Review of Seagate's 750Gb Hard Drive

Zoxed writes "The Tech Report have a comprehensive review of Seagate's Barracuda-7200.10 'perpendicular' drive, including a primer on the technology. They ran performance tests against 10 other drives, checking the noise and power consumption levels. The Seagate fared pretty well, even on cost (per Gigabyte)." From the article: "Perpendicular recording does wonders for storage capacity, and thanks to denser platters, it can also improve drive performance. Couple those benefits with support for 300 MB/s Serial ATA transfer rates, Native Command Queuing, and up to 16 MB of cache, and the Barracuda 7200.10 starts to look pretty appealing. Throw in an industry-leading five year warranty and a cost per gigabyte that's competitive with 500 GB drives, and you may quickly find yourself scrambling to justify a need for 750 GB of storage capacity."

Comments

The justification for more space

[Dude McDude]

One word: PORN

Re:The justification for more space

[thelonestranger]

Where? I cant see it.

Re:The justification for more space

[lakin]

Well duh, you dont own one of these drives yet!

Re:The justification for more space

[I Like Pudding]

I wonder if there is a correlation between hard drive size and blindness

Re:The justification for more space

[Skroggtar]

I must question why one would actually enjoy listening to porn music. What a silly world we live in.

Re:The justification for more space

[ericdano]

I meant Porn & Music not Porn Music.........yikes.

No no....no no. Well. No. Absolutely not. Porn Music no. Porn AND Music, yes.

Re:The justification for more space

[Doctor Faustus]

I must question why one would actually enjoy listening to porn music.
Do you mean music from porn movies, or audio porn? The latter is good to work out to; somehow, I find I don't get tired nearly as fast when I'm turned on.

Re:The justification for more space

[geekoid]

One Word: Internet.

Really, why are you storing porn for any length of time locally? sheesh.

Re:The justification for more space

[herauthon]

i am trying to dig up this feeling i had from that moment that i bought my first harddisk of the enormous cap of 32Mb (MFM !) this feeling...feeling... more then a feeling..

Re:The justification for more space

[Anonymous Coward]

One word: BitTorrent

Why limit yourself to just porn?

Re:The justification for more space

[creepynut]

Do your coworkers know about this? ;)

Re:The justification for more space

[Doctor Faustus]

Work out, not work. Gym, not office.

Re:The justification for more space

[Menkhaf]

So... did that make it any better? You just admitted that you like listening to porn in the gym... Hmm...

Re:The justification for more space

[pete6677]

While watching big sweaty guys lift weights.

Re:The justification for more space

[eosp]

Nts...nts...nts

Re:The justification for more space

[kazem]

Well, in our lab we take images of particles that we watch with our microscopes. So we have hundreds of gigabytes of data that need to be backed up for future use. For the time being we've been backing up on DVD since you can't beat a DVD's cost per megabyte (10-20 cents).

Re:The justification for more space

[Doctor Faustus]

Nah, I am bi, but I'm not into sweaty.

This is for the room with exercise bikes and rowing machines. There's usually no one else in there when I go in an hour before the gym closes.

Re:The justification for more space

[ErikTheRed]

I wonder if there is a correlation between hard drive size and blindness

Nah, just hairy palms.

Re:The justification for more space

[BiggyP]

all this talk of cheesy porn music and hard discs reminds me of http://www.afrotechmods.com/cheap/hdspeakers/hdspe akers.htm

Re:The justification for more space

[Anonymous Coward]

Bittorrent is no longer safe, Newsgroups is where it's at.

Re:The justification for more space

[Mozk]

10 * 4800 / 100 = $480 per DVD?

I hope you mean cost per gigabyte...

Re:The justification for more space

[NeuralSpike]

...and hairy palms.

Re:The justification for more space

[Anonymous Coward]

This is what immediately came to my mind: http://www.bash.org/?23114

Re:The justification for more space

[gardyloo]

I'm wondering if this whole "blindness" thing happens all at once, or if one's vision slowly degrades. You'd think that someone slowly losing his sight would ruminate, "Gee, pr0n or the ability to see sunsets, great works of art, my grandchildren, my wife..."
      Then again, 750 GB is a LOT of pictures of grandchildren and sunsets. Gotta fill it up somehow.

Re:The justification for more space

[Anonymous Coward]

I, for one, welcome our new porn-music playing overlords!

Re:The justification for more space

[Fulcrum of Evil]

Just one question: Sex or weightlifting?

Re:The justification for more space

[QuaZar666]

I'm thinking they mean the cost of one dvd being 10 - 20 cents

Re:The justification for more space

[syousef]

Do your coworkers know about this? ;)

Cow-orkers? Sounds rather dangerous and kinky to me.

Re:The justification for more space

[Skroggtar]

hard disks eh? i prefer to use wet storage

Re:The justification for more space

[hellmitre]

BWAHAHAHAHAHA.

Scrambling?

[_Hellfire_]

"...and you may quickly find yourself scrambling to justify a need for 750 GB of storage capacity."

With the amount of media stored on my server I can already justify a disk this size. The only downside is of course that you're going to need two of these for your mirror :(

Re:Scrambling?

[rts008]

Same here, but wish I could afford a "6-pack" of these instead of just wanting one! Oh well, with second paycheck from new job, one of these WILL be mine!

Re:Scrambling?

[ericdano]

Which bring up the question, do existing RAID controllers support this drive?

And, do firewire enclosures support them?

Re:Scrambling?

[drinkypoo]

Any firewire bridge that has the right interface to speak to the drive should be able to talk to it just fine. This isn't the old dark days of DOS where you needed extender software just to talk to fancy new drives. Since drives use logical geometry to talk to the host adapter, this just isn't an issue any more.

Re:Scrambling?

[Fweeky]

Anything supporting LBA48 should handle it just fine, although we're rapidly approaching the 2TB limit many controllers have on a single disk/array. LBA48 supports drives up to 128PB (512 byte blocks * 2^48), but of course we're still in a largely 32bit world, so it's more like 512*2^32 unless you're careful.

Re:Scrambling?

[WuphonsReach]

With the amount of media stored on my server I can already justify a disk this size. The only downside is of course that you're going to need two of these for your mirror :(

Plus a 3rd for near-line backup... and a 4th for a hot-spare...

=)

Re:Scrambling?

[Anonymous Coward]

Or 3 for a RAID5 array and a 4th as hotspare

Re:Scrambling?

[Surt]

I would sincerely hope that any sata raid controller supports this drive, that's pretty much the point of having a sata spec. Most sata raid controllers are only going to support the 150 MB/sec speed, though, but if you're buying a raid setup that can do more than 150MB/sec you bought scsi instead anyway.

Re:Scrambling?

[Doctor Faustus]

Which bring up the question, do existing RAID controllers support this drive?
Just think, 4 disk RAID-0. A three terabyte volume in a standard PC. *drools*

Re:Scrambling?

[ericdano]

And how sad you'd be when one of the drives fails and you lose all that data....

Re:Scrambling?

[Anonymous Coward]

so true, do a 5, or 10

Re:Scrambling?

[Doctor Faustus]

I'm not going to spend $1600 on drives, anyway. Still, it's a neat idea.

If you seriously wanted to do something like that, I think some of the more recent nForce 4 motherboards support RAID 0+1.

Re:Scrambling?

[Anonymous Coward]

Except marketing uses 10^3 units so 750 GB * 4 (marketing) = 3 trillion bytes = 2793.96 GB using binary units = 2.728 TB using binary units

Still quite nice ;)

Re:Scrambling?

[d3ac0n]

My Asus A8V-Deluxe had 4 SATA Channels. They have an optional setup for RAID 10 (a mirror with one parity drive) and a hot spare.

I'm not currently using it, as I have no need for that kind of data redundancy on my desktop. But once this machine gets older and is demoted to server status to replace my current server, it's RAID 10 all the way baby!

Re:Scrambling?

[jregel]

RAID10 (by most normal definitions) does not have parity.

RAID10 (aka RAID1+0) is a stripe of at least two mirrored pairs.

Re:Scrambling?

[masklinn]

nForce4 mobos have always supported Raid 0+1 (mine does and it's more than a year old). The change is that recent nForce4 motherboards also support raid5 while older motherboard needed such things as Promise chips to do so. Promise chips and stuff give you 4 additional SATA connectors though (for a total of 8)

Re:Scrambling?

[dbIII]

although we're rapidly approaching the 2TB limit many controllers have on a single disk/array

We are well past it - I had to put two partitions on an array because ext2/3 doesn't want to be on partitions larger than 2TB. It is with a 3ware controller, but I'm sure that other manufacturers can also handle more than 2TB. With twelve disks on a card you certainly would want it to be able to handle more than 2TB.

Re:Scrambling?

[yppiz]

I have a 3Ware 8506-8 -- an eight-drive controller -- and I think it has a 2TB limit per array. The newer 3Ware 95xx controllers go to 3+ TB per array.

--Pat

Now all I need...

[smaerd]

...are perpendicular STRIPES.

Actually, a think two in JBOD would work as well... until I lose 750GB of data in one fell swoop.

Get perpendicular :D

[LiquidCoooled]

http://www.hitachigst.com/hdd/research/recording_h ead/pr/PerpendicularAnimation.html

Watch out for the superparamagnetic effect though.

Re:Get perpendicular :D

[MrSquirrel]

oh man, I just posted the same link. my question is, how did the funding for that animation get approved?

Re:Get perpendicular :D

[JanneM]

my question is, how did the funding for that animation get approved?

Note how the story is about a Seagate product, but the link everyone is posting is for Hitachi. I don't know how it got approved, but whoever pushed for it deserves a raise.

Re:Get perpendicular :D

[svkal]

Seeing as approximately 75% of the comments on this story seems to be links to that very advertisement, I'd imagine funding that animation was a very good business decision(unless the animation somehow appears to have been very expensive to make; though I find that hard to believe, I'll leave the possibility open as I haven't got Flash or whatever plugin the animation uses installed on this computer and thus I haven't actually seen it).

Re:Get perpendicular :D

[LiquidCoooled]

Look at the sibling comment to yours, he points out that the animation isn't even from Seagate!

my head feels like its gonna explode.

And if you get the chance watch the animation, its very clever and explains the storage capacity issues very nicely.

Re:Get perpendicular :D

[MrSquirrel]

I'm not saying it's a bad business decision, I just can't imagine the project proposal... *picture businessmen gathered around a cliche table with big leather chairs acting all serious, then imaging them talking about "getting perpendicular"*

Re:Get perpendicular :D

[edzillion]

Did you notice that the ipod-like mp3 player the character was holding said 1 of 30,000 songs. 30 thousand! Does anyone else get the feeling of overload with this avalanche of content? I have noticed that the more music I have ripped on my pc the less I listen to each song. If consumers are said to empathise with their purchases - for instance it has been noted that people value items more when they own them - then having 30k songs or 50k episodes of the daily show surely means that each will get less attention. In these circumstances I find it hard to believe that these items will still hold their value.

Re:Get perpendicular :D

[Jesus_666]

Hello? We're talking marketing people here. The same kind of people who think that Germans like products more when the slogan is in English and who are thus using English slogans even for products never sold outside Germany - even though a poll has shown that many Germans don't understand what those slogans are supposed to mean. No other country that I know of has so many slogans in a foreign language.

Soem examples for slogans that just don't work in Germany:
RWE Group: One Group, Multi Utilities - only eight percent of all participants in the poll claimed to have fully understood the slogan, 15% claimed to have mostly gotten it. I guess I'm in neither of those groups - I don't quite know how to utilize a corporation.
Mitsubishi: Drive alive - Most people correctly translated it to "fahre lebendig", but couldn't tell what that's supposed to mean as you can't drive when you're dead. 18/25%.
SAT.1: Powered by emotion - This is a German TV station. Many Germans have never encountered powered by and would probably understand it as "has its eletrical energy provided by". At least it has 33/49%.
Douglas: Come in and find out - Apparently this perfume store chain has such a confusing floor layout that it's difficult to leave the store. When people translate "find out" to German they can either translate it to "finde es heraus" (find out about it) or to "finde heraus" (find a way out). The latter translation is the more intuitive one. 34/54%.


Marketers are so far detached from reality, if we'd put them all on a huge pile causality itself would rupture and a gate would open into an alternative dimension where strawberries are sentient and the color purple just got arrested for driving too fast on the Milky Way.

Re:Get perpendicular :D

[WuphonsReach]

Well... 750GB (let's say 700GB once we remove the overhead) holds:

200 DVD movies (3.5GB each) or 100 DVD9 movies
500 days of music (128kbps)
1400 TV episodes (44 min, MPEG4)
500 HDTV episodes (MPEG4, 1.4GB/show)

So yes, we're probably getting past that point with music, but not with video yet.

And, IIRC, Project Gutenberg has something like 300-400GB of text files in their library.

Re:Get perpendicular :D

[HolyCrapSCOsux]

I have recently found that I only listen to about 850 of the 12,000 mp3s I have. OTOH, that's what playlists are for. sometimes I'm in the mood to listen to funk. Best to have them on tap. Sometimes I go on a dead milkmen binge. good to have them too. Just cuz you have a copy of every song ever made, doesn't mean that they all need equal playtime.

Re:Get perpendicular :D

[argStyopa]

TIVO is already making a hash of the 'free television' model.

What happens when someone can have locally an mp3 playlist that rivals that of a local radio station? At least with TV, there is a constant flow of new content - good radio stations too. But most radio is just replaying over and over a list of probably well under 300 songs, with a weekly turnover of what, 5% or less?

Re:Get perpendicular :D

[Jah-Wren Ryel]

or instance it has been noted that people value items more when they own them

Its not when they own them, it is when they have paid for them - and the more dearly they paid, the more they value them - and payment not need be in the form of currency either. That's why so many frat boys are fanatical about their fraternity - they paid dearly in hazing rituals in order to attain membership - thus they attribute a higher than realistic value to membership.

Since, 30K songs, would be $30,000 at itunes-level pricing and potentially even more at retail-CD pricing, chances are someone with that many songs did not actually pay for them - either via raising the old skull & crossbones or using one of those all-you-can-eat music services. If they could afford full pricing, then they are probably so rich that the price is immaterial to them anyway.

So, unless they are insane, chances are they 'paid' so little for all those songs that they don't really care about any of them that much. Which is about the same conclusion you reached, but I felt like babbling on about the paid-vs-own thing because I think it is an interesting facet to human psychology.

Re:Get perpendicular :D

[Anonymous Coward]

FYI: Utilities means things like gas, water, electricity, sewers. RWE means "Rheinisch-Westfaelische Elektrizitaetswerke", so they think they need to tell you that they sell more than just electricity now. It's amazing what good marketing people can make you pay for.

Re:Get perpendicular :D

[tjw]

What happens when someone can have locally an mp3 playlist that rivals that of a local radio station? You mean 20 songs?

Re:Get perpendicular :D

[hackstraw]

I have noticed that the more music I have ripped on my pc the less I listen to each song.

Duh. Especially if you have it on random play, the odds of it being hit, are, well lower with the more content you have.

There is the 90/10 or 80/20 or 99/1 or whatever rules, depending on the situation, but what those guys say is that 90% of the time you will be listening to 10% of the material you have.

Its generally true. However, its still good to have those other 90% laying around for those times when you "really need them".

Other rough examples. You read 10% of your books 90% of the time. 99% of the world's money is owned by 1% of the population. 90-95% of the alcohol consumed in the US is drank by 5-10% of the population. 95% of my complaints/problems/issues from my users comes from 5% of them. Etc, etc, etc.

Re:Get perpendicular :D

[MooUK]

Thirty thousand at relatively low-quality lossy formats - or quite a lot less than that at full quality lossless rates. Besides that, I currently have a couple thousand tracks on this computer, not all of which will fit on my media player simultaneously, and I wish I had both more music and more space for them. I often use my mp3 player to provide music at various events I run, and I like to be able to fulfill requests people make. Whilst the average user might not have any real use for high-capacity portable players, DJs and similar can always use the space.

Re:Get perpendicular :D

[Jace of Fuse!]

I have noticed that the more music I have ripped on my pc the less I listen to each song.

That's what Smart Playlists are for. Playlists for different types of music. Playlists that choose based on rating. Playlists that play songs that haven't been played recently or very often. With the right software(ahem, iTunes) that problem is more under your control. Sure, you'll listen to each song less often, but the diversity is good and you can take greater control of the music it is you're likely to hear given your current mood.

Re:Get perpendicular :D

[Anonymous Coward]

Gutenberg's only got about 150GB of text.

Re:Get perpendicular :D

[cjb-nc]

Uncompressed, DVD-resolution (720x480) color video, no sound, 24 fps = 23.73 MB/sec, or about 8.4 hrs of video in 700 GB.

I could see where security users would find that quite useful. Add some low-loss compression and you've got a nice, large, random-access video vault for some casino in Vegas, for one example.

Re:Get perpendicular :D

[PCM2]

No, he means one where the major record labels pay him to play their music.

Re:Get perpendicular :D

[Isotopian]

Actually, REAL radio stations, especially ones that never play music, often seem to have huge libraries of music. I recall the station I used to volunteer at in Juneau, Alaska, having a vast room with wheeled floor to ceiling racks of music, and I could find anything I could think of. Too bad they just played NPR all day.

Re:Get perpendicular :D

[coofercat]

Is there such a thing as a sort of Moore's Law for music and disk space? Maybe I can coin one...

I'm thinking there must be a point at which you can buy a disk big enough to store every song ever produced thus far (let's say in Flac format, for the sake of argument). Indeed, at some point, I presume disks will grow faster than record companies can turn the handle to crank out more tat.

In other words, save your money, wait for the zillion GB drive, because then you'll be able to store every track ever produced, plus all those likely to be produced in the five year warranty period. Imagine that - go a step further and imagine buying a disk that comes pre-loaded with every musical production ever recorded. The *AA may have something to say about it, but that's another story...

Re:Get perpendicular :D

[kettch]

I have noticed that the more music I have ripped on my pc the less I listen to each song

For me that's not entirely true. I still have music that I like to listen to. I make sure everything is tagged with the genre, and some days I just feel like one kind of music or another. My philosophy isn't that it's overload, but that it's having a song for every situation. It's being able to hit play on "Viva Las Vegas" (ZZ Top version) as you pass the welcome sign, or queueing up "Teenage Wasteland" when my friends' kids are having a "teen" moment (that didn't help the situation any, but it was funny).

Re:Get perpendicular :D

[aquowf]

well, its good to know that hitachi's recources are being well spent.

no, seriously. what would we do if we didnt see a hard drive singing "get perpendicular"?

the would we be a scary place...

Re:Get perpendicular :D

[CODiNE]

And half of that space is their copywrite notice at the beginning of every file.

Re:Get perpendicular :D

[Anonymous Coward]

Well... 750GB (let's say 700GB once we remove the overhead) holds:

200 DVD movies (3.5GB each) or 100 DVD9 movies
500 days of music (128kbps)

In these days of 500GB+ hard drives and fast-encoding CPUs/GPUs, why would buyers of this hard drive encode their music at a lossy 128kbps (instead of ~900kbps FLAC) or DVD movies in 3.5GB+ MPEG-2 (instead of 1.5GB XviD/MPEG4)?

Re:Get perpendicular :D

[evilviper]

Watch out for the superparamagnetic effect though.

After reading the article, I think I can sum it up in one word:

superparamagneticarealdensitiesallowidocious

Is it just me, or did it seem like the writer is getting paid by the sheer number of letters, and kept saying the same dammed things (in the most verbose way possible) over and over again?

Re:Get perpendicular :D

[Anonymous Coward]

baba o'reily

Re:Get perpendicular :D

[Anonymous Coward]

Hmmmmmmmmm it's like its related to a bell curve, or something...

Re:Get perpendicular :D

[Anonymous Coward]

it's that is

Re:Get perpendicular :D

[methano]

And 95% of chain smokers prefer cigarettes to chains.

