Sun Unveils Thumper Data Storage

zdzichu writes "At today's press conference, Sun Microsystems is showing off a few new systems. One of them is the Sun Fire x4500, known previously under the 'Thumper' codename. It's a compact dual Opteron rack server, 4U high, packed with 48 SATA-II drives. Yes, when standard for 4U server is four to eight hard disks, Thumper delivers forty-eight HDDs with 24 TB of raw storage. And it will double within the year, when 1TB drives will be sold. More information is also available at Jonathan Schwartz's blog."

Comments

:O

[joe 155]

24TB... thats almost enough to hold all my pr0n!

Re::O

[UnknowingFool]

Don't lie. It's really full of Star Trek episodes.

Re::O

[FuturePastNow]

28 seasons of Star Trek + all the movies = 250GB.

Re::O

[Don_dumb]

28 seasons of Star Trek + all the movies = 250GB.

That is the most geeky post I have seen on this site for a while, you should be proud.

Re::O

[joe 155]

what season are you missing... by my count there are 29 seasons of star trek... next you'll be saying that you have all the movies except the wrath of Khan!

Re::O

[fm6]

The other 23.75 TBs is his personal archive of all the posts from Star Trek Usenet Groups.

Re::O

[Chris Burke]

I nominate "Star Trek Collections" as a new unit of storage measurement! Quick, somebody work out the conversion into the standard libraries' of congress.

Re::O

[Anonymous Coward]

They should call this drive the "Fapper," not the "Thumper."

Re::O

[FuturePastNow]

TOS= 3 seasons
TNG= 7 seasons
DS9= 7 seasons
VOY= 7 seasons
ENT= 4 seasons

Re::O

[joe 155]

touche, but in my defence tv.com lied to me about how many there was of the original... also who would spend too long watching that when we all know Picard is the best captain any enterprise has ever had... infact the NCC-1701-D had the best all round crew, but I digress

Re::O

[simcop2387]

you forget The Animted Series!

Re::O

[Duhavid]

Perhaps you have seen my porn collection.

I keep it on all the servers on the internet.

( Apologies to Steven Wright ).

Re::O

[nacturation]

That is the most geeky post I have seen on this site for a while, you should be proud.

Of course, the geekiest post would be a reply to this one containing a link to the torrent. Who is brave enough to claim the title of King of the Geeks?

Re::O

[Rob86TA]

I'm trying to figure out why it has yet to be modded Informative!

Re::O

[Silver Gryphon]

1 http://en.wikipedia.org/wiki/Library_of_CongressLo C ( ~20TB ) ... 80 STCs
1 Thumper @ 24TB ... 1.2LoC

Being named Slashdot's biggest geek for knowing how many bytes are in a Star Trek Collection ... priceless.

Re::O

[Anonymous Coward]

"...tv.com lied to me about how many there was of the original...."

How could they have watched the singing space hippies and ever thought there was any more seasons after that?

Re::O

[FuturePastNow]

D'oh! I don't have that. But it wouldn't take up much space.

Re::O

[hotdiggitydawg]

28 seasons of Star Trek + all the movies = 250GB.
Daily bandwidth used for trollind slashdot = 17Mb
Proving your geek-manliness to the largest online geek community = Priceless.

Thumper: If you...

[avirrey]

can't say something nice, don't say nothin' at all. Keep talking thumper because I like the nicities (specs). ------ Drools @ Technology

I want one!

[andrewman327]

This is perfect for the space constraints applied to many server rooms now days. I wonder how they managed to control the heat output. My laptop only has one HDD and it gets pretty warm. I am very impressed that (according to Sun) costs $2 per gig! As always, I hope it works as promised.

Re:I want one!

[cyanics]

and they are especially showing off the low power usage in that kind of space..

48 Hds, 2CPUs, and still less than 1200 Watts.

Oh many. Datafarm in a single rack.

Re:I want one!

[ScottLindner]

Maybe they didn't do much at all to control the heat? It wouldn't be the first time a vendor left heat issues to the end user to resolve.

I doubt this is the case though. Sun tends to make pretty good hardware. At least that's my limited experience.

Re:I want one!

[bcat24]

Didn't you hear? They're specially designed for use in cold climates. They store tons of data and keep your building at a comfortable temperature at the same time!

Re:I want one!

[Jeff DeMaagd]

It's not that big of a problem. A 7200RPM drive might take 15W max. 48 drives brings the total up to 675W. Not that bad in the server world, especially given the capacity.

Re:I want one!

[andrewman327]

From TFA (the last one): "We're still figuring out what to call the product, 'open source storage' or 'a data server,' but by running a general purpose OS on a general purpose server platform, packed to the gills with storage capacity, you can actually run databases, video pumps or business intelligence apps directly on the device itself, and get absolutely stunning performance. Without custom hardware (ZFS puts into software what was historically done with specialized hardware). All for around $2.50/gigabyte - with all software included."

This device is very interesting. It is poised to slash costs in data centers. It consumes less space, uses less power, costs $2/gig, and is managed just like any other server. Instead of calling it a data storage device, they should be marketing it as "DAS UBER SERVER!"

Re:I want one!

[andrewman327]

"Didn't you hear? They're specially designed for use in cold climates. They store tons of data and keep your building at a comfortable temperature at the same time!"

I am suddenly begining to doubt that greenhouse gases are responsible for global warming. Al Gore needs to make a movie about 48 drive servers.

Re:I want one!

[andrewman327]

7200 rpm? What fun is that? I want 48 15k rpm drives running in RAID 0! Now THAT is fast! (Drive failure? What's that?)

Re:I want one!

[dfn_deux]

2|=2.5 that 50 cents a gig makes a big price difference... when you are talking about 2.50*24*1024 that 50 cents adds 12 grand to the final price...

Re:I want one!

[Anonymous Coward]

This is perfect for the space constraints applied to many server rooms now days.

Starting at $40K a pop, a rack of these babies can buy more than a decent expansion of the server room space for many places. And with expansion you can get space for more than an extra rack. So need to do some math to justify these (as always).

Re:I want one!

[civik]

Im sorry, but in my opinion this thing is next to worthless. I cant really think of a single application that you would want 20 TB concentrated into one server chassis. I mean, its not like you can attach this thing to a SAN and carve off LUNS to utilize the space from other devices in the fabric. The only interesting thing about this server is the shock factor of 20TB in it. Once you realize that you can get the same thing in a SAN attached array from vendors like Nexsan for way cheaper, it really becomes sort of another stupid Sun gimmick.

Re:I want one!

[hpavc]

The heat and energy solution is the amazing part of that product, use of sinks and pipes it works well.

Re:I want one!

[Wesley Felter]

I mean, its not like you can attach this thing to a SAN and carve off LUNS to utilize the space from other devices in the fabric.

Actually, you can do that with iSCSI.

Re:I want one!

[dlasley]

SAN is not always the answer to large storage. Take Oracle 10G RAQ, for example: say you have a multi-master setup in several datacenters, and you want dedicated high-speed fault-tolerant local storage for each instance. You set up a 4900 with internal storage for the OS and Oracle, then add your data partition(s) on a Thumper in each location. Even the best fiber mpxio connection from the 4900 would be hard-pressed to match the speed, reliability, and responsiveness of a tuned array.

Given the disasters I've seen with SAN storage, I would happily spend 25% more to get the Thumper and know I could rely on Sun to have someone onsite fixing whatever problems we have within four hours - and know that the person who shows up knows not only what to do with the Thumper, but also what to do with the 4900 once things are fixed.

&laz;

Re:I want one!

[Anonymous Coward]

I'm going to hold out for couple more years before building my video library and see if seagate can really get a 10x density gain. :)

http://hardware.slashdot.org/article.pl?sid=06/07/ 06/1546235

Re:I want one!

[ScottLindner]

You actually know something more about how this product works? I didn't do any research on it. But in general, I tend to have respect for Sun hardware. Sun software on the other hand (Solaris) I tend to have very little respect for. It works and works well, but is clunky and very outdated IMHO. Yah.. I'm a Linux guy.

Re:I want one!

[andrewman327]

My apologies for using bad data. After all, GIGO. My point is, however, that this device offers the ultimate in economies of scale and very good price for the storage.

Re:I want one!

[trollogic]

Sun software on the other hand (Solaris) I tend to have very little respect for. It works and works well, but is clunky and very outdated IMHO.


You've gotta be kidding me or you are the kind that thinks my embraer EJ-145 gets me from SFO to Dallas (because thats what your realm activities are limited to) and you can do your job so all the Boeing 747s or Airbus 300 are overatted in your IMHO ... Linux is no-where near the heavy lifting capabilites of Solaris 10. Just because bash isn't installed as the default root shell does not make Solaris outdated. It doesn't wear the lipstick or the new hot mini-skirt that linux flashes .. but it certainly can do the job in day and night over and over, sideways and whichever ways and still be willing to service more ..

Re:I want one!

[ScottLindner]

Depends on your reasons for using the OS I guess.

At least I'm logical about it... unlike you. Try to be objective, please?

Re:I want one!

[ScottLindner]

Wow.. really objective rating there. I'm a troll and the other guy isn't? Objectivity makes me a troll.. and the other guy being a prick gives him an Insightful raiting?

WTF? Some of your Slashdot readers need to drink more beer.

Re:I want one!

[trollogic]

I fail to see your objectivity .. what is your basis for calling Solaris clunky ? You haven't explained anything towards your comment .. I've been using both Linux and Solaris for over 12 years now .. I can share this with you, I haven't had a kernel panic due to a driver in Solaris since 1998. Not the case with Linux. I don't need support for all the funky hardware in Solaris. What I need is sure-shot reliability and predicatibility. And clear information when stuff does break. I can tell you Linux today is no where near being reliable/predictable or debuggable. I run a ton of Linux boxes and for those purposes which I don't care for reliability or predictability or where I don't need to troubleshoot. Would I trust Linux with my mission critical applications .. NO. Would I trust Linux with my webserver .. Yes. Choose it, Use it.

Interesting

[Dark Paladin]

I've been talking to the wife about getting a NAS for the house - but now a 1 to 2 terabyte system seems so...puny.

Hey, honey - remember how I said I wanted to store *all* the movies on the server? Get a load of this ;).

Re:Interesting

[Anonymous Coward]

You misspelled PR0N as "movies". In addition, I did get the double meaning of the last sentance.

Holy SHIT!

