An In-Depth Look At Seagate's 1.5TB Barracuda

theraindog writes "More than a year and a half after the first terabyte hard drives became widely available, Seagate has reached the next storage capacity milestone. With 1.5 terabytes, the latest Barracuda 7200.11 serves up 50% more capacity than its peers, and at a surprisingly affordable $0.12 per gigabyte. But Seagate's decision to drop new platters into an old Barracuda shell may not have been a wise one. The Tech Report's in-depth review of the world's first 1.5TB hard drive shows that while the latest 'cuda is screaming fast in synthetic throughput drag races, poor real world write speeds ultimately tarnish its appeal."

Comments

Write speed

[qoncept]

How important is throughput? I'd be interested in knowing what percentage of these drives are going in external enclosures. For the time being, 1.5tb is much larger than you'd need to be running any applications off of and I'd guess the majority of these drives are going to be storing movies, mp3s and photos, where the speed hardly matters at all.

Re:Write speed

[morgan_greywolf]

How important is throughput?

Ask your swapfile or swap partition (as the case may be).

Re:Write speed

[Anonymous Coward]

It matter on workstations where these drives are probably most likely to the used. Think software development, artists (3D, animation, video, sound, etc), VMware machines, etc... Plenty of people need more space.

It's weird, I just just thinking about how my 7200.11 500 GB drive seems rather slow compared to the pair of 7200.7 200 GB drives I was using in my last workstation. The throughput is similar in synthetic benchmarks but the old pair of 7200.7's was much faster when I was using VMware and such. This new 7200.11 drive stalls all the time while writing.

Re:Write speed

[BorgAssimilator]

Well, I wouldn't count out speed completely. I'd agree that it's not as important for storing media files than like playing games, but the last thing you want is to have a movie that randomly skips every so often because of a hard drive. That being said, it looks like the drive is pretty good when you're reading stuff from it, so that shouldn't be a problem.

But you're right; I'd be more interested in seeing stress tests about how much usage these things can take before they give out.

Re:Write speed

[qoncept]

My point was I want to know how often the hard drive is going to be the bottleneck instead rather than USB or firewire, where all of them would perform (even more so than they already do) virtually identically.

That aside, this drive actually performed near the top in most of the tests and middle of the pack in most of the others, so the author talking bad about its performance was pretty unfounded. And I didn't see anything in any of the tests that would make me choose from the drives tested on anything other than cost and capacity. The truth is, in the "real world" everyone is clammering to compare the drives in, you'd never have a clue which drive was in your computer unless you opened up the case and looked.

Re:Write speed

[ShadowRangerRIT]

Speed can matter for video, particularly on higher quality settings. For example, using Windows Media Center with "Best" quality, according to MS's website, 60 GB will hold about 22 hours of video, which equates to about ~0.8 MB/s. If you are recording two shows with a dual tuner and watching a third that you already recorded, you're up to ~2.4 MB/s total throughput. If you aren't swimming in RAM for the disk cache, the HD head is going to be losing a lot to seek and rotational latency.

Since a lot of home users only have one HD, the OS is going to be chewing throughput, as will the page file, and both will increase latency further. It all adds up. Not everyone installs 6 HDs with specialized purposes the way a geek (such as myself) does.

Re:Write speed

[Endo13]

I'm inclined to agree with you. Also worthy of note is that most of the other drives in the test are actually more expensive, despite having less space. And guess what, most of the ones on the test that come in at a lower price are also ones that are outperformed by the new drive on virtually every test. So yes please, I'll take 50% more space for better read speeds and less money, not to mention a 5-year warranty. I've purchased Seagate drives exclusively for about 4 years now, and have yet to have one fail on me. Also worked at a computer service/build for a year recently where we used Seagate drives exclusively, and saw maybe as high as a 1% failure rate on new drives.

Re:Write speed

[Znork]

How important is throughput?

For what I'll use them for? Not very. Looks like they've got great stats for bulk storage, and any more demanding segments I can stripe and/or cache anyway (with memory prices where they are, it's not like you hit swap anymore).

Lack of speed can be overcome. In the worst case by patience. Lack of capacity, not quite so easily. So several of these are definitely on the shopping list. (Mmm, mythtv storage...)

Re:Write speed

[mikael_j]

Because obviously any disk not used for your operating system or applications would be connected using USB or Firewire, couldn't be that some people actually connect their SATA drives directly to the SATA bus in their computers, right?

/Mikael

Re:Write speed

[fifedrum]

speed always matters, just not as much as $/GB most times, at least where these drives are destined it doesn't matter as much as speed

Re:Write speed

[LWATCDR]

There are databases that could use these. Three in a RAID 5 would give you 3TB of storage. Or you could do four in RAID 1+0 for 2TB of storage.
Also people doing video editing can use the space. For those users the speed is an issue. I guess they will just have to stick with tiny 1 TB drives for now.
Actually for really high performance databases and video editing systems they will use RAIDs of really fast drives like the Western Digital VelociRaptor or if money is no issue at all 15k RPM SAS drives.

Re:Write speed

[thepotoo]

But do you really need 1.5 TB of space for software development?

Even 3D animation and video editing aren't going to eat up 1.5 TB until you hit the professional level, at which point you can probably afford a RAID solution.

Having said that, this tech will improve over time, and I would just like to point out that we now have more storage space available on a single hard drive than in any one human brain (well, probably. Unless there's some sort of "intron/alternate splicing" analogue going on in the brain).

Re:Write speed

[Bandman]

I came into this thread to post the same message for the most part. Though as size increases, so does risk of failure, as I'm finding out.

Re:Write speed

[Walpurgiss]

I have 7 of the WD 5400-7200RPM "GreenPower" 1TB HDs in a raid5 array that I access only through FTP and SMB.

I suppose 5400RPM is slow in terms of transfer and seek time, and being a software RAID5 set managed in software via mdadm likely also reduces the speed of the array. However none of that speed decrease is readily apparent due to the relative bottleneck of the 1GBPS ethernet connection.
I assume that drives of this size primarily would see similar use as the drives I use. Given the experience I've had, I agree that the speed of the drive probably doesn't matter so much. I doubt many people would use a 1.5TB drive for their OS or swap space, especially if speed mattered.
The speed people probably would be using some ultra wide scsi drives or some other speed oriented drive, perhaps the raptor line.

Re:Write speed

[tgd]

If you need a 1.5 TB swapfile, I suggest you start clicking some of those X's on your windows.

Re:Write speed

[Anonymous Coward]

I actually just picked one of these up from newegg, I've had absolutely no problems with speed, it replaced two 250 gig drives that used to be in my system, it streams HD video just fine. Heck, even the article shows this, with large single files it beat every drive in read speed.

Re:Write speed

[zsouthboy]

A whopping 2.4 MB/s (+ overhead, as you say)?

You realize that most modern drives are able to handle 60 MB/s with ease, even the low end ones, right?

You don't need 6 hard drives RAIDed to *watch* video...

Re:Write speed

[Chabil Ha']

I would argue that you need a RAID solution regardless. Assuming you use the bulk of that drive, backing up ~1TB of data in event of drive failure can be a pain. With those sized drives, I'd only go with RAID 1/5/6/combo anyway.

Re:Write speed

[Belial6]

Lack of speed can be overcome. In the worst case by patience.

That is just true. So from now on, it should be written...

Lack of speed can be overcome. In the worst case by patience.

--Znork

Re:Write speed

[LWATCDR]

A raid doesn't replace backups. It is an aid in maintaining up time and increase speed.
The problem is that backup media is now really lagging storage. I use external drives as back ups but they are not my first choice. Right now they are my only one for some data sets.
Another back up I have become really fond of are flash drives. I keep one attached to my PC all the time as an extra backup. Just copy the source directory to it as well as the server.

Re:Write speed

[Anonymous Coward]

nyarhg... must... correct... DB Newb...

No competent DB admin would use RAID-5 for a database unless it was read-only, the write performance of RAID-5 is known to be pretty bad. As well 'really high performance' databases would skip right past WD Velociraptors and go straight to SCSI or SAS (Where 10/15k RPM has been standard for decades), the WD Raptors are marketed towards gaming enthusiasts and power users, not high end servers.

Re:Write speed

[eln]

Swap performance is going to suck no matter where you put it, except maybe solid state. If you're hitting swap so hard that the performance of said swap is a real issue of concern for you, you really ought to consider buying more RAM.

Re:Write speed

[morgan_greywolf]

True. But my point is this: think about image or audio processing. Or video processing. How many temporary files are written during common processing operations? And it doesn't matter how much memory you have, they all do it.

swapfiles have the same consideration.

Re:Write speed

[Hao Wu]

I'd be interested in knowing what percentage of these drives are going in external enclosures. ... I'd guess the majority of these drives are going to be storing movies, mp3s and photos, where the speed hardly matters at all.

I will not consider another purchase until USB 3 or Firewire3200 is available.

Re:Write speed

[nabsltd]

Mod parent up.

Ask any MythTV user about how they can record a pair of HD streams while watching a third with no skips. That's about 7MB/second, and drives don't break a sweat on that.

Re:Write speed

[neumayr]

Most artsy things that take up a lot of space don't need much disk speed, as it's all so computationally expensive the disk usually doesn't have any trouble keeping up with the speed at which the cpu crunches through that data.

Re:Write speed

[MikeDirnt69]

Even 2xSCSI on Raid? If anyone have numbers, please post.

Re:Write speed

[Idiomatick]

Who has only 1 harddrive that is tech savvy enough to need a 1.5TB drive and uses it to record two shows at once while watching a show? A non geek should be fine with a 1TB drive... or 500 probably.

Re:Write speed

[Sancho]

No, he asked in his first post how many would be connected to a low-speed bus, and he clarified his point when someone else who couldn't read mentioned swap files. Here, I'll quote it for you:

I'd be interested in knowing what percentage of these drives are going in external enclosures. For the time being, 1.5tb is much larger than you'd need to be running any applications off of and I'd guess the majority of these drives are going to be storing movies, mp3s and photos, where the speed hardly matters at all.

http://hardware.slashdot.org/comments.pl?sid=1003109&cid=25457241

So if you weren't intentionally trolling, it definitely came off that way.

There's a bit of truth to what he says, too. Lots of people use drives this size for what is effectively long-term storage. They use it for their movie collections, their music, their HD TV shows, etc. Without that, in fact, the market for these drives would be really, really small--limited, if I were guessing, to people working with video. Write-performance will have a pretty big impact in that market, but just about anywhere else where this kind of massive storage is used, it's probably going to be negligible.

Re:Write speed

[nabsltd]

Though as size increases, so does risk of failure, as I'm finding out.

That blog post forgets one thing: sector remapping.

With any actual redundant system (i.e., not RAID-0), you increase the likelyhood that the data is still there somehow. The drive with the unrecoverable read error re-maps the sector and the RAID software/firmware uses the redundancy to recover the correct data and write it back to the re-mapped sector.

Re:Write speed

[Sancho]

RAID 1/5/6 will have worse write speed, anyway. RAID 0 is what you're looking for, but you want to use it only as a working drive, and transfer completed (and to some extent, even intermediary) work to safer storage as soon as possible.

Of course, we're mostly talking about raw video here. Working with anything compressed, and you can probably get away with simpler, safer solutions from the get-go.

Re:Write speed

[LWATCDR]

You mean like I said here.
"Actually for really high performance databases and video editing systems they will use RAIDs of really fast drives like the Western Digital VelociRaptor or if money is no issue at all 15k RPM SAS drives."

You are without question correct in an enterprise setting. Raid 5 isn't too bad with a good hardware controller and offers a good compromise of speed and size for a small departmental server. I would still use RAID 1+0 but I know people that are just too cheap for their own good.
The raptors I think will be game changers. They really push down the cost of a very high performance raid. Again not enterprise level but dang fast.
Notice also I was also talking about video editing at the same time. Video editing is about the only use of RAID 0 I can think of and to be honest I would still go for RAID 1+0. I will also bet you will see a lot of raptors in that space.

Re:Write speed

[Sancho]

Uh oh, are you running Linux? Are you aware of the head parking problem with these drives?
http://kerneltrap.org/node/14912

Re:Write speed

[jedidiah]

Anyone with a Tivo...

