Slashdot Mirror
Seagate Confirms 3TB Hard Drive
Stoobalou writes "After a few weeks of rumours, Seagate's senior product manager Barbara Craig has confirmed that the company is announcing a 3TB drive later this year, but the move to 3TB of storage space apparently involves a lot more work than simply upping the areal density. The ancient foundations of the PC's three-decade legacy has once again reared its DOS-era head, revealing that many of today's PCs are simply incapable of coping with hard drives that have a larger capacity than 2.1TB."
i have to know because i have a Win98 PC that i use to play some old games and i wanted to upgrade it
If you move to 4K sectors, that should change the limit to 16TB, meaning that this shouldn't be an issue for several years. Why would you want .5K sectors on such a large drive anyway?
I ran into that a few years ago when I added a 4TB hardware RAID5 to my Linux server. The partition table that is made by fdisk can't handle it. I was forced to use parted to make an EFI partition table instead. It was a little different but completely doable. Took me about 2 minutes on Google to find a howto.
If the problem is BIOS-related, or a legacy of DOS, or both, does this mean it will work fine on Mac OS X? All new Macs use EFI, and OS X Snow Leopard is a 64-bit OS.
I can see the reviews coming in now stating that "3 terabytes is all you'll ever need to store your documents and information."
The Kai's Semi-Updated Website Thingy
Why would anyone running XP fork over the dough for a 3TB HDD? XP is 9 years old and Win7 is a very good replacement for newer machines, particularly ones who's owners might want a 3TB drive for.
FTFA:
According to Seagate, this includes the 64-bit versions of Windows 7 and Vista, as well as modified versions of Linux, but it doesn’t include Windows XP. Not only that, but you may not even be able to see 2.1TB of a 3TB drive when using Windows XP.
Sure, Windows XP won't allow it - but your grandmother who is still running XP isn't going out to buy a 3 TB drive. The early adopters who want or need this are the ones who are already running a compatible OS.
I had a sudden and vivid memory of the little blue ASCII box displayed by the special Western Digital bootloader I needed to bypass my old BIOS's 2GB drive size limit.
End of lesson. You may press the button.
Maybe you could make multiple partitions, rather than treat the whole disk as a single partition?
No folly is more costly than the folly of intolerant idealism. - Winston Churchill
The problem lies with fat32. http://en.wikipedia.org/wiki/File_Allocation_Table
Fat32 has other problems but the real issue here is volume size. Easy to overcome, just partition the drive. (I think I remember doing the same thing to 1.2GB drives too and fat16.)
Besides SSD/flash that is used in camera/mp3 players/camcorders, why would anybody be using fat32 on a drive that massive? Common file access on dual/tri boot computers can be an issue, but folks smart enough to do that are smart enough to build a file-server.
Some legacy components are wonderful because they "just work" (ps/2 vs USB). But trying to shoehorn a new tech into an old standard just leads to problems.
One other issue with this announcement; why did they bother with 3TB? Should the next step be 4TB? We are counting in binary are we not?
"The price good men pay for indifference to public affairs is to be ruled by evil men." ~Plato (427-347 BC)
I'm aware that hard disk capacity follows a trend similar to Moore's law in that capacity roughly doubles every two years or thereabouts, but much like the CPU industry, does anyone know how far into the future magnetic storage will continue to scale at that pace? Even though solid state drives are becoming more affordable and the performance issues are being ironed out, when magnetic storage is only $70 / TB, it's hard to pass up. I'm just interested in how much longer we can expect to see capacity gains like this.
Is there anyone who currently works in that area or has a background in magnetic storage who has a better idea?
for now :)
seriously standard designer know those limits are going to crop up... and that applies to hard disk space as well as ram. why can't they design around those limits where the limit would be variable. some people say it will waste some space/ cause performance problems... not necessarily.
Never antropomorphize computers, they do not like that
Not this. I don't want my single 4TB drive being split in half then recombined in Windows! I want a solid 4TB.. The more places the data is manipulated, the more places a fault can exist. I want to be able to pull the drive out and stick it in another computer without worrying about what hardware is needed unless the benefits exceed the complexity addition (a hardware RAID will be tied to the hardware, but you get massive speed increases). With this "multiplier", what do you gain? Nothing. So why not fix the problem rather than putting a band-aid on a gunshot wound...
If a man isn't willing to take some risk for his opinions, either his opinions are no good or he's no good
Why can't we just move on and apply the sweet R&D money on the SSDs?
