put them into, not just slide in nicely like any Seagate (or most other brands). This goes for desktop and laptop. No wonder they are making their own external drives. Generic ones may not even fit their drives.

I have had much success with Seagate (lasts 5 years or more) and Hitachi (louder than most HDDs but they last). I do not know the warranty of WD, but the warranty for both Seagate and Hitachi are great (especially the Seagate one).

I am sure some people have luck, but after 2 dead hard drives (and many DOAs at a shop I worked at) and physical size problems, I will probably never give WD another chance, no matter what the price.

Re:Western Digital past experience

[ccool]

I'm not sure we should start talking about Seagate!!! Latetly, I would'nt buy another drive from them. I guess we will have to wait and see what happens with their new firmware.

Re:Western Digital past experience

[Tatsh]

True, but I am not affected. Largest drive I have from Seagate is a 500 GB. Beyond that, as a consumer (and not an enterprise), since when do you update firmware of your hard drives? I have never done that before and would be very weary to do so.

Re:Western Digital past experience

[makomk]

True, but I am not affected. Largest drive I have from Seagate is a 500 GB.

Their newer 500 GB drives are affected too.

Beyond that, as a consumer (and not an enterprise), since when do you update firmware of your hard drives? I have never done that before and would be very weary to do so.

When the existing firmware has a bug causing the drives to randomly brick themselves on boot, like the Seagate drives in question do. Officially, it's a low risk, but I think in practice if you're booting often the risk is quite high.

Re:Western Digital past experience

[bill_mcgonigle]

I do not know the warranty of WD, but the warranty for both Seagate and Hitachi are great (especially the Seagate one).

I recently warranty returned two each of WD and Seagate drives. The WD process was easier, better website, easier label to print, etc. The WD warranty is only 3 years, but that seems to be sufficient for my needs (3 year old disks tend to be recycled due to capacity).

I had a batch of 4 WD's come in recently, two died after a month. Looks like some of the Seagates on the market are doing the same and Seagate recently had horrible firmware trouble.

I am sure some people have luck, but after 2 dead hard drives (and many DOAs at a shop I worked at) and physical size problems, I will probably never give WD another chance, no matter what the price.

None of these vendors are the leader for very long. Seagate is on the decline, WD apparently is ascending at the moment. That ranking will likely change a few times over the next few years.

Moral of the story: mix your RAID mirrors and monitor your drive health. You're bound to wind up with bad drives.

Re:Western Digital past experience

[Anonymous Coward]

I recently had not one, but 2 400gb WD harddrives mechanically fail on me. It's to the point where I can't mount them or do anything with them. I will not be buying WD any more. And, conveniently they are both out of warranty, just barely.

Street date?

[Enderandrew]

Does anyone know the street date when these will be out, or price?

I've been worried about all the problems I'm been reading about the 1.5TB Barracuda drives. I've been waiting and waiting for something new to come along, but my current dig died the other day, so I bit the bullet and bought a new rig on NewEgg, including 4 of the Barracuda drives. If these are coming out soon, and at a reasonable price, I'm shipping the Barracudas back.

Re:Street date?

[Joce640k]

FTA: Price is $299, street date is "as fast as the trucks can get them there". THe first ones will probably be online.

PS: 2Tb for $299...!

I remember paying that much for 200Mb and thinking it was an incredible bargain compared to the old 20Mb drives (which cost thousands).

Re:Street date?

[phozz bare]

Already on sale in Australia, at about US$250.

Re:Street date?

[Enderandrew]

I'm torn. I just paid $130 for a 1.5TB drive that I don't feel entirely comfortable with (it has been plagued with firmware issues, which are SUPPOSEDLY resolved) by a company that purchased the Maxtor factory (which is almost the antonym of quality), and will use more power.

Yet a $299 price tag for a 2TB Green drive is almost double the cost, and I purchased 4 drives. I'd have to pay an extra $680, and as much as I'd prefer a Green WD drive, I don't think I can shell out the extra cash at the moment.

That figures...

[amigabill]

I jsut got a 1TB predecessor to this drive. They were waiting for me to buy before they announced the double size of it weren't they! I got it for my Tivo HD, I understand it's supposed to be nice and quiet and relatively low power compared to other drives. Others will likely enjoy it in DVR boxes when that kind of capacity is supported. (I understand Tivo's kernel currently won't do mroe than 1TB anyway)

Caviar

[Anonymous Coward]

After all these years, I'm still weirded out by that name.