Re:Write speed

[mikey1134]

I would have to disagree. My company uses a RAID-5 for DB's and a RAID-1 for logs, we have yet to run into a performance barrier with this configuration. Mind you, a larger DB than ours (60GB) with more users than ours (100-500 connected users) might require a faster setup. But I wouldn't say that it's incompetent to place a DB on a RAID-5.

Re:Write speed

[gravis777]

I agree. I got a 750 gig drive in my desktop, and its actually partitioned out, for three OSes, and one large storage area. All my other drives are used for storing photos and video, and are ALL externals. I care much more about how well the data holds up over time, and how well the drive handles shock, than the throughput. Shoot, my 750 gig external is hooked up via USB2 to my Dish Network HD reciever. Even when I am storing broadcast HD, you are only dealing with 17Mbps (a little more than 2MBps).

The exception would be if I was dealing with Uncompressed HD video - THEN throughput would be an issue for me.

Re:Write speed

[lysergic.acid]

speed definitely matters. i don't like watching a 2 GB HD video and having it skip constantly.

but the other thing that's important and less often discussed in reviews (or benchmarked) is reliability. with these large 750 GB+ drives, how many full disk re-writes can the drive go through before it starts crapping out (data becoming corrupted, or simply dying)?

are these high-density platters just as reliable as those of lower-capacity hard drives? for reliability/longevity, is a single 1 TB drive preferable to 2x 500 GB drives or 4x 250 GB drives?

Re:Write speed

[Anonymous Coward]

Just a nitpick... that 1Gbps. Little b bit, not BIG B Byte.

Re:Write speed

[afidel]

Another big market for big drives is archival systems and disk to disk to tape systems. I don't always need 15K FC performance, but if it was significantly slower than comparable 1TB drives I might just go with more 1TB drives to make up the capacity difference.

Re:Write speed

[afidel]

Come on, any SATA drive can play an HD movie, even a BluRay rip comes out at what 45Mbit/s max, that's a punny 4.5MB/s, something IDE drives could do almost two decades ago.

Re:Write speed

[afidel]

I have servers with swap files about 10% of that size and I can expect my next DB server will be 20% of that. If your server has 128GB of ram then a 256GB swap file is 'normal'. Of course I would never use a single SATA drive for that, it goes on the faster part of the SAN =)

Re:Write speed

[qoncept]

Why? Don't you store anything that doesn't have to be read fast? People don't seen to understand that your mp3s don't care if they are on a fast drive or not. The only reason external drives are 7200rpm instead of 5400rpm is because those are the drives they are already making anyway.

Re:Write speed

[afidel]

Yes, even 110xFC disks sucks for performance of swap, the problem is swap is orders of magnitude slower than ram yet it's made to stand in for ram at times. I wish there was an OS where the VM system was tuned to not need swap, but I've seen both Windows and Linux systems where enabling swap significantly increased performance even though the boxes shouldn't need it (IE a box with 64GB of ram and using ~17GB of it).

Re:Write speed

[Hyppy]

You are without question correct in an enterprise setting. Raid 5 isn't too bad with a good hardware controller and offers a good compromise of speed and size for a small departmental server. I would still use RAID 10 but I know people that are just too cheap for their own good.

"Just too cheap" is a bit harsh. Let's compare, say, 3TB (decimal scale) of usable space, assuming 200 per drive for roundness' sake. We can break the redundancy into minimum and maximum drive failures (Min/Max) Obviously, having a higher minimum is best, but having a higher max can't hurt either.

RAID 5 - you need to have 3 drives, which costs $600. In order to lose your data, you need to lose any two drives. That's a 2/2 min/max out of 3 drives for $600.
RAID 1+0 - you need to have 4 drives, which costs $800. You can lose your data with the failure of 2 drives within the same mirror, which there are two of. That gives us a min/max of 2/3, for $800.

In all likelihood, you will have time between the failure of the first and second drives to replace the failed drive. So, RAID 10 only buys you a 66% chance to keep your data in case of a second drive failure during that time window. On the other hand, the initial investment is also 33% higher, and it won't necessarily be available on the same controllers. Even then, the controller might actually be doing RAID 01 (damned mislabeling), which downgrades you back to a 2/2 min/max drive failure tolerance.

Of course, if you're REALLY concerned about write speed, RAID 10 or 01 is the way to go. Otherwise, if you're on a budget (who isn't), RAID 5 will do just fine.

Re:Write speed

[afidel]

RAID+snapshot's IS an effective first level of backup. For my dad's business I supplement that with local and remote copies to their workstations and then the local workstation also backs up to an offsite provider, but I've never needed anything more than the snapshots so far in 5 years of running this setup. Even in the S&P 500 company I admin there's very few non-legal requests to go back farther than my snapshots cover (just less than 3 months).

Re:Write speed

[LWATCDR]

But you are loosing a lot of performance to save a few dollars.
But if the database is read a lot and write a little in nature then as I said RAID 5 isn't terrible. Just not ideal.

Re:Write speed

[Hao Wu]

Occasionally I still use drag-n-drop to backup whole hard drives. 'Time Machine' made this less common, but it's still an overnight task when I copy to new drives.

Re:Write speed

[Walpurgiss]

Wasn't aware of that, thanks.
I'll have to dig into it. At least they're in RAID5 so unless I lose 2 together or a second before I can recover the first I'm ok. From the thread it seems like the expectable lifetime of the drives always on would be around a year. And as below, yeah it should have been Gbps not GBps. :p

Re:Write speed

[Instine]

I have an application for such a drive (a few thereof), and indeed it is storing audio data...

Re:Write speed

[fm6]

What about eSATA? Fast, places less strain on the host CPU, and is available now.

Re:Write speed

[PitaBred]

Swap? You need to buy more RAM. Swap should never be in regular use, only extraordinary use, and a few MB/s won't matter once the system grinds to a near halt once it starts hitting swap.

Re:Write speed

[Anonymous Coward]

Someone constructing a fresh computer and switching connections?

Someone not bothering to switch outdated hard drives over from an older machine and/or preferring the older machine still to be usable?

Re:Write speed

[lysergic.acid]

come to think of it, you're absolutely right.

i haven't had that problem with my new hard drive, but on my old IDE drive i regularly encountered skipping when watching hi-def videos. i mean, i have a 2.79 GHz P4 with 1 GB of RAM running Windows XP. so i don't think it's my processor. i even tried turning off all other applications in case they were sucking up the IDE bus.

eventually i just got fed up with the hard drive and had it replaced. by then data was becoming corrupted and the drive became less and less reliable. the thing is, it was only about 3-4 years old. which to me, is way too early for a drive to start failing.

but even with my new SATA drive it takes a while to seek on hi-def videos, and if i move the tracking cursor (whatever it's called) too much the player will freeze up or the audio will get out of sync. i guess when you fast-forward or rewind the player needs to read much faster than the video's normal bitrate rate, but it still shouldn't have that much of a problem.

Re:Write speed

[jherekc]

surprised no-one else has already picked this up, but four 1.5TB disks in RAID 1+0 would give you 3TB of usable space (slightly less as HD manufacturers count in 1000's rather than 1024's or whatever)

Re:Write speed

[frieko]

The answer there is also more RAM. File writes don't go directly to disk, they go directly to RAM and then are written at the OS's convenience.

Re:Write speed

[speedingant]

It's definitely your processor. 2Ghz C2D is absolute minimum for playing 1080p, even more for high bitrate media. Try using XBMC for playback, it's very efficient and may help you until you can afford to purchase another computer.

Re:Write speed

[afidel]

It's called eSATA, much better than either of those for external storage and it's here today.

Re:Write speed

[ATMD]

There are two criteria for a media centre PC: lots of storage space and small size. Oh, and minimal heat production, so loud fans aren't required.

All of this points to a single, high capacity disk as the optimal solution.

Re:Write speed

[Fulcrum of Evil]

If your server has 128GB of ram then a 256GB swap file is 'normal'.

No it isn't. If you're using more than about 100M of swap, you probably need more memory. Agreed on the multidisk thing, but it's mostly a matter of reliability - raid5 is cheap insurance, raid1 even cheaper.

Re:Write speed

[LWATCDR]

First level of back up... enough said.

Re:Write speed

[turbidostato]

"I would argue that you need a RAID solution regardless. Assuming you use the bulk of that drive, backing up ~1TB of data in event of drive failure can be a pain."

And here comes Chabil Ha' summing to the vast multitude that confuses avaliability with recoverability.

Re:Write speed

[GigaplexNZ]

And for those users, the 10,000rpm drives are a better solution. This isn't supposed to be the be all and end all drive.

Re:Write speed

[afabbro]

If your server has 128GB of ram then a 256GB swap file is 'normal'.

Only if you're pedantically following advice from 10 years ago. Swap "must be" 2x RAM was a suggestion at one time, but hardly required, and perhaps not even universally agreed upon best practice.

Re:Write speed

[afabbro]

I would have to disagree. My company uses a RAID-5 for DB's and a RAID-1 for logs

The DBA in me had a heart attack until I realized you were talking about logfiles and not redo logs for DBs ;-)

Re:Write speed

[zx-15]

Shouldn't 1Gbps saturate the drives. 1Gbps = 125MBps, I doubt that RAID 5 array could perform that fast. Maybe SMB protocol is the real bottleneck. I've tried copying files from an NFS drive over a gigabit and I got maximum speed of about 22MBps which should leave the link not all that busy.

Re:Write speed

[afidel]

For a home user a first level may be all they need, in fact it's probably more than they have today....

Re:Write speed

[afidel]

That's why the quotes, but I have things like Oracle that will copy everything the do at initial load into swap, so I need at least 1x ram, and 1.5x ram being 80% used isn't out of the ordinary. Since the result of running out of swap space is the server falling over I go conservative and go with 2x ram.

Re:Write speed

[nxtw]

I suppose 5400RPM is slow in terms of transfer and seek time, and being a software RAID5 set managed in software via mdadm likely also reduces the speed of the array.

Actually, it should be faster than the speed of a single drive.

On my five drive RAID5 running on CentOS 5.2 with the default kernel, no tuning, and the ext3 file system with default mount options, I'm getting write speeds of 109 mbytes/sec. It is possible to get faster speeds; there are paramaters and configuration choices that affect performance, and I haven't tuned them since installing this operating system. When things are set up correctly, the maximum read speed should be close to 250 mbytes/sec.

Re:Write speed

[Bandman]

As long as there's some recent copy of the data that doesn't rely on the controller not puking all over it's drives

Re:Write speed

[Barny]

No you don't, try grabbing CoreCodec.

I can't say it will definitely play 1080p on a 2.7 p4, but its the fastest h.264 decoder out there :)

Re:Write speed

[glitch23]

For what I'll use them for? Not very. Looks like they've got great stats for bulk storage, and any more demanding segments I can stripe and/or cache anyway (with memory prices where they are, it's not like you hit swap anymore).

You'd think so. Using Windows XP though it will use swap whether you like it or not. I haven't yet tried to configure my system with no swap at all (to force it to not be used) but with 2GB of RAM I never use it all up but yet my swap file is still used for some reason by Windows.

Re:Write speed

[toddestan]

I just ordered three of them, and truthfully the throughput is not going to be too terribly important. I'm using them as a mass storage of media - music, videos, pictures, and the like. Two of them are going into the PC in RAID1, and the third is going into a USB/Firewire/eSata enclosure as the back up. Throughput is only going to matter in the first few days, as I anticipate the time needed to copy everything to the new drives, then to copy everything *again* to the external is going to take many hours, possibly even a day. But after that the demands on them are going to be pretty light, as most media doesn't require much throughput. I'm still planning on leaving the old 300GB drive in place as the OS/Programs/extra free space drive.

Re:Write speed

[speedingant]

Personally, I'd rather put the $14 you have to pay for it into another CPU. ; )

Re:Write speed

[Ilgaz]

1080p videos and the sound accompanying them (generally 5.1) are gigantic. Soon the uncompressed sound will be an option too as they race with BluRay.

Even the trailers at Quicktime which are massively 2 pass H264 encoded are huge, about 120 MB each.

Of course I wouldn't buy a noisy, unreliable and slow 1.5 TB drive. 5 drives as single volume (for non critical data) using OS X'es own softraid functionality (or ZFS if they implement) makes more sense to me.