[gasmonso]

Did you see how tightly packed the drives were? Is heat a concern or is there a tornado cooling system in place?

http://religiousfreaks.com/

Re:Holy SHIT!

[IflyRC]

That's the Bambi Cooling Add-On system.

Re:Holy SHIT!

[Anonymous Coward]

Could you please put the link to your stupid website in your sig, so those of us who are uninterested don't need to read it a dozen times in every story? KTHX...

Re:Holy SHIT!

[imsabbel]

Why does everybody here get so up with "The HEAT!!111".
Its 48 hds in a 4U case. 48HDs is about 600W under full load.
If you compare this to the fact that there are dual-socket - dual core servers out there that push 300W through a 1U case, thats nothing.

Also, a 4U case allows the use of nice fat 12cm fans in the front, while the horizontal backplane allows for free airflow (in contrast to vertical ones like used before)

Re:Holy SHIT!

[UberLame]

It might have allowed for 12cm fans, but it you had looked, you would see that they are using 10 much smaller fans. Ick.

Meanwhile, the x4600 (8 dual core Opteron system) does apparently use 2 12cm fans.

With all those disks, I suppose it might not make much difference, but I would have rather seen them using 12cm fans on the x4500 as well.

Re:Holy SHIT!

[buysse]

Sun typically worries more about redundancy than noise. The 10 small fans are hot-swappable and run at ridiculous speeds (and yes, sound like a A320 revving up for takeoff), but I bet the thermal budget allows four of them to be dead at any given time.

Re:Holy SHIT!

[esaxe]

Have you seen the array of fans on the front of the system?

http://blogs.sun.com/roller/page/jonathan?entry=th e_rise_of_the_general

They are forcing the air along the gaps existing between drive columns out the back.
As you can see, the gaps between the columns are fairly narrow, so air is moving
*quickly* along the disks. Having the top cover (not shown in the picture) in place
is important for the system's cooling, so that the air taken in by the fans along the front
has no place to go except between the disk columns, and out the back.

Re:Holy SHIT!

[Anonymous Coward]

I bet the thermal budget allows four of them to be dead at any given time.

If their piece of shit Netra V210 servers are any indication you'll desperately need that redundancy. Our server was less than 18 months old when it lost a power supply fan and another 4 months before the system disk started popping up tons of read errors. I know I'll never buy another Sun box after our experiences with this piece of shit. It is a far cry from the quality we had with the e450 and e420R and even the Netra AC200 systems. Fuck, our Ultra 10s were more reliable than this thing.

Re:Holy SHIT!

[imsabbel]

Well, the problem with those small fans and high speeds is that most likely they WILL be dead after a few months.

12cm fans are low rpm and _extremely_, and i mean _REALLY_ extremely reliable (when not bought from the latest casemod supplier. I mean those nice metal ones). I have seen electrical installations where those things have been running for 20 years, nearly burried in dust, without breaking down.

Re:Holy SHIT!

[queenb**ch]

So, my question is do you need ear protection to be in the room with this thing? Whats the dB rating on those fans?

Queen B

$42,000

[enitime]

"starting as low as $2 per GB"

Doesn't sound like much.. but that's $42,000 for the top 24TB model.

Perhaps it's time to start using "per TB" costs for these things. Surely no one sells sub-terabyte storage servers anymore.

Re:$42,000

[Fredwick.com]

$42,000 for 24TB is dirt cheap.

Re:$42,000

[avirrey]

Remember the telco's are going to want to prioritize the HD usage... may not be that sweet of a deal. LOL.

Re:$42,000

[Homology]

> Surely no one sells sub-terabyte storage servers anymore.

Most "storage" servers sold today have less than a terabyte
of capacity.

Re:$42,000

[bryerton]

Is it? I recognize that the system itself is impressive. But to buy 48 750GB SATA-II 3.5" drives costs around $24,000, and gives you ~36TB. If you notice the pricing, it becomes obvious SUN is drastically over-pricing the drives. The only diff I noticed at a first glance between the $40k and the $90k option was the size of the drives. Perhaps I missed something...

If I didn't, only a fool would buy the more expensive version. Just go in for the cheap array, and purchase 750GB drives yourself, re-sell the original 48x250GB ones, and you'll save yourself a rather large sum of money.

Re:$42,000

[Anonymous Coward]

If I didn't, only a fool would buy the more expensive version. Just go in for the cheap array, and purchase 750GB drives yourself, re-sell the original 48x250GB ones, and you'll save yourself a rather large sum of money.

Pretty friggin' obvious you haven't worked at a datacenter. I know no administrator who would do that to save tiny amount of money and lose the guarantee. No, even if these systems use same drives as your home computer, these are not used as toys.

Re:$42,000

[99BottlesOfBeerInMyF]

$42,000 for 24TB is dirt cheap.

It is a pretty good price, but not insane. Apple's cheap RAID servers cost about $1.85 per gig. (I know, Apple + low price = crazy.) Mind you, you'd need 8u not 4u to fit them. That is really where I see the the advantage here for companies that need more storage but have real space constraints.

Re:$42,000

[OldeTimeGeek]

Perhaps I missed something...

If you were looking at putting one in your basement, yeah, you could do it cheaper. For a datacenter, however, it's more than just a bunch of drives. Stuff like maintenance, 24-hour support, firmware upgrades, monitoring software, making sure it works with all of your systems and being there to help when it doesn't are important to large enterprises...

Re:$42,000

[jacksonj04]

Eh? My home server has 2, where sells sub-TB storage servers?

Re:$42,000

[bryerton]

True. I have never worked at a *datacenter*. I have worked for some companies that require high reliability. I'm not saying build it yourself. I'm saying, don't pay SUN for the higher capacity HDDs. Obviously the enclosure and hardware is the key to running that many disks in such a small space properly. Is there some extra guarantee that comes from buying the HDDs from them? Would one have to buy every replacement drive from SUN as well? I'm pricing out 7500 rpm drives from seagate here that as far as I can tell match their specs listed (7500rpm, SATA-II, 3.5"). Love to hear back on from someone who purchases items at this level.

Re:$42,000

[fimbulvetr]

If you have data on there you need, and 4 years down the road and the place you bought the drives from says "Hey man, we don't stock those 250s anymore. They're ancient history, but we'll sell you this 2TB ZATA drive." you'll be screwed. You might be able to pick one up on ebay.
If you bought it from sun + contract, you get the same geometry drive in less than 4 hours.

Re:$42,000

[mieses]

Sun will probably not provide drive sleds for empty drive slots.

It's ridiculous to have to buy the exact same Seagate part for 2x or 3x more. No Sun warranty is going to replace lost data. Besides, the whole point of ZFS is to provide fault tolerance and recovery.

Also, the same storage concept has been on the market for over a year, albeit with 42 drives:
http://www.nexsan.com/products/ATAbeastlow.pdf

Re:$42,000

[Anonymous Coward]

Fortunately ZFS doesn't care about the geometry as long as the replacement drive is at least as large as the original.

Re:$42,000

[gstoddart]

Doesn't sound like much.. but that's $42,000 for the top 24TB model.

They used to say that about gigabytes, and people still bought 'em. Anyone who *has* 24TB probably thinks $50K or so for a Sun Thumper is a damned good bargain. Some of them might even fill the rest of the rack with more of them, you could get some crazy storage going on with a small herd of those suckers. ;-)

Perhaps it's time to start using "per TB" costs for these things. Surely no one sells sub-terabyte storage servers anymore.

Bring 'em on. The home market is still quite away from TB storage becoming common. I'm waiting for things to fall a little more before I get one for myself. =)

Cheers

Re:$42,000

[1lus10n]

No. Most consumer storage servers have less than 1TB of space.

Okay...

[Vo0k]

...but how good is it at repelling the antlions?

Re:Okay...

[PhrostyMcByte]

Pick your poison: antlions or Shai-Hulud.

Re:Okay...

[Propagandhi]

Can't believe I had to scroll this far to find a Dune reference! Shame on you Slashdot.. Shame.. on.. you!

Re:Okay...

[Vo0k]

Considering its size, weight, power consumption, I'd say it's the anti-antlion model.

Re:Okay...

[mobby_6kl]

Pick your poison: antlions or Shai-Hulud.

Thumpers have the opposite effect on the maker, IIRC.

Re:Okay...

[Neoncow]

Thumpers have the opposite effect on the maker, IIRC.

Exactly. Don't use the thumper; the antlions come. Use the thumper; Shai-hulud comes. Make your choice.

Re:Okay...

[Vo0k]

place it in damp terrain?

Re:Okay...

[fusion9290991]

My money's on Bambi :)

Re:Okay...

[mikael]

I dunno - but it sure attracts sandworms...

Ah, yes...a machine labelled "Sun"... appropriate!

[soren42]

No doubt that fully loaded it generates almost as much heat as the sun, too!

Really snazzy tech, but that's a lot moving parts in a little space... and probably too hot to touch. Could you imagine the cooling required for a densely-packed data center of these things?

Or am I way off base here?

Dune..

[WizADSL]

Thumper? I hope the sand worms stay away...

Re:Dune..

[__aaclcg7560]

That's the whole beauty of this new product. It's in an air conditioned room and outside the sand worms keep all physical intruders away. The only thing you need to worry about are the virtual sand worms coming over the Net.

Re:Dune..

[nacturation]

24TB in only 4U... And how can this be? Because it is the Kwisatz Serverrack.

Re:Dune..

[mikaelhg]

Because it is the Kwisatz Serverrack.

And here I stand, with no +1 Funny to grant.

Re:Dune..

[Anonymous Coward]

Actually, a thumper was used by the Fremen to call worms.

cooling

[Zheng Yi Quan]

Heat output from all those drives is a concern, but if you look at the photo on the ponytailed hippie's blog, you can see that the box has 20 fans in the front and probably more in the back. Makes you wonder what the thrust-to-weight ratio is. This box is going to make a screaming database server. 2GB/sec throughput to the internal disk beats anything out there, -and- the customer doesn't need to invest in SAN hardware to do it.

Re:cooling

[eln]

Great, so it's throwing all that heat out to raise the ambient temperature in the rack and force you to invest in more air conditioning power and more specialized airflow in the rack to keep this thing from damaging your other systems.

I'd be interested to see how your actual overall power consumption within a rack and within a data center is affected by this thing.

Re:cooling

[Wesley Felter]

Database performance is generally more related to IO/s, not GB/s. Thumper may still win an equal-cost comparison against eterprisey SAN equipment because it gets more spindles.