Re:That was quick, but normal

[Dogtanian]

Storage generally doubles every year.

As far as the traditional hard drives under discussion go- not any more. There were around 18 months between the launch of the first 1TB drive and the first 1.5TB model. Hard drive size increases are nowhere near what they were during the 1990s and early 2000s where they would increase by 2.5-3x (if not more) during a similar length of time. Over longer periods, the exponential effects make this a massive difference.

(This is the speed that flash memory devices are roughly growing at at the present time; they're increasing way faster than hard drives).

That's it? This made /.'s front page?

[Anonymous Coward]

Hmmm... A "pre-release engineering sample," no comparisons with other drives, and only a couple benchmarks? And this made /. front page? How is this preview useful? Why not wait just a week when some more reliable data on shipping retail samples is published?

Re:That was quick, but normal

[TwistedSymmetry]

[Citation needed]

This is getting modded up? I have never heard this anywhere.

Besides, from the "I hate hard-drives" comment it is obvious that the AC has an anti-hard-drive bias. ;-)

KISS - Keep It Simple, Stupid!

[TheMiddleRoad]

If you need 2TB of storage, a single drive's speed is enough, and you want real and always ready backup, do a RAID 0 for live backup and off-site incremental disk imaging for real backup. You can move up to a RAID if you need more bandwidth. Go ahead and use two different batches or brands if you want. It probably won't make a difference. Keep a spare drive around if you're paranoid. Smile as your power bill stays reasonable, your case doesn't resemble a monster, the fans don't break OSHA rules, and you don't waste money on maintaining ten drives. Recovering from a drive failure will be easier than doing it from a multi-drive RAID. Dealing with one disk is just plain easier, and there are more and cheaper tools available.

If you need massive bandwidth, like for uncompressed HD video, you'll need a bunch of huge drives. The same original comment, that 2TB is too dangerous for one drive, still won't hold.

Individual pornographic item value is ephemeral

[TheMiddleRoad]

It's my understanding that the value of any individual piece of porn is ephemeral. People don't keep going back to a good old favorite. Quite the opposite, they move on to the newest available. Are you still excited by the Sports Illustrated Swimsuit Issue that made you a man?

Re:That was quick, but normal

[floodo1]

only some drives over 640GB have reliability issues.

$1,400 movie & music backup solution? Yea, oka

[urbanriot]

You're missing the point and fear mongering - these drives aren't used in data centers for mission critical data, they're home user SATA drives. I'm not going to go out and buy a $1,400 LTO-4 solution to backup my movies and music when the likelihood that propagated corruption across my entire RAID array is extremely unlikely and somewhat ridiculous for you to suggest. I could easily come up with my own 'what-if' scenarios in the opposite direction, but that's silly. RAID-5 was "good enough" for me for many years, RAID-6 is "even better" now. I've had many drives fail over the years, and I've never had a catastrophic issue.

What's the next escalation above "Monster?"

[PolygamousRanchKid]

So I read on "The Onion" about Monster Trucks killing folks, and now on /. we have Monster Drives.

Alas, I lack the English eloquence of diction, but, pray tell, what comes after Monster?

Supreme Being drives and trucks? Phantasmal Force drives and trucks.

Any suggestions?

Re:What's the next escalation above "Monster?"

[mudshark]

Gigantumous.

Here's a question...

[ZarathustraDK]

When in time could you fit all internet-pr0n on a 2GB-disk?

Really, it's one of those facts that are fun to know. Screw Bill Gates sitting on top a mountain of paper holding a cd, or the comparison between a modern-day computer and the gym-sized old ones.

Re:That was quick, but normal

[__aasqbs9791]

About a year ago I had to load just under 4 TB of tiff images onto 750GB drives, and the first two out of the group of drives were bad. I'd get them part way loaded, and they'd start mentioning problems writing. I turned off caching, etc as part of the troubleshooting the manufacturer wanted me to do, wasting a good chunk of my day because it took so long to get anything done with drives that large, and I had to load this same data for 4 different clients (for about 15 TB total effectively). Once I got past the first couple of drives things went along much better and I didn't have any other problems. We just sent the bad drives back to the store (I think it was Newegg, but I'm not certain).