Re:Write speed

[FuturePastNow]

Indeed. I have a 2.8GHz P4 with 2GB of RAM, and an old Geforce 6600GT card. It plays 720p fine, but 1080p is out of the question.

Perhaps a more modern video card, that decodes xyz codec in hardware, would solve that problem. But it would have to be an AGP card and, frankly, a PC with a P4 isn't worth the investment.

Re:Write speed

[flappinbooger]

Video editing. It's all about video.

Media creation is basically THE giant black hole of computer technology. That is, no matter how fast the CPU, how large the HD, or how fast any one particular part is, content creation such as video editing will always be able to use something faster or bigger. Perhaps some day when I can render an entire 1 hour video file in a second, I'll be happy.

Re:Write speed

[Walpurgiss]

My usual speed via FTP transferring onto the RAID array is between 45 and 55MBps. Not saturating the link as you said. So I was wrong above saying it was a bottleneck. But its still fast enough for any and all i/o I use the array for. Definitely faster than a 100Mbps link though, which is fast enough to stream most non-raw video streams.

Re:Write speed

[Barny]

Your prerogative, $14 doesn't buy you the "next model up" cpu, hell it doesn't even cover the cost of good thermal paste :/

It will get you a h.264 decoder that can run on damn near anything (yes, projects to get it under linux are available) and, unlike most of the free decoders out there, it supports SMP.

Re:Write speed

[speedingant]

You're looking way too far into what I said. I did not mean to be prerogative, I was just stating facts. $14 is better put towards a new CPU. If his MB will take one that is. A low range C2D will run you in the range of $50 on a deal, not really out of reach? A new CPU would not only give power to play high definition media, but it will make his computing experience more satisfactory across the board. It helps to be open minded about such things.

Re:Write speed

[Anonymous Coward]

I meant provocative. Spellcheckers eh? And just a note, 1080p media doesn't just come in H.264.

Re:Write speed

[Barny]

Fair enough, but every bit of high def content I have ever DLed has always been matroska wrapped h.264 + ogg.

As for that decoder, as an idea how much better it plays 1080p, I can watch it fine on my 3600+ amd X2 cpu, it sits at around 60% on each core, on my 3GHz quad it barely hits 2% :)

Re:Write speed

[Barny]

I used prerogative as in its meaning of preference, nothing against you, yours is a valid opinion, its just one that I don't agree with :)

Shit, is this a reasonable argument on teh intrawebz?

Uh, OH YEAH AND YOUR MOM!!!!1!!!one

Re:Write speed

[YourExperiment]

There are two criteria for a media centre PC: lots of storage space and small size. Oh, and minimal heat production, so loud fans aren't required.

And a fanatical devotion to the Pope. There are four criteria for a media centre PC!

Re:Write speed

[Kjella]

There are two criteria for a media centre PC: lots of storage space and small size. Oh, and minimal heat production, so loud fans aren't required.

All of this points to a single, high capacity disk as the optimal solution.

Optimal would be:
1. Extremely low-wattage processor
2. Low-power integrated graphics chipset w/1080p H.264/VC-1/MPEG2 hardware decoding
3. SSD/CF/flash chip boot + cable/wifi to file server of cheap TB disks far, far, away
4. Bit streaming/multichannel LPCM over HDMI
5. Wrapped in a fanless small form factor

1. Means atom-class processor at 2-3W
2. Means something like an upgraded Poulsbo at 2-3W
3. A cheap small one without the stuttering problem at ~1W
4. LPCM is here but not everywhere, bitstreaming not yet but minor detail
5. Follows from the above

If they just get their act together it should be possible to produce an entirely fanless fully HD capable HTPC on a 15W budget, maybe even 10W but that's pushing it. You might have to wait until 2010 or so for that one though.

Re:Write speed

[ATMD]

And wouldn't it be nice if we could all afford that?

I don't have the money for a big fileserver with RAID or whatever, or indeed for most of those other cutting-edge things you mentioned. Provided you factor in cost, what I suggested is indeed the optimal solution for ~90% of people.

Re:Write speed

[Laurence0]

In that case, wouldn't the best solution be to make a 20 gig RAM drive and use that as swap? Granted, the OS should be smart enough to do that itself, but you could work around it.

Re:Write speed

[LWATCDR]

I wasn't thinking about home use.

Re:Write speed

[BLKMGK]

XBMC, the Linux version at least, will want closer to a 3Ghz C2D unless your idea of HD is low bitrate 720P stuff. The Killer Sample clip that everyone tests with is pretty insane but until I bumped to 3GHZ I saw skipping in BD rips I'd encoded too. The CABAC patch that they have used on FFMPEG really helps!

One of these days GPU acceleration will be available but for now brute force CPU is required. A P4 doesn't have a chance I'm afraid - not with XBMC anyway. I think on Winders they aren't using video acceleration either so that won't make a big diff either.

anyway, on a 100meg connection I stream this stuff no problem. SATA is 100% NOT required. In fact my dual tuner HDHomerun only has a 100meg connection and can stream TWO HD streams...

Re:Write speed

[BLKMGK]

CABAC patched FFMPEG does SMP VERY well in my experience. Unfortunately it won't use more than two cores right now.

Decent front-ends like XBMC won't utilize non-free CODECS so even if you get it installed you'll still have to find a good front-end to use it.

Re:Write speed

[psergiu]

http://www.popcornhour.com/onlinestore/

low wattage CPU - check
1080p hardware decoding - check
flash, network & optional HDD - check
LPCM - don't know
power - 36W

Not quite there but close. The only downsides are the remote with a bajilion buttons and a dead-brain UI

Re:Write speed

[nerdbert]

Not really. I work on these things. Drive manufacturers maintain many tracks of spare blocks. When a TA or other error occurs we use an escalating set of recovery techniques -- I've seen as many as 20 strategies applied to a drive. If the recovery was successful, and generally it is given the RS and other ECC used, the sector is remapped to one of the spare tracks and the bad sector is marked as not to be used.

And this is why I tell folks that if they start to see read errors for any reason, it's time to buy a new drive since yours is not long for this world.

Re:Write speed

[BLKMGK]

It might also be nice if something like oh say FFMPEG supported that H.264 decoding chipset too huh? So far it seems that high powered CPU decoding of HD is the order of the day if you're running free software like XBMC. It would also be most awesome if ALSA supported audio out over HDMI on a wide range of modern chipsets too. Some of this stuff is coming for sure but it looks to be a VERY long time before I'm going to have something powered by something as lowly as an Atom in my HTPC....

Re:Write speed

[BLKMGK]

If you have the money for a cheap motherboard, a celeron CPU, a gig of memory, a decent case, a decent P/S, and a pile of HDs laying around you *DO* have the money to build as big a NAS as you have drives for.

Look at unRAID from Lime Technology. It protects your data pretty well, spins down drives not in use, can be built on cheap hardware, and the software for it is pretty cheap too - Linux based. I have two of these beasts running and umm "many" TBs worth of disk space. When I built my first one the msot expensive part of it was the silly case! They even have a free version that will run 2 data drives and a parity drive that you can test with. When I run out of space I buy a drive - they are supporting up to 16disks total!

Sorry, but this is way easier than you make it out to be and not so cutting edge that you couldn't do it if you decided to try. Hell you've probably got half the hardware already laying around if you're anything like the rest of us.

Personally I want my HTPC running the smallest quietest drive possible. I currently use a small 320Gig laptop SATA drive I took out of a WD enclosure. I am considering building a small SSD instead just to knock the noise down further and allow me to shut off a noisy fan. I could boot it from a CD and store data files to a USB maybe but an SSD is more flexible I think. Big drives in a small HTPC case make no sense IMO.

Servers do not belong in the living room and 1.5TB isn't enough anyway - all of my noisy stuff is far far away and I try to minimize my HTPC machine....

Re:Write speed

[BLKMGK]

Define massive - my BD rips from disk prior to compression top out at around 40Gigs each with DTS sound. The sound track is normally about a half gig and with an uncompressed soundtrack I might see MAYBE 1.5Gigs. So, figure 41gigs or so per movie prior to compressing them. When compressed figure about HALF that even with really high bitrates - I've done this about 60 times now so I think those numbers are accurate. Bear in mind some BD rips are also UNDER 15Gigs prior to compression too

So, why wouldn't a 1.5TB drive work? I use WD green 1TB drives now in an unRAID system and they are GREAT! Hell I don't even use Gig ethernet as I bought the wrong cable and the run is too long. 100meg has ZERO issues handling HD streams with 5.1 surround on my setup! I have NEVER had an issue with throughput nor have I come close to maxxing out my read speeds while streaming even from my IDE drives - not even EIDE.

You won't need the speed gotten from reading from parallel\striped disks in order to watch HD movies. unRAID doesn't work that way and no one I know has issues watching HD video - even multiple streams from the same drive at the same time. You WAY overestimate what's required and a RAID5 system is unnecessary, uses too much power, and IMO risks the data too much over something simpler like an unRAID that uses a single parity drive and standard F/S formatting.

Re:Write speed

[Kjella]

Not quite there but close. The only downsides are the remote with a bajilion buttons and a dead-brain UI

Downsides:
Must hook HDMI directly to TV, not even image pass-through works (PC-Amp-TV works, Popcorn-Amp-TV doesn't)
Very annoying to get to the files on a shared drive, particularly the password prompt
Doesn't play HDDVD/BD-rips
Can't double as a PC (maybe some consider that a plus, for me it's a minus)

The HDMI trouble is the annoying one, means I have to unhook the PC, rehook the Popcorn, use a separate audio cable that doesn't do 7.1 plus navigating the lame UI with the controller. Give me a real PC where I can have a wireless mouse to click my way and I'd probably use it anyway, but for me it's annoying enough I usually just fire it up on the rather noisy PC. It's good, but could have been better.

Re:Write speed

[Penguinoflight]

It's a moot point anyway. New computers will have eSATA connections built in (my latest motherboard did, and I didn't buy it for that). Sure, most enclosures aren't there yet, but they will be once the need arrives with faster general purpose drives. Since even the fastest most expensive drives now can barely keep up with sata1, sata2 should be ok for some time.

Re:Write speed

[ShadowRangerRIT]

60 MB/s is usually for sequential reads, not random (which is what you'd be looking at in the situation I described if you're under RAM pressure which limits your disk cache). It's only a problem if the speed really is *slow*, which the summary seemed to imply, even though on reading the article it appears to be quite competitive. As for the six hard drives, that's for gaming and data backup, not for video.

Re:Write speed

[speedingant]

Fair enough. That codec sounds good anyway ; ) I'd like to think that Slashdot is a better place to argue that say, digg...

Re:Write speed

[speedingant]

I use Plex (a side project from XBMC) on the Mac. It works swimmingly on a headless MacBook I repaired ($100 baby!). It's a C2D with a GMA950 and it plays HD content without skipping from 720p right through to 30-40GB HD 1080p. It just plays through an Unraid box through ethernet. You could easily stream HD through 802.11.N, through my experience. Of course it does depend on the codec... H.264 is extremely efficient in my experience, as XBMC uses a codec that makes use of two CPUs instead of one.

Re:Write speed

[Barny]

Its supported (and comes with) the haali media splitter, so long as your player of choice uses that, its golden.

Re:Write speed

[BLKMGK]

Nope, XBMC won't work with it I'm pretty sure. Asked about and on the forums and the answer was negative if memory serves. Their player is custom and uses FFMPEG, no Haali...

Re:Write speed

[Barny]

Bleh, that sucks, was building a media machine today for a customer in a thermaltake mozart 102 with 7" lcd, it uses its own custome playback software which instantly picked up that (my pre installer for vista premium loads some useful extras) haali was installed and asked if I wanted to use it for all playback jobs... winz.

Some of the big name makers are starting to get their shit together, this thing, once fully loaded, was a piece of cake to drive.

Re:Write speed

[Barny]

Btw, have 2 1.5 TB drives on their way to my windows home server box, the data density is just great.