I guess that the "classic" hard drives will reach some sort of physical limit sometime in the (not so distant) future.
Why won't the big boys start to work hard on the SSDs?
It's almost as reading a headline like this "New awesome floppies will be released in a new 10 MB size! - 'USB flash disks are overrated and expensive, nothing beats a good old floppy disk' a spokesperson for a floppy disk manufacturer said"
"A sysadmin is a cross between a detective, a police officer, a gardener, a doctor and a fireman"
It's not a filesystem problem, it's an LBA problem.
The problem isn't so much the partition size limitation. It's a problem that the partition schema (fdisk in this case?) can't describe the entire drive (partition layout) of that size.
I suppose they need to be using GPT at this point.
The problem is probably a combination of (1) too many blocks on the drive, and/or nonstandard block sizes. (1024/2048/4096 instead of 512 byte) - one or the other has to increase to grow the size of the device. There's been a lot of noise lately about OSs that won't like drives that ship with nonstandard (greater than 512 byte) block sizes. Of course for anyone to consider 512 as the only 'standard size' nowadays for block sizes, those are the people that are part of the problem.
I work for the Department of Redundancy Department.
Yes, let's design and manipulate an entire standard in order to work around a problem based entirely on the fact that people won't upgrade. And then you just need them to buy this and upgrade their compu...
waaaaaaaaaaaaaait a second.
Canada: The US's more awesome sibling.
You need to be able to use larger than 32-bit block addresses. This is possible since the LBA spec uses 48-bit addresses. But the internals of some OSes use only 32-bit block addresses. The solution to this is to use a 64-bit OS. That allows 48-bit addresses (and then some).
Another other problem is that the MBR disk partitioning scheme uses 32-bit block addresses, so you can't partition a disk larger than 2TB. But the answer to this is to use GUID disk partitioning.
Finally there's your BIOS, it probably only supports MBR and 32-bit LBA. GUID disk partitioning supports making your disk look like an MBR disk so you can boot off it. You'll have to boot off a partition that starts within the first 2TB of the disk, but other than that you should be okay. Just make sure to never use any tools that think your disk is an MBR disk when you are repartitioning it or otherwise accessing it directly.
Windows 7 (or Vista) 64-bit supports >32-bit LBA and GUID disk partitioning.
And then finally hope you don't get any nasty viruses that try to go around the OS to access the disk (so as not to be detected) and mess it up when the calculations overflow 32 bits.
http://lkml.org/lkml/2005/8/20/95
...Because we all know that everyone is going to get 3 TB of space?
Most people can't even fill a 500 GB disk even with all the music they love, their photos and a few movies. Granted, if you digitize -everything- torrent -massive- amounts of music and want lots and lots of HD movies, you might fill one of these disks, however, the number of people who would say "oh, I'm now switching to Windows 7" is low because most of the time people would either A) partition it and put it on a NAS box or B) upgrade to a different OS
Taxation is legalized theft, no more, no less.
Hard drive manufacturers have long used the 1000 instead of the 1024 measurement, so your former guess is most likely correct.
You misunderstand GPT: if the capacity of the media exceeds 2TB, then GPT structure is required regardless how that capacity is partitioned. XP 32-bit's Disk Management will "see" a GPT disk, but only as a fake restricted type that is part of the GPT spec to guard against modifications by software that isn't GPT-aware.
I also think you underestimate just how quickly the new capacity will become mainstream. Were you were saying the same things about 1 terabyte drives five years ago?
that capability, much like RAM greater than 4GB, is arbitrarily restricted by license to XP 64-bit...
Whew!
I suspect the final nail in the hard disk's coffin won't come from technical advances in SSDs, but when people start to realise they're being ripped off by an exponentially growing amount. We're up to 10% already.
Calling this a 'DOS-era limitation' just shows a lack of knowledge about computers of the past. My MicroVAX 3100 will not boot from a hard drive larger than 1GB. I suppose it is fair to call it a 'DOS-era' thing if you were just another kid with a PeeCee clone and his own phillips screwdriver ( the only tool a 'leet computer expert' need own!!)
Hardware/firmware limits for hard drive topologies were a reality for much equipment of the past.
Nobody should be surprised, I expect we will be hitting 5-7TB in the same form factor in another year or two. The real question is how to best make use of the space since access overheads are not going to change much. Linear read speeds will be able to make use of the higher densities but the game is over the moment you have to seek.