Indeed

[dfm3]

I can only speak from personal experience, but it is a very bad idea to use more than one of the same make/model/batch of hard drives simultaneously because if one goes, the other seems likely to fail as well.

It's happened to me twice... install two identical drives purchased at the same time, a drive fails, and within a short timeframe the other one fails too. One pair was only a month old, and both failed within 48 hours of each other. I suspect that it has something to do with manufacturing tolerances and defects in a particular batch.

I think I remember seeing some sort of study that examined this and concluded that it was more than just coincidence, but, alas, I'm too lazy to look it up.

Re:That was quick, but normal

[Plug]

640GB ought to be enough for anybody?

Re:$1,400 movie & music backup solution? Yea,

[lgw]

If you don't have a copy of your data offsite, epic fail is waiting for you.

Tape is great if you need multiple prior copies of your data to protect against corruption, mett auditiing requirements, or other such needs that may be overkill for home use. But you should still copy your data to a cheap HDD that you store somewhere other than your house, to protect against theft, natural disaster, offspring-induced disaster, or just accidentally deleting something you care about.

Maxtorman, are you reading this?

[Something Witty Here]

Maxtorman,

First, THANK YOU for posting info about what is really going
on with the 7200.11 and related drives. That helps a lot.

A few questions for you (or anyone who knows the answers):

Re: The 320, or a multiple of (320 + x*256) counter.
How can we end-users find out how many log entries our disk has?
Is this the SMART error log or some other log counter?
(Smartmontools will provide the SMART error log.)

If the counter is, say 320, how can we change it to get
our drive out of danger. (short term fix only) What
events advance the counter?

I've see reference to an ISO that boots FreeDOS.
How does one update the firmware on a disk that is attached
to a non-x86 machine? (Alpha, Sparc, PPC, etc.) Which
might be running *BSD, Plan-9, Opensolaris, Penguinix, etc.

Assuming Seagate eventually comes out with a firmware rev that
really works, is updating the firmware fail-safe, or would a
power failure partway through "brick" the drive?

The web page http://support.seagate.com/sncheck.html
does NOT work properly. It says "If you use a popup blocker,
please disable it to use the serial number checker." Well,
I'm not using a popup blocker, Seagate's web page is broken.
Web pages need to work with all possible web browsers.
Could Seagate please just post the list of drive models with firmware
revs affected? (and serial number ranges, etc. if necessary)

How do we read the special location to see if the test pattern was
left there or not? If we have a drive with the test pattern, how
do we zero it?

Thank you.

To all government employees..

[slashmojo]

Please DO NOT lose any of these on the train!

Seriously.. 1TB disks full of sensitive data should be more than enough for anyone.. ;)

WD Layoffs

[frog_strat]

They had a round of layoffs in December. Another one coming in the spring. They are getting a little short handed. I hope the quality stays up.

RAID != backup

[Anonymous Coward]

What the hell do you do to back up your 2TB drive?

2 other 2TB drives?

Just remember to do copies / rsyncs to those other drives, and not use RAID.

RAID (mirroring / RAID-5,6) won't help you if you accidentally delete or corrupt a file, as that action will be automatically copied to the other drives.

Remember folks, RAID != backups.

No One

[Anonymous Coward]

"keeps pace with similar capacity drives from"

from... whom? From TFA, "Today marks the day that standard rotational media breaks the 2TB barrier." In other words, they're the only maker of a drive in that capacity. 33% larger than the single closest competitor and 100% larger than the next group is not similar. For anyone that thinks it is, let me borrow some money, and I'll pay you back a "similar" amount tomorrow.

Hrm. Google site checker says HotHardware is bad.

[Anonymous Freak]

First time I've ever heard about it. They've been a good site in the past. Maybe one of their advertising providers is eee-vil?

Re:That was quick, but normal

[Kent Recal]

We're getting there. Seriously who is using up 2TB and in what timeframe?
There's always use cases for big drives (video editing and databases) but the average consumer will probably stop caring soon.

Re:That was quick, but normal

[cenc]

yea, I bought 5 western digitals of the same make, model, batch, store, shelf, and 3 failed or where simply DOA. Now we can curse at the TB of data lost.