Re:Get perpendicular :D

[matt21811]

I looked into this.
One online music store claimed to have over a million songs in its library. I take this to mean that if you were to collect the most popular 98% (number of sales wise)of all music then it would probably be less than 2 million songs.
2,000,000 * about 4 MB per song is 8 terabytes.
8 terabyte dives should be here in less than 10 years.
The encrypting of songs on the disk waiting for a purchased unlock code is actually fairly easy to manage to the satisfaction of the record industry.
The real problem is that a lot of the music market is made up of current music. I took a quick poll of my local record stores and they said about 60% of music sold is new, leaving on 40% of back catalog stuff to be found on the hard disk after just 1 year of ownership.
This means that even with a lot of music still on the drive most people will still have to download most music from the net. It also means you are probably wasting 8 gigs of storage capacity filled with encrypted music you cant access and are never going to buy.

Re:Get perpendicular :D

[StikyPad]

Or 1 Director's Special International Extended LOTR Trilogy in SuperDeluxeHD with discreet 7.1-channel Surround Sound, featuring soundtracks in each of: all currently spoken languages, 3 dead ones, orcish, elvish, and -- controversially -- Klingon.

Re:Get perpendicular :D

[thogard]

A few years ago a radio station in Melbourne ran a contest where bands sent in CD's that they had made within the year. The result was from a listening population area of about 3 million people, they got about 3000 CD each with a normal amount of songs (about 10 I'm guessing). That was a few years ago and before the advent of a real cheap way to make good audio CDs in your basement but if the figures hold true, I expect that there are about about 1000 bands who can make a CD per million people in most developed countries. Since there are nearly a billion people who fit into that group, I expect that there are about 10 million new songs recorded every year. On a small fraction of them ever get to the local record store.

Re:Get perpendicular :D

[matt21811]

I feel very confident that all but 1 or two songs of that 3000 CDs fits in the least popular 2%, by sales volume, that I have already excluded.

Re:Get perpendicular :D

[Petersson]

Watch out for the superparamagnetic effect though. After reading the article, I think I can sum it up in one word: superparamagneticarealdensitiesallowidocious

I still don't get it, could anyone explain it by means of tachyons pulses and subspace distortions?

Whoah

[MobileTatsu-NJG]

"Throw in an industry-leading five year warranty..."

Wow, thought those days were gone.

Re:Whoah

[Martin Blank]

You just have to look for them a bit. I just picked up a 300GB Maxtor SATA-2 with 16MB cache and NCQ that has a 5-year warranty, and it only cost me about $6 more than the 3-year warranty version with identical specs. Other companies may also offer them. (Of course, Maxtor is now a part of Seagate.)

Re:Whoah

[Jace of Fuse!]

Wow, thought those days were gone.

A warranty that long on a hard drive is kind of silly anyway.

Let's say you buy a hard drive today for $300. In 2 years it dies. A replacement of that drive is probably less than $100, but your data is still lost.

And the price of data recovery is sky high for most users. I haven't found a service that will touch a mechanically failed drive for under $500.

In a couple of years when your hard disk crashes (if it does) if you don't have back-ups the last thing on earth you're going to worry about is a hard drive warranty.

Granted, backups of course mean your data is safe and all you need is a replacement hard drive. But if you think about it, the time spent recovering from backups, turning in the warranty claim, and getting a replacement is probably worth more than the value of that exact same drive. And while having your device replaced after a failure is nice, if you do have a hard disk failure the most important thing for you will be getting your data back either off of backups or from a data recovery service.

And while we're on the subject, how well do data recovery services fare when attempting to rescue information off of these new perpendicular drives? Does anyone know?

Re:Whoah

[Zoxed]

> that has a 5-year warranty, and it only cost me about $6 more than the 3-year warranty version with identical specs

I assume that the extra $6 is just the insurance on the replacement cost: some drives will be chucked before 3 years, or passed on to someone without the proof of purchase and a lot will reach the 5 years. The replacement few that fail and the owners successfully claim will be covered by all those extra $6's.

So rememeber folks: a longer guarentee does not *necessarily* mean a better product: just that the costs are covered !

Re:Whoah

[sixteenvolt]

Warranties on hard drives. Yeah, a fat lot of good my IBM Deskstar's 5 year warranty did me.

At least they'd send me a new one for free!

Wait...

Re:Whoah

[Anonymous Coward]

Well... when dealing with hard drives you always need to plan for failure. And a 5 year warranty means you typically end up with a larger replacement drive.

Re:Whoah

[WuphonsReach]

For a while they were, seemed like every manuf' was moving towards 1 year warranties on anything except their premium SCSI drives. Some of that may also have been the free-fall pricing due to hard drive sizes doubling every 12-15 months.

But lately, it's been getting easier to get 3/5 year warranty drives. (Again, possibly because drive size increases have stagnated over the past few years.) There's not much of a price differential either between 1/3/5 year warranty drives, which makes it worthwhile to go ahead and look for a 5 year warranty.

Re:Whoah

[loraksus]

Sure, the 5 year warranty is nice, but I submited a RMA on 4/1/2006 for a 300 gig seagate drive and I still don't have a replacement because they are apparantly out of stock. Apparantly sata versions were oos as well... Kind of annoying.

Myth boxes and the like

[debest]

Some keep saying that there's no point to ever-increasing drive storage numbers. I disagree. Huge drives will always be appreciated in media PCs, where good-quality video (even if compressed) takes up a good chunk of storage space. Since these devices are preferably low noise, low power, and small in size, you obviously can't just keep throwing more drives in the box: a single drive is the best solution.

Keep the size increases coming, I've got a mountain of content on DVD and VHS that I'd love to be able to rip to an online media library!

Re:Myth boxes and the like

[MrSquirrel]

I remember 7 years ago when I got 20 GB harddrive and thought "I'm never going to fill this up"... needless to say, 3 and a half years later I had a nice big 80 GB drive. Again, I thought to myself "I don't think I'll ever fill up 80 GB -- even with all my music and games, if I ever run out of space I can always just uninstall something". 2 years ago I bought a second drive for my main machine (160 GB) to act as a music/movie/game storage device... The 240 GB of space on my main machine are about 70% used, in addition to the 750 GB of harddrive space currently on my file server being almost 85% used, it's safe to say that a 750 GB drive could be extremely useful to me (then I could throw 4 of them in my file server). "If you give them the disk space, they will fill it... with pr0n!".

Re:Myth boxes and the like

[shawb]

In addition to flat out larger hard drive storage sizes, using the perpendicular method will also allow physically smaller hard drives of decent storage ability. The potential benefit to laptops, mp3 players and all other manner of portable devices is quite real.

Re:Myth boxes and the like

[Slashcrap]

Some keep saying that there's no point to ever-increasing drive storage numbers.

Yes, well some people are just fucking retarded. To be honest, I had kind of suspected that before today.

Re:Myth boxes and the like

[Jeff DeMaagd]

The Seagates aren't very loud, I think their noise is negligible except for the fussiest people, and improving case accoustics helps a lot more than reducing the number of drives.

I have five Seagates in my workstation here, 1x 15kRPM and 4x 7.2kRPM, and the sound really isn't objectionable.

I have one single 10k RPM drive in my HTPC, and it's not a problem, and wouldn't mind adding more. My refrigerator and video projector are both louder.

Re:Myth boxes and the like

[brianosaurus]

Geeks fit into hard drives like goldfish fit in bowls; they grow to fill the space...

Re:Myth boxes and the like

[poot_rootbeer]

Since these devices are preferably low noise, low power, and small in size, you obviously can't just keep throwing more drives in the box: a single drive is the best solution.

Actually, I'd say the BEST solution is to put a box with minimal or no drive capacity in the living room, and put in on a home network that connects to an arbitrarily large network storage box in the other room. But if that's not feasible for whatever reason, a single large disk in the media PC itself is the next-best solution.

Re:Myth boxes and the like

[LordKazan]

while i know that is meant to be a joke it should be noted that "goldfish grow to fit bowls" by poisoning themselves with their own wastes when they're in too little of a space - goldfish actually need a minimum of 30 gallons (this minimum size can hold like two though, three maybe with excellent filtration)

Re:Myth boxes and the like

[Anonymous Coward]

You really truly underestimate the fussiest people.

The fussiest people go to the lengths of using notebook (2.5") drives in their desktop computers, and add soft suspensions (i'm not talking about rubber grommets here, i'm talking hanging your hard drive to rubber bands in order to fully isolate it from the case). The fussiest people, the kind of guys that roam SPCR, don't go for "not objectionable noise", in the best case they go for quiet (quiet as in "you don't realise they're powered on during the day") and in the worst case they shoot for silent (as in "you don't realize they're powered on during the night when going to sleep").

A 10k RPM HDD wouldn't even be considered by the "fussiest people" (much less a 15k one), even 7.2k are usually considered unbearably loud (even when idling).

Re:Myth boxes and the like

[drasfr]

I totally agree with the usefullness of disk storage... I have personally so many uses...

I am a photographer. raw photos I take are easily 100MB. In a photoshoot i easily take 50 to ... 400... Once I work on them in photoshop I do like keeping the originals PSDs, which very often amounts to 500MB or more PER PHOTO. I try to usually limit to 500MB because of disk space and my computer/drive gets too slow. Image if each photo I work are 500MB. I could only store 2 of my originals per GB of space... Now imagine I want to store different revisions... I used to have a 250GB drive, mirrored. Got filled up very quickly.

I don't care about music, but I am also into video editing and archiving. 12GB/hour is what my camera require for disk space.

Oh yes, I do have a mythtv box that I am currently building, I can't imagine how much diskspace I may want to this, and in total.

so yes. bring on disk space. I will never have enough and will always find a way to fill it up - LEGALLY. I am probably going to upgrade my 900GB file server (4x300GB, raid 5) i have at home and replace all the drives with 750GB. It is probably going to last me 6 months to a year?

but it is true that for most people there is no need for such amount of space. my parents have plenty of space... and they only have 80GB...

Re:Myth boxes and the like

[Pharmboy]

Actually, I'd say the BEST solution is to put a box with minimal or no drive capacity in the living virtualroom, and put in on a home network that connects to an arbitrarily large network storage box in the other room.

This is exactly why I am waiting for 10-20gb solid state drives to get competitively priced. Plenty for the basic OS and a swap file (and then some), and network everything else. Already have a firewall at the house anyway with all my old 100-200gb drives.

For many offices that don't gigs of local applications (like mine), this would also be a great solution. 5gb drives would work great for us. xp/office/database front end only locally. All "user" files stored on the main server, which is RAID 5 anyway.

Totally silent, uses less power, no moving parts. A bit slower to load games (or faster if you store only a few locally). Also, the idea of a solid state drive for swap pretty much kicks ass. 10x slower than ram, but 10x faster than a hard drive. a TB or two of virtual memory would actually be somewhat usable without the trashing.

Re:Myth boxes and the like

[TubeSteak]

You'd think we'd start figuring out how to better use the space we have, in addition to upping capacities.

CPUs & GPUs are fast enough now that programs like photoshop should be able to start with your 100MB RAW file, save a set of instructions & give you your end result on the fly (or with a small wait).

Sure it's a tradeoff; render time vs file space. But at 500MB per file, I imagine someone with relatively limited resources would spend the extra time.

Re:Myth boxes and the like

[MechaStreisand]

The idea of a solid state drive for swap is actually pretty terrible. The most common solid state technology, flash, has only a limited number of writes available to it. Swap would use 'em all up pretty quickly, and then you'd have a dead drive in no time. You'd be much better off just adding more RAM and having no swap on the solid state drive at all.

Re:Myth boxes and the like

[Sloppy]

That's what I was thinking, but some people disagree (read the replies).

Big HUGE warnings

[VincenzoRomano]

1. The more data you pack in a volume, the higher the risk for data loss due to mechanical breaks.
2. 7 100 Gb disks (that would cost less than USD 430) will be at least 7 times more reliable than the 7200.10 with possibile similar performances.

Re:Big HUGE warnings

[Kandenshi]

How much physical space/heat would those 7 drives produce as compared to a single 750 gig drive?

Re:Big HUGE warnings

[TheSkepticalOptimist]

Or, you can get three 750gb hard drives and RAID them for security. Put it this way, 3 drives will fit in your box, 7 won't.

Re:Big HUGE warnings

[gelfling]

Maybe yes maybe no. If the likelihood of disk failure is .01%, just an example, then the risk of a failure in any of your 7 disks is .07% so instead of a 99.9% reliability you have 99.3%.

Re:Big HUGE warnings

[muhgcee]

With 7 100GB disks, we have that little problem of power consumption to deal with. And noise. And heat.

Re:Big HUGE warnings

[BillyBlaze]

How will having lots of smaller disks be more reliable? If they mirror each other, yes, but if they are striped, they'll be a lot less reliable. Plus who has room (and power) in their case for 7 drives? Datacenters, sure, but not home users.

Re:Big HUGE warnings

[Dan Ost]

No, this isn't true. If the failure rate of drives is constant (pretty close to reality), then
if you've got 7 drives and I've got 1, you're seven times more likely to lose a drive than
I am.

Granted, you only lose 1/7th if your drive fails, and I lose all of it, but since we're both
making backups (you ARE making backups, right?), you're paying 7 times the space, electricity,
heat, and noise costs for less reliable storage than I am. Assuming that we both run out systems
long enough for drives to fail, you're also paying 7 times as much of your time replacing drives
than I am.

What sense does that make?

Re:Big HUGE warnings

[ID000001]

Only if you assume the new drive, with new technology, is still just as reliable as those older drive in smaller sizes.

And only if you discount the fact that, while it is 7 times as likely to fails at once, it is also 7 times more likely to fail at all.

Also, 7 x 100 is less then 750. If you want some back up in place you will need the the extra controller and hardware to configure it into a RAID, which, even if in software mode, still take times to set up.

Lastly, I doublt 7 100 gig drives can consume just as much energy as 1 750 gigs drive, nor will they get the same amount of heat, or noise, or, density. I fail to find any consumer case that can take 7 drives comfortably, that means hold it, make it small, and cool them quietly.

As you can see, there are many many benifies to one bigger drive.

Re:Big HUGE warnings

[misleb]

7 times more reliable? Seems awefully simplistic to me. Most people are going to want to put those drives in an array. With RAID 5 (to get the performance you talk about) you'd have 6 disks worth of space and could only afford 1 drive failure. How do you get 7 times more reliable out of that? It is no more reliable than 2 mirrored 750GB disks. Besides, what does your average user put 7 drives in?

-matthew

Re:Big HUGE warnings

[ebh]

I remember thinking the same thing when I heard about the first 1GB drives. "Who would dare trust that much data to a single device?" (This was 1987, the same year that I paid US$775 for a 71MB drive, amazed that disk space was almost down to $10/MB.)

Re:Big HUGE warnings

[VincenzoRomano]

OK. OK.
If you fill the disk(s), the 750 GB one will have the arms and platters move 7 times more that each single 100 Gb disk.
Even if one of the 7 breaks, you'll loose only a mere 14% of the data.
My box fits 10 disks because I like to play it safe and some 7 times more heat (I blow away) is worth the increase in reliability.
And yes, mine was a joke.

Re:Big HUGE warnings

[mrchaotica]

Plus who has room (and power) in their case for 7 drives? Datacenters, sure, but not home users.

I've seen cheap ($30) mid-tower cases that had about 8 internal 3.5" drive bays -- they just had the mounting rails go the entire height of the case. Combine that with a decent power supply and you're set.

I just wish they made high-quality cases with that many drive bays, but I haven't found any for some reason.

Re:Big HUGE warnings

[Anonymous Coward]

Do you really believe that disk failure rates are anywhere near 0.01%? In my admittedly small sample over the years, it's more like about 5% over five years - and that's without ever buying Maxtors. If that is anywhere near the true mean of the disk failure distribution, then you could expect a 30.2% probability that at least one of those 7 disks would fail within five years.

Re:Big HUGE warnings

[Intron]

With RAID 5 and 7 drives you would probably want to stripe across 6 drives, giving 500 GB useable space and keep 1 hot spare. The performance of a 6-drive RAID with one failed drive would be terrible, so you need to have a spare drive to rebuild your RAID as quickly as possible.

Re:Big HUGE warnings

[muhgcee]

And 19 years later, you remember how much you spent on a hard drive. Impressive.

Re:Big HUGE warnings

[gelfling]

I wasn't claiming it is or isn't. They're just examples for numerical purposes. I tend to think that some people seem to have much worse experience than others, for whatever reason. I can tell you that in 15 years of home computing and 20 years of at-work PC's I've never lost a drive completely. I've had some bad sectors but I've never crashed a drive. I've blown out CPU's and motherboards and had several CDRW's crash and burn but never a hard drive. Maybe it's that I don't use particularly high performance drives ever - I always go for cheap.

Re:Big HUGE warnings

[Anonymous Coward]

No, the risk of failure in any of the seven drives is .01, the risk of losing all the drives is 0.1^7 - the drives are rather independent events.

Re:Big HUGE warnings

[shawb]

You'd better make sure that you have an excellent power supply and superb cooling if you're running that many drives on a computer. High heat and spotty power run havoc an a hard drive's life expectancy. I really don't see 10 smallish disks being any more reliable than 1 big one.

And then... how are you storing the data? Having the data split up among many different drives causes usability issues where you must make a conscious decision where to put your data each time you save a file. Software is horrible at saving to different locations (not that it can't do it, but requires a lot of user interaction.) This puts multiple drives out of the reach of a large number of users. The possible reliability gains really are not going to be that much.

So the other option is to stripe against several volumes. Here you come into more problems. If you stripe just for size, you are DECREASING your reliability by spreading the data across multiple disks. If one disk fails, the whole thing goes. If you are striping simply for reliability and end up with... say single mirrors of all data there is some increase in reliability, although I have heard many horror stories about what happens when a proprietary RAID card goes out and is not replacable. The data is still sitting on the hardware just fine, but you simply can not get to it without the (now defunct) controller card. Pure software RAID will improve your reliability a little bit, but with a huger performance hit as you are not using the data bus nearly as efficiently as a hardware solution would be.

Or you could just drop the cash on a big drive, and selectively back up any important data. Maybe even get two big drives and mirror them, which would be a much simpler and robust system than having a patchwork of 10 small disks.

Re:Big HUGE warnings

[Monster_Juice]

If you spent $775 on 71 megs you would remember it also.

Re:Big HUGE warnings

[WuphonsReach]

Define "many"?

I've gotten (8) drives into an Antec p160 case. Four below and four above (in a 4:3 CoolerMaster stacker unit that lets me put 4 drives into 3 5.25" bays with a 120mm cooling fan). I even had room to use the 4th 5.25" bay for the optical drive. The only unused bay in the unit is the 2nd 3.5" external bay.

I think with the newer p180/p180b case I could probably cram (10) drives into the unit.

Re:Big HUGE warnings

[Hillgiant]

you ARE making backups, right?

With what? 88+ dual layer DVD's. *shudder* I am rapidly approaching the point where I just mirror everything rather than make backups. Removable media is just not keeping up in either capacity or affordability.

Re:Big HUGE warnings

[syphax]

Back up "unlimited" data at carbonite.com for $5 a month. Software client is windows only at present, but works great. I haven't seen anything else even close for the price.

I have no affiliation.

Or, get a big USB/Firewire drive and back up occasionally to that.

Re:Big HUGE warnings

[StrawberryFrog]

If the failure rate of drives is constant (pretty close to reality), then
if you've got 7 drives and I've got 1, you're seven times more likely to lose a drive than I am.

Unfortunately, like many bad things (e.g. road accidents at a traffic black spot) the occurences are not constant and evenly distributed: they clump. For an example of how disastrously bad the consequences of this falacy can be, see here

If I have seven drives and one fails, then it raises the odds of a failure in the other six, assuming that they were bought together, and have the same make/model and date of manufacture.

The safest way is that if one fails, replace the lot.

Re:Big HUGE warnings

[mrchaotica]

Define "many"?

More than six internal 3.5" drive bays.

My Antec Plus 880 came with 3 5.25" bays, 2 external 3.5" bays, and 3 3.5" internal bays (in a removable cage thingy). I've since "upgraded" it to the equivalent of a 1080 by adding a second "cage thingy" to bring the total up to 6 internal drive bays, but I'd prefer to have even more.

Part of the reason for this is that I only keep 2 drives in each 3-bay group, so that there's space for airflow. In fact, each group has an 80mm fan directly in front of it, which is nice -- and this is the feature I've been unable to find in cheap cases.

I don't like using 5.25" bays for 3.5" disks.

Re:Big HUGE warnings

[mobby_6kl]

I just wish they made high-quality cases with that many [8?] drive bays

You're not looking very carefully then, here's one Chieftec case with 8 3.5" bays, here's another with 6. Based on personal experience, they're quality stuff.