There are lots of uses, even for home users. Mid-range digital cameras now pack well in excess of 16MBits and high-end cameras generate raw files in the 30MB per picture range. The newer Canon's (actually they've been around for almost two years now) can shoot video in full HD resolution at 30 frames a second and generate a 35+ MBit/s H.264 video stream. Just 3 minutes of video is a gigabyte of output. Terrabytes get eaten up very quickly under these conditions.
It gets even better for people serious about making backups and snapshots. I already use a 2TB drive for backups and it can hold about one year's worth of (efficiently stored) daily snapshots from around half a terrabyte of base use. Backup requirements WITH snapshots increase non-linearly, so as the base storage needs go up the backup requirements go up even more.
The biggest problem facing people using these larger capacity drives is being able to manage the filesystem meta-data efficiently. You would be just fine if all the files were huge but there are many situations where the number of directory entries can multiply (snapshots being one good example) and once you get enough that the meta-data cannot be reasonably held in ram any significant manipulation of the filesystem can simply take too long to be practical. I once had 40 million inodes on a Reiser filesystem and wanted to remove half of them. Two weeks later (yes, the rm -rf was running for two weeks straight!) it had only gotten rid of a few million and there was no end in sight. I wound up reformatting the filesystem. Ext is only marginally better but, really, any filesystem is going to have serious problems past a certain number of directory entries when the storage medium is a single hard drive.
This is where having a small SSD in the system to help cache the meta-data really helps. We are able to use fewer, larger drives only because we have a little 40G SSD on each system caching all the filesystem meta-data. So on DragonFly, a HAMMER filesystem + a small SSD does the job quite nicely.
-Matt
Seagate used to be the go-to disk drive maker. But in the last few years their quality has slipped and Western Digital became the better manufacturer.
But I seem to detect that quality and reliability is returning to Seagate's devices. Does anybody have any recent experience with Seagate to share?
Best regards.
Back in the day: 'Whoa, a 1 GB disc! A whole thousand megabytes! I doubt I'll ever get this thing full.'
Circumcision is child abuse.
I used to whine about this too, and my advice is just to let it go :)
SI prefixes were never meant to be used like that (1024), and we should let the concept die.
(a hardware RAID ..... but you get massive speed increases).
Hasn't been true since like the 90s. Linux software raid is always quicker, and far more interoperable and standardized.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
Insightful? WTF. Memory limit is not in license, but in 32-bit architecture.
Design a SATA controller that allows one physical 4.2TB drive to be presented as two 2.1TB disks, behind a SATA port multiplier. Then it's simple... plug your HD in... OS sees two drives, but you have 4TB of storage, once your volume manager does its thing and carves a single 4TB volume out of two LUNs.
Only if I could RAID0 them in software. ;)
Am I the only one who misread this as "Senate Confirms 3TB Hard Drive"? I didn't even know that machines could be nominated to the Supreme Court!
http://alternatives.rzero.com/
1920's peephole porn ... 1 pixel x 1 pixel x 1 frame every 5 minutes
Were you were saying the same things about 1 terabyte drives five years ago?
And we're still saying the same things about 1 terabyte drives today. A VERY small percentage of people have anything close to 1 TB of data. In fact, I'd speculate that a very small percentage of people have over 100 GB, and the majority of people are certainly under 10 GB.
Sometimes it's best to just let stupid people be stupid.
Linux software RAID 5 is faster than a dedicated parity processor on a hardware card? Maybe in pure benchmarks, but try comparing the speed when CPU utilization from other tasks is non-trivial... As for interoperable and standardized, I can agree with both of them...
If a man isn't willing to take some risk for his opinions, either his opinions are no good or he's no good
I don't get all the bellyaching about companies moving forward quickly with improvements and changes?
If they didn't do this, people would complain that buying a new machine from them sucks, because it doesn't feature items X, Y or Z.
For as long as I've done anything related to corporate I.T. - businesses have been depreciating their computer workstation assets on a 3 year basis (maybe 4 if they stretch things out). If they believe a computer has "zero value" after 3-4 years, it's no wonder you're finding a 5 year old one to be less functional than you'd like ....