Re:cooling

[Anonymous Coward]

So 24TB of storage servers otherwise doesn't generate heat?

Put doors on the rack and chimney the damn thing. Use plastic wrap if you have to. Ain't nothing any data center doesn't do for blades anyway. Drives just don't generate the amount of heat you think they do tho.

Re:cooling

[EbbTide]

This box is going to make a screaming database server. 2GB/sec throughput to the internal disk beats anything out there, -and- the customer doesn't need to invest in SAN hardware to do it.

Yes, but if you want any sort of redundancy in your database, you're going to have to go to the SAN. Unless you choose replication, but that's not an optimal solution for availability.

Re:cooling

[oliana]

Picture gallery shows 5 hot-swappable dual fans for a total of 10. And the back doesn't appear to have any (unless there are some hiding in the power supplies, but it didn't look like it.)

But real cool 360 tour complete with the ability to take out drives, fans, CPUs, power supplies.

Re:cooling

[Anonymous Coward]

.... and what happens when a disk fails? You have to stop the whole array to open the top to replace one disk - not very hot-swap! I seem to remember a very similar product called Kashia (Cashia, Cassia?... sorry, old brain in serious decline) that died a death as soon as datacenter managers asked "And what about online drive replacement?"

Imagine the scene - DB admin is on the phone to the FD:
ADMIN: "Yes, sir, we've found the problem with the order system, but I'm going to have to shut down all the applications - billing, shipping, payroll, all of 'em - with databases on the same storage array so I can swap out the one faulty disk..."
FD:
FD's ASSISTANT: "So, which address do you want me to send your pink slip to?"

Re:cooling

[Anonymous Coward]

erm ... you just swap it out ?

Who said the machine needs to be powered down ?
You unmount the drive or remove it from the ZFS pool, remove, replace and reintroduce the new disk into the pool.

ZFS (using RAID-Z) will carry on as if nothing has happened

Re:cooling

[CYDVicious]

"Makes you wonder what the thrust-to-weight ratio is." That's what she said last night. Boo-Yah!!

Re:cooling

[eln]

The problem is it appears you have to remove the top of the case to take the drives out. In your typical rack environment, the tops of cases are not going to be accessible without removing the server from the rack, which will likely involve shutting it down.

Re:cooling

[fimbulvetr]

p.s.: rails

Re:cooling

[Firehed]

Think of the CPU-to-HDD ratio. 24:1 in this case. If your lower-density boxes have anything less, they'll be putting out more heat per gigabyte, since all drives of x RPM are going to put out nearly identical amounts of heat, and I'm assuming that the type/speed of the CPU would be a constant between them (and when you're talking that many hard drives, a few watts in either direction on the CPU will be about negligable). I don't have a clue what an averate datacenter's rack is going to hold, but I'm pretty confident that you won't need as many of these to hold the same amount of data.

What really perplexes me is why they didn't use 750GB drives instead and up the storage by 50%.

Re:cooling

[Anonymous Coward]

Nope, the systems has cable management arms and a slide rail system to allow it to be taken out to service point without killing the power.

It is a heavy system, but no heavier than, for example, a v1280 or v890 which also have slide rail systems.

Re:cooling

[Anonymous Coward]

LOL!!! So, with all your disks spinning away with critical data, you want to yank them around sliding them out of the rack? I think NOT! First rule of hard-drive use - don't move the box whilst the platter is spinning, and that even goes for shiney and modern SATA drives. You may have guessed by now that's why you usually have hot-swap disks in bays at the front and the back of a system so that you can get the disks in and out without having to disturb the rest of the system.....

Re:cooling

[Anonymous Coward]

????
Are the discs on seperate SATA chains then? Am I getting confused or would you have to disable all the discs on the same chain (presumabley the ones in the same row?) in order to get the one disc out?? Can you actually take the cover off with the system running??
I remember a sales guy from IBM telling me how the memory on the x440 was hotswap and I bought it, but the reality was you had to take the top level of memory out to get to the second layer, which meant stopping the whole system. Not appreciated by management!!

Re:cooling

[nessus42]

So, with all your disks spinning away with critical data, you want to yank them around sliding them out of the rack? I think NOT!

It's done all the time -- you just have to be careful. The Thumper chasis and rails is designed for this sort of thing.

|>oug

Ok. Now if I could just afford it.

[kabocox]

I would love to be able to spend $33K on that. I'd be lucky to get something in the $3-$4K range approved though. Do you have anything in that price range that I might actually get past my boss?

Re:Ok. Now if I could just afford it.

[bearl]

I don't work for these guys, but we've bought stuff from them before and they've been reasonable.

NAS servers

Nowhere near those Suns in capacity or performance, but they are less expensive.

Re:Ok. Now if I could just afford it.

[guillebot]

Take a look at: http://www.rackmountpro.com/. We have several systems from these people and they are great.

Re:Ok. Now if I could just afford it.

[DAldredge]

A few GPS trackers, film camera with zoom lens and some remote audio equiptment + time should allow you to get enought information on your boss so that getting anything approved will not be a problem.

Twitterpated.

[Tackhead]

> 24TB... thats almost enough to hold all my pr0n!

Orly Owl: Why, don't you know? He's twitterpated.
Thumper: Twitterpated?
Orly Owl: Yes. Nearly everybody gets twitterpated in the Thumper room. For example: You're walking along, minding your own business. You're looking neither to the left, nor to the right, when all of a sudden you run smack into a pretty rack holding 24 TBs of pretty racks! Woo-woo!

Software RAID only, plus 7200 RPM no10k or 15k

[belrick]

This is a more low-end or small-business grade disk configuration. Mid-range or high-end would havwe HW RAID. I've only seen 15,000 RPM in 146 GB disks and smaller, and 10,000 RPM in 300 GB disk, but I'd hoped the 250 GB option would have had 10,000 RPM disk.

Re:Software RAID only, plus 7200 RPM no10k or 15k

[ratboy666]

Actually, software RAID is an advantage, performance-wise.

The old-time "big-ticket" was checksum calculation, but that is now an "also-ran". Distributing the i/o? Software can do it as well as hardware.

Both hardware and software have to be familiar with the blocking factor.

Where software wins is that it can be aware of, and skip reading to fill blocks if the block has never been used (or is not PRESENTLY in use). Which hardware RAID controllers cannot avoid doing.

The idea is to tie the RAID more tightly into the filesystem.

As to lower speed drives -- did you count the heads? Each is active at the same time. Yes, an individual i/o would complete faster with 10k or 15k spin, but the total throughput is based on the number of heads. For RAID5, reading multiple blocks will give you pretty much all the read performance you can stomach.

Write performance for an individual write operation would be improved; but generally application buffering deals with it. The tradeoff is number of heads, spin rate, and heat. The right balance? For you, write performance up, and, keeping heat constant, number of heads down (I presume that you are dealing with transactional loads, with commits). For me? tends to go the other way (my workload is general storage, with a bit of database).

As always, YMMV
Ratboy

Re:Software RAID only, plus 7200 RPM no10k or 15k

[E-Lad]

This box is 100% designed to be used in mutual full advantage with ZFS. Thumper is what you would call a modern RAID array, as ZFS in this case blurs the destinction between hardware and software RAID. The CPU and memory horsepower is there for RAID-Z.

From this box, one can serve out file systems with NFS and/or SMB/CIFS (aka a traditional NAS), and in future releases of Solaris 10, also serve out LUNs over iSCSI and FCP while having all that data backed by the performance, reliability, and features of ZFS. The only thing it's missing is a consolidated, centralized CLI for manipulating storage, a la NetApp and ONTAP... but all the requisite pieces are there to turn Solaris, and especially Solaris-on-Thumper, into a NetApp killer at less cost.

Re:Software RAID only, plus 7200 RPM no10k or 15k

[ChrisA90278]

Of couse it is software RAID. Every single last RAID is software. OK you might think there is such thing as hardware raid but if you look at the controller card you will find some kind of computer and some RAID software running on it. The only difference is that the software is burned into ROM on the card. If you buy this RAID system from sun you will never see the dual Operon or have need to know what software runs on it. You should think of these two Operons are a very, very powerful controller card.

Re:Software RAID only, plus 7200 RPM no10k or 15k

[setantae]

Actually you will need to deal with the software, because it just runs Solaris.
Hell, you fanboys could probably throw that linucks thing on there if you like ;)

Re:Software RAID only, plus 7200 RPM no10k or 15k

[TinyManCan]

I 100% agree with you.

The other hidden advantage here is storage density. If for some reason you needed 1PB of data storage in as small a space as possible, this is a big win for you. You would need about 45 of these servers to get 1PB of capacity. That would fit nicely into less than 5 racks of space, with room to spare for your networking and monitoring gear. A 1PB EMC Symmetrix is going to be a _LOT_ bigger.

No other storage platform has higher density (that I am aware of). Power use is good but not amazing (look at Petabox) and price is excellent for the size, but loses out as you scale.

Overall, I am stoked on them and want to try using them as backup servers. Attach one or two LTO3s and a couple 10gbs ethernet cards and you have everything you need! You can spew data over the network from the clients and then spend the whole day making very good use of your tape drive resources.

Re:Software RAID only, plus 7200 RPM no10k or 15k

[dwater]

> As to lower speed drives -- did you count the heads? Each is active at the same time. Yes, an individual i/o would complete faster with 10k or 15k spin, but the total throughput is based on the number of heads.

Total throughput, as I understand it, is based on :

1) rpm - the faster a disk rotates, the more data goes under the head(s) per second
2) number of heads/surfaces - more heads can read more data concurrently
3) data density - the higher the density, the more data travels under the head per second

I guess this perpendicular tech adds a dimension to the density part too, but I'm not really up on that.

ATA over Eithernet

[bhima]

I am a worng thinking that AoE is a little more flexible / interesting than this?

Re:ATA over Eithernet

[DavidRawling]

Possibly, but I'm certain you could add iSCSI target software to this and you get the benefit of multiple spindles and SAN-like behaviour with cheap disks, clustering and shared storage, heterogenous host systems (Windows, Linux, Unix, etc), single point of backup ...

And here I was salivating over 24 spindles in 4RU ... damn!