Re:Big HUGE warnings

[mrchaotica]

Oh, sorry, I should have clarified: I'm looking for a mid tower case, not a full tower. The cheap cases I've seen are less than 2' tall.

Incidentally, my case is a similar design, but the midtower workstation version (without a stupid door over the 5.25" bays). I'd like a case with even more bays than it, but in the same size.

Re:Big HUGE warnings

[WuphonsReach]

Hmm... at that point you're looking at server-level cases, which are about 12-15" wide (usually 5u or 8u cases which are either rack or tower mountable). Sometimes called pedestal cases. I'd guesss that SuperMicro, Lian Li or Antec carry them, but prices for server-level cases are typically $200-$500.

Re:Big HUGE warnings

[ckaminski]

Well, I just lost two this morning. Two disks of a 4 disk RAID5 set.

Complacency and "It'll Never Happen To Me" will result in some very stinky gooey egg on your face someday. I've lost more than I care to remember. Some gave some indication that they were failing. Others just died.

Re:Big HUGE warnings

[Zathrus]

Plus who has room (and power) in their case for 7 drives? Datacenters, sure, but not home users.

My server currently has 5 drives - 1 40 GB, 1 120 GB, 2x250 GB (RAID1), and 1 300 GB. Plus a DVD-ROM. I still have space for another 3 or 4 drives (I've left space between drives for now, to improve circulation). The system already has 2 additional IDE controllers, and if I add more drives then I'll add another one (everything is on master) or a SATA controller.

And yes, this is a home server sitting under my stairs, all of it running on a 350W PSU. I could safely add several more drives without worrying about overtaxing the PSU. People really, really don't have any idea how much power their systems draw, either at boot (the only really stressful time for drives, esp. if you can't/don't stagger the power up) or during operation.

Personally, I'd love a couple of 750GB drives. Or even another 500 GB drive. I'm really uncomfy with not having the data on the 300GB drive backed up (and there's only 5GB free right now), and would like an offsite copy of the entire system for that matter.

Re:Big HUGE warnings

[Fweeky]

8 drive 4U Supermicro case. And here's a 7 drive one, which I have running in a cupboard a few ft away. SCA, fixed bays and more modest variants abound.

And here is a 12 drive Lian-Li, although you need to be very careful with cable routing and PSU/fan choice to fit them all in and keep them well cooled (the 6 drive one is awful for this). They also lack a reset button for some reason.

Re:Big HUGE warnings

[Anonymous Coward]

I recently lost 2 drives also.. both died at the same exact time. I thought for a while that something strange was going on to have 2 drives fail at exactly the same time, but I'm hearing more and more stories like this. Is it common to lose multiple hard drives at the same time, and if so, what the hell is the cause?

Re:Big HUGE warnings

[NeuralSpike]

True, but in order to gain similar capacity from a seven disk array of 100 GB drives you cannot have any redundancy. Thus, while the failure rate any individual drive is an independent event, the failure of any single drive in the array represents a failure of the array. In other words, if you flip seven coins what is the chance you'll come up with at least one heads. This is a situation of additive probability if one ignores any factors other that could affect a moderately sized array such as the example.

Re:Big HUGE warnings

[mrchaotica]

Not really -- the cheap cases I mentioned (that do have enough drive bays) are normal mid-tower cases, not server ones by any means. Surely there exist cases that are the same size, but just not so flimsy?

Re:Big HUGE warnings

[gelfling]

I can't explain that. If I had to guess I'd say that high end drives sacrifice reliability for performance and/or the usage patterns pushes them far beyond their designed parameters. After all if you're using a RAID array one could guess that the array is being very heavility utilized and that each drive is probably being pushed much harder than the requirements of having only one drive on a system - so they're just wearing out faster because of much harder use. Why do they fail at the same time? Good engineering tolerances for drives in the same production run that fail at precisely the same projected design limits.

Re:Big HUGE warnings

[ChrisA90278]

If you assume that all drive have equal probibilty of failure then if yo own seven 100Gb drives you have a 7X greater chance of failure then if you owned one drive.

But if you organized those seven 100GB drives into a 500GB RAID5 then you would have to loose two drives before you would loose data.

Re:Big HUGE warnings

[Fulcrum of Evil]

I can't explain that. If I had to guess I'd say that high end drives sacrifice reliability for performance and/or the usage patterns pushes them far beyond their designed parameters.

I can offer pointers, but the seagate drives are by design within their specs.

1. Heat: most drive failures are from heat
2. systemic failure - if one drive from a batch fails, then another from that batch is more likely to fail.

Re:Big HUGE warnings

[StikyPad]

That's why my "hard drive" is a RAIF 0. Just 177,777 floppy drives give me 250GB of rock solid storage, and in the unlikely event that one disk fails I can just replace it for a few cents. With a transfer rate of a typical floppy at around 500Kb/s, the theoretical peak transfer rate is around 10.6GB/s. PCI bottlenecks prevent me from testing actual peak throughput.

Naturally I'd like to get set up with RAIF 0+1, but I'm waiting for the kids to move out so I have somewhere to store the second drive bank. I had previously considered building an extension onto the house, but my research indicates that that is not a cost effective solution.

Re:Big HUGE warnings

[StikyPad]

Just forgot to mention: If anyone's thinking of emulating my RAIF solution, make sure to enclose the storage area in sound dampening material. Nothing's quite as startling as 178 thousand floppy drives making disk-seeking sounds in unison.

Re:Big HUGE warnings

[ebh]

Back in those days, spending $775 on anything was memorable. That disk drive cost more than my car.

Warranty

[Anonymous Coward]

I noticed the warranty(5 years) is still the same as it's predecessor. This is pretty good news as with any new technology, especially 1st editions, there are issues. Plus, I thought I read somewhere that there was a concern because it was perpendicular, there could potentially be more frequent data loss, but don't quote me on that.

Uh... no

[93 Escort Wagon]

"Throw in an industry-leading five year warranty and a cost per gigabyte that's competitive with 500 GB drives, and you may quickly find yourself scrambling to justify a need for 750 GB of storage capacity."

Maybe I hang around with normal people a bit too much, but I can't see myself getting hot and bothered over a new hard drive. If you need the capacity, then sure - this is great. But c'mon! As far as the "lust after" quotient goes, this isn't exactly in the same league as some new piece of Apple hardware. Heck, it's probably not even in the same league as a low-end Dell box.

Re:Uh... no

[Anonymous Coward]

>the "lust after" quotient
what do you think people will be storing 750GB of?

Re:Uh... no

[Randle_Revar]

Heck, it's probably not even in the same league as a low-end Dell box.
Definitely not in the same league - it is much more exciting than a low-end Dell.

Maybe I hang around with normal people a bit too much
Yes, I think that is the root of the problem here :-)

Re:Uh... no

[laffer1]

Hot and bothered is right. If you look at the specs, this thing uses more power and most likely gets hotter than its predecessors. I'm starting to care about power consumption with processors, and it seems only reasonable that be applied to other devices like drives and video cards as well. The largest hard drive in my home is 160gb right now. I could use a 700GB disk, but I'd rather have a drive that uses less power at say 300GB.

Re:Uh... no

[DrKC9N]

I can lust after the perpendicular technology alone...how much more for 750GB at $420! Yeah, I think you're a bit too normal to get excited about this one. ;)

Re:Uh... no

[Anonymous Coward]

Maybe I hang around with normal people a bit too much, but I can't see myself getting hot and bothered over a new hard drive.

Speak for yourself, I've never been so hard in my life.

Re:Uh... no

[Anonymous Coward]

As far as the "lust after" quotient goes, this isn't exactly in the same league as some new piece of Apple hardware. Heck, it's probably not even in the same league as a low-end Dell box.

Had to do it :)

Vincent Vega:
It's laying your hands in a familiar way on Marsellus' new wife. Is it as bad as eating her pussy out? No, but it's the same fucking ballpark.
Jules Winnfield:
Whoa, whoa, whoa, whoa. Stop right there. Eating the bitch out and giving the bitch a foot massage ain't even the same fucking thing.
Vincent Vega:
It's not. It's the same ballpark.
Jules Winnfield:
Ain't no fucking ballpark neither. Now look, maybe your method of massage differs from mine, but you know, touchin' his wife's feet and sticking your tongue and the holiest of holies ain't "the same fucking ballpark." It ain't the same league, it ain't even the same fucking sport. Look, foot massages don't mean shit.
Vincent Vega:
Have you ever given a foot massage?
Jules Winnfield:
Don't be tellin' me about foot massages. I'm the foot fucking master.
Vincent Vega:
You've given a lot of them?
Jules Winnfield:
Shit yeah. Got my technique down and everything. I don't be tickling or nothing.
Vincent Vega:
Would you give a guy a foot massage?
Jules Winnfield:
Fuck you.

Re:Now all I need...is a backup perhaps?

[WidescreenFreak]

You've never heard of this thing called a "backup", I take it?

Seriously, there is no reason whatsoever for anyone to lose any data. Even if it means forking over the money for a tape backup and tapes, if you lose any data due to a drive failure you have no one to blame but yourself. If it's important, build a RAID. If its critical, build a RAID with some kind of tape or other backup.

Yeah, I know, this is "Well, no shit, Sherlock" territory, but it always irks me when someone talks about losing data because there's no real need for losing any data, particularly if it's important.

Of course, if getting that data back is a simple task of downloading (again) from alt.binaries.multimedia.erotica, that's a different situation. :)

Re:Now all I need...is a backup perhaps?

[Kandenshi]

"Only wimps use tape backup: real men just upload their important stuff on ftp, and let the rest of the world mirror it." - Linus B. Torvalds

I suppose the same principle applies to newsgroups and eMule, etc... :P

Re:Big HUGE warnings - Not quite true

[drhamad]

7x100GB is not 7 times more reliable than one 750 GB drive. It is 7x more reliable at not losing ALL of your data, perhaps, since you could only lose 100GB at a time. But it is not any more reliable for retaining ALL of your data, either. The big advantage in reliability to high capacity drives is the ability to RAID them in a relatively small enclosure - RAIDing 7 or 8 drives would be quite a task, while doing 2-4 drives is relatively easy.

only 187 million times cheaper per bit

[Ancient_Hacker]

$413 sounds a bit pricey, but then I thought back to my fiurst disk srive, a DEC DF-32. Only 32,768 12-bit words!

Price I don't know, definitely no less than $5000 of 1972 dollars. That's about 78 bits per dollar.

This new disk is about 14634146341.463414634146341463415 bits per dollar that's an improvement of about 187 million times .

but wait those old dolalrs were at least 4 times more studly than today's, so that's about 600 million times better over the last 34 years. An annual rate of about 183% !

Re:only 187 million times cheaper per bit

[enrevanche]

is that compounded daily?

Re:only 187 million times cheaper per bit

[ChrisGilliard]

so that's about 600 million times better over the last 34 years. An annual rate of about 183%!

Yeah, the Moore's Law equiv. for storage has been pretty amazing on a dollar for dollar basis. It's actually an 83% increase per year, not 183%, but that's still pretty amazing. That means if we keep up this rate, by the year 2020, we should be discussing a 3 PetaByte (note: peta, not tera) drive for under $500. More likely, it would be a 1 PetaByte drive selling for between $100 - $200 because prices seem to decline as capacity increases. But, hey a 1 PetaByte drive ain't bad.

Re:only 187 million times cheaper per bit

[WuphonsReach]

$413 sounds pricey until, as you noted, you do the math for the $/GB amount. For being a leading-edge drive, the price per GB is rather competitive.

The following prices are estimates based on www.pricescan.com. There could be as much as +/- 10% variation in prices.

PATA drive prices
120GB $64 - $0.53/GB
160GB $70 - $0.44/GB
200GB $75 - $0.38/GB
250GB $80 - $0.32/GB
300GB $105 - $0.35/GB
400GB $195 - $0.49/GB
500GB $260 - $0.52/GB
750GB $490 - $0.65/GB

SATA Drive prices ($/GB)
120GB $68 - $0.57/GB
160GB $65 - $0.41/GB
200GB $76 - $0.38/GB
250GB $80 - $0.32/GB
300GB $105 - $0.35/GB
400GB $175 - $0.44/GB
500GB $250 - $0.50/GB
750GB $434 - $0.58/GB

Re:only 187 million times cheaper per bit

[ponden]

In addition to bits/dollar, contents/bits is also growing.

Drop of bits/dollar is caused by not only the technology improvemnt, but also the importance/bits is dropping.

Though improvement rate seems actually very high, something may be missed...

Re:only 187 million times cheaper per bit

[houghi]

An annual rate of about 183% !

That is also the annual increase in weight for most hackers. Coincidence? I think not.

Re:only 187 million times cheaper per bit

[Anonymous Coward]

DVD media is even cheaper; 100 TDK 4.7 gigabyte DVD-R disks costs $35 at the local Fry's. That's 7.4 cents/GB.

Re:only 187 million times cheaper per bit

[Anonymous Coward]

You know, Moore's "law", isn't a law?

It was (at best) a prediction using observations to date at the time.

Why do some people have to cling to such technobabble wreckage as if it's cast in stone like this?

Re:only 187 million times cheaper per bit

[ChrisGilliard]

Why do some people have to cling to such technobabble wreckage as if it's cast in stone like this?

I didn't cast it in stone. I did say "if this rate keeps up...". Maybe you missed that part. People probably discuss what you call technobabble wreckage because it's been remarkably accurate for somewhere around 30 years since the trend was noticed. Maybe you're right and the trend will not continue. Maybe this Seagate hard drive will be the highest capacity disk drive ever made. I think it's more likely that the trend will continue though even though I made no such prediction in the previous post. I wonder why this kind of progress troubles some people?

Re:only 187 million times cheaper per bit

[matt21811]

You are right. Paying double per gig doesnt seem like a rediculous premium for the drive.
It suprised me but your data doesnt show anything unusual. New large drives have always been around twice the megsper $ figure of the drive in the sweet spot.
Well, acording to my research anyway:
http://www.mattscomputertrends.com/harddiskdata.ht ml

Re:only 187 million times cheaper per bit

[matt21811]

My research shows the 85% average trend hasn't held for the last 3 years:
http://www.mattscomputertrends.com/harddrives.html
It has been closer to 30%.
It also shows that the trend seems to be slowing down.

Re:only 187 million times cheaper per bit

[WuphonsReach]

I dug up my prices from May 9th's check:

IDE/PATA
200GB $85 - $0.42/GB
250GB $95 - $0.38/GB
300GB $120 - $0.40/GB
400GB $220 - $0.55/GB
500GB $290 - $0.58/GB
750GB $495 - $0.66/GB

SATA
200GB $85 - $0.42/GB
250GB $95 - $0.38/GB
300GB $120 - $0.40/GB
400GB $180 - $0.45/GB
500GB $275 - $0.55/GB
750GB $480 - $0.64/GB

Most sizes show a ~$0.05/GB price drop in just a few weeks. Probably not unusual when a new, larger sized drive hits the market.

They probably could've charged a higher premium for those 750GB drives and they'd still sell. Price trend graphs are still pretty flat (PATA and SATA).

Re:only 187 million times cheaper per bit

[ChrisGilliard]

I took a look at the graphs. This is very good data. But, I'm not sure I agree with your conclusions. I don't think three years is enough to determine that there is a slow down. According to your data, there's a lot of volatility in this trend. Some years, it goes up by as much as 150%, others it only goes up by 6% (note: not in the last 3 years). Also, 2006 isn't even half way over yet. According to your data, we've already seen a 25% increase in 2006 and we're just begining June! By year end, the compound rate of increase will be greater than 70%. I also wouldn't be surprised if the rate increases for the rest of this year and brings us above the 81% mark. This is because the Segate hard drive has a new technology break through as this article indicated. New breakthroughs are expensive at first, but it will get cheaper fast. With that said, we really only have two slow years (2004, 2005). This is not very unusual. In '98 and '99, there were two slow years. These were followed by 4 years of fast growth. I would not be surprised to see a similar occurance in the remaining years in this decade.

Re:only 187 million times cheaper per bit

[matt21811]

Three years isnt a huge amount of time but it is the certainly longest period of poor growth in the survey and is hard to ignore.
Also, I know it is not clear from my graph but the increase that I have labeled "06" is the growth from March 05 to March 06. So it really is a full 3 years of poor growth.
I am also not expecting 4 years of fast growth from perpendicular recording. These 750 gigdrives are currently amost 200 Gigabits per square inch and they dont expect to push that past 1000 gigabits/sq in using perpendicular. Thats only a 5 fold increase or two and a bit doublings. Hitachi themselves say they are expecting to hit 230 Gbits/sq inch in 07. Perpendicular looks like will take a long time to play out meaning a continuation of ordinary results.

After looking at my data, interviews and data from manufacturers I have come to the conclusion that the hey day of hard disk is over. Hard drive based personal music players will all be replaced by flashed based devices in just 3 years. Disk drives in laptops will be going in 5 years. 3.5 inch drives will be the last hold out at about 10 years. These dates are taken from a simple graph that extrapolates the current 3 year trends. Sure some new break through could save the disk drive, but it had better hurry.

Re:only 187 million times cheaper per bit

[ChrisGilliard]

Hard drive based personal music players will all be replaced by flashed based devices in just 3 years. Disk drives in laptops will be going in 5 years. 3.5 inch drives will be the last hold out at about 10 years.

I agree with you on this point. Actually, the mp3 player I own uses flash and I'm happy with it. With the advent of the internet, the only place where mass storage will be needed is on servers. Even today, you can buy a 32 gB flash drive. For most people that is enough. I would definitly consider using a laptop with a 32 gB flash drive and have my media files stored on a server. If you have a fast internet connection, that's a pretty nice setup.

Why do airplanes only have 2 engines?

[Marx_Mrvelous]

Because with only 2, there is *less* risk of engine failure.

Having 7 drives increases your risk of failure by a factor of seven. Unless you mirror every drive, but then, you now have 14 disks v 2...

Re:Why do airplanes only have 2 engines?

[MBGMorden]

Ummm, WTF are you talking about? The phrase "Why do airplanes only have 2 engines?" doesn't make a bit of sense. SOME planes have 2 engines. Plenty have just 1. Some have 4. Some have even had 8 of them. And towed gliders don't have any (though they're not technically considered an airplane).

A plane will have multiple engines because it needs additional thrust, and that's often more efficiently gained by hanging on extra engines rather than just increasing the power of a single one.

Re:Why do airplanes only have 2 engines?

[jizmonkey]

Because with only 2, there is *less* risk of engine failure.

You aren't serious, are you? ETOPS ("engines turn or passengers swim") standards are only met by certain of the newest twin-engine airplane designs. Historically four-engined airplanes were allowed to go much further away from diversion airports than two-engined airplanes.

Re:Why do airplanes only have 2 engines?

[angrist]

No, there is MORE risk of engine failure. But when the first engine dies, you can still land on one. Just like a RAID mirror.

Re:Why do airplanes only have 2 engines?

[ars]

You can't compare engins and hard disks.

You can fly with just one engine, but if any of your drives die you are lost.

If you raid5, then 1 drive can die, vs all but one of them in an airplane.

I suppose you could raid0 all 7 drives, then you could compare engines and drives.

Re:Why do airplanes only have 2 engines?

[LWATCDR]

"Because with only 2, there is *less* risk of engine failure."
Actually light twins have a much worse safety record than light single engine aircraft.
For exactly the same reason that more drives are less reliable than a single drive.
If an air plane requires both engines to keep it in the air and each engine has a 1% chance of failing then you are twice as likely to have to make crash landing in a twin than a single. Most light twins can hardly climb with a single engine.
This is much the same situation as a computer using two drives as a RAID-0.
Now if the twin can fly fine on a single engine then you have less of a chance of a crash landing but you still have a greater chance of an emergency landing. Most airliners and all fighters are in this category.
This is the very like a computer using two drives but as a RAID-1.
Adding more engines or drives just makes the problem worse. The the B-29 lost on engine on take off you where going to have a really bad day. This is like having a four drive RAID-0.
What is interesting is the trade offs are almost identical. If you want the biggest load or the most space you have the highest risk.

Re:Why do airplanes only have 2 engines?

[mha]

Unfortunately waaaaayyyy off the topic of this discussion, but I've no choice! As a pilot I am forced to humbly but strongly object to two of your statements...