Here's the thing; Mac users have taken a certain pride in bragging about the longevity of their machines, vs. their Windows counterparts. For quite some time, that had validity, for a few reasons. For one thing, the Mac systems really were built with above-average quality of construction. (So were the original IBM PC's and PC XT's, with those solid metal 5lb. keyboards and steel cases, etc. But the PC clone market quickly became commodity-based, so people were more interested in something with cheap sleeve-bearing cooling fans, a power supply JUST adequate to power it up, and a thin plastic or sheet metal case that cost as little as possible.) So you had a lot of cheap PC clone desktops out there that were literally beat up and falling apart after they got a few years of use in a school computer lab or what-not, while the comparable Macs just needed a good scrubbing with a Brillo pad and some 409 cleaner, and they looked good again, and kept on chugging. Additionally, the MacOS had a more limited selection of software -- and the "staple item" apps were usually updated regularly for a long period of time, but not written to push the limits of available hardware. (Their authors knew their sales numbers were going to be roughly 1/10th. of what a Windows counterpart program might net ... so they weren't going to whittle that down any further by saying "only works with the latest Mac with such-and-such specs!") Therefore, people with older Macs could keep running the latest software for longer than most Windows users could, and that gave them a feeling of getting more "value" out of their Mac purchases.
But times have changed, and today's Macs use the same Intel processors that the latest Windows PCs get (sometimes even a little bit earlier, thanks to exclusive deals Apple and Intel worked out). Like it or not, OS X and modern Macs are on the same upgrade path/treadmill as everything else out there now. As I say though, I actually LIKE it, because it means a new Mac purchase gives you much more of a cutting edge computer while still getting all the benefits of their operating system (which I vastly prefer to Windows). Just resell your old hardware on a 3 or 4 year cycle, while you can still get some money out of the old stuff, and you'll do fine. (The money you get from the used system sales will offset the cost of upgrading more often, basically.)
The magnetic domains and track widths are on the order of 20 nanometers these days, so considering that the average molecule is a nanometer or two in diameter, disks will have to use some other method of storage to get beyond a petabyte per drive.
The determined Real Programmer can write Fortran programs in any language.
Spin power is proportional to the square of the radius only if the power is applied at the centre spindle.
Circumference driven motors exist. It's only a matter of time before the idea is applied to hard-drive platters. Lower power consumption and enormously reduced spin-up time, resulting in even more power consumption as the drive can be more aggressively idled.
When a disk that big crashes, it will probably set off nearby car alarms.
http://www.miracleas.com/BAARF/BAARF2.html
Wow, sent an e-mail as suggested when clicking on "use classic" banner, and got a fast response that addressed my msg
The story has been tagged "getarealos", presumably in response to the DOS-based limitation. But yet DOS has OS right in it - can you really be more real than an OS with OS in the name?
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
From TFA:
Basically, with the original LBA limit set at 2.1TB, it seemed pointless for anyone else to prepare for any capacity beyond this, so we now have a situation where many hard drive controllers, BIOSes, drivers and operating systems are all set with caps of 2.1TB, and this is going to take an industry-wide overhaul to overturn.
Heh! They ARE the industry. Good ideas normally float to the top. Bad ideas have to be forced down customer throats by greedy marketers, making things industry standards.
So, if nothing has been standardized yet, then here's what I see:
Just like 56K being the absolute max in dial-up speeds, a competitor develops a "good enough" incompatible standard to push down people's throats, like DSL. They begin profitting, and others follow suit, even if slowly. Contrast to ISDN being the losing choice to DSL --speedwise it was competitive in the early DSL days, but home pricing and 2-line inconvenience killed it.
I can't recall seeing the peer-reviewed study verifying those facts mentioned on Slashdot. ;-)
It hardly matters, though... there are enough people numerically that DO to still make it a marketplace. Microsoft can afford to be patient and wait for the remainder to hit the wall, which they will eventually.
There's a simple workaround for the MBR problem with 2TB+ hard drives - for Linux.
1) Store your boot files on a small secondary device, e.g. a USB stick. That includes: Master Boot Record, boot loader (e.g. GRUB or LILO), kernel, initrd - about 10 megabytes are easily enough. /sbin/init process.
2) Boot the USB stick. The initrd mounts your hard drive and starts the actual
This comes with several nice options: /.
- Encrypt every single byte of your hard drive. A script in the initial ram disk asks for the passphrase, creates the dm-crypt device and mounts
- You don't need a partition table. Just use the LVM2 inside the encrypted block device.