Now because of the sheer size of drives and read / write speeds are not really keeping up so that it would take a lot of time to restore a full backup, it is the same as saying your data is not important unless you use Raid. Raid is a must, if you wish to see your data again.

I would be almost inclined to think this is some sort of conspiracy by the drive makers, as forced use of RAID solutions means selling more drives.

What is wrong with this data protection solution?

[cenc]

So, can anyone tell me what the vulnerability is to this backup solution?

2 prime drives in raid 1. Another drive that is an rsnapshot backup drive, hot swap, connected by a USB cable through the back of a fireproof safe bolted to the floor next to the server. Once a week or so I swap that drive and place it in another safe. Once a month, I rotate that drive out of the region. Total cost for this solution by the way was about $400, to protect 500 gigs of data (2 TB with all the drives).

What could go wrong with this, assuming the frequency of backups are sufficient? Here I am not just trying to protect from data corruption, but fire, theft, natural disaster.

Christ! Slashdot, get a brain!

[Master of Transhuman]

I looked to the first page to see if anybody is complaining about the speed variations due to the "green" power balancing - and the whole first page is a bunch of morons arguing about powers of 2!

Morons.

Who gives a shit if the drive is only 1.8TB instead of precisely 2TB? The question is the performance and reliability of the drive. "Green" drives suck at performance, and the Seagate 1TB and 1,5TB drives are unreliable. So what's the story on the WD?

Fuck powers of two!

They could name it DeadBabiByte

[coryking]

No matter how stupid, ill-thought, ugly, racist, or sexist a product or standard is named, there exists at least one Slashdot poster who will defend it on the grounds that if you are truly pure, names never matter.

I call this new law The Greasy Nerd Law. It might sound like a silly name, but what, really, is in a name?

We are talking about a 2TB drive

[coryking]

Unless you have fiber pulled into your house, 2TB is a hell of a lot of data to backup offsite. Even if it is a differential backup, the time it would take to initially back up even 1TB of data on a home internet connection would be measured in weeks.

The proper answer to the original question "how do I backup a 2TB disk" is "there is no good answer".

Here in the real world

[coryking]

We dont want to spend a bazillion dollars to buy a tape drive for home use. "Offsite" is not an answer either because our internet isn't fast enough, and there exists no cheap media to backup 2TB of data. DVD only holds like 4 gigs so you'd need about 500 or so of those. Blu-ray might make that 50 disks or so.

The proper answer for the backup of any media at home is "buy another hard drive and copy to that". Any other suggestion, such as a $3,000 tape drive that still only holds 800GB (!!!) is academic, pedantic, impractical or all three.

Short answer: basically, there is no good answer. Buying a second drive is the best we have these days.

Re:Here in the real world

[lgw]

Buy a second hard driuve and store it off site and you have a backup solution. As I said, tape only makes sense if you need dozens of old copies of your data. But if your "backup" HDD is in the same room, or worse mirrored, you don't have a backup solution.

You guys do no help

[coryking]

You get your jollies off harping on "RAID != BACKUP" and yet offer no cheap, low-cost way to back up several TB's worth of media. A software RAID1 array is plenty fine for backing up 2TB's of video and music. Obviously you back up important stuff like the 500 megs of documents and the family photos offsite. But that just isn't sensible for 2TB of videos your Sage/MythTV recorded.

Until one of you pedants suggest a better way to kinda-sorta-protect 2TB worth of data for under $200, you add nothing to the conversation.

Re:You guys do no help

[corsec67]

Instead of a RAID 1 array, buy 2 disks, and 1 USB case.

Put one drive in the computer, use the other as a backup that is only connected when actually backing stuff up, then remove it, and put it in a different building, not connected to anything, and maybe in a media rated firesafe.

RAID only helps you if 1 drive breaks, and automatically destroys the data if there is a software error that corrupts the data.

Having 1 drive live and one spare gives you much, much better data reliability, and gives you at least one "oh shit, ReiserFS just murdered my data" checkpoint.

Problem

[coryking]

What case holds ten 200GB disks? What kind of power supply do I need when none of them can fall asleep? How will I keep said drives cool and quiet when they are supposed to be in a HTPC in the living room?

Small is the future. I don't buy big tower computers any more. They were cool when I was 16 and 386's with turbo buttons roamed the earth, but I've grown out of that state.