First, gliders ARE airplanes in every sense of the word including the legal one. It's not the engine that makes an airplane, it's the method of flight. Balloons are different, but gliders and a Boeing 747 use exactly the same principle and are therefore both called an airplane.

Second, having more than one engine IS done to increase the likelihood of survival when an engine quits e.g. while flying over water or mountains, where it's hard/impossible to find an area emergency landing, not to mention IFR flight conditions (little or no sight so for landing anywhere you need landing guidance systems).

There are quite a few small two-engine airplanes (and the worldwide fleet of 2-8 seat airplanes is much larger than in the 100+ seat category, of course most people don't notice them because when "normal" people fly it's in the large ones) that would be a lot more efficient, cheaper (to buy, to fly and to maintain), even faster and able to carry more load, etc. if the two engines would be replaced by a single one. Still, a lot of people buy them, and not just for introductary "multiple engine" rating training where the only reason to have two engines is the thing itself. They're wanted for IFR flying, for redundancy, to avoid having to land under conditions mentioned above when one engine quits.

Also, I'm not sure comparing RAID and multiple engines on aircrafts makes that much sense...

Re:Why do airplanes only have 2 engines?

[YrWrstNtmr]

Because with only 2, there is *less* risk of engine failure.

No, there is more risk of engine failure. It is, however, less catastrophic if you lose one engine, and still have the other one.

"I'd rather lose an engine, not the engine"

Re:Why do airplanes only have 2 engines?

[MBGMorden]

I'm a pilot too (though admittedly only PPL-ASEL). I said gliders are not considered airplanes because they aren't. A glider cert is completely seperate from airplane (for instance you can solo a glider at 14 years old. you have to be 16 to solo in an airplane). Also check the regs for right of way for example. Airplanes (and rotorcraft) must always give way to gliders (and balloons). Airplanes fall under ASEL, ASES, AMEL, AMES (with various other signoffs required for high perf, complex, tailwheel, etc, as I'm sure you know). Same principles yes, but they're not legally an airplane. Heck even motorgliders that DO have an engine are not legally airplanes, even though they can take off under their own power.

Also, as I'm sure you're aware, multi-engine airplanes aren't always more survivable in the event of an engine out. While virtually all twins can fly with a single engine, they can usually only do so at a significantly reduced altitude (at sea level), which may or may not be below ground level. It does you no good to be able to maintain 3000 feet if the ground is at 3500.

That being said, I'd agree that this topic has very little to do with RAID comparisons.

Re:Why do airplanes only have 2 engines?

[NeuralSpike]

Um, most of you guys replying to this post didn't read it, did you. The poster said planes only have two engines, implying the reason they don't have more than two engines. Large jet liners used to four engines, each of which was capable of allow a controlled landing because, while the chance of one engine failing was greater with more engines, the probability of having a remaining engine to control an emergency landing. As jet reliability has increased, large commercial jets have moved towards a two engine design since the probability of losing both engines is now sufficiently low as to not require four (or three) engines (as two engines having other economic benefits).

Re:Why do airplanes only have 2 engines?

[mha]

Right of way has nothing to do with airplane or not. What you say about gliders not being airplanes is simply not correct, period.

Re:Why do airplanes only have 2 engines?

[mha]

... and I'm well aware that many twins cannot fly anywhere near their altitude with just one engine, nethertheless that doesn't change any of the reasons why they are used instead of singl-engine airplanes, which almost always win in terms of efficiency at least in all the small airplanes (i.e. Boeing-size is an entirely different story). So I did not mention it in order to not increase size of the off-topic posting even further, which I now do anyway... ;-)

Re:Why do airplanes only have 2 engines?

[mha]

Okay, last comment - let's just call it an "aircraft" and I think we can both agree on that term :-)

On the twins, as I said it doesn't change the reasons for having them if they're pretty bad with one engine off - the performance with NO engine is even worse and that's what a twin avoids ;-)

Re:Why do airplanes only have 2 engines?

[MBGMorden]

You're going to have to do better than "simply not correct" on the glider point. Gliders are certified COMPLETELY seperately from airplanes. They're legally not the same thing. Heck a remote controlled electric toy can use the same physics of flight as an airplane (using a propeller to generate thrust and wings as a lifting surface), but they're certainly not legally considered an actual airplane anymore than a glider is.

Yes, from a layman's POV, it's an airplane in that it has wings and the same basic control surfaces, but from the FAA's POV, it's classified a glider, not an airplane.

Proven technology

[Anonymous Coward]

Until these drives have been in production longer, I wouldn't trust them.

Some of the early perpendicular storage drives by Hitachi exhibited bit-flipping problems. For important data, I'd much rather use the older technology.

Sure, they have a 5-year warranty - but the warranty doesn't cover your data. If it fails, you will get a new drive, but that won't do anything for your 750GB of TV shows. (or whatever you do with that much space)

I'd like the underlying technology to be at least half as old as I want to trust my data on the drive.

That's a lot of DVDs

[waif69]

If I were to backup a 750GB drive to DVD as I do currently, that would take quite some time and quite a few disks. I am too lazy right now to calculate how man or how much it would cost, but I think I would want to move to a different method of data backup.

Re:That's a lot of DVDs

[ericdano]

As in getting TWO of them, and mirroring them. When you get into 100s of Gigabytes, it doesn't make sense to use DVDs (right now, unless you have BluRay or something) to make backups. Get another drive of the same size, or two of them, and mirror them.

Re:That's a lot of DVDs

[Anonymous Coward]

For home/home-office use there is really only one practical way to back up huge hard drives - and that's onto other hard drives. However, for good security, you really need the backup drive(s) to be in another PC - and preferably in another building - and the network performance ends up killing you there.

So right now - I use sledded drives that I can remove from the PC and put somewhere safe.

Re:That's a lot of DVDs

[Anonymous Coward]

Assuming there is no loss in formatting (which we know isn't true) it would take 160 DVDs (4.7 GB). If you have dual layer disks, it would "only" take 80 disks to back up 750 GB. Because of marketing, this drive is probably closer to 700 GB in size, which would take about 150 (4.7 GB) DVDs. Time to go buy a couple spindles of DVDs, eh? Or...by two of these, and stick the second one in another computer somewhere and pray you have a nice upload speed.

Re:That's a lot of DVDs

[mrchaotica]

Mirroring provides redundancy, but not a backup. If anything happens to the box itself or the building it's stored in (e.g. a fire), you're still screwed.

You're on the right track with the "buy another hard drive" suggestion, except that you want to put it in a removable cage and store it in a safe deposit box at the bank. (And preferably get two backup discs, so that you can be copying to one while the other is safe in the vault.)

Re:That's a lot of DVDs

[Anonymous Coward]

It would take a similar amount of time and DVD's to backup 750GB of data from multiple disks too :)

It's really a matter of having that much data and which disk configuration you want to go with based on your speed, redundancy, availability needs...and, of course, budget.

Re:That's a lot of DVDs

[Anonymous Coward]

Keeping hard disk in storage as a backup device is said to be very risky.

As I've read in an article from a well known recording producer, entire tour recordings were lost on hard drives, kept in storage for an extended period of time. Supposedly your hard drive will loose data if you don't hook it up every two months or so "to keep it in shape". I was quite surprized to learn this, but it's coming from a guy, who's life line depends on reliable backup and he had tried all available tape, etc. solutions.

What is your experience with off-site hard drive as a backup solution?

If that does not work, off-site backup for huge hard drives is a huge problem, since tape backup solutions are not getting similarly cheaper per gigabyte.

Re:That's a lot of DVDs

[abb3w]

So right now - I use sledded drives that I can remove from the PC and put somewhere safe.

External SATA enclosures can be another practical alternative, if you have a machine with external SATA connectors.

Re:That's a lot of DVDs

[mrchaotica]

Supposedly your hard drive will loose data if you don't hook it up every two months or so "to keep it in shape".

The idea would be to do weekly or monthly backups or whatever, rotating the disks each time. E.g. back up on one and put it in the bank, then a month later back up on the other and put it in the bank while taking the first one out to be used the next month. In other words, do it like you would with tapes.

What is your experience with off-site hard drive as a backup solution?

Absolutely none -- if one of my hard drives dies, I'm screwed. I'm a college student, and (until I get another paycheck or two from my summer job) I'm too poor to afford proper backups. And I certainly can't afford tape!

Re:That's a lot of DVDs

[hackstraw]

Mirroring provides redundancy, but not a backup. If anything happens to the box itself or the building it's stored in (e.g. a fire), you're still screwed.

You're right. Its so important to have offsite backups for reliability.

Think about it. If you're diligent about it, and take offsite backups from your house to your work or neighbors or whatnot, when you're house burns down you can still have all of your MP3s, TV shows, and everything!

The bitch is that you forgot to do redundancy and offsite backups of your couch, TV, computer, stereo, wife, kids, and pets.

Re:That's a lot of DVDs

[mrchaotica]

It's harder to replace a whole lot of downloaded information than it is a TV or a couch. Also, unlike data, the wife, kids, and pets can get out under their own power.

Re:That's a lot of DVDs

[Anonymous Coward]

that's why you have another family on the other coast

Re:That's a lot of DVDs

[Anonymous Coward]

The bitch is that you forgot to do redundancy and offsite backups of your couch, TV, computer, stereo, wife, kids, and pets.

No I didn't. I'm a mormon.

Re:That's a lot of DVDs

[QuoteMstr]

I actually just implemented a DVD-based backup solution for my employer. While DVDs are much smaller than tapes, they're much cheaper, especially when buying in bulk. They're so cheap, in fact, that we can afford to do nothing but level-0 backups, which eliminates a great deal of complexity. We deal with the optical degradation problem by using error-correcting codes to gaurantee against the loss of one disc from any given backup set --- and we usually make multiple sets anyway. All in all, it's not a terrible system.

Bad math..

[JMZero]

If you've got 7 drives and I've got 1, you're seven times more likely to lose a drive than
I am.

Let's say each drive has a 20% chance of failing. So if you have seven of them, do you have a 140% chance of one failing? Of course not. What you really have is 80%^7 percent chance of them all remaining OK. 80%^7 = 21%. Thus you have around a 79% chance of failure with 7 drives (if they all have 20% failure rate).

Your point still stands - but I noticed pretty much all of the replies to this guy used the same bad math.

Re:Bad math..

[LFS.Morpheus]

Uh, you're still 7 times more likely to lose a drive.

Re:Bad math..

[Anonymous Coward]

Er, you're not. As the post you replied to showed you, you're ~4 times as likely to lose a drive.

Re:Bad math..

[Zathrus]

I noticed pretty much all of the replies to this guy used the same bad math

Except that your math only stands if you're calculating the loss of the entire drive (or drive set) as opposed to any data loss. The chance of data loss (of any kind) is greater with 7 drives than it is with a single drive. The difference is merely how much data is lost. Which is of little comfort if the drive that fails is the "really important" one.

You're still playing Russian Roulette, but this time you're aiming at various body parts instead of just your head.

Re:Bad math..

[enrevanche]

Not true, if you flip a coin 3 times are you 3 times as likely to get a tail. i.e. do you have a 150% chance of getting at least one tail? no you have 1 - 0.5^3 = 1 - 0.125 = 87.5% chance. There is 12.5% chance that you would get no tails. With drive failure it works the same way, you have a chance at no failures and also a chance of multiple failures.

7 times as many failures (over a large number of samples) is not the same as 7 times the chance to have a failure.

Re:Bad math..

[pla]

Thus you have around a 79% chance of failure with 7 drives (if they all have 20% failure rate).

IF you have a 20% failure rate.

It cheats somewhat to use that as an example, however, because with the real probabilities involved, you approach a linear trend with the number of drives.

Let's try an MBTF of 50k hours. That gives us a 0.002% chance of failure per hour. Take 0.99998 to the seventh, and we get 0.9998600084... Or "seven times as likely", accurate to better than one part in a thousand.


Though, I will admit doubt that the GP explicitly took that into consideration in his statement. ;-)

Re:Bad math..

[LFS.Morpheus]

Thank you for explaining with math what is quite obvious.

No, not being sarcastic - thank you.

Re:Bad math..

[RedWizzard]

Thank you for explaining with math what is quite obvious.

It's only quite obvious if you don't understand statistics. In fact it only works out to be so close to 7 times more likely to lose data because the chance of a failure is so low in the first place.

Re:Bad math..

[RedWizzard]

Except that your math only stands if you're calculating the loss of the entire drive (or drive set) as opposed to any data loss.

No. The chance of losing all 7 drives (using the 20% per drive chance of failure) is 0.2^7 = 0.00128%.

The chance of data loss (of any kind) is greater with 7 drives than it is with a single drive.

Yes, it is greater. 79% vs 20%, just as the OP said. But it's not 7 times greater. Let me go through it again:

The chance of losing at least one drive is equal to 100% minus the chance of not losing any drive. Each drive has an 80% chance of surviving (1-0.2). So the chance of all drives surviving is 0.8^7 = 21%. Therefore the chance of at least one drive failing is 79%. This is basic statistics.

As someone else pointed out, using real failure rates it does in fact work out to be almost exactly 7 times the failure rate of one drive. However that is only because the failure rates are so low. Most of the people in this discussion clearly don't understand the mathematics at all.

Re:Bad math..

[MechaStreisand]

It's not bad math, though, if the percentages are small: say, around .5%. Then simply adding the percentages makes for a pretty good approximation. Doing it that way you get 3.5%, and doing it the mathematically correct way you get 3.45%. Pretty close. And the smaller the percentages, the closer the approximation is.

Re:Bad math..

[Fulcrum of Evil]

Let's try an MBTF of 50k hours. That gives us a 0.002% chance of failure per hour. Take 0.99998 to the seventh, and we get 0.9998600084... Or "seven times as likely", accurate to better than one part in a thousand.

Looks closer to ten times as likely to me.

Re:Bad math..

[pla]

Looks closer to ten times as likely to me.

1 / 50000 = 0.00002
1 - 0.00002 = 0.99998
0.99998 ^ 7 = 0.9998600084...
1 - 0.9998600084... = 0.000139991600...

0.000139991600... / 0.00002 = 6.99958...


Where did you get 10x from?

I can't pronounce Bacaruda

[MrDoh!]

"Seagate's Barracuda-7200.10" Would that be Babararacucudada

and only 10K rpm

[Anonymous Coward]

Nice size, but at only 10K rpms, you're leaving 33% performance on the table.

The 15K drives really do make a difference in many workloads.
Maybe we should compare 7x 120GB drives in RAID0 stripe array?
About the same risk of data loss, and maybe even better streaming performance?

Re:Now all I need...is a backup perhaps?

[gid13]

"there is no reason whatsoever for anyone to lose any data. Even if it means forking over the money"

Psst... Money is a reason.

Re:Big HUGE warnings - Not quite true

[mrchaotica]

RAIDing 7 or 8 drives would be quite a task, while doing 2-4 drives is relatively easy.

On the other hand, more drives would give better performance and capacity, since you could use fancier RAID levels (e.g. 5 instead of 1).

Well...

[GmAz]

Technically, I already have 830gb of storage, but it consists of a 500gb Raid0 array, a 200gb backup drive and an 80gb external. Personally, I would say to go out and buy two 400gb drives and put them in Raid0. In my experience, it is much faster than a single drive. My boot time is under 20 seconds easy after the CMOS finishes loading.

Re:Well...

[drinkypoo]

I have two 80GB 7200 RPM disks in a RAID 0 and yes, it's dramatically faster. I did some benchmarks (we all know what those are worth, but...) and I gained about 90% speed by adding the second drive.

Unfortunately, I don't have any way to back up 150GB (actual usable space) and if I lose one drive, I lose everything...

I'm thinking the best idea for the way I use my system is to get like five drives and make four of them into a RAID 5, using the last one as a scratch volume for photoshop and such, until such a time as some other drive dies and I need it for a hot spare.

Re:Well...

[CharlieHedlin]

So you just doubled your power requirements, heat output and chance of failure. I would much rather have 2 750GB drives with Raid 1.

I just wish there was an affordable removable media alternative. If I want to have 750GB of storage I have to buy it twice, probably 3 times (online raid 1 for reliability, and an offline drive for backups). In a datacenter enviroment, a nice robotic LTO2 system helps, but I can buy a lot of hard drives for the price of one of those.

Re:Well...

[Anonymous Coward]

Stick a 750gig in a external usb case.

Re:Well...

[Vancorps]

LTO3 is far and away the better deal. Looking at library system to backup about 20tb and you can do it for about 10k from Quantum. Combine that with an additional drive for performance and you've got yourself a good way to backup data. At $70 per 400/800 gig tape thats a pretty good price point. So build your array of 12 of these 750gig joys and then all is well. Of course my new arrays have 38 drives in them so this will get even more important.

I'm liking this trend, this is a bigger step than previous tries although my concern is over reliability still much like how drives over 160gigs failed more often that larger 250gig and later 300 gig drives. Now reliability is up on them and you can have 400gig drives in your array without sweating too much. Not sure about the 500s yet. Still this 250gig leap might make it worth while to sweat it out a bit as long as the library keeps churning away.

Re:Well...

[WuphonsReach]

Unfortunately, I don't have any way to back up 150GB (actual usable space) and if I lose one drive, I lose everything...

Why not? Prices on 200-300GB drives are well below $100 now. Even if you factor in the cost of an external USB case for those drives you can still backup for fairly cheap.

Some of the drives are now as low as $0.32/GB (the sweet spot is the 250GB drives).

Re:Well...

[egarland]

I wouldn't bother leaving the last drive out of the Raid5 array. I have a software RAID 6 array and it's speed is limited by the speed of the drives, not by the parity calculations. Just get a nice dual-core cpu so one can do the RAID math while the other runs your application. RAID parity calculations are very easy work for CPUs these days and so generally the more drives in an array the faster it is, especially with bulk data transfer. Another good option is to just get more RAM so you don't really use that scratch space as much. I'm planning on 4 GB in my next box since with DDR2 that's only about $300. That's a lot of pixels.

Also consider mixing protection schemes. My current desktop box has 2 drives with 3 partitions on each. The first partition on each drive is a solo partition (1 OS on each drive so I have a backup in case one has trouble), the second is a striped set for things I want to be fast and don't care about reliability (games, swap, etc) and the third partition is a mirrored set for data I want to keep safe. This combination gives me everything I need without having to power and have room for a whole array of drives in my desktop. The only issue I had was having to hack XP to support mirroring since it's crippled by default (shame on you Microsoft) but that was made much easier by having the second OS partition around. Hopefully with Vista they won't be so dumb.

I'm very happy with this mixed protection scheme and I'll probably do the same thing on all the computers I build soon. I'd like to see retail boxes configured like this by default. It makes a ton of sense, even for casual users.

Re:Well...

[drinkypoo]

As soon as I have "under" $100 that's not going straight into car parts, I'll set it aside for computer crap :P Just because I'm employed in a technical field doesn't mean I have money...

Re:Well...

[Emetophobe]

Raid0 offers no redundancy, it's only good for speed. If either one of the two drives dies, you're hosed. Raid1 or higher would be better. My boot time is also under 20 seconds, and I only have 1 250gig hard drive.

Re:Well...

[GmAz]

As for power consumption, its no biggie. I have a 500w PSU in there that takes care of everything. My video card sitting there idle probably takes up more power than the three hard drives. As for redundancy, yes, I would like to go Raid5, but I would need to get at least one more HD and that cost $$$. If I were to lose a drive, I woulnd't be losing anything major. I do regular backups of my important data and if its not backup'd, then its all good. Its not like this is a business computer or tracks my vast finances...if I had them that is. I have known many people that have had their Raid0 config for several years before any issues and to me, that is a good amount of time since I upgrade every two years max.

This will be great for...

[Go MSFT, stop Linux!]

Microsoft Windows Vista :D

Re:They left out...

[Anonymous Coward]

holy shit, Hitachi's completely on acid!

its funny...

[Connie_Lingus]

..but those 5-year warrenties don't really help you much if you FORGET THE BACKUP THE FRIGGING DRIVE!

customer: "my drive failed...i would like it replaced"

company: "sure..here is your new one!"

customer: "uhhh...what happened to my data?"

Re:its funny...

[Slashcrap]

customer: "uhhh...what happened to my data?"