Re:Write speed

[BLKMGK]

Take a look at unRAID by Lime Technology. Maxxes out at 15 data drives, 1 parity. Lots of advantages too numerous to list here to include spindown of drives not in use. Plenty fast enough for streaming multiple HD streams and runs on low end hardware easily with good data protection. I'll take it over standard RAID for home use any day. My two servers together are over 7TB and growing :-)

Re:Write speed

[Anonymous Coward]

I've got a drive or two that are large but slow. I use Linux only, which I'm sure helps, but maxing out at even 40MB/sec over USB2 really doesn't make disk usage slow (one of my older 250GB drives tops at 47MB/sec read speed over IDE, and it's also pretty decent.) (My other drives get like 75-80). If I have to sacrifice speed or size to save money, I'll drop speed first for a disk, even a slow disk is decent.

Re:Write speed

[Barny]

Home server doesn't use raid, it clusters drives into a storage array, making sure to keep data (you flag it if you want it to do it per share) on at least 2 drives at once.

Its also a great backup system, can act as a streaming server as well as web server.

And it sits in the corner hoovering dust through its filters and I don't have to think about it till it warns me I am running low on space, then I yank out a drive and throw a new one in :)

Re:Write speed

[BLKMGK]

So to protect data it must be flagged and the protection consists of making a second copy? That's pretty inefficient and a PITA. unRAID would do it better although depending on what you are streaming to you might have to do a little more work. To each his own but I think you'd find that there are better solutions out there....

Re:Write speed

[Ost99]

3. SSD/CF/flash chip boot + cable/wifi to file server of cheap TB disks far, far, away

Just boot over net from the server. No SSD/CF/flash needed.

Re:Write speed

[Ost99]

There are disks with better URE than 1 in 10^14.
You get rather cheap 7200rpm disks with 1 in 10^15 and fast SCSI and SAS drives tend to have URE of 1 in 10^16.

When will hardware manufacturers get a clue?

[morgan_greywolf]

We know most benchmarks are useless and a lie. What matters is not how well your product performs on a benchmark, but how well it performs for real-word usage. If you're selling a desktop hard drive, well, damnit, benchmark it for typical usage scenarios and then look at your results. If you they suck, your product sucks. Go back to the drawing board. Pulling the wool over everyone's eyes with synthetic benchmarks tailor-made to make your product look good is just downright dishonest. And I hate doing business with dishonest people more than most things.

Re:When will hardware manufacturers get a clue?

[CaptainPatent]

Why would they ever go back to the drawing board with this drive?

This is a Consumer-level SATA drive. IF someone even gets by the size of the drive the next thing they'll see is the price. At ~180 bucks it's honestly a steal!

Even the most hardcore of gamers / power users would see that price point and say RAID Array, here I come!

Re:When will hardware manufacturers get a clue?

[Endo13]

Even the most hardcore of gamers / power users would see that price point and say RAID Array, here I come!

Not me. I hate RAID even more than morgan hates dishonest people.

Re:When will hardware manufacturers get a clue?

[Gat0r30y]

Hardware manufacturers generally don't get to decide what 3rd party benchmarks get run on their drives. At least not from my experience. Unfortunately, performance on these benchmarks is what ultimately determines what we can charge vs. our competitors through OEM's (most of the distribution channel). So while the benchmarks may not be particularly indicative of real world performance, HDD's are tested against them because Dell, HP, Apple and the like use the benchmarks.

Re:When will hardware manufacturers get a clue?

[Anonymous Coward]

RAIDz FTW!

Re:When will hardware manufacturers get a clue?

[Anonymous Coward]

Are you a roach? What are you smoking that RAID seems like a bad thing? RAID 1 on a gaming system is great, real short load times with great performance.

Re:When will hardware manufacturers get a clue?

[Endo13]

I seriously hope you just posted that for the lulz.

Re:When will hardware manufacturers get a clue?

[morgan_greywolf]

Yeah, RAID1 isn't much of a performance boost. In fact, in most cases it causes a performance HIT. RAID5, OTOH is really a good balance between performance and redundancy for most things.

An In-Depth Look At Seagate's 1.5TB Barracuda:

[Warll]

I didn't RTFA but I can guess it:

0|1|1|1|0|1|0|0|0|1|1|0|1|0|0|1|0|1|1|0|1|1|1|0|0|1|1|1|1|0|0|1|0|0|1|0|0|0|0|0

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[speroni]

I don't get it, and it wasn't conducive to just go google it...?

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[Anonymous Coward]

Who you calling tiny?

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[egomaniac]

I'm guessing it's a reference to "an in-depth look". That's supposed to be the binary stored on the platters.

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[Warll]

The joke is that "an In-Depth Look At Seagate's 1.5TB Barracuda" would really jsut be some really tiny bits, the binary example I used was just "Tiny eh" in ASCII binary. So I take it the jokes no longer funny since I had to explain it?

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[speroni]

Ahhh....

I was hoping it was a futurama reference.

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[Anonymous Coward]

It only matters when that's what s/he said!

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[noidentity]

That's because they use perpendicular recoding technology. Other drives with classic, lower-density recording look like this:

<>|--|--|--|<>|--|<>|<>|<>|--|--|<>|--|<>|<>|--|<>|--|--|<>|--|--|--|<>|<>|--|--|--|--|<>|<>|--|<>|<>|--|<>|<>|<>|<>|<>

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[tylerni7]

You actually only wrote "tiny " unless I am missing something...

/me kills the joke even more

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[Warll]

Hhhhmm, guess the eh got cut off.

Re:An In-Depth Look At Seagate's 1.5TB Barracuda:

[gardyloo]

Hhhhmm, guess the eh got cut off.

Lorena? Is that you?

Capacity is hardly news anymore

[iminplaya]

How reliable is the thing? I never had very good luck with Seagate.

Re:Capacity is hardly news anymore

[MightyYar]

How reliable is the thing?

Buy me one and I can promise status updates.

Re:Capacity is hardly news anymore

[Firemouth]

I've always had good luck with these drives. It's the only brand I'll buy and recommend to another person. The fact they will warranty their drives for 5 years where most others will only do 1 - 3 years says something about them. If they're betting their drives will last 5 years, who am i to argue?

Re:Capacity is hardly news anymore

[MBGMorden]

I've had the same experience - Seagate has consistently outlasted every the drive brand I've seen. Based on past experiences, I'd rank them, from least reliable to most, as:

Hitachi
Western Digital
Maxtor
Samsung
Seagate

Drive brands not listed I either have no experience with or not enough to form an opinion.

Re:Capacity is hardly news anymore

[Mashiki]

Been a long time since I was in the business as a reseller, but we used to have more WDC failures then Seagate. But we'd get cases of both that had 20-30% of drives that were sealed from the factory, that were either DOA or had cascading bad sectors. But that was back in the days of absolute crap when everyone was in the size race.

Things change in 10 years, I do like the current brand of Samsung drives.

Re:Capacity is hardly news anymore

[Dan667]

and Seagates are quieter than almost all others

Re:Capacity is hardly news anymore

[Cynonamous Anoward]

Ditto on this...Had really bad experiences with Hitachi and Western Digital. Swore by Maxtor for a while in the early-90's, and then a got several in a row that died within 2 years. Never used a samsung, but I've been sticking with seagate for about 15 years now, and they are incredibly reliable. testament to reliability: with only a little care and maintenance, I have now gone a whopping 12(!!!!) years with out losing a single byte of important data. The only problem for me actually is size...in 12 years without information loss, you really do accumulate a massive amount of data. Even with regularly cleaning out unneeded data and archiving stuff I don't need instant access to, I've managed to fill 2/3rds of my 1TB storage drive already.

Re:Capacity is hardly news anymore

[X0563511]

Now, is the barracuda line, or there other drives as well?

Re:Capacity is hardly news anymore

[Anonymous Coward]

I've managed to fill 2/3rds of my 1TB storage drive already.

Your wrist must be tired!

Re:Capacity is hardly news anymore

[mc900ftjesus]

I only use Seagate. The only company offering consumer hard drives with a 5 year warranty.

They put their money where their mouth is, and their RMA process is easy and fast. You'd really be crazy to buy from anyone else unless you replace your drives every year.

Re:Capacity is hardly news anymore

[thepotoo]

I've managed to fill 2/3rds of my 1TB storage drive already.

640 GB of porn ought to be enough for anybody!

Seriously, my experience has been remarkably similar to yours, except without the little care and maintenance bit. I once dropped one of these down the stairs at a LAN party. Worked just fine afterward. I've also had good luck with Maxtor, however one of these died after only 10 years of use (in an incident perhaps relating to a Coke set on it), so I generally buy Seagate.

Re:Capacity is hardly news anymore

[AllynM]

I'd venture a guess that WD gets a bad rap from a bad start into the 'RPM race'. I built an array when the 400GB drives came out. I had even sprung for the RE2 server class drives. Over the course of 2 years I had to RMA 5 out of 6 of them. The problem was they ran extremely hot even with above average cooling. Their lineup around that period must have been built on a bad thermal design.

I've since switched to the new GP (Green Power) drives and not had a single problem in a full year of use, using 10 of them. They run *significantly* cooler (hottest drive is 100F and that is with weak airflow) and appear to be much more reliable as a result of it.

Also on WD: I've raid-0 striped a pair of nearly every generation of Raptor they made and never had so much as a hiccup.

So yeah, they had a rough spot, but I'd say they would be higher on the list if you go on their current product lines.

Re:Capacity is hardly news anymore

[idiotnot]

They're certainly quieter than WDs, but nowhere near as quiet as the Samsungs I have.

Re:Capacity is hardly news anymore

[KenjiFinster]

Try to tell that to anyone with a MacBook. I see at least 5 dead MacBooks with Seagate drives every day at work.

Re:Capacity is hardly news anymore

[Pentium100]

I have a very old Seagate drive (well, it says Seagate ST41200N on the top, but windows recognizes it as Imprimis 94601-15). It is a 1.2GB (991MB) 5.25" full height drive and it works perfectly. I have another one, a bit younger (ST34520N) ~4GB, it also works very well. All the new ones also work well, so when I buy a hard drive, I buy Seagate.

I wonder why nobody is making 5.25" hard drives anymore... With current technology they could have at least 10TB capacity...

Re:Capacity is hardly news anymore

[MadTinfoilHatter]

Seagates are quieter than almost all others

That may have been true 6-7 years ago, but after the legendary Barracuda IV and (even better - if very short-lived) V-series, Seagate has gone badly downhill when it comes to silence. If you check out www.silentpcreview.com you'll see that Seagate hasn't been on top for years - Samsung has, and Western Digital has also been a good choice for the last two years, or so. My own experience also supports this assessment.

Re:Capacity is hardly news anymore

[Yewbert]

Another pretty-close-to-ditto. I've had Seagates die, WDs die, and I've had (more) Maxtors die, BUT, it's mostly because I mistreat the drives badly, running them bare, sitting upended on my desktop, outside the computer they're connected into, sometimes with and sometimes without additional cooling. They get hot, they get ESD, they get dusty,... and they occasionally die. The big difference (since I have all data backed up on DVD-R anyway) is that the Seagates are still under warranty and I get the hardware replaced for the cost of shipping. The Maxtors are just paperweights.

(At work, I have some responsibility over ~300 workstations, and "dead Maxtor drive" has become a running joke. We can't wait till we've replaced all of them with Seagates or Samsungs.)

Incidentally, having just acquired my first-ever Samsung AND Hitachi drives (other than the original Hitachis in ThinkPad laptops,...), and having just started restoring data from numerous smaller drives that have been sitting offline, powered down, in a cabinet for up to three years or so, I'm finding very weird *read* problems from one particular Maxtor drive, i.e., I copy a bunch of files off it onto the new drive, a Hitachi 1TB, verify the data set, find about 1% corrupt files that don't match their MD5, recopy the same files from the Maxtor again, and *most* of the recopied files are fine. What the hell? File copying from other source drives worked A-100%-OK-spiffy, just that one bad source.

Re:Capacity is hardly news anymore

[noc007]

This is roughly my experience as well. The Hitachi (formerly branded as IBM) DeskStars we've called DeathStars for quite some time for good reason. Each and every Western Digital HDD that I've had under my roof has died on me within a year or two. Quantum was the worst on my list and I was quite concerned about Maxtor drives when they bought out Quantum. So far my first/current Maxtor drives are still running after a year.

I swear by Seagate and love having drives that are warrantied for five years. I've only had one die on me and that was due to over-voltage from the USB external HDD enclosure.