Re:ATA over Eithernet

[E-Lad]

Sun is indeed currently working on a iSCSI Target (and possibly FCP target as well) server.

http://www.opensolaris.org/os/project/iscsitgt/

Re:ATA over Eithernet

[Anonymous Coward]

Yes AoE damage is great, but you should try to be a bit less WOW addicted and follow the discussion here : )

Re:ATA over Eithernet

[Octorian]

What, exactly, is the point of ATA-over-Ethernet? Honestly it sounds like a stupid idea, whose only merit is being implemented by someone and getting a Slashdot post about it. (oh, and by supporting Linux first)

If you want a real external storage solution, there are plenty of better technologies... If you must use Ethernet for your interconnect, just use iSCSI (who cares what the drives are). Remember, you're not having drives talking naked over the wire to the server anyways most of the time.

Wow

[bepolite]

If my math is right... that's 50,331,648MB / 295,734,134 (US Population) = 174.27683 kilobytes for every man woman and child in the US. In one box!

Re:Wow

[Anonymous Coward]

Wow, that much? 174.27683k should be enough for anyone.

Re:Wow

[Frightening]

Are you trying to (politely) say that we should welcome our new..never mind.

Re:Wow

[mobby_6kl]

174.27683 kilobytes for every man woman and child in the US. In one box!

Holy shit, this can easily be one 1024x768 jpeg image per person! If we automatically throw out the dudes and everyone under 18* or over 35**, there would be enough space for a small gallery!

--
*, ** - adjust to your preference; further filtering depends on available data

Re:Wow

[diskis]

How much space does the human genome take again?
In a few ten years you can fit every single american on a storage server.

(And archive it on antarctica :)

Re:Wow

[pimpimpim]

Still less than the 640 k that should be enough for everybody, though! But with four of these you might be ok :)

Re:Wow

[Hektor_Troy]

Well, $640k should be enough for any one person for a long time ... that'll buy you 10x24 TB of storage and 4x12 TB of storage for a grand total of 288 TB of storage (including the two racks needed) AND a two year service agreement. And that's list price.

288 TB of storage. Hrmm ... that's what? 5,760 filled to the brim Blue Ray discs (50 GB for dual layer). That's a pretty decent movie collection right there.

And if MY math is right

[jeffmeden]

296,344,308,438,456,234 * 349,000,000 = Who the hell cares about that statistic??? Seriously, lets compare the bit count per 1U to the number of chicken eggs laid per year in the US.

Variable redundancy?

[Jerk City Troll]

It would be nice if the system had a setting where you could transparently specify a redundancy factor in sacrifice of capasity. For example, I could set a ratio of 1:3 where each bit is stored on three separate disks. This ratio could increase to the number of disks in the system. And of course, little red lights appear on failed disks, at which point you simply swap it out and everything operates as if nothing happened (duH). Sure, we have a degree of this already, but managing redundant arrays is still a very manual process and when we start talking about tens or soon hundreds of terabytes, increased automation becomes a necessity.

Re:Variable redundancy?

[Anonymous Coward]

Check out ZFS-- http://www.opensolaris.org/os/community/zfs

It makes managing this sort of storage box a snap, and allows you to dial up or down the level of redundancy by using either mirroring (2-way, 3-way, or more) or RAIDZ. And soon, RAIDZ2.

Additionally, Solaris running on the machine has fault managment support for the drives, and can work with the SMART data to predict drive failures, and exposes the drives to inspection via IPMI and other management interfaces. Fault LEDs light when drives experience failures, making them a snap to find and replace.

Re:Variable redundancy?

[Wesley Felter]

ZFS can provide anywhere between 200% and 10% redundancy depending on what mode and stripe size you use. It should also automatically repair when failed disks are replaced.

Re:Variable redundancy?

[Anonymous Coward]

It would be nice if the system had a setting where you could transparently specify a redundancy factor in sacrifice of capasity. For example, I could set a ratio of 1:3 where each bit is stored on three separate disks. This ratio could increase to the number of disks in the system. And of course, little red lights appear on failed disks, at which point you simply swap it out and everything operates as if nothing happened (duH).

with ZFS it's as easy as that.

I saw a demonstration at LinuxTag in Wiesbaden, germany. The used files instead of harddisks, and just filled one of them with random bytes. Everything worked as if nothing had happend...

--
http://moritz.faui2k3.org/

Re:Variable redundancy?

[Trifthen]

And of course, little red lights appear on failed disks, at which point you simply swap it out and everything operates as if nothing happened...

I was thinking about that. With that case design, you have to pull the entire server and pop off the cover to yank one drive. I couldn't tell from the pictures how the enclosures worked, either; handles didn't seem evident. The idea is interesting, though. It almost looks like they could even squeeze it down to 2U with some creative cooling.

Since they have 12-rows of 4 drives each, I wonder if they could have mounted each column on rails so you could slide out the column of the offending drive and pop it out from the front. If they could manage to integrate the data cable into one or both of the rails, it would be much more accessible. Here's hoping for next-gen!

Re:Variable redundancy?

[C-Shalom]

I don't remember where in the specs I read it, but it is rail mounted and has a place for the cords when you push it back in.

Re:Variable redundancy?

[peterpi]

10%?

Is that a missing zero, or have I misunderstood something?

(I'm really not sniping, it sounds interesting)

Re:Variable redundancy?

[eclectus]

The drives do pop out the front. They are 2.5" sata drives.

Re:Variable redundancy?

[Wesley Felter]

The way I see it, an 11-disk RAIDZ stripe gives you 10% redundancy - 10 disks worth of data and 1 disk worth of parity.

Re:Variable redundancy?

[Tpenta]

In the current development kernel, Raid-Z2 gives you double redundancy. This should be in one of the forthcoming updates to Solaris 10 (it didn't make update 2).

Tp.

Re:Variable redundancy?

[ArsonSmith]

No, they all fall out the bottom when they are bad. they are all Blutooth controlled CF Card drives.

We are just making stuff up right?

Re:Variable redundancy?

[Cyno]

Standard RAID 5 XORs data between all disks, resulting in an automatic 1 disk of parity. Additional parity disks can be added easily. So unless ZFS offers live dynamic parity scaling or something new here, its no different than any other RAID..

The main benefits to having ZFS are its volume management and extended file/partition size limits. However, nobody I know has exceeded the ext3, NTFS, XFS, or HFS+ file or partition size limits. And volume management for workstations is a piece of cake.

At this time I could not see the benefit of ZFS being worth the additional cost of a Sun system. but 24TB in 4U? Now that's something to talk about.

Re:Variable redundancy?

[EndlessNameless]

I believe the percentage refers to number of bits providing redundancy relative to the original data set. Hence, RAID 5 would use approximately 10% more bits than the size of a to write that file to disk. This equates to using about three of 24 drives for parity data. Standard RAID 1 with two mirrored disks would be rate as being 100% redundant because it writes 100% more bits to disk than the file contains.

This is the oddest method of measuring redundancy that I've heard, but it's the only way I can read that post and have it make any sense. Personally, I usually measure redundancy the quick and dirty way... by saying what proportion of drives can fail before data loss occurs, which is what most people are interested in anyway. Knowing that you can lose 1/3 or 2/5 drives (or whatever the actual ratio is) gives them something concrete to chew on.

Re:Variable redundancy?

[Anonymous Coward]

Bad math... 10% from that perspective is more like 2 drives out of 24. 22 drives with an extra 2 for parity = approx 9.1% additional bits for each bit of data.

Re:Variable redundancy?

[Trifthen]

Right, the entire thing is rail mounted. You still have to pull out the entire enclosure to replace one drive. And it weighs 170 pounds... pulling out such a beast frequently doesn't seem very safe to me, personally.

Re:Variable redundancy?

[Paul Jakma]

I couldn't tell from the pictures how the enclosures worked, either; handles didn't seem evident.

They pop-up partially out of the bay. You then slide the bare drive in or out of the protruding enclosure - the disk cage/enclosure is actually part of the drive bay.

If they could manage to integrate the data cable

No cables. It uses a backplane with standard SATA+power connectors, the normal way for hot-swap drive bays.

Re:Variable redundancy?

[killjoe]

IT's only available as a part of solaris so I guess it's useless to me.

Re:Variable redundancy?

[sammy baby]

IT's only available as a part of solaris so I guess it's useless to me.

If Solaris isn't an option for you, I guess so. At least, right now. From the Wikipedia entry on ZFS:

Sun has said it is investigating the porting of ZFS to Linux, but there are no plans to port it to HP-UX or AIX.[4] Since the code for ZFS is open source, a port to other platforms can be produced without Sun's involvement. Matt Dillon from the DragonFly BSD project started porting ZFS to DragonFly BSD as a plan for their 1.5 release.[5]

It also appears that Apple is interested in porting ZFS to their Mac OS X operating system, according to a post by a Sun employee on the opensolaris.org zfs-discuss mailing list.[6]

Re:Variable redundancy?

[Trifthen]

Right, but I was speaking of my proposed design where each column pulls out of the chassis on a rail. In order to keep the drives live, you'd need to integrate cables into the rail to keep the backplanes active.

Kids these days... what ever happened to reading comprehension? ;)

Re:Variable redundancy?

[Anonymous Coward]

I like the transactional writes and the end-to-end checksums. The checksums let you actually tell when the data you've read from one side of a mirror or a RAID5 member is bad, and then to read or reconstruct a good copy. Traditional RAID arrays and software only protect you from complete hardware failure and will unknowingly return the bad data to an eagerly awaiting app. Few applications checksum everything they write and read from disk, let alone their own code.

Re:Variable redundancy?

[Basje]

Well, linux is out of the question, at least in the kernel.

If they argue over the new ReiserFS and the layer seperation, imagine the fight they would have this. It has to much layer integration for the kernel guru's tastes.

Re:Variable redundancy?

[Cyno]

Yes, the transactional writes are cool.. but the checksum stuff concerns me a bit, those extra CPU cycles aren't free..

Re:Variable redundancy?

[Anonymous Coward]

I doubt there's an excessive amount of cross layer code here, it's the tools zfs and zpool which makes it look and feel that way. zfs and zpool are fantastically well designed apps, as easy to use as anything I've experienced before.

Re:Variable redundancy?

[nessus42]

Sun claims that the CPU cost of ZFS is only a few percent of a modern CPU, even when going full-tilt. They claim that even if you use a compressed ZFS filesystem, you will often see higher disk bandwidth due to decreasing the disk bandwidth bottleneck.

|>oug

Other metrics...

[dreddnott]

And how much more quickly would 10,000RPM drives skeletonise a cow, anyways?

Pfft...