BOFH: "You will have to re-enter it all manually. And remember that hard drives store data in binary, so you will have to use only the 0 and 1 keys."

Re:its funny...

[DDLKermit007]

company: "n00b"

Re:its funny...

[Anonymous Coward]

Keys?!

How about "Here is your perpendicular electro-magnet"

Re:Now all I need...is a backup perhaps?

[SatanicPuppy]

I use my gmail account for that. Encrypt it, mail it to yourself, and let Google back it up. I can fit almost all the really important stuff in 2 gigs. As for the rest of it, I'll just have to hope my house doesn't burn down.

Re:Big HUGE warnings - Not quite true

[Jeff DeMaagd]

RAID 5 isn't necessarily faster, but it's a lot better for data redundancy, so you can keep going when a drive fails. I wouldn't bother with RAID 0 unless you have a heavy backup system (like 0+1), and mirroring is a bit inefficient.

I just set up a 4x400GB drive array because Best Buy had the Seagates on sale, and then bought an old PCI-X 3Ware Escalade and put it in a 64bit / 66MHz PCI slot, giving me a pretty fast 1.2TB volume.

Re:Now all I need...is a backup perhaps?

[Kandenshi]

If you need more, just give yourself another gmail invite?

Have two accounts for storage then...

I've got a pair of gmail accounts, one for more business stuff, and one for family, friends, and fun. I don't see why you couldn't have storagemule1@gmail.com and storagemule2@gmail.com if space is an issue for 'the rest of it'

why not?

[Skadet]

Since these devices are preferably low noise, low power, and small in size...



Um...preferred by whom? I work in media and I can tell you, space is preferred, period. Nobody cares about the power draw of one more drive or the whirling of another disk.

Of course, if you're working with a lot of video media, you're probably not storing it locally, anyway. In fact, we don't even store our audio locally.

Re:why not?

[mrchaotica]

He's talking about PVRs in people's homes, not anything for professional use.

Re:Now all I need...is a backup perhaps?

[graemecoates]

Seriously, there is no reason whatsoever for anyone to lose any data. Even if it means forking over the money for a tape backup and tapes, if you lose any data due to a drive failure you have no one to blame but yourself. If it's important, build a RAID. If its critical, build a RAID with some kind of tape or other backup.

Personally, I would not want to be doing tape backups of a 750GB drive that is even a quarter full (unless I/company have serious wads of cash to spend). A tape drive with enough capacity to handle the entire drive in one go will cost about 20 times the price of the hd (Seagate is ~400USD from TFA). If you go for smaller tapes, you're probably still talking multiple times the cost of the Seagate for the tape drive itself + media costs. And there's the time for doing the backup, swapping discs, etc.

If you're serious about backing up your files/music/pr0n, then you are probably better off buying multiple drives and sticking them in a mirrored RAID array of some sort (RAID 1,5,51,etc) and hoping they don't all kark it at once if you're protecting yourself against failure.

If you want to protect against stupidity and the bad luck of having multiple drives die at once, then having another removable drive as a mirror that won't get blown away if you do a "rm -rf /" is helpful...

Of course, as you say, if it is critical, it's likely the data belongs to someone else and they'll pay for the kit to back it all up ;)

Re:Big HUGE warnings - Not quite true

[drinkypoo]

RAID 5 is typically pretty slow on writes, but about as fast as a RAID 0 on reads. As you likely know, it also sacrifices one drive's worth of space for parity information regardless of the number of drives in the stripe. Anyway if you are running a multiuser environment with a bunch of ordinary users, most of whom are reading more than writing, a RAID 5 makes sense. If you have a database with lots of reads and few writes, likewise. If you are doing video, where you are reading in a lot of data, and writing out a lot of data, it makes more sense to use a RAID 0, and make backups. If you cannot afford to drop back to the last backup, then you need a RAID 0+1.

"Myth boxes and the like"

[debest]

It was in the subject, and the body. I was referring specifically to "media PCs" (as well as DVR/PVR boxes), where size, noise, and power DO matter! I don't have a data centre, I want a solution appropriate for my home.

Re:"Myth boxes and the like"

[d3ac0n]

If I hadn't already replied to someone else in this story, I would use one of my 5 Mod points to mod you up JUST for having a Veggietales related sig.

Now, where IS my hairbrush?...

Re:"Myth boxes and the like"

[debest]

Wow, I've been cycling through assorted bizarre VT quotes for months, and no one notices. Now, I've had two shout-outs in four days! (Here (at the bottom) is Friday's reply, just before I changed sigs on the weekend.

It is great to know that there are other VT fans out there in /. land, though.

Ah, what the hey, I'll change it again!

Re:"Myth boxes and the like"

[Anonymous Coward]

Ah veggie 'lets brainwash our kids with christian propaganda' tales

Re:"Myth boxes and the like"

[xwipeoutx]

I didn't recognize it until you pointed it out... Thanks :)

Lame excuse.

[WidescreenFreak]

Then you shouldn't be buying such a large hard drive if you can't afford to lose the data that's on it without redundancy or archiving capabilities. That's like buying a luxury car when you can't afford insurance for it.

Re:Lame excuse.

[Anonymous Coward]

Not really. Drives are cheap, getting a good backup working is considerably more expensive and time consuming. i.e. buying a drive, plugging in into your motherboard, done, vs buying a drive, an enclosure and a fireproof safe, and spending at least 5 minutes per day setting up, running your backup script, and locking it away again.
If the data is worth thousands it justifies the employee time involved. Hundreds, nope.

Re:Lame excuse.

[WidescreenFreak]

Even a redundant RAID would be inexpensive enough for most people. Frankly, I question any corporate entitiy that uses IDE as an enterprise, data storage solution, but if it's the most feasible alternative, two drives in a mirrored configuration is still more acceptable than a drive by itself.

Re:Lame excuse.

[Jim_Maryland]

In a corporate environment, I agree with you entirely. Personally, I like NetApp for the NAS, Veritas or NetBackup for software, and Exabyte for tapes (although it's been a while since I've had to deal with the tape hardware). For personal use, most people are of the mindset to get the most CPU/memory/hard disk/etc... they can afford. They aren't likely to be thinking to back it up unless they've had a failure of some sort in the past. Come to think of it, I haven't exactly seen much reference to redundancy on any of the hard disk I've bought recently nor did any of the sales associates mention redundancy or backups when I talked to them about hard disk. I'm guilty myself of not doing regular backups though as I just had a problem with my mother-in-law bumping the reset switch while running a disk defrag this weekend. My backups were about 1 year old for a fairly complete backup. For home use, generally I run a backup when I start noticing disk errors indicating failure is imminent. Yeah, not the best practice but I kept holding off buying a new hard disk to mirror as I was planning on building a new system in which case I figured to mirror the OS at a minimum. I didn't have a plan yet for a data share but was considering a simple 3 disk RAID 5 but I still hadn't thought out a backup to media.

Re:Lame excuse.

[Anonymous Coward]

What? Man fuck you.

Get back to us when get in touch with reality.

Re:Lame excuse.

[PeelBoy]

So if I can afford a car I can afford a house with a garage to put it in?

Re:Lame excuse.

[buck_wild]

If you bought the luxury car BEFORE you bought the garage to keep it in, then that's just sad.

FIVE WORDS : ANIME, MUSIC, PORN, GAMES, MOVIES

[unity100]

And that is the end of that. Keep it comin

Re:FIVE WORDS : ANIME, MUSIC, PORN, GAMES, MOVIES

[meringuoid]

ANIME, MUSIC, PORN, GAMES, MOVIES And that is the end of that. Keep it comin

Hey! Who are you, and how do you know what's on my hard drive?

Re:FIVE WORDS : ANIME, MUSIC, PORN, GAMES, MOVIES

[unity100]

We have been watching you, citizen. You better repent and redeem before our lawyers come to get you - they are coming, they have no mercy, and lo ! Here is RIAA, MPAA and NSA, the omnipotent collective of 'A's that are shaping your world !

Re:FIVE WORDS : ANIME, MUSIC, PORN, GAMES, MOVIES

[jollyroger1210]

He is from the NSA. They are tapping your DSL/Cable too.

Not just RAID controllers...

[Anonymous Coward]

...but how do these 300MB/s SATA NCQ drives actually fare against U160/U320 SCSI drives for sustained thruput in something like a database server that normally benefits from the multithreaded i/o capability of SCSI? The "300MB/s" is pretty close to the "U320" rating of peak data xfer rate, but as we all know, the absolute very best and fastest disks themselves can generally only stream a continuous ~ 80MB/s due to mechanical limits of the hard drive regardless of the electrical interface, and most commodity-grade disk drives on the market today actually do well to reach and sustain ~50-60MB/s continuous stream rate, with ~30-40MB/s being common for low-end cheap drives.

I'm betting that in a situation where you need the utmost in high-traffic-load, direct-attached storage like on a heavily loaded transactional database server running Oracle or similar, that the U320 SCSI disks connected to a good hardware-caching raid controller card still are the unbeatable king daddy paw-paw of sustained thruput.

Re:Now all I need...is a backup perhaps?

[SatanicPuppy]

I like to think of it as my safety deposit box...I do some thinking about the stuff that I absolutely have to have...Everything else stays on my home network, which is pretty solid. I don't worry too much about home stuff. I'm not a video/music/graphic guy, so my stuff is low-bulk enough that I can burn a cd every month or so to back up important stuff, and not really worry about losing much.

Big Big Drives are great...but backup is a problem

[haplo21112]

This drive increases the ever widening gap between available storage and backup media. Great I can buy a 750GB drive...however how the hell am I gonna back this thing up...actually even with many many dics how am I gonna backup 750GB. There is a huge disparity in the amount of data we can store these days and the stuff we have to back it up. There is no afforadable backup solution for this much data.

you failed prob and stat didn't you?

[Anonymous Coward]

then the risk of a failure in any of your 7 disks is .07%

You failed statistics didn't you.

If you flip a coin 100 times and it comes up heads 98 times what is the probability it will be heads on the next flip?

dumbass.

Re:you failed prob and stat didn't you?

[adamdeprince]

Now be nice. He is right, 0.07% is correct.

The chance of any one failing is the same as 1 - the chance of none failing. If you have n gadgets with a risk of failure of m, the risk of any one or more failing is 1-((1-m)^n). For 0.01% (0.0001) and 7, that puts your risk of any disk failing at about 0.07%. 1 - (1 - 0.0001 ) ^ 7) = 0.0006997901 ... or about 0.07%.

The idea is the chance of any one or more failing is really the flip side of the chance of none failing. If any item has a p chance of happening, then the chance of them all happening is p1 * p2 * ... pn. So, if we have a chance of one drives survival at .9999, the likelyhood that all will survive is that 0.9999 ** 7.

Re:you failed prob and stat didn't you?

[Anonymous Coward]

Wrong scenario. You need to flip 7 coins and then ask what is the probability that they will
all be heads or all tails, or 1 head or 1 tails, etc. For each individual flip the probability
is the same, but for the whole group...that's a little different.

Magneto

[OYAHHH]

Do,

You suppose this drive uses technology similar to that Magneto uses to achieve all of his Mutant feats?

Re:Magneto

[VinB]

+1 for interesting punctuation. Reminds me of a Haiku.

Re:Now all I need...is a backup perhaps?

[Anonymous Coward]

Side note: rm -rf / will wipe out all the mounted drives too, make sure you unplug the backup drive when you're done with it.

750 GB

[Mister Jimm]

Men are like gas, they fill the space available. -- Red Green

The real reason why this drive is great

[shodanx]

while buying this product doesn't make much sense economically (unless you consider the cost per "bay" in your server and you overpaid your server), there is one reason this drive is great for the rest of us since it came out on the retail market about 2-3 weeks ago the price of almost all the other drives dropped significantly here are the price per GB at my fav wholesaler (eprom.com) 40gb 1.2$CAD/gb 80gb 0.675$CAD/gb 120gb 0.608$CAD/gb 160gb 0.481$CAD/gb 200gb 0.425$CAD/gb 250gb 0.356$CAD/gb 300gb 0.386$CAD/gb 400gb 0.472$CAD/gb 500gb 0.546$CAD/gb 750gb 0.705$CAD/gb as you can see the 750gb has the second worst price per gb, of course part of this price is the extended warranty but from my experience the very high reliability of seagate drives makes warranty not all that valuable so unless you server has high cost per bay it's not really worth it (you server has to cost over 130$ per bay which is ludicrous considering my lastest 12 bay system costed only 41.08$CAD per bay) and don't get me started on electricity usage & heat & noise, since a proper case won't let much noise escape or will just drown the noise with it's fans or will be in a location where noise isn't important (it's a networked server who cares if it's in the attic) and the heat being only 40 btu per drive (or a third of the heat of a fluorescent tube not even counting it's ballast) and finally electricity usage is just as insignificant at 6$CAD per year of utilisation however this product is great because suckers will buy it at great markup for seagate which will cover a part of the cost of developpement of this technology which will last until the 5 terabyte drives and from now on seagate will probably start releasing higher capacity model more regularly meaning an even lower cost per gb in the near future (disk capacities more or less stalled in the past few months this will easily put hard drives back on track of the "capacity schedule") but I do have a question , how do you backup this much data at a low cost? tapes are out afaik since they're way more expensive per gb (reusability isn't helping much since you need at least 1:1 the storage capacity of your server at all time, but probably even more than that to compensate for failures and redundancy) than harddrives and also a lot more cumbersome to use dvd are even more annoying but this is offset by the fact that they only cost 0.0744$CAD/gb which make a disk burning robot probably a economically viable option next generation disk won't be below the cost of dvds for a long long time (if ever) so they're out too for now so is there any other option or if not what's the market for disk burning (and maybe disk loading, like a automated dvd carroussel with an integrated reader) robot these days ?

Re:The real reason why this drive is great

[Anonymous Coward]

Wow! The drive is so great it even breaks the keyboard's enter key!!

Re:The real reason why this drive is great

[shodanx]

that's weird when I posted my text was properly formatted damn this forum software sucks !

Re:The real reason why this drive is great

[HolyCrapSCOsux]

(it's a networked server who cares if it's in the attic)

This may be a good idea up there in the great north, but down here in Nevada, My attic regularly reaches 160degF+ in the summer.

I keep mine next to the first floor AC collector. gets lotsa air and if it catches fire, everybody smells it in a hurry!

Re:The real reason why this drive is great

[ars]

I'll give tou two tips:

1. Preview

2. Go into your prefs and change Comment Post Mode to 'Plain Old Text'.

Re:The real reason why this drive is great

[roscivs]

Also note that HTML tags (like bold, italics, etc.) still work properly when 'Plain Old Text' is the comment post mode. I don't know why it's not the default for everyone.

Re:The real reason why this drive is great

[Anonymous Coward]

Well, it's default for ACs. I can't speak for users, since I don't have an account yet; I'm planning to get one soon.

I'm afraid you've misunderstood my post.

[JMZero]

But I don't think I can make it any clearer. Maybe I'll try: Your chance of losing one or more of the drives is around 79% if you have 7 drives (and they each fail at a rate of 20%). That's definitely more than the rate of 1 drive - but it's not nearly 7 times the rate (which would be 140%).

Re:I'm afraid you've misunderstood my post.

[SmittyTheBold]

This is a ridiculous way of measuring it anyway, and is NOT what the pros do. So, 79% chance of failure, when? Today? In the next year? The September 23rd? You're all arguing about pointless numbers.

You know that MTBF number that drive manufacturers provide? It's there for a reason.

Wayback archive of IBM's MTBF documentation

Re:Big HUGE warnings - Not quite true

[swillden]

RAIDing 7 or 8 drives would be quite a task, while doing 2-4 drives is relatively easy.

For those who don't see the difference: Most boxes don't have controller capacity for more than four drives (two PATA channels and two SATA channels) and seven or eight drives will also strain your PSU and your cooling capacity. Might be hard to fit in your case, too.

Of course, once you solve those problems, actually setting up the RAID is no different whether you have three drives or 30. A little more typing, maybe.

My file server has six disks in it, BTW, so I've worked through all of this. I can easily add a seventh without any trouble. An eighth would require a new controller card. I'm not sure how many drives I can add before my 550W PSU starts to have trouble. My cooling solution is low-tech, loud and very effective: The side of the case is off and I have a 30-inch box fan (the kind you mount in a window to cool your house) blowing into it.

One nifty trick I discovered is that if you slice all of your disks up into many small partitions, then create many RAID-5 arrays (using partition 1 on each disk to create the first array, etc.), then use LVM to bind all the arrays together you can add additional disks and rebuild the arrays without having to find some way to back up all of the data first.

I just added a 500GB drive to my system and I'm in the process of changing all of my four-disk RAID-5 arrays to five-disk RAID-5 arrays. The process works like this:

1. Use pvmove to migrate all of the data off of an existing four-disk array.
2. Use vgreduce to remove the now-unused array from the volume group.
3. Use pvremove to remove the LVM superblock from the array.
4. Use mdadm to stop the array and clear the md superblocks on the partitions.
5. Use mdadm to construct a new five-disk array from the four partitions that made up the old array, plus a fifth partition from the new disk.
6. Use pvcreate to add an LVM superblock to the array.
7. Use vgextend to add the array into the volume group.
8. Go back to step 1 with the next four-disk array, until they've all been converted.

This assumes Linux, obviously, is a bit tedious and requires that your LVM volume have enough free space so you can drop an array out of it. It's a whole lot easier than trying to figure out how to back up a TB+ of data so that you can rebuild your array, though. In my case, there's an additional step right after step four -- because my new drive is SATA and Linux doesn't support more than 15 partitions on an SATA drive, I'm moving from using 20GB partitions to 40GB partitions. So after I kill each pair of four-disk arrays, I repartition the drive to merge the partitions.

Let me tell you... repartitioning all of the disks holding my data made me more than a little nervous at first :-) I kept backups of the partition tables, just in case, but it actually worked just fine. Next time, though, I think I'll just create a single partition and use LVM to chop it into pieces which I can RAID together. So I'll have LVM over RAID over LVM. Sounds weird, but it makes a lot of practical sense.

Re:Big Big Drives are great...but backup is a prob

[Chirs]

The most effective thing I've found is to buy two disks and use the second to back up the first. Not so good for offsite backups, but protects against disk failure at least.

Re:Now all I need...is a backup perhaps?

[scarolan]

Money is only a valid reason if the data is worth LESS than the cost of a backup solution. In many cases, the nominal cost of a backup solution is far less than the value of the data on the computer.

Someone may say "I can't afford a new $80 drive to back up my data." But when they lose years of family photos and other documents, that $80 doesn't seem like so much compared to the hundreds or thousands of dollars it costs to do data recovery on a broken hard drive.

RAID1 mirroring is the mosty cost-effective way

[Anonymous Coward]

Q: What's the fastest way to backup a hard drive?
A: Well, to another hard drive, of course.
Q: And how should you back up one hard drive to another?
A: In real-time, of course. (e.g. mirror them)

Now of course, that doesn't protect against things like user bonehead f-ups, and virus/malware damage etc, but all hard drives will fail eventually (and that's guaranteed to happen at the worst possible moment), and when it does, it sure is nice to have a complete backup copy at hand.

I've been RAID1 mirroring my hard drives on a Promise card (or now on the mobo's integrated raid) since I got my first pair of 40GB drives when they first hit the market a few years ago, and have always just considered when it's time to buy a bigger drive to upgrade, then I must buy at least two of them as just part of the cost of owning and operating a computer. I use my machine for development and business, and all my documents and development projects are too valuable to risk to a drive failure. Just recently I've upgraded to 250GB SATA drives, and now I have three of them... I always keep a mirrored pair in the machine, and once a month, I pull one drive out, breaking the mirror, and put the extra one in and resync the mirror. I then take the drive I pulled, which has a month-end snapshot of my whole system on it, and put it in a safety deposit box at my bank. I'm going to do this every month from now on as part of my regular scheduled business process just in case my house gets hit by a tornado/burns/gets burglarized/or whatever-disaster.

diskseek

[Anonymous Coward]

still bad 11ms diskseek time?

Re:Big Big Drives are great...but backup is a prob

[tomstdenis]

RAID-5 [or 6]. If you're running something where you have 750GB of information chances are you can justify spending 2-3K on reliable storage.

3x750 in RAID-5 would net you about 1.3TiB of storage and would allow upto one drive to completely die without losing data. If you're more paranoid you could use 4x750 and have upto two drives die.