Re:Capacity is hardly news anymore

[Urza9814]

And based on my experiences I'd put Maxtor at the top (Well, the old Quantum at the top, but since I don't believe they make consumer drives anymore...), followed by Seagate, with WD in dead last. Of all the hard drives I've owned, the only ones that have failed were WDs. Two of 'em. Neither lasted more than a year.

Re:Capacity is hardly news anymore

[dhanson865]

You think Seagate is the only company to offer consumer hard drives with a 5 year warranty?

http://www.wdc.com/en/products/products.asp?driveid=488&language=en

It's not that hard to find 5 year warranty mentioned on the WDC website.

Black = WD6401AALS 5 year warranty
Blue = WD6400AAKS 3 year warranty
Green = WD6400AACS 3 year warranty

And to add to the fun the Black has twice the cache and is only about $10 more than the Blue at 640GB.

Seagate is by no means a bad company but they aren't the only game in town.

Re:Capacity is hardly news anymore

[sheldon]

Interesting.

I've only ever had bad experience with Maxtor drives. Can't say I've used Samsung, and the only Hitachi I ever had was a 7200 rpm drive in my notebook.

Western Digital and Seagate I'd rank pretty even. The only caveat is never buy the first generation of any Seagate technology. So say a 1.5TB drive... wait until the second generation models come out and they got the bugs worked out.

Re:Capacity is hardly news anymore

[nabsltd]

Of course, everyone has their own drive horror stories, and there are many people who swear by a brand that others swear at.

Overall, I've had every brand die in every stage of their lifetimes, and I've found that I've RMA'd far more Seagate drives than any other brand. It's not that they are any worse, it's just that with the 5-year warranty, they are far more likely to still be in warranty.

So, I tend to buy the drive that best fits my needs and has a 5-year warranty. I've got Maxtor, Western Digital, and Seagate at this point in multiple arrays with a total of about 15TB.

The only things I have learned for sure is that I'll only use RAID-1, true hardware RAID-5 or 6 (no Intel ICHx "RAID-5"), and Linux software RAID-5 or 6...anything else is too dangerous for recovery.

Re:Capacity is hardly news anymore

[nabsltd]

I only use Seagate. The only company offering consumer hard drives with a 5 year warranty.

I'm not sure what you mean by "consumer hard drives", but both Western Digital and Maxtor (which admittedly is now part of Seagate) offer 5-year warranties on at least most of their drives.

One thing to watch out for is that drives sold in external enclosures typically have only a one year warranty, even though they use the same drive that when purchased bare has a 5-year.

Re:Capacity is hardly news anymore

[bigstrat2003]

Seagates are solid, but I'd have to say you've had bad luck with WD. I've had one drive fail on me in 5 years of using their drives (for work and personal use, not just personal), I swear by 'em. The only manufacturer I don't trust at all is Maxtor. 90% of all dead drives I've ever seen have been Maxtor.

Re:Capacity is hardly news anymore

[dgatwood]

Agreed. My MacBook Seagate drive died (click of death) at a mere 9 months of age under relatively light use. The same week, I had two Seagate desktop drives bite it (one head crash, one bizarre constant data corruption problem with no read errors reported). Hard drives suck, period. Statistically, with the exception of the drives I bought week before last, I think that every hard drive I've purchased in the past couple of years has already failed.

I had a WD die a couple of weeks ago. A couple of years ago, I lost a Hitachi drive, and I have a second Hitachi (well, IBM, I think) drive in semi-permanent storage that I cloned off onto another drive out of paranoia after it failed to spin up at about the same time. (Powered it off and back on again and it spun up, so it was probably just poor bearing lubrication distribution, but....) You get the picture. Oh, and two WD drives, one of which also went bad, the other of which I pulled because it sounded like a chainsaw at about the same time. I'd estimate that in spite of proper cooling, clean, UPS-backed power, etc. I'm finding a failure rate for heavy use drives that approaches 100% in the first three years. I'm even seeing spotty failures in occasional use drives like my MacBook.

At this point, I'm convinced that the craptastic manufacturing in third world countries has driven the quality of hard drives down to the point that they no longer adequately serve their intended function for me, and I'm sure this 1.5 GB Barracuda will be no exception. Just keep at least three copies of anything you care about and you should be fine. For now. At the rate of reliability decline I'm seeing, though, the rule will likely be five copies within ten years. Bring on the cheap flash drives. The Winchester disk design has outlived its usefulness, IMHO.

Re:Capacity is hardly news anymore

[networkBoy]

I wonder why nobody is making 5.25" hard drives anymore... With current technology they could have at least 10TB capacity...

Two words:
Angular Momentum

At the outside of the disk there would be an incredible amount of stress on the rotating media.
The head seek times would go up as well....
Though, while 7200+ RPM would certainly be out, and likely 5400 RPM as well (remember the old drives ran <= 3600RPM, I would consider a 4200 RPM 10 TB drive for near-line storage...
even 5.25/FH that would be a decent volumetric density (equivelent to 5x 3.5" drives).
-nB

Re:Capacity is hardly news anymore

[Pentium100]

One of my old drives runs at 3600RPM, and there were others which run at 5400RPM (I think the newest version of Quantum BigFoot was 5400RPM). It seems that 5400RPM is the max, but at 5400RPM the edge of a 5.25" platter is traveling faster than the edge of a 3.5" platter, rotating at 7200RPM.

I don't really need a fast high capacity drive (I can use 36GB 15kRPM drive for the OS and a big slow one for everything else), so I could live with 40ms seek times and, say, 40MB/s linear read if that meant having a cheap high capacity drive.

Re:Capacity is hardly news anymore

[Anonymous Coward]

Are they the same 5 dead MacBooks that you see everyday, or are they different ones?

That makes a difference....

Re:Capacity is hardly news anymore

[jcrousedotcom]

I remember when WD was the only drive to buy - they had lifetime warranties and lasted forever. I ran into several Maxtor drives in a row that failed inside a two year window but I've got two 60Gb drives that have been running 4 or more years (I don't usually turn my home system off).

What I am saying, I guess, is drive quality is a moving target. What used to be a fantastic drive now isn't quite as reliable (WD) and drives that used to be junk (Maxtor) have improved quite a bit.

I remember when IBM quit using their own branded drives and switched to Maxtor and I thought is was the worst thing in the world. Now I wouldn't view it so negatively.

What drive is the best depends on when it was manufactured.

Re:Capacity is hardly news anymore

[pyite]

The only things I have learned for sure is that I'll only use RAID-1, true hardware RAID-5 or 6 (no Intel ICHx "RAID-5"), and Linux software RAID-5 or 6...anything else is too dangerous for recovery.

I think hardware RAID is more dangerous than software RAID for recovery. You can always install the identical Linux kernel and mdtools you had running to recover a Linux software raid array. Or you can always install the same version of Solaris to recover a zfs array. What you can't always do is purchase the same exact card needed to recover an array if it's the card that fails.

Unless you have a managed storage solution (e.g. from EMC) where you can guarantee you can get replacement parts, I would have no warm and fuzzy feelings from hardware based RAID. ZFS all the way.

Re:Capacity is hardly news anymore

[CrazedWalrus]

I don't really know, but I'd venture a guess at bad seek times caused by longer head travel.

Re:Capacity is hardly news anymore

[nabsltd]

I have several different LSI hardware cards, and I don't need the exact same card as a replacement.

As long as I can plug the drives in (and since they are all SATA/SAS controller cards, I can), the array shows up. When you connect the drives to another card, it will complain that the drives don't match the configuration in the NVRAM, but you can just tell it to copy the drive config into NVRAM and use it.

I don't know how other controller manufacturers build their cards, but LSI seems to do the right thing with this.

Re:Capacity is hardly news anymore

[toddestan]

In my mind WD's rap goes back years - their 100MB-1GB drives were no more reliable than their more recent stuff. It actually amazes me that they manage to stay in business this long, though admittedly their Raptor line is quite good and does not seem to suffer from the same problems as their regular stuff.

It doesn't surprise me that you found their GP drives run cool. They're basically 5400RPM drives - WD couldn't manage to make a 7200RPM 1TB drive that was considered reliable even by their standards, so they let marketing rebrand them as low-power "Green" drives instead.

Re:Capacity is hardly news anymore

[Anonymous Coward]

samsung no good

Re:Capacity is hardly news anymore

[Ost99]

There might be something wrong with your PSU(s) if you're getting 100% failure rate.

Even more impressive...

[ratnerstar]

Did you hear about Seagate's new Sarahcuda drive? Not only does it also have 1.5 terabytes of capacity, but it scrambles your data so as it make it completely incomprehensible. Plus you get a free one if you vote for John McCain

Re:Even more impressive...

[Anonymous Coward]

is it loud, obnoxious and inflammatory?

Re:Even more impressive...

[Killer Orca]

Hey if it's as hot then sign me up!

Re:Even more impressive...

[Yewbert]

... I was waiting for that.

I hear, though, that even when it's filled to full 1.5TB capacity, it contains only 250GB of actual information, what with all the repeating itself and random $maverick insertions, you betcha.

Re:Even more impressive...

[ktappe]

Did you hear about Seagate's new Sarahcuda drive?

If you stand on top of it you can see Russia?

Re:Even more impressive...

[Archangel Michael]

That's NOTHING, There's the new Barackuda model that promises more than 1.5 trillion in new storage, and will give poor people free money, won't raise taxes on anyone(except Joe's Plumbing service, tax cheater extraordinaire).

After that is the Barracuda. In other news, Ralph Nader is pissed.

Re:Even more impressive...

[up2ng]

Yeah but files either grow unexpectedly for 9 months after writing them or look a little...off

Re:Even more impressive...

[HTH NE1]

Did you hear about Seagate's new Sarahcuda drive? Not only does it also have 1.5 terabytes of capacity, but it scrambles your data so as it make it completely incomprehensible. Plus you get a free one if you vote for John McCain

Would that qualify as a DILF (Drive I'd Like to Fsck)?

hmm

[poetmatt]

I seem to recall someone saying many times over that this was not the first 1.5TB, but that it's claimed anyway (with more specifics, like "first consumer") etc.

Beyond that, insert 1.5TB ought to be good enough for anyone, and will it blend jokes here.

Don't always need fast write

[Anonymous Coward]

Slower write speeds are fine for some projects. A 1.5 GB on which to keep all my FLAC files would be nice and slow write wouldn't deter me. I'm more concerned about heat and reliability.

True Terabyte?

[Anonymous Coward]

So this thing will have a true terabyte of 1024+ Gigs?

Re:True Terabyte?

[Warll]

You mean will the stated 1.5TB magical become 1.5TiB just because that's how Windows calculates capacity?

Re:True Terabyte?

[hechacker1]

it's getting even worse by crossing the 1.0TB barrier. Not only are they using the base10 to 2 conversion to cheat you out of space, but they are advertising this drive as 1.5TB when it's actiually 1500TB. That's 36GB missing. Still though his drive was worth it for it's low GB per dollar ratio.

Re:True Terabyte?

[Warll]

I take it by "1500TB" you mean 1500Tb? is that really what they do? Wouldn't the lose be much greater than 36GB?

Re:True Terabyte?

[hechacker1]

I meant 1500GB.

Re:True Terabyte?

[egomaniac]

...but they are advertising this drive as 1.5TB when it's actiually 1500TB. That's 36GB missing.

I am sick of this stupid fucking argument. A 1.5TB drive storing 1,500,000,000,000 bytes is a lot more sensible than a 1.5TB drive storing 1,649,267,441,664 bytes (actually 1.5TiB).

Do you really CARE about the exact number of bytes on the drive? Do you lovingly count each and every one of them? Or do you just care "1.5TB holds 50% more than 1TB, let me buy that one". Since all of the drive manufacturers use the same units and all of the units are consistent with the metric prefixes, why are you complaining?

Oh, it's because your stupid operating system reports your 1.5TB drive as a 1.36TB drive? Let me spell this out for you: all that means is that the operating system is wrong. Seriously, if the entire hard drive industry has accepted that 1TB = 10^12 bytes, why on earth is your operating system persisting in using an obsolete and incorrect definition of the unit? If it suddenly started reporting it as a 1.5TB drive, would that make you happy? The number of bytes on the drive doesn't change either way, of course.

You can't possibly tell me that you really NEED those extra bytes a 1.5TiB drive would have compared to a 1.5TB drive, so really you're just bitching about the OS not reporting the same number that the manufacturer does. And that's the OS's fault for being stupid.