Every generation of hard drive is more quiet and more power efficient then the prior generation. 200GB IDE drives are about 2 generations old now days. Those are some nasty, noisy drives.

My solution would be 2 2TB drives on a software raid. "OMG IT ISN'T BACKUP". So what. If one dies, I'll probably have a day or so to go get a new drive. The risk of "rm -rf" is pretty low too, the only software touching it is SageTV.

Ehh

[coryking]

I'm not worried about offsite. My specific case is a bunch of recorded TV shows. If those go, the lady will kill me. I'm not interested in thinking too much about rotating backups and juggling disks for this :-)

Re:Ehh

[corsec67]

Then just have 1 live disk in a server, and 1 floating backup that is in a USB/Firewire enclosure.

As a bonus, you could take the floating backup to a laptop or somewhere else if you want to watch those shows.

(Personally I have a RAID 1 array AND 2 floating backups, one of which gets stored in a different building in a concrete room. I have had more issues with the software RAID than anything, and if I built my computer again, would skip the RAID. If a hard drive has failed, I can't just keep my computer up, it is too annoying to listen to the broken drive, so the computer is going to go down for a while. Having backups is necessary, so I can't skip that.)

Re:That was quick, but normal

[tpz]

Buying 5 drives from any manufacturer that were the "same make, model, batch, store", and "shelf" is exactly how people get into massive RAID (or JBOD or even just multiple single-drive machine) failure situations in the first place. Just. Don't. Do. It. Especially when it comes to getting multiple drives from the same production batch. You created your own failure.

Re:What is wrong with this data protection solutio

[toddestan]

It seems unlikely that you'd lose all the copies, as you've spread them around pretty well. The biggest danger to me seems to be that some file you don't often look at would get corrupted, deleted, overwritten, or something without you realizing something happened to it. If I understand your system right, if it took you more than two months to find out something bad happened, it would have propagated to all your drives by then and the original file would be lost. You might want to consider writing burning everything to DVDs every once in a while, or dump it all to tape or another drive that is stored and never rotated back in.

Re:That was quick, but normal

[Splintax]

The average consumer stopped caring a long time ago - I don't know any 'average' users with more than 100GB of data. I know lots of people with over 1TB of pirated movies and music, though. It's a legitimate use case for the storage, even if it is illegal.

Re:That was quick, but normal

[zipherx]

Same experience, i have had 3 1TB drives for personal use, the first 2 was broken when i got them.....! one was a WD green drive, another was a Samsung.. the one working in the end was the most expensive of the 3, a hitachi.
I for one was actually experiencing better durability of ide/sata drives for a long time, but the huge failure rate of 2/3 got me wondering.

Re:What is wrong with this data protection solutio

[cenc]

yea, thanks that would be the thing I believe I am missing. The size of the data has reached the point where burning dvd's becomes a serious hassle.

Re:That was quick, but normal

[FTWinston]

On my older computer, I've filled up about 900Gb across 3 drives (video, music, lots of large code projects, games) and seriously want more space to expand my kubuntu partition. A 1Tb disk wouldn't improve things unless I left the 500Gb disk in place also. 2Tb, on the other hand... yummeh.

Re:That was quick, but normal

[Kent Recal]

Well, I think it's a bit far fetched to call kids that go through their archival phase (which is a compensation behaviour) a valid use case.
Yes it's nice that they can. But it normally cools down with age and generally even those kids have a hard time explaining to anyone why they need a video and audio collection that rivals video-stores or radio-stations.

1TB holds 250 dvd-rips without compression. Or roughly 1000 full-length movies in the more common compressed form (divx etc.).
That's good for one year of constant consumption of three movies per day (about 6 hours), every day...

Consequently I think even these part-time pirates will stop caring at some point in the near future, simply because they won't be able to fill the drive up before their spleen ends.

Re:That was quick, but normal

[Splintax]

Consequently I think even these part-time pirates will stop caring at some point in the near future

Probably, but by then other people will be doing it. Why is it far-fetched to call it a valid use case? These people have money and they're using it to buy storage.

1TB holds 250 dvd-rips without compression.

Or 20 uncompressed Blu-ray rips. People will find a way to use the storage. It's very easy to fill a 1TB disk right now by downloading HD content.