The RAID access will be automatic so effectively you're always backing data up.

Tom

Re:Big Big Drives are great...but backup is a prob

[nuggz]

Buy 2 drives, use the second one for backup.

Put it in a USB adapter and use rysync.
Quick, easy, cheap.

Re:Big Big Drives are great...but backup is a prob

[geekoid]

I amassuming you are talking for home personal use.

Must of the space used on disks is stuff thats been instaled, as opposed to created.

My wife is an avid digital photograher, but we not only keep it on HD, but also burn it to disk, at the time she downloads the pictures.
Everything else could fit on a couple of dvd's. I find that's true with most people.
Point in fact, most computer users I know who do not have a digital hobby wouldn't even fill a 20 gig drive.

AS for your problem you ahve two reasonable choices.
1, by a second drive and mirro.
2 Get a webstite host and just use it to archive your data, you can probably get one for a few bucks a month.

High-definition MythTV box is *wonderful*

[Yeechang Lee]

Some keep saying that there's no point to ever-increasing drive storage numbers. I disagree. Huge drives will always be appreciated in media PCs, where good-quality video (even if compressed) takes up a good chunk of storage space.

As the owner of a MythTV box equipped with dual HD cable boxes (*and* fortunate enough to have a cable provider that doesn't 5C encode its HD premium movie channels) and a HD over-the-air capture card, all of which I can use simultaneously, I can testify to that.

Here's my experience with bandwidth use:
* Digital non-HDTV channels generate the smallest files at about 900-1000MB/hour for a movie channel and up to 1200MB/hour for a cartoon (with probably a lower-quality feed).
* Analog channels such as TCM generate about 2900MB/hour due to the extra noise.
* HDTV premium movie channels generate about 4400MB-4700MB/hour.
* A high-bandwidth HDTV channel (defined as HDNet or Discovery HD Theater and most network affiliates over cable or over-the-air) generates 7400-7700MB/hour . . .
* Except for ABC and Fox, whose 720p programs record at about 5.8GB/hour.

On the MythTV box's dedicated NAS, I have (according to MythWeb) 176 programs, using 1.6 TB (324 hrs 32 mins) out of 1.8 TB (111 GB free). Almost all of the programs are high-definition movies. Examples:

* The Untouchables, 125 minutes, 16GB
* St. Elmo's Fire, 120 minutes, 15GB
* Shakespeare in Love, 125 minutes, 16GB
* Ben-Hur, 215 minutes, 15GB
* The Matrix Revolutions, 135 minutes, 11GB
* A Passage to India, 165 minutes, 21GB
* La Bamba, 110 minutes, 14GB
* Mona Lisa Smile, 120 minutes, 6.1GB (Commercials transencoded out)
* Spider-Man 2, 135 minutes, 12GB
* Batman Begins, 150 minutes, 11GB
* Seabiscuit, 180 minutes, 10GB (Commercials transencoded out)
* Witness, 115 minutes, 11GB
* The Passion of the Christ, 135 minutes, 9.8GB
* The Lord of the Rings: The Return of the King, 205 minutes, 19GB
* Doctor Zhivago, 215 minutes, 14GB
* Emma, 129 minutes, 12GB
* Bye Bye Birdie, 124 minutes, 16GB
* Giant, 204 minutes, 26GB
* GoodFellas, 154 minutes, 12GB
* Bullitt, 124 minutes, 16GB
* Real Genius, 119 minutes, 11GB
* Pulp Fiction, 164 minutes, 12GB

. . . etc., etc. Many of the larger-sized films were recorded off of HDnet Movies, which is an especial godsend for any movie lover. (I *can't wait* for the day TCM starts broadcasting in HD!) My all-time champion, now unfortunately lost in a box rebuild, was NBC's The Sound of Music annual broadcast. Four hours, including commercials, and 28GB!

Re:High-definition MythTV box is *wonderful*

[runlevel 5]

Do you ever use compression?

Re:High-definition MythTV box is *wonderful*

[Jester6641]

part of me wants to bow before you for your system. the other part of me wants to mock you for telling all of slashdot that you recorded "mona lisa smile."

Re:High-definition MythTV box is *wonderful*

[illumin8]

Do you ever use compression?

The HDTV broadcasts which he records are already compressed using MPEG2 compression. It's much better quality (and exactly the same as when you watch it at home) to just record the MPEG bitstream coming in off the firewire cable, rather than convert it to analog and then back to digital again.

Re:High-definition MythTV box is *wonderful*

[runlevel 5]

MPEG2 requires a quite a bit of space relative to newer formats. I believe upwards of 2.5mb/sec. I was thinking MPEG4 a la DivX or Xvid could give similar picture quality with a smaller footprint.

Re:High-definition MythTV box is *wonderful*

[evilviper]

Analog channels such as TCM generate about 2900MB/hour due to the extra noise.

GAH! Information... lacking... all... context...

A high-bandwidth HDTV channel (defined as HDNet or Discovery HD Theater and most network affiliates over cable or over-the-air) generates 7400-7700MB/hour . . .

HDTV streams have HORRIBLY poor compression. They encode with a constant bitrate, and use a very, very small GOP size (so you don't have to wait very long for the picture to appear when channel-surfing).

Using a better codec (eg. lavc, Xvid, x264) with a much larger keyint, varible bitrate (2-pass) encoding, etc., you can get that down to at least 1/4th the size, with really no quality loss at all. Throw some good denoising into the mix (lavc's "nr" denoiser is great, and takes almost 0 CPU time) and you'll get it significantly smaller, still, and it will look *better* than the original.

In addition, commercials are very fast, flashy, etc., and use-up much more than their fair-share of the bitrate. Editing them out will reduce the video length by 1/3rd, and reduce the overall bitrate even more (assuming VBR re-encoding).

If you don't have a very fast CPU (~3GHz/3000+) h.264/x264 is out-of-the-question. However, MPEG-4 decoding is actually FASTER than MPEG-2 decoding with a decent codec.

*And if your system is about 2GHz/2000+ or so, hardware decoding (XVMC) will use up as much or more CPU-time than decoding in software, unless you've got an AGP2x bus/card, or DMA doesn't work on your motherboard/videocard.

Re:High-definition MythTV box is *wonderful*

[Yeechang Lee]

I wrote:
Analog channels such as TCM generate about 2900MB/hour due to the extra noise.

evilviper wrote:
GAH! Information... lacking... all... context...

Nice imitation of Comic Book Guy's death-in-Lucite speech. Otherwise, not sure what you mean here. My understanding is that, over FireWire, *all* the channels on the cable box--whether HD digital, regular digital, or analog--are sent in MPEG2 format. The fact remains that in my experience analog channels take up much more space per hour than non-HD digital channels, whose signals are of course much cleaner.

Using a better codec (eg. lavc, Xvid, x264) with a much larger keyint, varible bitrate (2-pass) encoding, etc., you can get that down to at least 1/4th the size, with really no quality loss at all.

You've inspired me to look further into MythTV's transcoding features. So far all I've done with it is to remove certain HD network TV shows' commercials (thanks to MythTV's pretty darn good commercial-detection feature) with MythTV's lossless MPEG2 encoding. (x264 isn't yet in the cards, alas, for Linux. The other codecs should be, though.)

That said, one reason I haven't bothered is because I rarely want to keep movies and TV shows after I've watched them and those I do want to keep I generally want to keep as pristine as possible, with no resolution loss. Not sure how denoising would help on my clean, no-analog-involved-anywhere MPEG2 HD recordings, anyway.

If you don't have a very fast CPU (~3GHz/3000+) h.264/x264 is out-of-the-question. However, MPEG-4 decoding is actually FASTER than MPEG-2 decoding with a decent codec.

*And if your system is about 2GHz/2000+ or so, hardware decoding (XVMC) will use up as much or more CPU-time than decoding in software, unless you've got an AGP2x bus/card, or DMA doesn't work on your motherboard/videocard.

My MythTV frontend/primary backend is indeed a 3GHz Pentium 4 and it indeed displays 1080p happily with Xv (not XvMC) and a Nvidia 6200. Bob deinterlacing is truly a sight to behold.

Re:High-definition MythTV box is *wonderful*

[Yeechang Lee]

part of me wants to bow before you for your system.

And you should! (As well as bowing before my four-digit Slashdot ID, you 900K+ peon.)

Seriously, I really am curious to hear whether the commercial PVR software packages like Windows MCE, SageTV, or EyeTV can, like my MythTV box, simultaneously record two FireWire streams and one ATSC stream. (I could of course record more if desired simply by using additional cable boxes and capture cards.) Having never owned any of them, I don't have any firsthand experience. I *do* know that no TiVo can equal what my box can do.

the other part of me wants to mock you for telling all of slashdot that you recorded "mona lisa smile."

Not the greatest movie in the world but not that bad either. The wife of a longtime role model graduated from Wellesley in almost exactly the same year the film's students did and says that it got the details all wrong but the larger themes all right.

More to the point, in my experience too many Slashdot types suffer from an intellectual poverty that causes them to never try any films outside SF, fantasy, and superheroes or any books outside . . . well, SF, fantasy, and superheroes. Don't get me wrong. I've read pretty much every Heinlein novel and short story and every Asimov robot story and Foundation series. I can recite the real names of every pre-Five Year Gap Legionnaire and have a healthy pull list at the local comics shop. I've watched the 1978 Superman a dozen times or so and can't wait for Superman Returns.

But I've also read most of Austen (I'm one of those who prefer Emma to Pride and Prejudice), am slowly working through Trollope, and have read through the entire Sherlock Holmes canon several times. I've been watching the new Battlestar Galactica since the miniseries, but I also enjoyed watching the first four seasons of Gilmore Girls (for some reason get out of the habit during the past two), and have watched The OC from the first episode, before it became huge. My MythTV collection has Mona Lisa Smile, yes. It also has three Star Trek films *and* Ben-Hur *and* LA Confidential *and* Ghostbusters *and* Master and Commander *and* The Music Man. I see nothing contradictory in juxtapositioning these works and, again, I wish more Slashdot types felt this way.

Re:High-definition MythTV box is *wonderful*

[evilviper]

The fact remains that in my experience analog channels take up much more space per hour than non-HD digital channels, whose signals are of course much cleaner.

Bitrate and filesize of an analog capture are traits that YOU SET, not some inherent property the video has. If you would have said "my cable box encodes them at..." or "they don't look very good below this size..." etc., I likely would have understood.

With (apparently) better hardware MPEG-2 capture cards, only perhaps a bitrate of half that would be necessary for practically perfect picture. With software encoding to MPEG-2, it's much, much lower still.

(x264 isn't yet in the cards, alas, for Linux. The other codecs should be, though.)

One of us is confused. x264 is the name of a GPLd h.264 implimentation, specifically for Linux, written by the VLC guys. If you can't select it with whatever tools Myth provides these days, you can certainly use it via programs like MPlayer/mencoder, which handle MythTV streams just fine, in fact.

Not sure how denoising would help on my clean, no-analog-involved-anywhere MPEG2 HD recordings, anyway.

I can assure you, it does. I use it encoding from any source, including fully digital sources like DVDs.

The CCDs, digital cameras or film sources produce noise... Hardware MPEG-2 encoders add noise... too-subtle-to-notice camera movements produce noise. etc. Even if you film a perfectly solid white or black wall, upon playback you'll see different glitches, color artifacts, aliasing... basically: noise.

My MythTV frontend/primary backend is indeed a 3GHz Pentium 4 and it indeed displays 1080p happily with Xv (not XvMC) and a Nvidia 6200.

The key here is that the H.264 video codec, which is very CPU-intensive, and used by the 1080p (24fps) trailers on apple.com. Even if your computer is fast enough to handle them with a few cycles to spare, it still might not be able to for 30fps content.

Incidentally, HD playback may be much faster with OpenGL, rather than Xv. Try "mplayer -nortc -dr -vo gl", and perhaps try adding "-vf format=bgr24" to that command-line.

Re:High-definition MythTV box is *wonderful*

[Yeechang Lee]

Bitrate and filesize of an analog capture are traits that YOU SET, not some inherent property the video has.

As I've made clear from the beginning, I don't have any analog capture cards. Neither FireWire cable boxes nor ATSC capture cards emit any analog signals to me from my perspective.

One of us is confused. x264 is the name of a GPLd h.264 implimentation, specifically for Linux, written by the VLC guys.

Sorry for the confusion; the last time I saw x264 discussed on mythtv-users it was said that the Linux GPL reimplementation wasn't up to snuff yet. I have recently used VLC to play back high-definition trailers from the Apple trailers site but haven't needed to pursue it any further.

Incidentally, HD playback may be much faster with OpenGL, rather than Xv. Try "mplayer -nortc -dr -vo gl", and perhaps try adding "-vf format=bgr24" to that command-line.

Will try this; in my experience -vo gl has been slower (or, at least, no faster) than -vo xv, but perhaps the additional switches will make a difference.

Re:High-definition MythTV box is *wonderful*

[Anonymous Coward]

But I've also read most of Austen (I'm one of those who prefer Emma to Pride and Prejudice), am slowly working through Trollope, and have read through the entire Sherlock Holmes canon several times. I've been watching the new Battlestar Galactica since the miniseries, but I also enjoyed watching the first four seasons of Gilmore Girls (for some reason get out of the habit during the past two), and have watched The OC from the first episode, before it became huge. My MythTV collection has Mona Lisa Smile, yes. It also has three Star Trek films *and* Ben-Hur *and* LA Confidential *and* Ghostbusters *and* Master and Commander *and* The Music Man. I see nothing contradictory in juxtapositioning these works and, again, I wish more Slashdot types felt this way.


*swoons*

You're so dreamy and better than everyone else! May I have you fantastically boring children?

Re:High-definition MythTV box is *wonderful*

[evilviper]

Will try this; in my experience -vo gl has been slower (or, at least, no faster) than -vo xv, but perhaps the additional switches will make a difference.

I should mention that -vo gl underwent some major changes after 1.0pre7, so you need a CVS snapshot or wait for 1.0pre8 (probably a couple weeks away).

Fabulous for scientific use...

[Dr. Zowie]

Solar scientific data is growing too large to handle. The SOHO data are almost small enough to ship around by internet (the whole dataset is something like 20-30 TB for 10 years of operation), though data mining and such are starting to fall back on SneakerNet as the SDAC is shipping around terabyte lunchbox drives as their preferred method of bulk data export.

But Solar Dynamics Observatory, which is currently being built, will generate about 3 TB of data per day. We're all a little worried about how to distribute, store, and use such vast quantities of data. Perpendicular-storage drives like these just might save the day...

Re:Fabulous for scientific use...

[WhiteWolf666]

How far are you distributing this data? Is it going places Internet2 doesn't go? Is it prohibitively expensive to hop on to Internet2, given the budgets of these sorts of projects?

Seems to me that needing to distribute this kind of data is _exactly_ the sort of impetus needed to kickstart next generation internet infrastructure. Of course, this does nothing for storage problems.....

One should be able to get ~ 1Gb/sec over fiber. Conservatively, assuming 500Mb/sec real throughput, that means 12 hours in transmission time, per day. That's faster than most sorts of not-too-expensive shipping techniques.

Heck, 10 of Verizon's FiOS connections would be able to handle the bandwidth, assuming you didn't have to deal with Verizon's bottlenecks, or could somehow get the data on to their network.

Keep in mind I'm not suggesting that the infrastructure exists right now to handle this sort of thing, but it seems that the technological barriers are long in the past, and the remaining barriers are fairly simple economic ones.

Re:Fabulous for scientific use...

[Peter Mork]

The interview is a few years old now, so I don't know how well the numbers hold up, but Jim Gray suggests that the postal service was a far cheaper mechanism for transferring terabytes of scientific data than any Internet: ACM interview.

Re:Fabulous for scientific use...

[Dr. Zowie]

Hmmm. If I understand correctly, the problem is that bandwidth doesn't scale well -- if you need to distribute data to 100 different sites, you need 100x the bandwidth. The MBONE, if it had ever been adopted, wouldn't solve that problem because not every institution looking at the data will be interested in the same data at the same time -- the distribution problem is more like video-on-demand than like broadcast television.

Re:Fabulous for scientific use...

[mako1138]

Over in the particle physics world, they're putting together some sort of grid for LHC. Seems to be a rather large undertaking.

Re:Fabulous for scientific use...

[hr raattgift]

The MBONE, if it had ever been adopted

Multicast technology has advanced considerably since the days of MBONE.

David Meyer, University of Oregon, and at various times with Sprint and Cisco, and on the Internet Architecture Board, spent a considerable amount of time herding cats in the development and deployment of scalable native multicast forwarding across large-scale ISP infrastructure.

Here is one result: no-fee Sprintlink native multicast (see also the FAQ).

Several other ISPs and university networks also exchange multicast network layer reachability information (mBGP), implying support for native multicast.

Currently most multicast deployment favours large dynamic groups of listeners subscribing and unsubscribing to traffic originated from a relatively small set of sources, rather than generalized many-to-many conversations. With current technology many-to-many "oligocasting" is most efficiently emulated by a single sender receiving (via unicast) messages which are then multicast to a tree of subscribers. This is generally referred to as SSM, Single-Sender Multicasting, which is suitable for the modern Internet in which switching equipment performance (or cost, if you will) is bounded by dynamic adaptiveness (control messaging, routing) rather than the amount of traffic being switched. SSM is also a good fit for applications similar to cable-TV-like broadcasts or distribution of large sets of data from a primary centre to subsidiary or replica centres.

MBONE's performance was bounded by the tunnelling and especially replication performance of the UNIX workstations that were the majority of MBONE's routers, as well as by the obsolete DVMRP and MOSPF multicast WAN routing protocols which were used even across ISP and other network boundaries.

not every institution looking at the data will be interested in the same data at the same time

You could combine SSM with a "TIVO-like" function, saving the multicasted data locally and extracting what you want, when you want it, from your local copy. If your shared multicast distribution tree significantly reduces the sender's bandwidth by moving your 100x duplication into "the network" such that you approach only one copy of a given datum sent once from the sender to the sender's ISP, it's a win. If you also approach one copy of a given datum sent once down any given branch of the tree rooted at the sender and branching at Internet routers between the sender and every receiver, you make a good trade-off, as the cost of small numbers of copies performed in routers (possibly one per interface) distributes well and for heavily-subscribed Single Sources the cost is likely to be smaller than the cost of carrying multiple copies of the same traffic.

With your example the question is whether the cost of locally recording likely-to-be-interesting data received via multicast is greater than the cost of smearing out any duplicate requests (from you or any of the other 100 sites) over time. Most of the cost in the latter "video-on-demand" style approach is borne by the primary sender.

3000 not 30K

[Pranjal]

Actually the ipod says 1 of 3,000 not 30,000 as the parent says.

Re:3000 not 30K

[Anonymous Coward]

In the intro to the clip it says 1 of 3,000, but after the animation has concluded that number jumps up to 30,000; a reference to the proposed 10x data storage density.

Re:3000 not 30K

[Tanktalus]

You need to watch the whole thing. At the beginning, it shows 3,000. At the end, due to the new "perpendicular technology", it increased to 30,000. I'm actually pretty impressed, however, that they managed to download 27,000 files in that short span of time ;-)

Practicality

[geekoid]

How much data n a person hard drive really needs to be backed up?
About a CD's worth in most cases. I am talking about critical files.

WHen you way that against the liklyhood of damage to the physical premesis(extremly low liklyhood) the level of backup you suggest is overkill.

Re:Practicality

[daft_one]

When you *weigh* that... not *way*. :-)

Re:Practicality

[Anonymous Coward]

I've got at the very, very least 15 dvd's *worth* of .raw and manipulated photography.

Re:Now all I need...is a backup perhaps?

[Anonymous Coward]

Even if it means forking over the money for a tape backup and tapes, if you lose any data due to a drive failure you have no one to blame but yourself. If it's important, build a RAID.

Have fun when you lose your important data due to a limited disaster of some sort. RAID is not a backup technology.

If its critical, build a RAID with some kind of tape or other backup.

Which the vast majority of household users will not do. And I'm betting you fall into that category, too.

But, please do keep feeling smug about having RAID for your important data. When that fails you as a backup approach, laugh at yourself for me.