- Add a Live CD image to the USB stick (separate FAT32 partition, convert isolinux config to syslinux, chainload from GRUB)
I'm using all 3 of the aforementioned options and you'd have to take them from my cold, dead hands ;)
...and make *reliable* drives. I've had 2 500GB and a 320 GB Seagates die, a 1TB Hitachi is beginning to act on its own, so far the only drives I have that seem to work OK are WD (Green, in the PVR). Seagate has a long history of having problems (firmware, hardware) since they bought Maxtor. Give me a reliable 1-2 TB drive instead. So far WD are the only drives that didn't die here.
I've got better things to do tonight than die.
IPv6, Hard drives, multiple cpu cores... just too many Hz and bits. [...]
Drive makers: Take a pause, catch your breath and work on access times, reliability, and pushing reforming drive technologies like GPT.
Parent is insightful, and hasn't been modded up. Our industry thrives in forced obsolecense of older capable tech... while skipping real features that should be priorities.
If these companies weren't American, craving "bigger is better (at the same $$$ margin)" we would see reliable drives and improved access times. Companies put the foot down and we lose anyway: the last time I heard of a race to reduce RGB dot pitch size (increasing monitor resolutions without an end in sight) was when the USA introduced HDTV and widescreen LCD's... This resolution listshows favor to HD and widescreen, though it didn't exist at all around 1998. It's just that profit margins on 4:3 became smaller.
HD fixed resolutions to an immovable size and forced resolutions to relatively poor movie-ready sizes, ignoring graphic work, increasingly cheap large monitors at the home and gaming or 3D editing at huge resolutions. Imagine the profit margins when you no more R&D to invest in. 1080p just doesn't compete with, er.. "1200p" monitors of yesteryear, but the industry took the latter away from my local stores.
High capacity drives should come in one enclosure that can be seen over its PC interface connection as multiple smaller drives the OS can handle. A "drive RAID".
And OSes should stop representing each drive as a separate object, except to lower level administration.
All longterm storage should be pooled as a RAID, even if there's only a single device. Presenting individual volumes to the user (and programmer) is more trouble than it's worth. It's the OS' job to present a simple, effective interface to complex underlying hardware.
--
make install -not war
It is insightful, but it wasn't MY insight, I'm just repeating someone else's discovery, which has been referenced on Slashdot:
http://www.geoffchappell.com/notes/windows/license/memory.htm
You have to keep reading to near the end, because he starts off discussing the similar restriction in Vista 32-bit.
Like I said, license restriction, not a physical one.
I have 14 2tb drives in my MacPro and Drobo. I hope that I will be able to swap them all out for the new 3tb drives. We can never have too much mass storage.
I'm sorry but I really don't see what the problem could possibly be... Assuming you were to actually use a 3TB HDD, I would think of only 2 scenarios : /OS X and these OS don't have that kinda of limitation (the 2.1TB)
A] A workstation with heavy files to manipulate => a somewhat recent hardware/software couple running most likely win7 / Linux
B] A NAS/fileserver/archiving/whatever.. and I really don't see why anyone would run windows XP/Me/98 on this except for the joyfull experience of instability and lack of features. Just use a good OS and be done with this nonsense !
I mean when I want to play FEAR 2, I don't complain about how it doesn't work on ubuntu 6.04 with 2006 opensource GPU drivers..
what is it with you people ?
Ars had a very nice article up a couple of months ago covering sector sizes in detail and the problem of the " 512-byte leaden albatross". Link: http://arstechnica.com/microsoft/news/2010/03/why-new-hard-disks-might-not-be-much-fun-for-xp-users.ars
Maybe in pure benchmarks, but try comparing the speed when CPU utilization from other tasks is non-trivial...
Drives are slow enough, and multi-core CPUs are fast enough, that processing drops into the noise. Google some recent testing... fully saturated drives at 1% CPU use are not unusual claims. Or rephrased, doing software raid under worst case CPU use and best case drive thruput situations might cause up to one percent slowdown, theoretically, although no one will ever see that in the real world.
The situation with software raid drives is exactly where network firewalling was maybe a decade ago... In the olden days a PC-XT 16 bit 8 MHz bus could be fairly well saturated by a 100 meg ethernet. So "everyone knows" that firewalls need special dedicated expensive gear. But the folks with PCI buses and early pentiums running linux knew that they could do line rate no problem. For years I ran a desktop 486 fanless and underclocked to 25 MHz to firewall line rate 10 meg ethernet with ipchains, never getting much above 33% CPU or so, and the old timers simply refused to believe it was computationally possible, or I was faking it or something...