[Miguelito]

I saw at least 2 different companies offering almost identical ideas at least 2 years ago. Sure the total storage wasn't as high as the disks weren't as big yet, but the big 4U chassis with a ton of disk installed vertically isn't anything new.

Re:Pfft...

[spun]

You can also buy commodity 3U server chassis that hold 16 drives. We built a number of these as ROCKS cluster head nodes for Los Alamos National Labs. Two 3ware SATA raid cards running 8 drive RAID 5 arrays, bonded together in software as a RAID 0 array. Decent performance relatively inexpensively. Which is after all what the I in RAID is supposed to stand for. If you do this, get the SATA backplane that uses 4 Infiniband cables instead of 16 SATA cables and the cards that support that. I've done it both ways, and trust me, your knuckles will thank you for the four-fold reduction in cables. As an interesting aside, the chassis we used has a space up top for a 2.5" laptop hard drive to use as the system disk. It's is the only way to fit a system disk in that chassis.

Re:Pfft...

[pimpimpim]

so, you can answer the question: how hot does this all get? Can the disks really stay cool enough during full use? And what about the swappability of the disks, in front loaders it'll be easy to reach, but to remove these vertical ones it seems you'll have to do a lot of hassle to get them out (and hope they're not hot :) )

Re:Pfft...

[toQDuj]

the trouble with a homebrew system is that it is made using the cheap(est) parts. so no redundancy in your power supply, difficult to replace disks, no redundancy in the system disks. Additionally there's the limited raid volumes (your volume size is the number of data disks in one cabinet divided by the redundancy factor), and low network throughput.
Not to mention the non-redundant powersupply's that break once every six months and take down an entire shelf.
Not to mention the sheer time one needs to spend to build and moreover maintain it...

honestly, I think you're better off with one of these sun's.

B.

Re:Pfft...

[Miguelito]

the trouble with a homebrew system is that it is made using the cheap(est) parts. so no redundancy in your power supply, difficult to replace disks, no redundancy in the system disks. Additionally there's the limited raid volumes (your volume size is the number of data disks in one cabinet divided by the redundancy factor), and low network throughput.
Not to mention the non-redundant powersupply's that break once every six months and take down an entire shelf.
Not to mention the sheer time one needs to spend to build and moreover maintain it...

honestly, I think you're better off with one of these sun's.

Who said homebrew? There are other vendors out there that have had similar systems for at least 2 years. With redundant and hot swappable power, etc..

And frankly, my experience with Sun's amd based systems thus far has been that they're nowhere near as good (stability wise) as their SPARC line, only on par with other reputable system vendors. For the price, I can usually get more systems and the same stability using another vendor.

Re:Pfft...

[spun]

The chassis I worked with had three large and three small fans in the middle, right behind the drives, as well as three fans at the back and two in the PSU. You can blowdry your hair standing behind one. The 3U, 16 disk chassis I used had four rows of four disks mounted horizontally, and they were very easy to hot swap. The SATA backplane comes with signal lights that are compatible with the 3ware cards, having power, access, and error lights which made it really fun to watch: when there was a lot of disk access, it looked like the WOPR from Wargames.

Re:Pfft...

[spun]

Ha! These came with three hot swappable PSUs, and easy to replace disks with locking tabs. But you are right about the system disk, there is only room for one 2.5" hard drive, but the controllers we used do allow you to set up more than one RAID array, so you could easily use two of the 16 in a RAID 1 mirror if you wanted to. These were scientific computing clusters, so not mission critical, and our clients didn't mind the possibility of the system disk going caput. So you take the cluster down for an hour to replace it, no data is lost and the thing keeps going from where it left off when you reboot.

The RAID volumes are pretty damn huge, we routinely built 4 and 8 terrabyte head nodes, and network throughput?!? We used Infiniband for interprocess communication and GigE for control, so, um, no. We had very, very high throughput and lower latency than you've ever seen.

Did I mention the three hot swappable, rock solid PSUs? Oh, and it took me, bare metal with no motherboard to complete system with ROCKS installed and ready to go, about three hours to set one up, and the guys in Los Alamos National Labs could maintain these things in their sleep. Every piece of hardware was triple checked and completely Linux compatible, and we offered on-site replacement within 24 hours.

Honestly, Sun can go hang. I can build a box that's just as good with one hand tied behind my back, for half the price.

The return of Sparc Storage Array

[Wiseleo]

I have an SSA 1000 in storage with a matching SparcStation 10 that has a fibre channel host bus adapter. The SSA had 30 SCSI disks in groups of 10 and a fibre channel interface. Once upon a time it made for a lot of fun with doing some database performance modeling with 30 1GB SCSI drives. The size was roughly the same.

Blades are more important!

[Anonymous Coward]

I think the bigger news is the sun blade system. They were miising it for so long and now they have something to compete with IBM and HP blade servers. And although sun's low end servers are something similar to blades ($1300 per unit duh!) they're not quite the same.

Vista

[Kazrael]

I wonder if it's requirements meet Vista's ungodly needs...

Re:Vista

[Tandoori Haggis]

Well almost. Given the increase in data transfer since broadband became commonplace, 24TB is probably just the requirement for Vista's swap partition ;-)

sun infatuated with sw-raid ?

[TobiasS]

Why does sun hate hardware raid solutions ? I guess they dont want it to compete with their SAN'ish products. I just can't see using this thing for anything other than workgroup file servers and disk-disk backups. It will probably do fine on linear reads, can't imagine the random access and write performance to be all too hot.

Re:sun infatuated with sw-raid ?

[Anonymous Coward]

> Why does sun hate hardware raid solutions ?

Do you think hardware RAID controllers have their transistors configured for the purpose of RAID alone? Know what they're running? SOFTWARE.

This thing IS a raid controller.

Re:sun infatuated with sw-raid ?

[Anonymous Coward]

Why does sun hate hardware raid solutions?

Remember that "hardware" raid is simply a dedicated low-end microcontroller (compared to an Opteron) running pretty much the same software that "software" raid runs -- but in an environment that's harder to apply patches&bug fixes (need to re-flash the firmware).

This extra dedicated microcontroller is nice when it's in a server that's doing a lot of other CPU intensive stuff (rendering); but on a dedicated file server you're far better off having the powerful main CPUs performing the RAID logic == software RAID.

Re:sun infatuated with sw-raid ?

[SETIGuy]

Why does sun hate hardware raid solutions?

I can give you a few reasons they might. Having been through some hardware RAID nightmares I have first hand experience with a few of them.

HW RAID makes you dependent upon the manufacturer of the card both for RAID implementation and for drivers. We once a a couple hardware RAID cards managing a large (at the time) RAID0+1 array that would occasionally glitch and fail a drive or two (or occasionally every drive on the controller). The driver and monitoring daemon wouldn't report anything until a second drive failed. Despite battery backup on the card cache, a single drive failure would often corrupt the data on the mirrored drive. The manufacturer was nowhere to be found when requesting updates or bug fixes.

We eventually switched to software RAID and found that in addition to making the array reliable it improved our performance. This was in part because the 6 CPUs on the machine were significantly faster than the 25MHz i960 managing the RAID cards. We could also mirror across controllers on the 4 separate PCI busses which gets rid of a major bottleneck (the I/O on a PCI bus can be easily saturated by a few drives)).

There are other benefits to being able to RAID across controllers. A RAID controller is a single point of failure. If a controller fails on a HW raid system, your array goes down. On SW RAID (done properly) a single controller can go away without a problem.

The most reliable storage system we have (a Network Appliance rack) is entirely software RAID. (RAID 4, a number you don't hear often).

Considering....

[cbiltcliffe]

Considering you can get 750GB drives now, shouldn't this thing be currently capable of 36 TB raw capacity?

Re:Considering....

[DAldredge]

Sun has to have time to test and certify the drives - they don't run down to fries and buy the biggest drives they can buy. People who buy this sort of thing kind of look down on the use of untested/unproven drives...

Re:Ah, yes...a machine labelled "Sun"... appropria

[geoff lane]

The (redundent) power supply is rated at 1800Watts which implies about 6300BTU/Hr heat out of the box. For 24Tb and a server that is remarkably low.

Is it even possible (even on Slashdot)

[The_REAL_DZA]

to imagine a Beowulf cluster of these?

Sorry, couldn't resist; I'm usually about a day late for that particular well-worn meme.

Apple did it first.

[Anonymous Coward]

Like all other wicked developments in computing, Apple did it first. Welcome to the club, Sun.

Re:Apple did it first.

[DAldredge]

Name three (3) major things apple did first. Please note that first means no one else did them before apple.

Re:Apple did it first.

[Monster_Juice]

1) They got Windows to run on their hardware. Nevermind that was someone else.
2) Made a dumb looking all in one computer that was not portable. Nevermind...
3) Used a piece of fruit as a computer logo. I think we have a winner here.

Re:Apple did it first.

[Anonymous Coward]

Just because you hate Steve Jobs because you think he 'looks' Jewish does not mean that you have to keep up your apple suxors bullshit all the time.

Re:Apple did it first.

[DAldredge]

yawn...

Re:Apple did it first.

[good soldier svejk]

Name three (3) major things apple did first. Please note that first means no one else did them before apple.

I guess "major" is subjective, but my recollection is they were first to...

Release a PC with a color display
Release a PC with a floppy drive
Release an OS with pull down menus
Release an OS with drag n' drop file manipulation
Release an OS with a clipboard
Support forked/streamed files
Release a PC with a color display
Release a PC with built in four-voice sound
Release a PC with wireless networking

Re:Apple did it first.

[10Ghz]

"Release a PC with a floppy drive"

IBM PC shipped with a floppy before Apple shipped theirs. First PC to use 3.5" floppies was HP-150.

"Release a PC with a color display"

Wasn't Mac II the first Mac to ship with a color display? Mac II was introduced in 1987, long after computer with color had been available (Amiga 1000 for example.).

Re:Apple did it first.

[good soldier svejk]

IBM PC shipped with a floppy before Apple shipped theirs.

Apple released the DiskII in 1978, three years before IBM started making PCs.

Wasn't Mac II the first Mac to ship with a color display? Mac II was introduced in 1987, long after computer with color had been available

The Apple ][ had color graphics in 1977.

Re:Apple did it first.

[10Ghz]

"Apple released the DiskII in 1978, three years before IBM started making PCs. "

But their computer didn't ship with one. Therefore you can't claim that "Apple was first to ship PC with a floppy-drive", since it didn't ship with a floppy, it was an add-on product. If we talk of add-on products, then I feel compelled to remind you that IBM was making floppy-drives back in the sixties.