Re:Big Big Drives are great...but backup is a prob

[markjo]

"The RAID access will be automatic so effectively you're always backing data up." RAID != Backup RAID will only protect the integrity of your data. It will not protect your data from a trojan/worm/whatever nastiness that can blow your data away. It will not allow your to recover that important document that you accidently deleted two weeks ago.

Don't spend it if you can't afford to back it up.

[Were-Rabbit]

You're buying a 750GB hard drive but you can't at least afford two for a mirror? If you can't afford to lose the data on that 750GB drive, what the hell are you doing by buying such a large drive in the first place?

You're obviously a religious man because you're putting a lot of faith in the reliability of that drive.

Fatter than Faster

[Doc Ruby]

The ratio of storage transfer-speed:capacity is favoring capacity much more as time goes on, especially among most media consumers. I'm not nearly as interested in my drive delivering multiple video streams as I am in keeping all my media, including movies, TV shows, and even security cameras available immediately across the Net. I'd prefer Seagate and other storage companies to funnel more money into storage capacity and simple, cheap RAID tech (for reliability of so much storage) than into faster seek and transfer time tech. Also higher priority than transfer speed is duping directly from old storage to new storage, even across interconnect technologies and data formats. Storage companies could make a lot more money from the consumer market selling us more drives to rotate into our RAIDs before MTBF.

The server market could still benefit from extra transfer speed R&D, for server drives at higher cost. This is true now, but the highest capacity "server" drives are well within the capacity needs of many consumers.

Re:Big HUGE warnings - Not quite true

[OrangeTide]

RAID 1 and 0+1 is mainly good if you expect total drive failure. (a detectable condition). but undetectable failures like data corruption don't show up as to which drive it happened to. It makes recovery of those situations a real pain. At least with raid 5 you can reconstruct the original data even if one is corrupted. you can even reconstruct with multiple corrupt drives if the corruptions don't overlap (nice!). If you're just getting corruption on one drive with raid 1 you can still manage to figure it out if you peak at the SMART data and see which drive has been complaining the most about failures, it's almost always that drive's fault. still not automatic though.

RAID 5's flaw is that you can choose to limp along with failing drives. Which is incredibly dangerous to your data. This is probably why a lot of people claim RAID 5 destroyed their data. 1. it doesn't replace backups 2. you need to order new drives and ship them overnight as soon as you have problems.

On-line storage will become popular soon

[BrentRJones]

Seems to me that eventually I will trust my backup to a transparent, automatic, secure system that puts it on the Net somewhere. I currently use FTP to back up and transport from work to home and back. The hard disk is the sameone my website is on; and it is co-located. But someone will make better software to automate everything. I hope it is open source freeware.

The most important factor of hdds...

[abigsmurf]

is not the noise power etc. but if it's going to die after 12 months and take your precious por...data with it.

Re:The most important factor of hdds...

[SysGoddess]

...is reliability. We're currently on our 3rd warranty replacement Seagate drive in my personal machine while the cheap lower capacity Samsung secondary drive has ticked on without a hiccup.

This last time, they replaced the previous 200GB hdd with a 250GB and the fscking bearings went out in the thing barely 8 months after I spent $25 for them to overnight the replacement so that I could get back up and running ASAP.

So all my data is gone in a blink. But wait you say, surely you have backups. I have an older partial backup and the notebook was somewhat recently shadowed thanks to my nimrod spouse who managed not only to toast my image on the server but then played tetris on my PDA until the batteries were depleted and then didn't charge it so I lost all my contacts, databases & spreadsheets that were there.
{insert very long list of personally favourite expletives here}

Thankfully we didn't put Seagates in our server RAID array.

Re:Now all I need...is a backup perhaps?

[TheGavster]

If you can afford one of these 750GB drives, you can probably afford two 250GB drives with dosh left for RAID hardware.

Nah unRAID!

[BLKMGK]

Lime Technologies unRAID system instead. One drive is Parity the rest are data drives, no striping. If a disk fails you get your data back, if two disks fail you lose those disks worth of data but nothing else. I have 3TB in one of these and while it has it's quirks I'm *not* forced to have a bunch of drives all the same size....

Re:Big HUGE warnings - Not quite true

[drinkypoo]

you need to order new drives and ship them overnight as soon as you have problems.

Not only that, but you NEED a hot spare. If you have a RAID made up of a bunch of the same drive, and one fails, it should be a warning sign to you. The other drives are substantially the same as that one, and they should now be considered likely to fail, unless it failed WELL within the MTBF window.

Urgggg. My brain is melting from the BS

[amcdiarmid]

Raid 5, in most applications means that if one drive fails you have no problems. If a second drive fails before the dead drive is replaced & rebuilt, you are screwed.

Raid 5 with spare(might be called 6 in some vendors terminology), is almost unseen outside the enterprise (read real raid controller, not home nas-box, or home-pc) means if one drive fails that drive will be rebuilt on the spare drive. If a second drive fails before that happens: HEHEHEHEHE (can you say $$$ to ontrack?)

In either case, a power-surge eating your controller will still shaft you. (You did purchase a second controller to sit on the shelf didn't you? Oh, you were using the controller on your motherboard. pffft. and no, it's not raid 6 unless you are a geek running Linux to handle the raid bits.)

There is a reason specialized backup devices exist. Recommended is still off site storage (tape recommended, HD sometimes used.) so if a fire eats your server (farm if you have one) - you still have your data. If you fall down, and want to get back up - you need backup.

Note: (I'm poor and only have a single drive NAS-box with duplicate data. I'm hoping that if it goes down, I can replace it before the machines with the data die. Or vice-versa)

Note2: It's 90 outside & my DSL bridge just melted. I'm cranky and need to get it out of my system before I go to client meeting this afternoon.)

Re:Urgggg. My brain is melting from the BS

[eh2o]

IANASE (I am not a storage expert) but according to "a comparison of disk drives for enterprise computing" (see latest issue of ;login vol 31 no 3 by kurt chan), the common idea that "its okay until two out of three drives fail" is not quite accurate.

Here is the dirt (as I understand it): even if only one drive fails, you still have a non-zero probability of a non-recoverable read error on one of the remaining "good" drives during the rebuild. With a UER of 10^-14 bits (typical for SATA drives), rebuilding a drive in a 5x500GB RAID-5 array has a 20% chance of an unrecoverable error (or ~ 1.5% probability for a 10^-15 UER eg FC or SAS).

Re:Big HUGE warnings - Not quite true

[Anonymous Coward]

For those who don't see the difference: Most boxes don't have controller capacity for more than four drives (two PATA channels and two SATA channels) and seven or eight drives will also strain your PSU and your cooling capacity. Might be hard to fit in your case, too.

• Any modern nForce4-based motherboard handles raid arrays on 4 SATA drives, any additional chip is gravy (my 1 year old DFI handles 4 form the nForce plus 4 from a Promise chip)
• Seven or eight drives won't strain a PSU (unless it's a cheap no-brand piece of crap), even a Raptor doesn't burn more than 15W when searchin, most 3.5" are at 8-12W tops. If we take 8 drives and 12W we're talking about top consumption of 96W. That's less than a modern CPU or GC alone.
• Cooling is likewise, while hard drives don't cool very well (they don't have heatsinks or anythink) they produce very few heat, just put a low-speed 120mm fan (low speed as in under 1kRPM, I'm talking Papst or Nexus here) in front of your drives (a fan for each 3 or 4 drives) and they'll keep well under 40C in a room at 25C.

Case room is the only real fact in here (most quality cases can only fit 4-5 drives, some go up to 6, some get as low as 2), and even then there are now several cases built specially for that kind of uses, such as Coolermaster's Stacker

My cooling solution is low-tech, loud and very effective: The side of the case is off and I have a 30-inch box fan (the kind you mount in a window to cool your house) blowing into it.

The only effective thing here is that your're slaughtering your case' airflow, while this is often the only way to cool crappy cases it doesn't work well in quality cases, unless you put so much brute strength in the cooling that the airflow doesn't actually matter anymore (which is what you're doing).

One nifty trick I discovered is that if you slice all of your disks up into many small partitions, then create many RAID-5 arrays (using partition 1 on each disk to create the first array, etc.), then use LVM to bind all the arrays together you can add additional disks and rebuild the arrays without having to find some way to back up all of the data first.

Some controllers are also able to extend (or even fully replace) arrays out of the box. You usually don't find them on consumer-grade motherboards though.

Re:Big Big Drives are great...but backup is a prob

[tomstdenis]

Yes, I know that. I'm saying though that you get immediate reliability.

Also, anyone who puts a 3000 dollar raid setup in a Windows box deserves what they get. :-)

Yes, you still need daily backups, off-site, etc. But you still need that first line of backup which is what RAID gives you.

Tom

Re:Big Big Drives are great...but backup is a prob

[Anonymous Coward]

Christ!

If RAID is your idea of backup I sure hope you aren't a sysadmin.

I'd love to watch you trying to restore from your "backups" after fire/theft/flood/etc though :)

Lol

[JMZero]

and is NOT what the pros do

I was using random numbers to illustrate the math in question. As I said in that first post, the original point still stands (ie. more drives = greater chance of one or more failing).

And the top post that prompted the reply I replied to was clearly a joke.

But I guess the pros do NOT joke around, or use easy to understand numbers in examples to explain a point. I'll try to be more PRO in the future.

Re:Lol

[Monster_Juice]

But I guess the pros do NOT joke around, or use easy to understand numbers in examples to explain a point. I'll try to be more PRO in the future.

Yes please do because this is slashdot where things like this matter. 7*100=750 We live and die by good math.

That and I am glad someone pointed out that this was a joke.
Personally I sit in the same room as my computer when I get home and I can tell you that if I have 7 drives sitting next to me it does not matter what the MTBF is they are going to fail the second I throw them out the window.

Re:Lol

[SmittyTheBold]

Could you tell me which post is "obviously a joke?" I'm still not spotting it, and I went up about five levels in the thread.

Re:Big Big Drives are great...but backup is a prob

[Surt]

Buy 2+ and raid-1 them? You'll be basically immune to data loss barring destruction of your computer. If that's a serious enough concern, buy a dual layer dvd writer and a bunch of disks (at $1 per 8gig disk, that runs you $100 to fully backup a full 750gb disk ... which is not outrageous considering you paid $400 for the disk in the first place). Using incremental backup software you'll not need many additional disks to maintain your backup.

Re:Big HUGE warnings - Not quite true

[operagost]

Replace "task" with "expense" and I would agree with you. RAID with seven drives is the same as RAID with three drives, assuming the same RAID level. You add the drives to the array using the supplied utility, optionally reserving spares, and you're done. Maybe it will take you an extra few minutes per drive to mount each in the enclosure.

Re:Big Big Drives are great...but backup is a prob

[tuffy]

Buy 2+ and raid-1 them? You'll be basically immune to data loss barring destruction of your computer.

...or accidentally deleting/overwriting something - which the RAID will happily propagate to all the drives instantly.

Get 2 drives, put one in an external enclosure and leave it off when not backing up/restoring. It's not a perfect substitute for tapes, but should be good enough for home use.

Re:Now all I need...is a backup perhaps?

[LurkerXXX]

RAID is not a backup. Say it with me. RAID is not a backup.

Not only won't RAID save you from a "rm -rf /", it won't save you if your power supply dies in an ugly way and takes all your attached components with it, your power line gets hit by lightening, and fries everything attached to that outlet, your house/biz gets burgled and they take the computer, you accidently delete a critical file you really need or realize you need it later, etc, etc, etc.

RAID is not a backup.

A backup is an offline copy that you can store at an off-site location just in case one of many many 'bad things' happen.

RAID is simply a way of increasing your uptime in case a single component fails. It's not a backup.

newer Seagate 7200.9's

[jagilbertvt]

Some of the newer 7200.9 models (80gb, 120gb, and 160gb) also feature the perpendicular technology. I'd like to see a comparison between these and the older 7200.9 models that don't feature it.

Nice try.

[WidescreenFreak]

I welcome you to come over here and see my Sun Blade 100 with both a DDS-4 and newly-added, DLT7000 drive attached to it that I use to back up my documents, photos, and other important files on a monthly basis with a three month retention. I guess these three boxes of DDS tapes and dozen DLT IV tapes are just an illusion.

Thou mayest go back to fornicating with thyself now.

Re:Nice try.

[Lawrence_Bird]

right.. so by the time you finish backing up your 750gb drive with those its past your retention time.  Sounds practical.  The bottom line is there are really no practical ways to backup at reasonable cost and speed.  RAID is great for a hardware failure, not so great for accidental deletions or files which are corrupted by software. 

Re:Big HUGE warnings - Not quite true

[ckaminski]

Which is why the market invented RAID6. Instead of having to rebuild your dead "parity drive" as in RAID5, you can use that hot-spare to hold another live copy of your "parity data", so that your RAID volume can suffer 2 disk failures and still be safe.

I've lost two volumes (600GB+) lately because a second disk in a RAID5 set failed before the replacement or the hotspare could regenerate the lost information.

Having disk arrays of more than 4 disks without RAID6 is needlessly risking your data today, IMHO, assuming your RAID controllers support it (which is rare).

Re:Big Big Drives are great...but backup is a prob

[hackstraw]

Great I can buy a 750GB drive...however how the hell am I gonna back this thing up...actually even with many many dics how am I gonna backup 750GB.

Real men buy two :)

Actually, I currently have 2 400 gig hds that are pretty full with no backup at all right now except for "important" data that I have replicated on different machines. These drives are about 1 year old, and I realize that I will need to get a backup for them sometime, or feel the pain.

So, now for about $400 I can backup both of my drives onto one disk. Not bad. Last year or maybe 18 months ago, I paid about $350 for one of those drives. Now, I can pay about the same money and have all my data backed up. Not bad.

Re:Big Big Drives are great...but backup is a prob

[termigan]

With the difficulty of backing up drives, it may be time for the return of the versioning file system. With the proper utilities and aging mechanisms, a versionning file system would be able to insure data content as well.

Useful support facilities beyond recovering old versions would include:
- Add the concept of 'vital' files to signal the filesytem to maximize the recoverability of a vital file.
- Perhaps pay attention to the sticky bit for 'vital' files so you could have a directory that contained vital files by default. Could be mis-used, but features are what you make of them.
- Allow configurable notifications when for when 'large' amounts of data have been changed recently, 'large' numbers of files, or when vital files have been changed.
- Enable the user to set up notifications of what history is in danger when free space gets low
- Allow the user to make some sort of off-line backups of only the history, only the active files, or vital files only.

This is the sort of innovation that Open Source should be doing. New concepts that solve problems that come up because the innovation of major OS vendors haven't kept up with the changing landscape.

You forgot....

[DrYak]

The impressivley huge repository of .ISOs of install DVDs for Microsoft Office Vista edition, Microsoft Visual Studio Vista edition, and the 12 subtely different flavors of Microsoft Windows Vista.

Re:You forgot....

[unity100]

But they would just burn the hd platter with pain, just sitting there. Why ? Why not more space for more porn, anime, games and movies instead ?

Re:Big HUGE warnings - Not quite true

[swillden]

Any modern nForce4-based motherboard handles raid arrays on 4 SATA drives, any additional chip is gravy (my 1 year old DFI handles 4 form the nForce plus 4 from a Promise chip)

This is perhaps the case. I don't use modern motherboards and processors for home file servers. Old stuff works fine -- but suffers from the drive controller capacity limitations.

Seven or eight drives won't strain a PSU (unless it's a cheap no-brand piece of crap)

My experience is different. I had a decent (but non-server) 350W PSU and when I added the fifth drive to my box I began getting intermittent failures reported by the drives. When pushing all of the drives hard, the failures were frequent and sever. I put a good 550W PSU in the box and the problems went away.

Cooling is likewise, while hard drives don't cool very well (they don't have heatsinks or anythink) they produce very few heat, just put a low-speed 120mm fan (low speed as in under 1kRPM, I'm talking Papst or Nexus here) in front of your drives (a fan for each 3 or 4 drives) and they'll keep well under 40C in a room at 25C.

But only if your case has airflow that is designed for that many drives. My case doesn't, and most cases not specifically designed for lots of drives don't, so they end up being placed in all sorts of odd locations, which means that cooling becomes problematic.

So, to sum up, if you're (a) using a newish motherboard, (b) a PSU designed for server, or at least high-end workstation use and (c) a case designed to hold and cool lots of drives, then you don't have any of the problems I mentioned, because you've already bought the components to solve them. If you're building a home file server the way most people do, though, out of the machine that was your desktop box several years ago, they're real problems.

Some controllers are also able to extend (or even fully replace) arrays out of the box. You usually don't find them on consumer-grade motherboards though.

Not to mention the fact that if you use hardware RAID you're tied to that particular hardware's implementation of RAID. If that controller dies (unlikely, but possible), you need to replace it with the same sort of controller. For a data center that's not an issue, since you just purchase a service contract that ensures the components will be replaced as needed. For my home use situation, it make me nervous to tie my data to any particular controller. Software RAID is the safer bet.

Re:Big Big Drives are great...but backup is a prob

[Anonymous Coward]

You can also leave it in the case, and then periodically spin up-rsync-spin down. Here's my crontab:

0 4 * * 0 sdparm --command=start /dev/sdc ; mount /dev/sdc1 ; rsync -av --delete /sdb/* /sdc ; umount /dev/sdc1 ; sdparm --command=stop /dev/sdc

Re:Big HUGE warnings - Not quite true

[blankoboy]

Heaven help the l337 fool that would RAID0 7 x 100GB drives for a 0.0005 fps improvement in BF2. He will have 7x more likelihood of losing all his data. =)

Which HDTV card?

[antdude]

Which HDTV card do you use? I use the Technisat Air2PC-ATSC-PCI in Windows XP Pro. SP2. Nice cheap card for ATSC. However, its software (DVB Viewer) is buggy and feature limited.

Re:Which HDTV card?

[ch-chuck]

Those are supported in the latest (2.6.16) linux kernels without having to add anything (I just upgraded Fedora core 4 to a 2.6.16 kernel) - then you can use dvb-apps to scan for stations and zap it to channels, plays with xine and mplayer, save the transport stream to disk, re-encode with mencoder, etc. I also have a Technisat Skystar2 card that works about the same for the free satellite stations.

Re:Which HDTV card?

[Yeechang Lee]

Which HDTV card do you use? I use the Technisat Air2PC-ATSC-PCI in Windows XP Pro. SP2. Nice cheap card for ATSC.

I have the Airstar HD5000, the Air2PC's successor. Not that having one versus the other matters in what we both use it for, ATSC. Linux supports both cards very well.

However, its software (DVB Viewer) is buggy and feature limited.

Get thee to MythTV immediately! Or, if you don't want to go to the trouble of installing a Linux box just for it, get yourself Sage TV or some commercial package. I can't say how well the Windows-based commercial packages work for FireWire, but it doesn't sound like that doesn't concern you, anyway.

Re:Which HDTV card?

[antdude]

The last time I checked, only MythTV (Linux) and DVB Viewer (Windows) only work with this card. :(

750G Disks are BAHD for Databases!!!

[vallee]

I can tell from the tone of this review that a lot of pointy-haired purchasing managers are going to be dying to use these for enterprise database applications. I can feel the tense discussions coming on strong now.

That's why I posted the following manifesto: 750G Disks are BAHD for DBS a few weeks ago when these disks were released. Find out why huge disks are the bane of DBAs everywhere. My manifesto has been signed by the Oracle DBA industry's leading lights, please, use these disks for the purpose they were designed for, whatever that may be (home movies from your Canon S2 IS? I've got one of those and the on-board video compression is TERRIBLE!), and not for databases.

This public service announcement has been brought to you by Pythian Remote DBA.

--
Paul Vallee
President, The Pythian Group, Inc.

Re:750G Disks are BAHD for Databases!!!

[Just Some Guy]

Find out why huge disks are the bane of DBAs everywhere.

I read your manifesto, but still don't understand your premise. You don't adequately explain why larger sizes are inherently bad, save for the seek time issue. Given two drives with identical performance but a 2x difference in size, why is the larger worse if it's holding the exact same data?

Re:750G Disks are BAHD for Databases!!!

[evilviper]

You don't adequately explain why larger sizes are inherently bad,

He's operating on the mistaken impression that you have to FILL every hard drive in your array.

Obviously, if you have twice as many disks in your array (each half as large) it will be much faster.