Now I will concede the point that small SOHO NAS boxes have notoriously underpowered CPUs... But good luck finding one that supports more than one installed drive, much less full RAID-5.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
One print page: http://www.thinq.co.uk/news/2010/5/17/exclusive-seagate-confirms-3tb-drive/?full
What about external HDDs? I only saw 2 TB and read that they can't go higher. :(
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
C) Use a file server or more than one hard drive. I have ~3.2TB of space but it's on 10 hard drives of 120-750GB capacity.
Then how come 32bit Server 2000 can access 32GB and 32bit Server 2003 can access 64GB?
Also, why my 11 year old server that's 32bit has maximum memory capacity of 16GB?
+1 to parent.
When I was a college freshman (1992) my pals and I made a pair of spreadsheets charting the CPU clock rate** and primary storage capacity of every computer any of us had owned, back to my first C-64. Both graphs made a nearly perfectly straight line on a semilog plot.
We were totally shocked that our graph predicted a 400 Mhz computer with a 40 GB hard drive in the year 2000 - the numbers seemed completely impossible.
And then 2000 rolled by, and those numbers were perfectly typical. Not much in the computer world has surprised me since then.
** Clock rate looks like it's flattened out since 2005, but only if you consider the clock rate of a single processor. If you multply rate * cores, the curve has continued more or less unabated, at least for computers I've owned.
I stole this sig from someone cleverer than me.
I'm confused... Can XP not access the entire drive or does it have problems booting off of it?
I have a Toshiba Libretto 110CT with a 160GB PATA drive with XP-SP3. I knew that the BIOS couldn't see above the 8GB barrier, but in researching how to resolve this, I found out that XP-SP1+ (or W2K-SP3+) can access the whole drive... The trick is to set the boot partition no larger than what the BIOS can read--2GB, 8GB, 32GB, etc. After it boots, NT uses its LBA-aware disk management driver, ignoring the BIOS. So I set it up with an 8GB primary partition and the rest as a secondary partition & it works fine...
Windows 3.1x calc: 3.11 - 3.10 = 0.00
have any data on how much CPU you need to run raid 5 or 6 using software on linux? can i use a low end via? how about an ARM7? say 6 disks, on sata2 getting as much performance as the drives can handle? how about reading from a RAID6 array while transcoding a video for a device? I know if i use a hardware raid card the CPU is fairly unimportant as far as disk access goes.
All of the above was encrypted with a Quad ROT-13 method. Unauthorized decryption is in violation of the DMCA.
A new drive, which isn't even available yet, will not be compatible on some older hardware and with older operating systems. BTW, in the future, they will release some 128 gig ram chips that aren't compatible with your current motherboard either. Everyone expects to upgrade hardware and software over time as we utilize new technology. I figure, by the time that Fry's has a huge pallet of them sitting by the front door, there will be enough customers that have sufficiently upgraded to be able to utilize one fully.
-Restil
Play with my webcams and lights here
Oh, ye of little mind... think of all the possibilities it would open.. like 3.5" full sized HDDs that have two 2.5" SFF hot-swappable HDDs inside them.
Think RAID1 or RAID0 controller integrated into the HD itself
Who needs more than 2.1TB on a NTFS filesystem anyways? Sounds like a major stability risk, seeing as one bit out of place, or power outage at the wrong time, and the entire medium becomes unusable, due partly to the inpetitude of the filesystem, its repair tools, and lack of data journaling.
> a standard that the industry has long since abandoned
Or rather, never adopted. IBM has specified disk capacity in powers of 10 since RAMAC.
Remember that disk drives aren't just logical storage containers, they're rotating machinery. If you're using a drive this big, you're probably trying to store lots of stuff on it, and you don't want that storage process to be interfered with by rotating or seeking the disk to some other location to grab programs or store random files from other applications, so either put the OS and application software on another physical disk (or mirrored pair), or put them on flash drives or even SSD if you're feeling spendy.
Also, from an operations standpoint, the way you handle large data storage is enough different from the way you handle infrastructure and applications that you probably want to keep them separate anyway. Maybe you're using the 3TB drives to build RAID arrays or databases, maybe the 3TB is in an external-SATA shoebox that you want to be able to replace, or whatever - it's a lot cleaner to keep it separate, and if you can afford the latest biggest drive on the market, you can not only afford the extra $50 for a drive for your OS, you'll save it in operations and maintenance costs.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
So, if we accept the approximations that the human brain stores around 500-1000 Tb of data, we're getting close. However, with the recent problems Seagate has experienced with poor reliability figures on their smaller 500Gb and 1Tb drives, I think we still have a long way to go.