Re:True Terabyte?

[Belial6]

While you sound a little more agitated about the subject than I am, you are pretty much spot on. It is hard to take anyone seriously about the 2.4% "missing" space when they can't even tell you 2.4% of the files on their computer. Sure they can tell you about groups of files, but I know I certainly cannot list off 2.4% of the files on my hard drive, and I doubt there are many other people who can. So, I could just as easily lose more space to worthless, unused files than to the difference in size calculations.

Re:True Terabyte?

[dsanfte]

"Do you really CARE about the exact number of bytes on the drive?"

Yes, I do. In every other domain I can think of, when we say GB or TB we refer to the binary, not the decimal, equivalent (powers of 2). Disk manufacturers took it unilaterally upon themselves to pervert this for their own purposes, mostly to sell smaller drives as larger ones. It was all about money.

There is absolutely NOTHING wrong with expecting 2^40 bytes on your hard drive when it says 1TB on the box. That's the standard in use and it OUGHT to be common sense.

Re:True Terabyte?

[toddestan]

Since all of the drive manufacturers use the same units and all of the units are consistent with the metric prefixes, why are you complaining?

So basically, the entire harddrive industry decides to go against convention in order to make their drives appear bigger than they are really are, and you don't expect people to get annoyed over it? Granted, it's been about 20 years since they made the switch, but still...

Re:True Terabyte?

[Ilgaz]

That is a hopeless situation since even MicroSD and mem stick people are lying. That fake thing really matters more when you buy a 2 GB (!) MicroSD. You end up like having ~1900 MB space. On a smart phone it really matters.

Re:True Terabyte?

[Anonymous Coward]

So, with Tb's they already sell a hdd of size ~9% smaller than one they would well with equivalent TB number.

But understandable, we all think in decimal values, and binary counting for them would be giving out more than what we think they give.

Re:True Terabyte?

[egomaniac]

I don't get how saying that a 1.5TB drive holds 1,500,000,000,000 bytes makes it appear "bigger than it really is". Who, other than a computer geek, wouldn't expect it to hold 1,500,000,000,000 bytes?

Plus you can't exactly say that they're going "against convention" to say that "mega" == 1,000,000, now can you? The metric conventions are far older and better established than the binary prefixes, so if anybody is going against convention it's the memory manufacturers.

Re:True Terabyte?

[egomaniac]

Memory is the only thing that is measured in binary units.
Hard disks, transfer rates... everywhere other than memory is using metric prefixes, as well they should -- the idea that "tera" == 1,099,511,627,776 is asinine. Computer geeks complaining about using tera to mean a trillion is really asinine.

Re:True Terabyte?

[toddestan]

Convention in the computer industry was that 2^10 = kilo, 2^20 = mega, etc. Yes, this didn't agree with the SI prefixes, but at the time there wasn't any SI unit for data storage so by definition "kilobyte" didn't make any sense anyway if you're a truly anal-retentive type anyway.

The thing that pisses people off is that the harddrive manufacturers didn't switch over to be righteous or anything like people make them out to be, they did it to be intentionally misleading as to the storage capacity of their offerings. That's why people are annoyed by it.

Re:True Terabyte?

[Anonymous Coward]

That might be because everything else in the computing world is measured in KB, MB, GB, and TB. I have 4GB of RAM, and 4GB of swap, but they're not the same size.

Also, 139GB is an insignificant amount of space to be losing due to HDD manufacturer counting peculiarities, right? (You are, of course, correct in saying that I don't /need/ them, but it'd be nice not to have to recalculate HDD space into "actual space" all the time.)

Issue a ticket...

[Besjon]

Only 50% in 18mos? They is disrespectin' Moore's authoritay

Re:Issue a ticket...

[Warll]

When the first 1TB came out they were priced at around $400 if I remember right. This one is retailing for $200 or so. Thus its almost beating moore's law. Not that moore's law applied to hard-drives anyway.

Re:fp?

[thepotoo]

I swear, if you were any more dense, you'd fall within your own Schwarzschild radius, become a black hole, and we wouldn't have to deal with you anymore.

Hey, everybody's thinking it, I'm just saying it. ;)

Re:True Tebibyte?

[corsec67]

You mean a Tebibyte, or 1024 Gibibytes?

I wonder . . .

[PeeAitchPee]

. . . if lack of "real-world throughput" might have to do with other parts of the system which haven't yet caught up or been optimized for these huge new drives. E.g., OS, disk controller, etc. Just my .02.

Re:I wonder . . .

[Anonymous Coward]

It's just the insanely poor seek time. Seek time when accurately measured (e.g. not the vendor given one) is usually the highest weighting indicator for real live performance.

But they didn't even do 1T right...

[pseudorand]

Has anyone else noticed that a large number of the Seagate 1T drives fail on you in 30 days. The same is true for samsung and WD. Even with the Hitachis I get 1/5 failed out of the box. I still buy all Hitachis though, because the ones that do work keep working. Why are we moving to 1.5T when the 1T are too buggy to be useful. (BTW, my epxerience is based on buying 100+ drives).

Re:But they didn't even do 1T right...

[iny0urbrain]

I bought 4 1TB Seagate drives within the past month, and one certainly has failed within the first two weeks... I've sent it back for replacement and already got a shipping confirmation. So hey, at least their returns system is running smoothly...

Re:But they didn't even do 1T right...

[Taibhsear]

Check to see if it was from Thailand. Not that I have anything against the country (their food is delicious!) but the manufacturing plant there has been churning out sub-par drives in certain models. Check the newegg reviews on your specific drive.

As far as this drive though, I recently got this exact drive (the 1.5TB). The write and read speeds, though not documented, seem right on par with my other sata drives (one is 300GB Maxtor with 32MB cache, the other is 320GB Seagate with 16MB cache. Both SATA with the limiter jumper removed.) I only use the 1.5TB drive (actual space is about 1.35TB) for media storage, formatted in NTFS but used mainly in Ubuntu 8.04. It, however, was from Thailand so I'm a little worried. I keep all the stuff I've backed up on it on other drives and plan to until a few months have passed the trial. Ran seagate tools and the drive passed all tests.

Re:But they didn't even do 1T right...

[Anonymous Coward]

You seriously just buzz-killed my new Seagate 1T drive.

Re:But they didn't even do 1T right...

[Maltheus]

It seems like they'll either fail in the first three months or not at all. Given that, the 5 year warranty probably doesn't cost Seagate any more than a 1 year one would. I generally stick with the ES drives though and I've only seen one early failure there.

Re:But they didn't even do 1T right...

[pla]

Has anyone else noticed that a large number of the Seagate 1T drives fail on you in 30 days

No, not really... And though not a statistically significant sample size, I currently have four (three different brands) in use, with a single failure that came DOA due to shipping damage.

I have noticed, however, that the 750+GB drives run a good bit hotter than their smaller counterparts, with the 7200RPM models even worse.

Once upon a time, I would merely mount HDDs in such as way as to passively encourage decent air flow, and that did the job. Now, I always mount them with a (slow and quiet) 120mm fan actively moving a decent volume of air directly over them, making sure to never mount them in adjacent bays (ie, leave some room for the air to flow completely around them). That makes all the difference, bringing them back down to a comfy sub-30C.


Of course, I also use to swear by DiamondMaxes, which most people considered garbage... So I have to suspect that paying attention to temperature makes all the difference between my experience and yours.

Re:But they didn't even do 1T right...

[Eil]

It shouldn't be a huge surprise that the higher-density drives have a greater failure rate than those with lower-density platters that have been in manufacture for years.

I don't mind buying Seagate because their 5-year warranty for every drive that leaves their factory, frankly, the best available.

Re:But they didn't even do 1T right...

[AllynM]

I've bought a total of 10 1TB WD GP drives. 6 when they first came out, 2 a month later, and 2 a month ago. None have been DOA, and not so much as a hiccup so far.

Just a data point for ya.

Note: the same cannot be said for the 400GBish 7200 RPM drives of theirs from a few years back. Those things were a head crash waiting to happen.

Re:But they didn't even do 1T right...

[Sleet01]

I don't want to give too many details, but at least one storage company has encountered 1% to 2% drive failure rate from spindle motor burnout *alone* on the Seagate 750GB drives. This is not counting bit rot.

Re:But they didn't even do 1T right...

[greg1104]

If you take a look at the newegg reviews, you'll find 16% of them give the 1TB 7200.11 drive a 1 star review, most of which are because of DOA or D shortly after A. So it's not just you who noticed.

Seagate's Barracuda line had a good run with high reliability for quite a while. If you check the reliability database at storagereview (unfortunately you have to go through some trouble to become a member and see the data), the Barracuda ATA III, IV, and V are ranked near the top--92, 90, and 96th percentile respectively. Then things went way downhile--7200.7 hits 88, the 7200.8 at 49, and the 7200.9 at 43. That matches my own anecdotal experience.

Sometime after the 750GB drives came out reliability took a further dive south. I believe that was caused by switching a large amount of production to a new plant in Thailand (the reliable models came out of Singapore). That seems to be the inevitable way hard drive manufacturing works--whenever some company moves to a new facility, quality dives for a few years afterward. I predict that 5 or 10 years from now talk will be about how reliable the old Thai drives were compared to the new junk coming out of [new country of origin].

Re:But they didn't even do 1T right...

[cawpin]

I currently have 2 Seagate 1TB drives in a RAID1 in my Ubuntu server and they have been fine for the 3-4 months since I got them.

Re:But they didn't even do 1T right...

[greg1104]

The data from Google's study say that lowering drive temperatures to below 35C increases their failure rate, particularly when they're new. I'm not sure I agree with the entirety of their methodology, but it's certainly persuasive enough that I've switched to aiming for 35-40C rather than sub-30C. That normally means the same basic approach you outlined, putting a single large and slow fan in front of the drives, but with some way to slow it down even further than the defaults if necessary. I don't hesitate like I used to in mounting drives in adjacent bays either.

I suspect the true cause of the correlation you suggest (drives >750GB fail more often) is mainly due to the switch to the perpendicular recording methods that started in larger capacity drive around that same time.

Re:But they didn't even do 1T right...

[Spoke]

Well, I don't have a lot of experience with them, only having had 3 of the 7200.11 1TB drives.

But I only bought two and one of the first two died after 3-4 days of operation with a huge number of sector read errors. I tried rewriting the entire drive and that helped for a day before it went kaput again.

Good thing I bought them specifically to use in a RAID1 array, but having to wait 4-5 days for my RMA to be processed did worry me while running in degraded mode. Luckily, I did have backups in case the other one failed.

That said, if you get a good one, they are fast and fairly quiet. At the fast end of the disk you can write over 100MB/s to the disk which drops to 60-65MB or so at the slow end of the disk.

I'm hoping the early deaths are just a residual of some manufacturing defect and doesn't mean that the drive will fail more often down the road. At least the drive has a 5 year warranty.

Re:But they didn't even do 1T right...

[estarriol]

Just to add my own anecdotal experience, I've had 2 out of 2 Seagate 750Gb 7200.11s fail on me within 6 weeks of receipt.

I also have 3 external FreeAgent drives with 500Gb capacity, which I've had for over a year and they're rock solid. I believe these use the 7200.11 internally.

My main case runs pretty hot - don't know if this did it. I've RMAed the 750Gb again and purchased an eSATA enclosure, will be interesting to see if it dies on me again or not.

My gut instinct says that the newer drives run very hot and burn out unless actively cooled. Certainly their reliability is WAY down on what I'd normally expect from Seagate - I'm happy to keep buying the excellent FreeAgent range but I think I might try Samsung for my next internal drive.

I heard a rumour that insufficient power inputs can also fritz these drives - has anyone heard anything about that?

Re:But they didn't even do 1T right...

[afidel]

Sub 30C? During the Summer my ambient room temp is barely sub 30C (~28-28.5C) how does even rapid airflow get the drive that close to ambient when it's using a couple Watts in a fairly small volume? I try to keep my drives around 35-36C which is well within their specified range.

Re:But they didn't even do 1T right...

[pla]

Sub 30C? During the Summer my ambient room temp is barely sub 30C (~28-28.5C) how does even rapid airflow get the drive that close to ambient when it's using a couple Watts in a fairly small volume?