Re:750G Disks are BAHD for Databases!!!

[Anonymous Coward]

He's operating on the mistaken impression that you have to FILL every hard drive in your array.

This is were practice deviates from theory. An experienced DBA looks at it the other way around. Try it: In theory you don't FILL every disk. In practice, every storage device in the vacinity is full.

In practice, the moment you turn your back some yahoo fills the disks with more OLAP data than anyone ever hinted to you would be necessary. The new queries saturate the controllers and you get called because it's too slow.

Re:750G Disks are BAHD for Databases!!!

[loraksus]

Did the fact that the drive is about 1/2 the cost of a 146 gig 15k cheetah give it away?

Justification ?

[DrSkwid]

With non-HD DV video at 12Gb per hour I can fill these disks a couple of times over with my DV tape store.

The summary sounds like a press release, another slashvert ?

Re:Justification ?

[Bruce Allen]

I agree, it's not real. Slashvert!

I do still kinda want one though.
Bruce Allen

Re:People rival the radio stations already

[bussdriver]

the typical station has 30-40 songs in their playlist and they alter only a few songs at a time. Some seem to have a few playlists that rotate daily and they don't change much either:
Look at iTunes -> Browse -> Radio Charts

Re:Big Big Drives are great...but backup is a prob

[debest]

Must of the space used on disks is stuff thats been instaled, as opposed to created.

Digital Video (like a Myth box)

Enough said.

(If it's not, look into how much storage an hour of high-quality video, even with a good codec, can require. Then imagine your video collection ripped to the media centre.)

Re:Big HUGE warnings - Not quite true

[drinkypoo]

Is there linux support for software RAID 6? I'm giving up on hardware RAID and going multi-core in the future. Specialized computing hardware rarely makes sense... Of course I'll have to boot off of a different disk/volume but I'm going laptop anyway, so all my external storage will be IEEE1394.

NewEgg

[blogeasy]

Only $440 from NewEgg.Com with 3-day shipping.

Re:NewEgg

[WeArab]

Nice stuff!

When will we see 3 TB then?:P

These drives are needed for a PetaByte array

[Terje Mathisen]

Oil companies which do a lot of Pre-Stack processing, i.e. raw seismic data, need an awful lot of disk storage: We're currently in the ~50 TB (geographically mirrored RAID5 servers) range, and this is with Post-Stack only.

Going to Pre-Stack will generate 10 to 100 times as much data, which means that 500 TB to multi-PB is where we'll be in a couple of years. Having 750 GB SATA drives in a Nexsan SATABeast http://nexsan.com/products/products/satabeast/sata beast.html enclosure results in about 27-28 TB of usable disk space in a single 4U rack unit.

Very nice!

Terje

Re:Now all I need...is a backup perhaps?

[accessdeniednsp]

I didn't catch that. Did you say "there's no reason to lose data?" This Slashdot, you have repeat yourself at least three times. Two won't do. Remember how it worked out for Dorothy?

Translate please

[accessdeniednsp]

Can someone put this into more familiar terms such as Libraries of Congress, hogsheads, volkswagens, etc.? I am having a hard time identifying with these contemporary measurements.

Re:Translate please

[Anonymous Coward]

The number of digitized books it can hold is the equivalent volume to olympic sized swimming pools stacked to the moon.

Re:750G Disks are BAD for Databases!!!

[poopie]

I remember when DBAs were screaming that they only wanted 1 and 2gb disks and as many spindles as possible, and at that time, it was the 9gb SCSI drives that were BAD because they we too big and people wanted more spindles

Then the DBAs wanted to horde 9gb drives because 36gb drives were too large and they wanted as many spindles as possible.

Now DBAs only want the 72gb drives because the 144s and 250s are too large and they want as many spindles as possible

I guarantee that a few years from now, we'll read about the DBAs wanting only 750gb drives because the 3tb drives are too large and they want as many spindles as possible

Surviving 350Gs ? (request for real world units)

[Zoxed]

One thing I noticed is that one of the photos shows the sticker on the drive and it includes a warning that the warranty is voided if the drive experiences greater than 350 Gs !! Can this drive really survice a 340 Gs impact ? I am not a scientist nor a mathematician but that sounds like a hell of a shock.

Can any Slashdotter convert 350 Gs to real world units (eg dropped 5m onto concrete) ?

Re:The perfect excuse...

[9Nails]

I cried. I cheared. I taped my toe to the beat and even clapped at the end. Bravo!

re: no point to large drives?

[King_TJ]

Anyone saying the ever-increasing storage capacity is "pointless" is simply being short-sighted. We're just starting the era of high-definition TV, for example. Editing HD content from an appropriate camcorder and storing footage for possible future use is going to burn up a *lot* of drive space. When prices of HD compatible camcorders fall and when most people have upgraded to HD compatible TV sets, you can bet people will be complaining that their 400GB drives just aren't big enough for everything.

Re:Big Big Drives are great...but backup is a prob

[jabuzz]

How about a DLT-S4 tape then. They hold 800GB of data native. At no time to the best of my knowledge in the last 10 years has the largest hard drive ever held more than the largest tape. Before that I don't know.

Yes they are expensive, but that is because people don't backup so the volumes are to small. Chicken and egg situtuation really. Only something like two million DLT drives of all types have ever shipped. In the same period it is probably more like two billion hard drives that have shipped. Hard drives are so cheap because the volumes are there to make it so.

Re:Big HUGE warnings - Not quite true

[Anonymous Coward]

And don't forget that 7 drives have 7 times the moving parts.

Re:People rival the radio stations already

[soupdevil]

Radio stations, at least the hit stations, have 20-30 songs in heavy rotation (at least once a day) and hundreds of others on "gold" or other rotation, that may play once a week, and hundreds or thousands that may make an appearance once a month or once a year, or when occasion calls: Artist dies, event mirrors lyrics, etc.

Re:Big HUGE warnings - Not quite true

[ckaminski]

There is, but last I heard it's spotty and not ready for prime-time (Jan timeframe). I've yet to experiment with it, but I have a VMware guest at home all prepped to do so.

I use a pair of 250GB drives with RAID1 and LVM. I was once burnt when my Promise SuperTrak failed, and left me high and dry. Turns out I also had a bad disk too, so off to software RAID I went. I don't need the throughput, but when my disks fail, I absolutely need to get to it.

Re:Surviving 350Gs ? (request for real world units

[Anonymous Coward]

This isn't terribly accurate...

The problem is not the drop, but he change in velocity (aka - acceleration).

Dropping a pen on your desk from your hand resting on on the desk can be about 25Gs. This is about a 1" fall.

Dropping it on a magazine reduces that to about 5Gs simply because the magazine provides a cushion and extends the decelaration time by a factor of about 2.

350Gs (depending on the MASS BEING DROPPED and what it FALLS ON) may translate into about a 3" drop.

Re:Surviving 350Gs ? (request for real world units

[Anonymous Coward]

350Gs (depending on the MASS BEING DROPPED and what it FALLS ON) may translate into about a 3" drop.

Because velocity in a fall depends on mass? What?

Sounds sexy..

[Drinking Bleach]

But my /home partition is only filled with 7.6GiB of data! That's the part I care about, I can reinstall OpenBSD any day; the rest of the filesystems have only about 2.6GiB of data on them (just the operating system, installed programs).

Oh yeah, I'm not just wiggling around on a small hard disk, I've got a 250GB (roughly 232GiB, I think) disk here. So much space unused, I just decided to partition only 120GiB to OpenBSD (/home being the largest filesystem) then screw around with other OSes with the rest of the space.

Re:Sounds sexy..

[gral]

I have a 160Gb with around 20Gb free. MythTV cuts down on free space QUICKLY. ;-)

Re:Big Big Drives are great...but backup is a prob

[UnknowingFool]

however how the hell am I gonna back this thing up

My God, man, that's what RAID is for! Or an NAS. Any self respecting geek wouldn't bother to ask that. I think at the next convention, you need turn in your geek card membership and your light sabre. You can still drool over Natalie Portman.

Re:Big Big Drives are great...but backup is a prob

[tomstdenis]

Why do you people troll as AC? Penis not big enough?

RAID is the best first step anyone can do. 99% of the time specially I've seen systems die because the single HD they store their entire CVS tree on dies. Not because the lab blows up or a virus infects the OS.

So getting a RAID-1 or RAID-5 is the best first thing to do. Yes, after that you get into nightly tapes and offsite storage.

But without the first step it's meaningless. You can't afford to have to force a full restore or even wait for partial restores [possibly corrupting existing files] to get back up. In RAID-1 if one HD dies you turn it off [if using typical retail gear without hotswap], replace the drive and boot back up. You can be back up and running fairly fast.

Getting two or even three drives in RAID is a hell of a lot cheaper than a DLT library.

Tom

Re:Big HUGE warnings - Not quite true

[soleblaze]

Yes, there's raid 6 support in the linux kernel, but it was found out that it was the major cause of Bush to be re-elected. That being the case, I hear it's a lot better now than it was then. Never used it though.

Re:Big HUGE warnings - Not quite true

[RedWizzard]

Not to mention the fact that if you use hardware RAID you're tied to that particular hardware's implementation of RAID. If that controller dies (unlikely, but possible), you need to replace it with the same sort of controller.

This is not too much of a concern if the controller is on an expansion card. But if it's on the motherboard it's a major issue. Motherboards fail much more often than RAID controllers, and new designs come out much more regularly so you're more likely to have trouble finding the particular motherboard you need a year or two down the track. I stay away from motherboard based RAID now (RAID 1 is fine, obviously).

LVM over RAID over LVM

[Anonymous Coward]

Dude

Perl expressed as storage...

You win. I will never subject myself or anyone else to something like this. I forfeit. U R TEH UBER DISC MASTOR!!!11

You have no right to expect any of that data to survive a reboot. If it does it's just luck. Seriously. If Linus knew what you were doing he would ridicule you publicly. :)

Have you observed any performance loss (including CPU overhead) with your setup?

Re:LVM over RAID over LVM

[swillden]

You have no right to expect any of that data to survive a reboot. If it does it's just luck.

LOL.

Actually, LVM and MD are both very resilient. It's pretty hard to lose data unless you do something really stupid. The fact that the metadata is stored on the disk partitions and that the tools can find it automatically has saved my butt any number of times.

Have you observed any performance loss (including CPU overhead) with your setup?

My current setup is just LVM over RAID and I haven't noticed much performance loss, if any. There is a bit more CPU utilization, but since the machine is pretty lightly loaded it doesn't affect the I/O rate much. I see what you'd expect to see from RAID-5 across disks on independent controllers -- very small performance degradation on writes, significant performance improvement on reads.

I don't expect to see any change by moving from "physical" partitions at the bottom layer to LVM logical volume pseudo partitions. The advantage of using LVM to create the bottom-level partitions is flexibility. I think. I haven't tried it yet.

Obviously, doing LVM over RAID over LVM will require some customization of the boot scripts, since most systems start MD devices, and then start LVM. I'll need to start some LVM devices, then start MD devices, then start the other LVM devices.

Re:Big HUGE warnings - Not quite true

[drinkypoo]

The simple fact is that in many cases, SW RAID will provide better results than HW RAID. HW RAID controllers have fairly low-power (as in ops, not consumption) processors on them. If you have a multi-core system, which is pretty normal, odds are you will have the processing time someplace available for doing your parity computations or whatever.

Re:Big HUGE warnings - Not quite true

[jfisherwa]

Neil Brown, mdadm maintainer, has added the ability to grow (i.e., add disks) RAID5 arrays to mdadm (2.5) and sent the related patches up the 2.6.17 tree. Compiling 2.6.17-rc5 now to give it a go. I've seen reports of success on usenet, so I'm cautiously optimistic. This could be a great time-saver.

Re:Big HUGE warnings - Not quite true

[swillden]

Yes, that would be very cool indeed.

Re:Nice try, strike two

[WidescreenFreak]

I have several points which shoot your statement down, at least with me:

A. I have absolutely no need for 750GB hard drives. I'm not into ripping movies to DIVX files nor am I looking to (for some odd reason) pirate every MP3 that's available on Usenet, like some Slashdotters something appear to like doing.

B. You're assuming that I would have need for all 750GB of storage to be backed up, which I don't. If I have need for that much storage that requires retension, I'll dump multiple copies them off to other media with parity files. If I ever did require that much storage, it would probably be for CDs or DVDs that I already own, so it would be no problem to fill up that much data again.

C. A DLT7000 drive backsup at about 10MB/sec. 750GB would take about 20 hours to backup locally, far less than my three-month retension time.

D. You're equating me with a Joe User. I'm careful about what I delete and I know what the "Recycle Bin" is for. It's been quite some time since I've deleted files accidentally that required a restore or rebuild. In fact ... let me think ... oh, I've *never* had to do a restore of my personal files!

E. I'd most likely purchase a second drive, enclose it in an external FireWire enclosure, and copy my new/modified data over to it once a month or so ... in combination with tape backups.

Perpendicular deathstars?!?

[OfNoAccount]

Marketing slogan: Try our new perpendicular "Gary Glitter" drives from HGST - they're perfect for those moments when you just can't destroy your data fast enough.

Not that I'm bitter about those 60GXPs or anything ;)

Re:The justification for more space - FLAC

[quiddity]

one word: FLAC

FreeNAS?

[MikeFM]

Given that I read Slashdot from the top down I have to ask what the FreeNAS support is.

I'd like a nice RAID 5 setup in a big ol case that can hold a dozen hdds. Of course maybe I should just wait until the drives cross the 1TB barrier. These would be nice for my servers that use a combination of ramdrive and CF-based drive for the files that need fast reliable storage and cheapie disk arrays for providing my bulk storage. I always like the occassional density upgrade that lets me put twice as much into the same amount of space.

Re:Big HUGE warnings - Not quite true

[OrangeTide]

Yea. I worked at a startup building a Linux-based NAS Filer appliance. We found it just made more sense to run LVM and have a cpu or two mainly dedicated to pushing bits to the drive controllers. It's faster and it gives you a lot more flexibility in vendors. Hardware raids are less configurable and often you can't do funny mixes of raid 0, 1, 5 and append data. Plus we wanted to plug our controllers straight into hypertransport and at that time all the SATA raid controllers were PCI-X at best, but a couple companies were making straight HT SATA/SAS controllers. (48 SATA drives in a 5U, 4 opterons, up to 26 gigE nics). Append "raid" is actually pretty fast because you can have some filesystems like XFS spread the accesses accross your appended volumes and get the same performance as a RAID-0, with the benefit of being able to grow volumes dynamically.

Re:Big HUGE warnings - Not quite true

[Sloppy]

So I'll have LVM over RAID over LVM.

You are my hero.

Re:Big Big Drives are great...but backup is a prob

[Sloppy]

Yeah, when is someone going to make a tape drive (it's ok if the drive is a little expensive) with this technology, so I can store a terabyte on a $10 tape?

Re:Big Big Drives are great...but backup is a prob

[darkmeridian]

Buy another drive. Or buy a lower capacity drive and save the stuff you *really* want. I would stick with mirroring, though.

Re:Nice try, strike two

[Lawrence_Bird]

You did reply in regards another poster's comments on backing up 750GB, no?
In that case your reply was off topic as clearly there are backup solutions
for those like you who only need to secure a few 100MB or a couple GB.

As to your 20hr backup, how many tapes, and will you be there to change
them?

Your original post advocated RAID for all but the most critical, I simply
pointed out the fact that RAID is not a 'backup' at all and does not
protect from user error or programs gone awry.  That you have (so far)
managed to avoid such a situation is likely as much to do with luck as
aptitude. Will your deleted/damaged file be part of your once a month or
so extra hard drive + some tapes backup? Maybe..maybe not.

Re:Now all I need...is a backup perhaps?

[Anonymous Coward]

The nominal cost for backing up family photos is a CD-R drive and a pack of CD-R disks. Do this a few times a year to keep the media fresh, and it works pretty well. I also avoid deleting the pictures on my SD card until it is full, because the card itself is a backup for a while.

BTW

[SmittyTheBold]

And no, that "2, Funny" great-great-great-(how great?)-grandparent barely counts as "obviously funny." I've seen essays on the topic of back-alley abortions in third-world countries funnier than that.

Re:Big Big Drives are great...but backup is a prob

[Eil]

Great I can buy a 750GB drive...however how the hell am I gonna back this thing up...

Easy: with a second 750GB drive. In my case, I have a file server with two 200GB drives. (Installed back when 200GB was widely considered a metric buttload of storage space.) All of my important data lives on one of the drives and the workstations mount a Samba drive exported by the server. Each night, the 200GB drive containing my valuable data is backed up to the other with 7 days worth of increments. (Off-site backups merely involve an extra machine and a decent network connection.)

Sure, this method assumes the added cost of a whole separate drive, but I challenge anyone here to come up with a cheaper and easier way to keep a large disk properly backed up.

That reminds me of a story...

[Atario]

from the raid-array-in-a-box dept.

I want to buy a RAID Array, but I went to the ATM Machine for money, but I forgot my PIN Number. Luckily, I keep a copy in my phone, but found that the LCD Display was broken! I went back to my computer, where I had yet another copy, but found that both the NIC Card and the SCSI Interface had gone bad!

That's my problem: not enough redundancy!

The real measure

[Aceticon]

How much do they hold in Library of Congress units?

ENIAC and vacuum tubes

[Circlotron]

ENIAC with it's 5000 odd vacuum tubes was a potential reliability nightmare. People argued that it would break down on the average every 30 minutes. It did rather better than that through running the tube filaments at a quite reduced voltage. The comparison with multiple HDD's should be obvious though. The more you have, the more likely you are to have a system (not individual drive) failure.

Use flac instead of mp3

[antikristian]

Convert your CD collection to Flac and you would end up with a few years less of music, it would on the otherhand sound great! Throw out that old Ipod video and exchange it for a 60 GB Iaudio flac compatible player, then we are talking!

Re:Use flac instead of mp3

[WuphonsReach]

Yeah, I sorta figured the 128kbps number would be controversial (which is why I documented it as 128kbps). Personally, I rip to FLAC but use 192/256kbps MP3 for day-to-day use. The FLACs get stored on DVD in case I ever have the desire to re-convert the collection to another format.

(One reason that I use 192/256Kbps MP3 instead of, say, 320Kbps MP3 is that portable CD/MP3 players have limited buffer space. So the higher bitrate tracks result in more skips if you're actively moving. The 192/256Kbps MP3s strike a balance between quality / size / skip resistance. Dunno if the skip issue applies to HD-based players and I made that choice a few years ago before flash-based players entered the multi-gigabyte range.)

Re:Big Big Drives are great...but backup is a prob

[haplo21112]

Ok, let me clarify...

Yes RAID5....Got that, however STILL have to backup the RAID array every so often! Double disk failures are rare, and remembering this is a home setup there are certain dangers that a real datacenter is pretty kmuch immune to that are almost unavoidable in a house. Everything is on APC power protection and such however I can't control the roof suddenly deciding to leak, or the foundation cracking and letting water in or a pipe bursting, the washer hose blowing or a thousand other things. I can do my best but there are situations that the RAID just isn't gonna do it.

And lets not forget the ultimate home setup issue...$$$...I love RAID but all and all its expensive for what it does and convincing the wife that its worth the money that you are not going to use (from her prespective, you bought 1.5 TB of storage but you can only store 1 TB worth of stuff?) that drive.

What I was really getting at is the old reliable backup methods seem to have fallen by the wayside. I liked tape it was affordable, and often in its heyday the tapes were bigger than some of the hard drives. It could be automated, it was mindless. The buy another drive method seems nice except i have to go pull it in and mount it each time I want to do a backup.

Re:Surviving 350Gs ? (request for real world units

[Anonymous Coward]

It could be referring to kinetic energy (KE=m(v^2)/2) or something similar. In that case, mass would factor in.

So no, velocity in a fall does not depend on mass. Kinetic energy does.

(I don't know what the GP was actually talking about, but it sounded to me like that's what he meant.)

Re:Surviving 350Gs ? (request for real world units

[Zoxed]

OK: but do you know the equation that is required ?: eg we know the mass of the disk drive, and if we assume a drop onto a concrete floor, then can we calculate the height that the drive has to be dropped from to suffer a 350G deceleration ? I honestly have no idea: would it be more like 3" or 10' ?

Thanks for any pointers !