"The Apple ][ had color graphics in 1977."

I don't see a color-screen in the picture you linked to.

Re:Apple did it first.

[good soldier svejk]

"But their computer didn't ship with one. Therefore you can't claim that "Apple was first to ship PC with a floppy-drive", since it didn't ship with a floppy, it was an add-on product. If we talk of add-on products, then I feel compelled to remind you that IBM was making floppy-drives back in the sixties.."

I'm sure the compulsion is strong, but it isn't relevant. IBM made floppies in the sixties, but they didn't work with PCs (obvously). I didn't claim Apple was the first to ship a PC with a floppy drive standard. I said they were the first to ship a PC with a floppy drive. In other words, They were the first company from whom you could buy a PC with a floppy. They were the first company from whom you could buy a floppy drive which worked with a PC and they were the first company from whom you could buy a PC that worked with a floppy drive. This was an important advance and Apple did it first.

"I don't see a color-screen in the picture you linked to."

What's your point? The Apple ][ had color graphics. Look it up.

24TB for $70k (Sun) or 24TB for $16k (generic)

[linuxbaby]

We were waiting anxiously for this item to be announced, because we have about 100TB of storage (now) and add about 8TB per month. Perfect customer for these.

But, unfortunately, they're not quite as cheap as I had thought. (Friend on the inside thought Sun was going to price them at $1.25 per GB, not $2 per GB)

Instead, we've been using these. Very good cooling:
http://www.rackmountpro.com/productpage.php?prodid =2348

32 SATA-II 750g drives = 24TB, same as the Sun X4500, but for only $16,000 for the entire system (chassis, mobo, ram, drives) instead of $70,000 for the Sun Thumper. Huge difference especially if you're ordering many of them.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[JungleBoy]

I agree, this Sun box is way over priced. In May, I received a simliar box for our lab from Atipa. Dual opteron, 24 x SATA II 500G SATA. I'm testing it in a RAID 60 configuration right now. I'm pulling over 350 MB/s at the application level. I'm using a pair of Areca raid 6 controllers (with real Open Source kernel support, thanks Eric Chen!) and striping them together with mdadm. It's amazingly fast. And with 500GB platters, I'm relieved to have N+2 redundency.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[larien]

If you buy 10 at a time, it comes down to around $47k each (http://store.sun.com/CMTemplate/CEServlet?process =SunStore&cmdViewProduct_CP&catid=151017). Also, if you're paying list price on Sun kit, you're doing something wrong.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[Hektor_Troy]

Sure, but if you're ordering many (ie in 10's), you're paying $47,099.50 a piece. Still more expensive tough. But as I understand it you also get the entire rack as well (no clue how cheap that is though).

Also, the one you're linking to is a 7U unit, whereas Sun's is a 4U unit. IOW you can mount I think 6 units from Rackmount or 10 units from Sun, for 144 TB/rack vs 240 TB/rack. (That's with a 42U rack, which I believe is standard).

I won't get into anything wrt servicability, management etc., as I've absolutely no clue about that, nor do I know what clustering is like and bla bla bla.

The point is that while Sun's offering is more expensive/GB, it's also a lot more compact.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[Anonymous Coward]

8Tb / month of *what*? p0rn?

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[ahl_at_sun]

FWIW, ZFS has N+2 redundancy: http://blogs.sun.com/roller/page/ahl?entry=double_ parity_raid_z

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[illumin8]

Instead, we've been using these. Very good cooling:

Unfortunately, with a generic motherboard and an off-the-shelf SATA RAID controller, good luck fixing the thing when a drive fails. What's that? The RAID controller is reporting a bad drive, but you have no idea which drive it is because there's no way to light it up without shutting down the server and going into the RAID controller BIOS and telling it to flash the drive light?

Tough luck. There is a reason why Sun is a little more expensive: RAS. RAS is Sun's main hardware principle. It stands for Reliability, Availability, and Serviceability. Sun hardware is truly built with these concepts in mind. Concepts like: A failed component should trigger a visible alert (warning light), as well as a human readable syslog message that calls out the exact part that failed. You will never see these things in a self-built beige box without some serious hardware hacking on your own, and at that point, you might as well hire a team of EEs to reinvent the wheel.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[twitchingbug]

Just curious, how much would you expect to pay for it then if list isn't the real price?

-don

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[Zheng Yi Quan]

Thumper is indeed overpriced, list, but hardly anyone pays list. This is a problem Sun continues to fight--list pricing has little relation to what you pay for it on the street. There's no way it'll get down to $16k, but any competent Sun reseller should be able to get a fairly deep discount, especially three or four months from now when Sun's not trying to control demand.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[swordgeek]

Typically with big orders from big vendors, count on 40% off the list price without any hassle. If they're fighting to make the sale, it might be more than that.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[Anonymous Coward]

FWIW, Sun claimed over 5 times that level of performance (of course, overhead, etc.) for their box.

The nice thing about this is that you can get this kind of capacity from a Tier-1 OEM. Easier to enterprise purchasing departments than many of the smaller vendors, better support, engineering, etc.

Also Sun's list pricing is not like Dell's list pricing.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[Anonymous Coward]

Its good they went through all this trouble because their hardware blows up all the time, especially those #*#$!!)%*# chinese power supplies ! It looks like one of our brand new T2000's needs a new power supply. The damn thing was just jumped yesterday too. Bah.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[Anonymous Coward]

Google's approach is to use lots of cheaper parts -- so you're good even if a whole system catches fire. Just plug in another one somewhere else, and it'll automatically get replicated to it eventually. The filesystem will see what's dead (or giving read errors), and stop using it.

I'm surprised companies are still going for the centralized approach. Of course, I'm not really surprised that Sun is. They've not been terribly innovative in years.

As for price, an XServe-RAID is $13,000 for 7 TB, or $39K for 21TB -- just over half the price of Sun's. Apple's hardware would fill 9U, compared to Sun's 4U, so if you need to save 8 inches, I guess it's not an option.

Though Schwartz calls the Sun box "a thing of beauty", I'm not seeing it. It looks like a 30-year-old air conditioner. Sun products simply never have that elegance that SGI's or Apple's have had. I guess that's not necessarily a reason to avoid buying it, but it makes me wonder if he's lying by repeated assertion, or simply has no taste.

Re:24TB for $70k (Sun) or 24TB for $16k (generic)

[Anonymous Coward]

I take it you have no actual experience with the new Sun x64 machines? We don't have any of these x4500's yet, but if our experience with the x2100's is any indicator, I wouldn't trust my data to them.

The X2100 and Solaris 10 has been a huge disaster so far, and this is coming from a big Sun shop. Unreliable, unimplemented features, hardware failures, unknown crashes..

Congratulations, Jonathan!

[Reverend528]

You got your Blog linked on the front page of Slashdot. Now get your butt upstairs, Mom needs help with the dishes!

Re:ATA over Ethernet

[dreddnott]

I suspect that the X4500 is just a vehicle for Sun's much-vaunted ZFS filesystem (hence RAID-Z in the specs).

uh oh

[slashdotnickname]

I smell a lawsuit from Disney around the corner...

Re:uh oh

[Anonymous Coward]

A thumper is a device that moonshiners added to their stills that increased the quality by adding a second distillation. It was inline before the main condenser and it ended up holding fusils and other crud from boil-overs. This explanation makes more sense to me than a bunny.

ZFS

[XNormal]

This fits nicely with Sun's new ZFS file system.

ZFS blurs the traditional boundaries between volume management, RAID and file systems. All disks are added into one big pool that can be carved out into either the native ZFS filesystem format or virtual volumes that can be formatted as other filesystem formats. It has many other interesting features like instantaneous snapshots and copy-on-write clones.

Re:ZFS

[Cyno]

sounds complicated.. I'd rather stick with what I know. ;)

Re:ZFS

[Sloppy]

Bloody hell, that looks so cool! I don't give a damn about this piece of hardware; it's ZFS (which I hadn't heard about until today) that is giving me a nerd-woody. ;-)

Re:ZFS

[XNormal]

zfs is administered via the zpool(1m) and zfs(1m) command line utilities. They're quite a bit more friendly than most tools for managing raid, volumes and filesystems I've encoutered so far.

True size?

[drkfce]

Is this 24TB where 1TB = 1,000,000,000,000 bytes or 1,099,511,627,776 bytes? If it is the former, that sure cuts down on the true size.

Yes, but...

[Niebieski]

...how many times can it store the Library of Congress? 24TB sounds so abstract...

hardly...

[dreddnott]

ZFS is probably a step ahead of most hardware RAID solutions. I believe Ratboy666 wrote an excellent post above that details Sun's likely reasoning.

That brings me back...

[bennomatic]

I remember when I got my 1541-compatible MSD SuperDrive for my Commodore 64. Those 5 1/4" floppies held an amazing 170KB of data. That was like the equivalent of 20 5-minute tapes!

You Must Be New Here

[darthservo]

296,344,308,438,456,234 * 349,000,000 = Who the hell cares about that statistic??? Seriously, lets compare the bit count per 1U to the number of chicken eggs laid per year in the US.

Welcome to Slashdot...News for nerds

more insightful than funny

[Anonymous Coward]

Not that funny - that's pretty much what we're evaluating using one -- a record on near every person in a state - with hopes of going nationwide.

Two and a Half Libraries of Congress

[monopole]

Or the complete text content of the Library of Congress, coupled with 6 Academic Research Libraries, with the capacity to dump the equivalent of 2 pickup trucks worth of books every second . In a 4U rack. For the price of several cars. Now that's my type of bookshelf system!

Re:Two and a Half Libraries of Congress

[ThousandStars]

That's all very interesting, but how many VW Bugs is that? And can you convert it to my preferred metric of football fields?

Re:Two and a Half Libraries of Congress

[rts008]

I still prefer the "how many naked Japanese girls in a phonebooth" metric myself (http://nakedworldrecords.com/phone.htm) when doing in-depth research!

Huh?

[Anonymous Coward]

Its compact dual Opteron rack server, 4U high, packed with 48 SATA-II drives.

OK, what does its compact dual Opteron rack server, 4U high, packed with 48 SATA-II drives do?