Well, for one, I keep my house around 70F (21C) year-round.

But fair enough, in the interest of accuracy, I should have said "around 10C above ambient" rather than an absolute temperature. I should also qualify that further with the fact that I only refer to mostly-idle (but not spun-down... I never spin my drives down) temps, not under heavy load.

Screw capacity, I want reliability

[damn_registrars]

I know I'm not the only person around who feels that Seagate's consumer-level drives have taken a turn for the worse in terms of QC, and their customer service is terrible at best. But it doesn't seem like the other manufacturers are doing a whole lot to try to take over the high-quality consumer-level niche.

Anyone have a recommendation for a drive manufacturer whose quality has improved over the years, and actually makes good consumer drives? I'm so disgusted with Seagate I'm even willing to consider Connor or Maxtor.

Re:Screw capacity, I want reliability

[CorporateSuit]

I'm so disgusted with Seagate I'm even willing to consider ... Maxtor

No! Don't!

Unless you WANT to hit F4 to finish booting up your computer every time you turn it on.

Re:Screw capacity, I want reliability

[geminidomino]

Err... Connor is long gone, dude. Been Seagate since 1996.

Re:Screw capacity, I want reliability

[damn_registrars]

Unless you WANT to hit F4 to finish booting up your computer every time you turn it on.

Well, that depends. Will my system remain booted afterwards? Because the crappy seagate drive I bought was causing my system to crash when it was otherwise running. If the system is reliable enough to not need to be rebooted, I'll take a minor booting inconvenience for that.

Re:Screw capacity, I want reliability

[mc900ftjesus]

Yeah, terrible customer support. I had to fill out this RMA form that took like 5 minutes, then I had to ship the old drive back, THEN I had to wait 1 whole week for a new drive to show up.

Man, that was hard.

Connor has been gone for years and Maxtor is the worst manufacturer of them all, where the hell did you come from?

Re:Screw capacity, I want reliability

[qoncept]

That's the nature of dirt cheap hardware. It happens in every industry. I remember a rant by the owner of an aftermarket parts store for Subarus where he complained about all of the 17 year old kids running drop ship operations from their parents basements selling $800 exhausts for $5 over cost. They make a marginal amount money, but not enough to continue to do business on, they are soon replaced by another kid, and the established, brick and mortar stores with good service can't compete.

That's also (part of) my theory on gas prices. Remember $1 gas 8 years ago? I said back then that it was getting so cheap that it wasn't going to even be worth selling and would cause a collapse in the economy.

Re:Screw capacity, I want reliability

[Anonymous Coward]

You may want to try getting hard drive cooling bays, current hard drives get very hot while running and your case cooling may not be enough to keep your drives from being damaged.

Re:Screw capacity, I want reliability

[ZerdZerd]

5 years warranty on Seagate. Suits me perfect.

Re:Screw capacity, I want reliability

[damn_registrars]

I had to fill out this RMA form that took like 5 minutes, then I had to ship the old drive back, THEN I had to wait 1 whole week for a new drive to show up.

Some of us remember a time when customer support was important enough that companies would be willing to cross-ship the drives - send you a replacement before you send back your defective one so you can try to get your data off.

And where did you get the shipping material from? When I was getting ready to send mine back they told me that if I did not use the very specific containers and packing methods, they would void the warranty and not ship anything back. And being as I had already thrown out my original box, I had nothing to pack it into that would meet their specification - nor was anything available within 100 miles of my house that would.

where the hell did you come from?

I came from the past, when customer support actually meant someone on the phone gave a damn and wanted customers to be ... supported!

Re:Screw capacity, I want reliability

[neumayr]

And while we're at it, so is Maxtor. Since '06.

Re:Screw capacity, I want reliability

[Anonymous Coward]

Um, Maxtor was bought out by Seagate.

Re:Screw capacity, I want reliability

[networkBoy]

see,
when I have a drive die and need to ship it back, I go buy a replacement drive (so to minimize downtime), then ship the failed unit in the replacement's box. When the RMA return comes in I add it to the JBOD cluster and the cycle continues.
-nB

Re:Screw capacity, I want reliability

[geminidomino]

And while we're at it, so is Maxtor. Since '06.

Yeah, I realized that after I posted. I think I missed the joke...

Swear by Seagates myself though. the only one ballsy enough to still offer 5 year warranties

Re:Screw capacity, I want reliability

[Anonymous Coward]

Our failure rates with the 1TB Barracudas are pretty close to the advertised 1.5%. There were a few issues in the beginning but overall they do just as specified. That's with close to 1,000/month.
Both Conner and Maxtor are Seagate now. Hitachi, WD and Samsung are the only other players left. Hitachi makes slightly higher quality drives at a significantly higher price IMHO. WD has a spotty quality record, we stopped working with them. Samsung has a bad reputation but I have no experience with them because we only use high end drives.

More like..

[m651612763]

More like a milestone-and-a-half.

wtf?

[bigdaddyhame]

LaCie's had 2TB models out for a while now. Why is 1.5TB important? http://www.lacie.com/ca/products/product.htm?pid=11111

Re:wtf?

[GenP]

I'm pretty sure that just has two 1TB drives striped together rather than one large 2TB drive.

Re:wtf?

[CaptainPatent]

Um... Because the LaCie model uses two 1TB hard drives to store externally.

The big news is the single-drive capacity reached 1.5TB meaning that external multi-drive enclosure could now reach 3TB.

Re:wtf?

[pushing-robot]

LaCie's had 2TB models out for a while now. Why is 1.5TB important?

...because the LaCie is just two 1TB drives in RAID 0?

Re:wtf?

[DarthJohn]

LaCie's had 2TB models out for a while now. Why is 1.5TB important?

http://www.lacie.com/ca/products/product.htm?pid=11111

That is not a single 2TB drive. It's two 1TB drives in one enclosure.

With that enclosure, and two of these drives in it, you would have 3TB.

Re:wtf?

[HTH NE1]

This means I'm going to have to backup my drobo so I can reconfigure it from a maximum 4 TB enclosure to a maximum 6 TB enclosure (5.5 TiB actual). Good thing it isn't nearly full yet (and only has three 500 GB drives in it).

Capacity is everything!

[Ngarrang]

Yowza! Bring it on! That 15-disk array just got much larger. Roughly, at the rate of growth of data at my company, we wouldn't run out of space for nearly 10 years. I think I can handle that.

Re:True Tebibyte?

[Piranhaa]

Please don't say that word. It sounds like something my 3 month old niece says. Rather, call it Decimal/fake terabyte (found on hard drives) or just a (real) 'terabyte'. I think it's pathetic people have come up with some new (baby sounding) word because hard drive manufacturers are too f'ing arrogant to make 'true' sizes. In marketing 1TB/1000GB sounds a little bit better than 931GB..

Re:True Tebibyte?

[Anonymous Coward]

I can't tell, is that ebonics, or do you have a lisp?

Customer Service and Cost

[Lead Butthead]

When everyone are in a tight race to the bottom (of price bracket) it's hard to have extra money to pay for decent support staff. I've always anticipated that at some point this mad drive to lower cost will have to halt, as surely the cost of material has only been going UP over the years (petro that is the basis of nearly everything hasn't exactly went down over the past eight years despite of its recent (short-term) fall); it's logically absurd to expect price of tech products to continue falling.

Re:True Tebibyte?

[Anonymous Coward]

Why the switching base numbers on the product? 32MB of cache that sounds like it came from base2. Why not use 25MB or 50MB if your counting your total storage in base 10?

Re:True Tebibyte?

[corsec67]

In what other field does "tera" mean anything other than 10^12?

How many Hertz in a Kilohertz as it relates to a computer? 1024? 1000?

storage capacity boggles the mind

[jollyreaper]

Wow. My first hard drive was 20mb. I bought a keychain flash drive the other day with 16gb of storage. I can go on youtube and watch playthrough recordings of games that had me going ZOMGWTF!!! 15 years before that phrase was even coined. I remember being blown away by how incredibly awesome the newer Sierra adventure games were once they supported VGA graphics.

I remember how cool I thought it was when I could dub my dad's old sabbath records off onto a tape and bring my tunes with me on the go. It boggles the mind that I can fit dozens of albums on a single mp3 player. The Internet makes Asimov's concept of the Encyclopedia Galactica appear small and pathetic, we're seeing more and more scifi computer technology made real each and every day. Snow Crash, anyone? With how the economy's tanking, I expect burbclaves are just a few years off.

Makes me wonder what I'll be thinking given another ten years of progress, what will be boggling my mind then?

Re:storage capacity boggles the mind

[flyingfsck]

Grampaw? Is that you!???

Re:storage capacity boggles the mind

[Anonymous Coward]

Makes me wonder what I'll be thinking given another ten years of hard-core marijuana smoking, what will be boggling my mind then?

There. Fixed it for you.

Who cares about speed?

[Godji]

Who cares about speed on a storage device? Show me some reliability data instead. Nobody that I know bases their purchasing decisions on speed. It's always relibability or cost/gigabyte or both.

Re:True Tebibyte?

[Pentium100]

And it also looks better to write "200GB" on a LTO-1 tape and then add the fine print on the other side of the box (assuming 2:1 compression).

Can anyone explain to me how this "tradition" (of writing double capacity) came to be?

Linux will freeze with these 1.5TB drives

[tivojafa]

I have a few of these drives... they are very fast for sequential read (>120MB/s sustained)

However, if write-cache is enabled (default) Linux will freeze intermittently reporting a SATA timeout executing a cache-flush command.
Tested with the 2.6.24 and 2.6.26 kernels. Other people have reported the same problem with the 2.6.27 kernel.
Tested with multiple drives and multiple SATA controllers (different chipsets). No SMART errors logged.

Thread on the Seagate support forum: http://forums.seagate.com/stx/board/message?board.id=ata_drives&thread.id=2390

The workaround is to disable write-cache on the drive.

Re:Linux will freeze with these 1.5TB drives

[Anonymous Coward]

I'm running 2.6.27 on the intel p43 chipset with no issues on one of these drives.

Re:Linux will freeze with these 1.5TB drives

[Anonymous Coward]

got the exact same problem, waiting for a fix :(

Re:True Tebibyte?

[Anonymous Coward]

No, it just that manufacturers don't want to explain base-2 to base-10 conversion. Most people don't care that 10^12 is not the same as 2^40, and they don't want to take the time to learn. It's easier to communicate to people who understand base-10 that a terabyte is 1 trillion bytes, even though it's really 1,099,511,627,776 bytes.

On the flipside, they shot themselves in the foot having to pay for average Joe customer calling about his "missing" space.

Terabyte Tebibyte?

[AliasMarlowe]

Please don't say that word. It sounds like something my 3 month old niece says. Rather, call it Decimal/fake terabyte (found on hard drives) or just a (real) 'terabyte'. I think it's pathetic people have come up with some new (baby sounding) word because hard drive manufacturers are too f'ing arrogant to make 'true' sizes. In marketing 1TB/1000GB sounds a little bit better than 931GB..

Please don't abuse the word Terabyte, or attempt to usurp any of the other base-10 prefixes which were defined long before computers were invented. It is the base-2 interpretation of these prefixes which is fake.

The abuse started with use of kilo to denote 2^10 instead of 10^3, often using K instead of k as prefix. This was relatively innocuous, since the case of the letter could ensure the prefixes were somewhat distinct. However, for 10^6, the prefix for mega is M (and m is also allocated for milli), and abusing this prefix to mean 2^20 is unconscionable.

The kibi, mebi, gibi, etc. prefixes were created to solve a real need. The base-10 prefixes were already assigned, and could not be usurped.

Re:Terabyte Tebibyte?

[Skelde]

The kibi, mebi, gibi, etc. prefixes were created to solve a real need. The base-10 prefixes were already assigned, and could not be usurped.

Shure, but that doesn't change that the sound uterly moronic.

Re:True Tebibyte?

[JesseMcDonald]

SI prefixed only have standardized meanings when used with SI base units. The byte is not an SI base unit. Actually, there is no official SI base unit for information, but if there were one it would most likely be the bit, which is already associated with base-10 SI prefixes. Mixed units (e.g. MB/s) vary depending on how the value is calculated, but are generally SI.

kilobits, megabits, terabits: SI prefixes
kilobytes, megabytes, terabytes: binary prefixes

The HDD manufacturers want to use real SI units they should say "12Tb" rather than redefining "1.5TB".