Linux's software RAID is quicker than fakeraid -- that is, the hardware RAID which is merely a bit of code which is executed on the CPU itself. However, it's definitely slower than a real RAID card -- decent ones run in the hundreds to thousands of dollars.
In addition to offloading the RAID function (saving a fraction of the CPU and I/O bandwidth), decent RAID cards will contain a large amount of battery backed cache RAM. Any operation which requires writes to be committed to disk will complete in the time it takes to write it to the cache RAM -- a huge savings. If/when the battery fails, these cards fail back to committing the write to disk. The performance impact seriously hurts our databases (to the point where we typically fail over to the standby as quickly as possible). This is all on a host with 60-80 2.5" SCSI drives.
The heart of the problem with backwards incompatibility is this scenario: Come into possession of a massive drive (or whatever future tech) and not have on hand a good enough computer (in spite of it having been faithful for over 10 years already) to achieve basic access.
Hardware designers hear this: Make a thunk, even if it doesn't promise the best feature set or performance. At the very least, please provide enough info for someone to make thunk eventually.
Know your pads. One time pad: good for cryptography. Two timing pad: where to take your mistress.
Insightful? WTF. Memory limit is not in license, but in 32-bit architecture.
Yes, that's why there has never been a 32bit machine with more than 4GB of memory. PAE doesn't exist (well it doesn't in XP that was the point he was trying to make). Please don't comment on computer architecture again as you will just make yourself look stupid, and then there'll be a long of people correcting you. It just wastes everyone's time and energy, whereas if you keep your mouth shut everything's fine. OK?
The biggest manufacturer of PC Operating systems in the world still seems to remember that there are 1024 bytes in a kilobyte etc.
This does not prove it any more than a bunch of rip-off clowns now overstating their wares by 10% proves it. If you want to have a nice 10^X series. Make up your own names. The existing ones are already defined..
Most of the world does not measure their vehicle speeds in kiloyards per hour. When they wanted to simplify the arithmentic, they adopted metric. The day is 86,400 seconds long. If you want it to be 100,000 you will need to think of a new unit name.
Why didn't they 'redefine' the number of bits in a byte as well? Because that would have made their wares seem smaller. That was all it was about.
I'll see your Constitution and raise you a Queen.
Many complaints about 1TB Seagates they made. If this one is half as bad as that series then I do not care.
The Fujitsu CEO reacted angrily to the news of the impending release of the Seagate 3TB hard drive. "Would someone tell me how this happened? We were the fucking vanguard of storage in this country. The Storagebird 1 TB was the drive to own. Then the other guy came out with a 2TB drive announcement. Were we scared? Hell, no. Because we hit back with a little thing called the MaxStorageBird 2. That's a drive with monster cache, for performance. But you know what happened next? Shut up, I'm telling you what happened--the bastards went to 3TB. Now we're standing around with our cocks in our hands, selling two gigs and a big cache. Monster cache or no, suddenly we're the chumps. Well, fuck it. We're going to five terabytes.
Sure, we could go to three terabytes next, like the competition. That seems like the logical thing to do. After all, three worked out pretty well, and four is the next number after three. So let's play it safe. Why innovate when we can follow? Oh, I know why: Because we're a business, that's why! . . .
We didn't claw our way to the top of the storage game by clinging to the two-terabyte industry standard. We got here by taking chances. Well, five terabytes is the biggest chance of all.
Here's the report from Engineering. Someone put it in the bathroom: I want to wipe my ass with it. They don't tell me what to invent--I tell them. And I'm telling them to stick two more terabytes in there. I don't care how. Make the platters so thin they're invisible. Put some on the motor magnets. I don't care if they have to cram the fifth terabyte in perpendicular to the other four, just do it!"
Dumbass. If the problem is people aren't upgrading their machines for one reason or another, why the fuck would designing a new standard that would require them to upgrade their machines make any fucking sense?
I'd ask when slashdot became overrun with morons, but I know they've always been around.
Canada: The US's more awesome sibling.
What would be the draw for the average consumer to actually own a 3TB hard drive? Is there a point where the size becomes too much?
I must say I have never had any luck with WD drives, with my general finding is they don't last long and tend to die without warning. And this goes back to the days when 200MB was a big drive.