Crazy

[Reality Master 101]

Yeh dern kids today are gawdamn spoiled. Back in mah day, we didn't have these FANcy tahrabyte arrays! My TRS-80 had 128K -- that's right, KAY-uh -- on a floppy! And the operating system took about 40K of that, leavin' me about 85K left! And I was happy to have it! I had tuh use a paper hole-puncher and cut a write-protect tab so I could flip the floppy over tuh get more space! Damn kids these days... -mumble- -grumble-

Re:Crazy

[sharkey]

Well OOOO-La-La Mr. Frenchman, with your fancy-schmancy 6 digit /. ID and your floppy disks. In MY day, we used paper tapes and used our hole punches to send men to the MOON! Damn kids.... GET OFF MY LAWN!





Where's my World's Fair spoon?

Re:Crazy

[Cid Highwind]

Well if it ain't mister Johnny-come-lately himself! In MY day, we wrote "code" in pencil, did computation on an abacus, and sent our cliche posts to slashdot by telegram. ...mumble mumble... MD5 sums by hand ...mumble... seige of Vicksburg ...mumble... TCP/IP over smoke signals... durn whipper-snappers

Beware of the 2.5" disk drives

[xenophrak]

I'm glad that they are at least offering a server in this class with 3.5" disks. The 2.5" 10K RPM SAS disks that are on the x4100 and x4200 are just junk pure and simple.

Re:Beware of the 2.5" disk drives

[bencc99]

The disks are ok - the controllers are a bit pap. It's the same in the T2000, sadly :(

See my benchmarks - I might get around to doing one of the X4100s we've got for comparison...

Re:Beware of the 2.5" disk drives

[h8sg8s]

When I saw my first 1.6 MB Diablo removable disk, the rumor was "they go bad, don't trust them. Not as reliable as drum memory.." When I saw my first CDC 300MB "washing machine" style disk (circa 1980) the rumor was "they crash - not as reliable as the Diablo.." When I saw my first Winchester (Fujitsu Eagle) the rumor was "they crash - not as reliable as the old CDC drives.." When I saw my first 5.25" Winchester I freaked. This couldn't be a good idea. I was wrong. When I saw my first 3.5" disk I thought "cute, but it'll never catch on." I was wrong. When I saw my first sub 1" high 3.5" disk, I wondered how it would perform. It blew me away. When I saw my first SFF drive in the Sun 4200 server I was cautiously optimistic. So far it's worked out great. Each one of these changes increased disk reliability and performance greatly. Change is good, conventional wisdom usually isn't.

Must ... fight ... compulsion...

[Anonymous Coward]

174.27683 kilobytes ought to be enough for everybody.

You call 24 TB storage? Bah!

[Sodki]

Thumper delivers forty-eight HDDs with 24 TB of raw storage. And it will double within the year, when 1TB drives will be sold.

Probably they'll just stick a Bacterial DVD on it.

Indeed, Sun's list prices are way too high

[this great guy]

The 12 TB config is sold at $33k, or $2.75/GB, but assembling such a server yourself is possible and can be done today for 1/3rd of this price:

• 1 x dual-Opteron mobo = 1 x $500
• 2 x Opteron 285 = 2 x $1100
• 8 x 2 GB DDR400 registred DIMM = 8 x $300
• 6 x 8-port PCI-X Marvell SATA card = 6 x $100
• 48 x 250 GB 7.2kRPM SATA disks = 48 x $110
• 1 x Chassis+PSU+Rails = 1 x $1000
• Total = $11980 or $1.00/GB

(I have actually slightly overestimated the above prices.) Of course people are going to say that such a server is not be as reliable as a Sun server, that it does not come with technical support, etc. But in most cases such arguments are invalid because you save so much money that you can afford assembling/maintaining the server and replacing faulty hardware parts yourself. Time is money, but by having saved money you can now afford time ;-) The living proof that such a model would be successful is Google: instead of buying Sun servers like most startups in their time, they built their servers themselves to save money.

Re:Indeed, Sun's list prices are way too high

[Anonymous Coward]

I can tell you, from first-hand experience, that only one who has too much time on his/her hands, or doesn't actually work for a company where data-operations are critical would propose such a thing. I built just such a dula-opteron/high-density storage white-box that I run ... and it is a *nightmare* of unreliability, etc. Yeah, it was dirt cheap, and yeah, time is money, but my time is worth more than this ...

Re:Indeed, Sun's list prices are way too high

[setantae]

So you're keeping those disks in a bucket and cooling them with slave-girl driven fans or something?

Re:Indeed, Sun's list prices are way too high

[oogbla]

I think you're missing one thing. Where would all the drives go? On the floor? Suspended in mid-air? I'd like to see you get a Chassis+PSU+rails for $1000 that holds not only your Opteron motherboard, but all 48 disks as well. Plus, with that many drives, cooling, a *real* power supply is required (at 15W per drive, that's 720W right there, plus the Opterons, memory, fans, etc. and you're talking about 1100W - not your average power supply).

Another problem is vibration. If you don't have a good mounting scheme for all these disks, cross-drive vibrational issues will adversly affect not only performance, but MTBF as well.

Lastly, what about performance? I've seen this machine sustain raw access to the disks at 3GB/s.
That's *bytes*. Through the filesystem (ZFS), you get close to 2GB/s if you're careful. The machine has 10 fully-independant PCI busses inside - not a bottleneck in sight. Let's see the PCI bridge of your $500 mobo take that.

Once you do all of this, you're not $1/GB anymore, you don't fit in 4RU anymore, and you certainly
won't get the same performance. So I think that to build a similar box, there's no way you can
significantly beat the price. Plus, you have to remember that almost nobody pays Sun's list price.
Most VARs that sell Sun gear will give you a good discount. Comparing Sun list price to We-won't-be-here-next-week computers is not a valid comparison, either.

Re:Indeed, Sun's list prices are way too high

[_damnit_]

Just curious how you are going to hook up those 48 SATA drives to your 6 8-port SATA hbas? Where are you going to find a $1000 chassis that fits 48 drives? In one MAINTAINABLE configuration? As far as I know (and could be wrong) SATA is not an external bus. The SATA cards you mentioned would have to run outside of the box to another unless you find that 48 drive chassis I mentioned. Even if you ran a long enough wire to get to the other box mounted above your standard opteron 1U box, there's a lot of slack that has to be on that connection unless you want to disconnect everything just to pull the server out of the rack for maintenance. There are limitations to the SATA cabling you're not taking into account. Also add a couple more power supplies on here for each of the boxes that hold you drives. Cooling is also an issue that tier 1 vendors model very seriously before they put together a kit. Most home baked kits have either dangerous hotspots that effect reliability or are overcooled which wastes money. You should also keep your drives mounted with dampening to avoid vibrations from each other which can cause early drive failure.

There's more to this than simply buying parts. This appears to be another viable option in the storage arena for apps that need very large local storage. The problem with using it for NAS storage is that Solaris has historically been pretty slow compared to NetApp. ZFS could improve the score here with simplified administration if anyone actually understood how it worked.

My $.02

Re:Indeed, Sun's list prices are way too high

[rainer_d]

> Where would all the drives go?

Here:

http://www.rackmountnet.com/Rackmount-Chassis/5U-R ackmount-Chassis/5U-Rackmount-Chassis-48-Multilane -SATA-hot-swap-34-depth-1350W-redundant-power-supp ly-RMC5D/

But you need 1 RU more. It's 5 RUs.
Amazing that SUN (or Andy B.'s Keralia) was able to cramp it into 4 RUs.

Re:Indeed, Sun's list prices are way too high

[tntguy]

I don't know about you, but I like when "the boss" can call the support number, rather than mine, when I'm on vacation.

Re:Indeed, Sun's list prices are way too high

[mieses]

i'd rather take an extra week of vacation than pay an extra $10K for Sun stickers on Seagate hard drives.

Re:Indeed, Sun's list prices are way too high

[mieses]

or here:

http://www.nexsan.com/products/products/satabeast/ satabeast.html

(only 42 drives. but it's a 4U and the drives appear to be more evenly spaced than in the generic 5u rackmount.)

Re:Indeed, Sun's list prices are way too high

[this great guy]

There are 2 kinds of IT engineers:

• Those who work 70-hour weeks for the first 6-8 years of their professional life, building custom-designed IT infrastructures out of cheap hardware parts, designing reliable software assuming servers will fail, saving tons of money for their company, accomplishing projects everyone else thought were impossible before, and then retiring in their early 30s, rich and most importantly ready to fully enjoy the rest of their life with friends & family members.
• Those who work 9-to-5 in a cubicle, wasting the company's money by making poor decisions and buying expensive servers from common hardware vendors, constantly looking for the least amount of technical responsibilities, blaming everybody else for the failure (or lack of success) of their company, and finally retiring in their 60s, tired by years of a boring life.

I choosed to belong to 1st one. Sorry to hear you saying I made a crazy choice.

Re:Indeed, Sun's list prices are way too high

[numbski]

I think the problem here is the rule of diminishing returns. There are very few people that would be able to garner full benefit of this configuration. There are large corporations that could, sure, and perhaps a few years down the road individuals would as well.

Another diminishing return is the benefit of cramming that many drives into 4u of space. I've recently had the benefit of trying out ATA over Ethernet and a company named Coraid sells what they call an SR1520. It's a 15 drive unit with a SATA backplane that fits into 3u, and then you can either buy their little 1u debian box (CLN20) to act as a NAS or BYO. That's 4u with 15 drives, decent performance, and the cost of buying from then runs ~$8000 for a full setup, and you get 7.5TB of raw storage, and all you need is a layer 2 switch, no fiber channel to mess with, and it's not over IP.

You could build that same unit on your own if you felt like hacking it together for about $2000. The neat thing about ATA over Ethernet is that you use LVM2 and you can just keep slapping more AoE drives into the setup. So let's say you had the cash to toss around, you get 3 of these SR1520's, and you take up 9u instead of 4u, get the same storage benefits. Downside is that you eat up more rack space and the whole setup eats more power. Again, rule of diminishing returns...how much is that dollar to you vs. the rack space and power requirements involved?

I could see considering using this, but I would prefer that at very least the box run linux or freebsd to keep our environment uniform. Adding yet another *nix to the mix adds another layer of management responsibility I'd like to avoid. (No flame wars, please.)

Re:Indeed, Sun's list prices are way too high

[Anonymous Coward]

Pure. Fucking. Genius. All of the SysAdmins I respect technically are more than capable of building a simple system like this. One that would be a perfect fit for any application this piece of gear might be good for. Most likely the one-time price would be cut in half. And there's little doubt that you maintenance costs would be ~10% of what Sun's contract run you over the lifetime of the unit.