Re:fp?

[Anonymous Coward]

Deal with you? He would suck you in, god damn it.

SATA can be used externally

[Nick Ives]

There are eSATA connectors. These days you just get a bog standard SATA drive and put it in an enclosure. The OP was mistaken in assuming that anyone buying one of these drives for external use (no mention of USB or Firewire in the OP) would be using a low speed bus.

So, GP was being sarcastic but not a troll.

Re:SATA can be used externally

[Sancho]

Eh, there are two main points which got conflated.

1) A drive this size will likely not be used for high-performance tasks. That is, it will probably be used for storage of music and movies rather than for applications and swap.

2) That enclosures will be slow.

Point 1 still hasn't been contested, and the first "troll" post didn't seem to care to discuss that--he just seemed to want to attack the idea that someone would only use a disk this size on a slow bus. The more I think about it, the more it sounds trollish--rather than attacking the thesis of the post, he nit-picked in order to argue.

Point 2--you're right. There are high-speed external enclosures, though in my experience, eSATA is fairly rare. I have a drive (used for backups) with both eSATA and USB, and I'd love to be able to use the eSATA, but so far, I have been unable to find an expansion card with suitable (if any) Linux support. Nonetheless, I'm not at all surprised that this was overlooked by the OP, but it's still irrelevant. If you don't need high write speeds, that should not deter you from buying this drive.

Re:SATA can be used externally

[GigaplexNZ]

My point was I want to know how often the hard drive is going to be the bottleneck instead rather than USB or firewire

The OP was mistaken in assuming that anyone buying one of these drives for external use (no mention of USB or Firewire in the OP)

Seems there was mention of Firewire and USB in the clarification at least, with a question on likelihood (didn't appear to be an assumption) which this whole thread inherited from.

Re:SATA can be used externally

[JET 666]

got a spare internal port?

http://www.usb-ware.com/sata-dual-internal-esata-adapter.htm

4 of these failing on me right now ....

[mrgermany21]

Very interested to know other people's experience.

Using 4 Seagate 1.5TB 7200.11 drives in an 8 tray enclosure by SansDigital for JBOD configuration for MacBookPro with a PC Express card. Essentially for external [video] media storage.

Can write to all 4 drives from a FW800 drive. Have been getting "Error -36" messages when reading the same data from one of the four drives and writing to another.

SansDigital not much help. Am considering RMA'ing the whole deal and starting over.

Thoughts anyone?

Re:4 of these failing on me right now ....

[actionbastard]

"Thoughts anyone?"

How big a set of tits does it take to crush a four-drive JBOD array?

Re:True Tebibyte?

[tchuladdiass]

That was due to the drives having built-in compression. And it turns out that 2:1 was about right at the time for a typical storage mix of code (which would get around 1.6:1) and data (text / spreadsheet files would get up to 5:1).

But now, most of the data on a large drive is already in a compressed format.

I've got 2 of 'em

[RecycledElectrons]

I ordered a pair of 1.5TB (1361GB) Barracudas from New Egg on the night of the VP debate. $189 each with free S/H was cheaper per TB than the 1TB drives I had planned to buy.

They work well for library storage on my aging Pentium D (GA-P65-DS3R, XP x64, 8GB RAM) machine.

Andy

Re:True Tebibyte?

[jedidiah]

In what other field to they mainly count by 1s and 0s?

SI units are meant to be COMPUTATIONALLY CONVENIENT, not merely arbitrary.

If we wanted units plucked out of someone's nether regions, we could just stick to English measurments.

With sufficient disclosure, no quantity is "ambiguous" or "confusing".

Re:True Tebibyte?

[zippthorne]

Hard drive manufacturers use SI nomenclature. Nomenclature that's in countless other industries and projects. How does that make them the arrogant ones?

rsync is your friend

[Anonymous Coward]

What you really want to do is setup your drives however you want to use them, and then have another set of drives that are periodically rsync'd to from your main set of drives.

Thus, you have a safe backup, and if you happen to be developing and accidentially delete a file, you have a recent copy.

Use multiple rsync sets, etc. for multiple backups. If you have hard drives to burn, do it to live drives in your box. Or across a LAN at night while you sleep. Etc.

Re:rsync is your friend

[Sancho]

Yeah, that sort of setup would probably work, too. Snapshots with rsync (and hard links) are nifty :)

Re:True Tebibyte?

[jedidiah]

The manufacturer didn't have to explain anything.

They merely had to disclose all the relevant information in un-ambiguous terms.

Of course they preferred to allow the unwary continue to think they were getting more than they actually were.

The idea that this sort of thing is wrong was adjudicated as such before the first Dutchman or Englishman set foot in the new world.

Re:True Tebibyte?

[Anonymous Coward]

In what other field does "tera" mean anything other than 10^12?

How many Hertz in a Kilohertz as it relates to a computer? 1024? 1000?

We are not IN another field. It is a marketing gimmick. They 'changed' the rules and OS people are still playing by the old rules. So which is it?

When it was a few meg it is not that big of a deal. At 300+ gig you are talking 20-30 gig 'missing'.

Re:True Tebibyte?

[Anonymous Coward]

SI units were developed to stand for binary usage, which a Byte is.. 8 bits = 1 byte, 1024 bytes = 1 Kibibyte (kilo-binary)...

Sorry, but in this case you're off base... The SI units have been standardized, just not adopted by drive manufacturers as it would make them look bad.

The measurement 1.5TB would equate to 1.39 TiB.

Re:True Tebibyte?

[GuyverDH]

For your own educational review...

http://physics.nist.gov/cuu/Units/binary.html

and yes, they refer to usage with Bytes (B) not just bits (b)...

Re:True Tebibyte?

[Anonymous Coward]

Rubbish.

BAARF

[tepples]

That blog post forgets one thing: sector remapping.

You can't remap a sector if an entire drive fails. By the time you have put in your spare and your RAID controller has finished rebuilding the arrays, there's a not insignificant chance that the stress of constant reading has killed one of the remaining good drives. This chance goes up as the time to write an entire drive (capacity divided by transfer rate) goes up. See Art S. Kagel's article that strongly recommends RAID 10 over RAID 5.

Re:BAARF

[afabbro]

there's a not insignificant chance that the stress of constant reading has killed one of the remaining good drives.

You are assuming that "stress" (high use) is a contributing factor to hard drive failure. This may not be so.

Re:BAARF

[tepples]

You are assuming that "stress" (high use) is a contributing factor to hard drive failure.

Even if "stress" isn't a contributing factor, as capacities go up, "random chance" becomes one. A recently posted article has the details.

Re:BAARF

[nxtw]

You can't remap a sector if an entire drive fails. By the time you have put in your spare and your RAID controller has finished rebuilding the arrays, there's a not insignificant chance that the stress of constant reading has killed one of the remaining good drives. This chance goes up as the time to write an entire drive (capacity divided by transfer rate) goes up. See Art S. Kagel's article that strongly recommends RAID 10 over RAID 5 [miracleas.com].

This argument is essentially: RAID 5 is more likely to fail during rebuild than RAID 10, so use RAID 10. But why not consider something else?

Consider an array of 1TB drives with a published failure rate (BER) of 1e-15 (one in every 10^-15 bits read will fail.) Assume that one drive fails. If there is a single mirror, there is a .11% chance that the mirror will have a read error while rebuilding. In a three-drive RAID5, there is a .21% chance that one of the remaining drives will have a read error while rebuilding. In a four-drive RAID6, the chanche of having read errors twice while rebuilding (essentially three drives failing) is less than .01%.

Re:BAARF

[Bandman]

I've thought about using RAID 5 as the bottom level of some other level, but I can't convince my self that it would be slow as molasses.

Re:True Tebibyte?

[JesseMcDonald]

Nice try, but the SI standard is defined by the BIPM, not NIST. The SI base and derived units do not include bits or bytes.

Redefining the prefixes for base-2 units (e.g. byte) as base-10 does no one any good, as all the historical literature on the subject is already using the existing prefixes as base-2. The new terms (KiB/MiB/TiB) are clear enough, but the old ones -- which used to be unambiguous in context -- now have two possible meanings. Given the ridiculous names suggested for the new prefixes, that isn't going to change any time soon.

Anyway, it makes no sense to pile a base-10 prefix on top of a base-2 multiple of the base unit (one bit). It's like measuring distance in kilofeet, or area in millihectares. If you want base-10 units just use bits, and leave our perfectly useful binary prefixes well enough alone.

Seagate 7200.11 and WDC Caviar Black

[rubeng]

Both the Seagate 7200.11 and Western Digital Caviar Black families have a 5-year warranty (see comparison).

Well, D'uh

[Kittenman]

"More than a year and a half after the first terabyte hard drives became widely available, Seagate has reached the next storage capacity milestone. With 1.5 terabytes, the latest Barracuda 7200.11 serves up 50% more capacity than its peers"

Hmmm. So .."1.5 terabytes is serving up to 50% more capacity than its peers". Peers that were up to a terabyte in size.

No shit, Sherlock.

Physical limit and other thoughts.

[EnsilZah]

The subject came up because I recently filled my new 1TB external WD drive, everyone seems to have a favorite brand by the way with only anecdotal support of their choice.
Anyway, I was filling it up with some stuff I've had on CDs and DVDs because I figured that it's more convenient and also actually getting cheaper to store that stuff on a large HDD than on optical disks when I've noticed that the top capacity drive has been occupying the sweet-spot for a while which wasn't my experience with previous generations.
Now I probably just wasn't paying as much attention previously and it's not a new trend, but I started thinking of how close are we to the physical limit for data density on magnetic platters.
I tried looking up some information on the subject but didn't find anything conclusive, found an article from 2001 titled "Hard drives bumping up against physical limits" so I'm guessing it's the same as the limits on ICs or Peak Oil, the prospect seems to be looming just a few years ahead but we never actually reach it.

So I would appreciate anyone who might shed some light on the matter.

Re:Physical limit and other thoughts.

[rrohbeck]

Watch Mark Kryder's CMU seminar from last year at http://www.ece.cmu.edu/news/seminar/2007/fall/kryder_11_29_07.asx.

Like Peak Oil, it's about diminishing returns and commercial viability.

Abuse of Barracuda naming

[Ilgaz]

Seagate Barracuda has been long known for excellent speed and reliability for high workstation workloads, not huge capacity. They made it a generic thing with range from 80 GB (the real one) to 1.5 TB.

They make up great model names and make it a generic thing with horrible speed issue. Look at this mess yourself: http://www.seagate.com/www/en-us/products/desktops/barracuda_hard_drives/

Lets hope we don't see a 2 TB Cheetah with horrible transfer speeds.

They should give up everything in hand and design an actual accessible hybrid drive with an innovative cache mechanism. I am not speaking about the 32 MB cache, something else... Like a cache which has clue about filesystems in use etc. With Laptops taking over the scene and UPS sells for $50, that dinosaur RAM cache mechanism must change.

Re:True Tebibyte?

[GuyverDH]

They aren't redefining base 2 units for base 10... they are still base 2 units - all the way through...

KiBi - as in Kilo-Binary
MiBi - as in Mega-Binary, etc...

It's more intuitive than everyone using the same word, with different meanings and forms, like megabyte -> (1024x1024) or (1024x1000) or even (1000x1000) depending on who designed the storage unit or memory module.

I guess I am a proponent of the exa-binary notation for all things computer related - it just makes sense.

7200 rpm limits write speeds.

[MtViewGuy]

I think the issue of the relatively slow write speed of the new 1.5 TB Barracuda drive comes down to this: they're still using 7200 rpm for the spindle speed of the platter drive motors. Because of this, we're not going to get much in the way of write speed improvements compared to the smaller 1 TB drive.

Given the state of materials technology nowadays, I'm surprised that most Serial ATA-2 drives haven't increased their spindle speed to 10,000 rpm outside of a few specialized drives like the Western Digital Velociraptor series. At 10,000 rpm, both sustained write and read speeds will be substantially higher than 7200 rpm drives.

Backups

[Jeppe Salvesen]

Poor write performance matters when you're using the disk for backups. Or databases. But it seems fine for hosting all your DVDs and FLACs. :)