This is for purchase by people who are not spending their own money, or people too incompetent or lazy to build an economically efficient solution themselves.

But that's why Sun is still around. Because of it's young nature the field of "IT" is _flooded_ with workers who would not get an interview if they were proportionately skilled 20 years from now.

Re:Indeed, Sun's list prices are way too high

[nacturation]

Nice false dilemma. You forgot the third kind:

• Those who work hard at a company (whether it has cubicles or not) making intelligent buying decisions for commodity products from established vendors rather than wasting their time reinventing the wheel, so that they can focus their time on the challenging work which will set their company apart, making tons of money for their company, accomplishing projects everyone else thought were impossible before, etc. etc..

Anyways, I'm glad to hear that you've done well with your life. Kudos... I'd like to hear more about it sometime.

Re:Indeed, Sun's list prices are way too high

[stonecypher]

Where are you going to find a $1000 chassis that fits 48 drives?

A 3U chassis that handles 11 drives is currently $140 on PriceWatch. Do the math.

As far as I know (and could be wrong) SATA is not an external bus

And, in fact, you are. First hit for external SATA. Rule of thumb: when saying something you don't actually know, it is no longer appropriate to say "I could be wrong." Now, you just check. By the by, I own a SIIG SATA card with both internal and external plugs. It cost me a whopping $45 at Fry's.

The SATA cards you mentioned would have to run outside of the box to another unless you find that 48 drive chassis I mentioned.

Never run a datacenter, have we?

There are limitations to the SATA cabling you're not taking into account.

Yeah, cabling arbitrary lengths of drives together has been easy since SCSI2 Fast Ultrawide. What you do is use seperate cables every few drives. Magic.

Also add a couple more power supplies on here for each of the boxes that hold you drives.

Depending on the drives, you can expect 48 drives to cost between 600-700 watts. There are 1000w power supplies sitting in your local CompUSA right now.

Cooling is also an issue that tier 1 vendors model very seriously before they put together a kit.

Cooling is a problem for CPUs primarily, not drives. Cooling large blocks of drives is relatively easy.

Most home baked kits have either dangerous hotspots that effect reliability or are overcooled which wastes money.

1) Affect, not effect.
2) Cooling a 48-drive box is going to cost less electricity than running a single CPU. These people put down a thousand bucks a month just for the privelege of being in a controlled room. Let's have a sense of scale for things, please: fans just aren't that much power.

You should also keep your drives mounted with dampening to avoid vibrations from each other which can cause early drive failure.

Wow, so you get wide mountings, and put silicone glue in for the inner rails. That's gonna cost like two dollars in caulking and maybe ten in rails and screws. Next?

There's more to this than simply buying parts.

Not really.

The problem with using it for NAS storage is that Solaris has historically been pretty slow compared to NetApp.

Er, speed is one of Solaris' big selling points, if you'd actually look before announcing.

ZFS could improve the score here with simplified administration if anyone actually understood how it worked.

Yeah, uh, ZFS takes like five minutes to set up. It's trivially simple. Why would you pretend otherwise? Have you even touched it?

Re:Indeed, Sun's list prices are way too high

[Anonymous Coward]

Ja, but the last price I've heard thrown around for these is around $1.75/GB. You'll easily hit > $2/GB when you throw in the host, FC adapters & infrastructure, and the performance is significantly less. (800MB/s max v 2GB/s).

Re:Indeed, Sun's list prices are way too high

[elmegil]

A 3U chassis that handles 11 drives is currently $140 on PriceWatch. Do the math.

Because, after all, 4 x 11 = 44...but wait, isn't that 12 RUs instead of 4?

Get real. This is heavily engineered; no, it's not for the average slashdotter surfing porn in their mother's basement, so fucking what?

Re:Indeed, Sun's list prices are way too high

[Anonymous Coward]

Driveby comment:

A 3U chassis that handles 11 drives is currently $140 on PriceWatch. Do the math.
lol, 15U+ v. 4U. also, link for a 11 drive SATA external enclosure for $140? Your other link indicates significantly more expensive.

I own a SIIG SATA card with both internal and external plugs. It cost me a whopping $45 at Fry's.
I personally hate SIIG chips; never had a good experience w/ them, but YMMV. Also, link? I'd like to see if this card could keep up.

These people put down a thousand bucks a month just for the privelege of being in a controlled room. Let's have a sense of scale for things, please: fans just aren't that much power.
When you're billed per U, 16U v. 4U makes a big difference. More importantly, 40U of this ~~ 160U of commodity hardware.

This box has it's place in the market, as does the hand-rolled system approach. Different segments for different folks.

Re:Indeed, Sun's list prices are way too high

[_damnit_]

A 3U chassis that handles 11 drives is currently $140 on PriceWatch. Do the math.

OK. Your works out to be 3x larger than the 4U Sun's comes in. Did you have some other math for me to do?

As far as External SATA, I was wrong. I'm OK with that sometimes in a congenial conversation. It's little different to say "I may be wrong" than posting with IANAL. This is /. It's a discussion forum not a whitepaper or thesis. Thank you for correcting me and providing useful information to everyone.

Yeah, cabling arbitrary lengths of drives together has been easy since SCSI2 Fast Ultrawide. What you do is use seperate cables every few drives. Magic.

Yes, that'll look nice out the back of the rack. How many cables?

put silicone glue in for the inner rails

You have a lot of spare time.

Er, speed is one of Solaris' big selling points, if you'd actually look before announcing.

Well, I worked at Sun for 7 years and know a few things about Solaris. If you think Solaris is fast at NFS compared to NetApp, well I beg to differ.

Yeah, uh, ZFS takes like five minutes to set up. It's trivially simple. Why would you pretend otherwise? Have you even touched it?

Yes, I spent two years explaining and demonstrating it to customers. It's a huge change and takes a big mindset change for those who are used to Vxvm or other LVMs.

Also, if your going to nitpick the use of effect/affect:

1) Affect, not effect.
2) Cooling a 48-drive box is going to cost less electricity than running a single CPU. These people put down a thousand bucks a month just for the privelege of being in a controlled room. Let's have a sense of scale for things, please: fans just aren't that much power.

At least spell "privilege" right in your nit-picking.

Also, if we're going to have a "sense of scale" here, there is a huge cost difference between 12 RU and 4 RU in datacenter costs.

Re:Indeed, Sun's list prices are way too high

[slashdot-me]

I'm actually in the process of building a cheap-ass file server. It's a dual opteron with 2 GB ram, twelve 400 GB disks, cheap marvell controllers, and diy drive mounts. Price was around $10k--roughly $2.25/GB. Cost per gig will decrease as I add more disks. It'll eventually hold 32 or 40 disks.

Since the core machine is a Sun X4200 running Solaris and ZFS, it's pretty comparable to Thumper, just a bit smaller. I'm still waiting on parts so I haven't any performance figures. I expect it to be pretty zippy.

Re:Indeed, Sun's list prices are way too high

[stonecypher]

Uh. Don't get angry until you read context. I didn't say they were the same thing at all. I simply responded to grandparent suggesting that you needed to spend $1000 on the chassis to mount 48 drives. Also, I'm not sure where you get 4x11.

Get angry on your own time. I wasn't talking to you, and I don't need you talking down to me when you don't even realize why I said what I did.

Re:Indeed, Sun's list prices are way too high

[stonecypher]

Yes, that'll look nice out the back of the rack. How many cables?

Well, the whole reason you'd be dealing with 3x larger was to make them all internal. And, depending on your arrangement, probably 16 total. If you're willing to deal with external enclosures, you can knock the whole thing down to 5U. Besides, take a look in a datacenter some time. You'll find the average machine has a dozen or more cables sticking out of it.

put silicone glue in for the inner rails You have a lot of spare time.

It takes like five seconds per screw. To knock $48k down to about $5k, yeah, I've got a few minutes. Bet you would too.

Er, speed is one of Solaris' big selling points, if you'd actually look before announcing. Well, I worked at Sun for 7 years and know a few things about Solaris. If you think Solaris is fast at NFS compared to NetApp, well I beg to differ.

I'm glad to know you're a former Sun engineer; otherwise my reply would be long-winded and probably seem condescending. Basically, the issue here isn't one of OS speed. The issue is of disk speed. The reason ZFS is near-platter speed is that almost all of the work is amortized into the write phase; it's a ridiculously read-focussed scheme, which is appropriate in the rare situations where it's deployed.

Granted that Solaris itself isn't the speed champion it once was, the throttlepoint is the disk subsystem, not the OS/CPU. Sure, if you're doing heavy simulation or something else machine-intensive, that might be a problem. But, in context it seems like we're talking about a data store, and the situations in which a data store has significant non-disk responsibilities are vanishingly rare.

Yeah, uh, ZFS takes like five minutes to set up. It's trivially simple. Why would you pretend otherwise? Have you even touched it?

Yes, I spent two years explaining and demonstrating it to customers. It's a huge change and takes a big mindset change for those who are used to Vxvm or other LVMs.

Well, my apologies; I had not previously realized your familiarity. That said, I'm not sure why that mindset change would come in, specifically from the perspective of an external data store. Perhaps you could enlighten me? Realizing your former position, I now wonder whether I'm missing something, or whether I just took for granted something customers often don't understand.

At least spell "privilege" right in your nit-picking.

I deserved that. Touché.

Also, if we're going to have a "sense of scale" here, there is a huge cost difference between 12 RU and 4 RU in datacenter costs.

Well, yes and no. 4U goes for $50/mo these days from a place like ColoPronto (it may be cheaper elsewhere - I haven't looked in a long time.) So, sure, you're looking at an extra $100/mo. On the other hand, the Thumper fully built out is $48k, whereas the equivalent machine stitched together is about $11k. So, you're looking at about a $37k difference, which at $100/mo is almost 30 years of datacenter rental to break positive.

I'm not saying the Thumper isn't worth it - it damned well is, and Sun isn't stupid. They wouldn't price it up there if it wasn't worth that. The reason isn't the cost of rental, though. The issue is paying the time for someone to put stuff together, to maintain stuff, to handle keeping the machines playing well together, all that jazz. For the little guy (and let's not kid ourselves, that means startups, and startups need these things too,) the Thumper isn't the right way to go. Too much cost up-front, and all the startup kids are dramatically underpaid anyway, so getting them to maintain the box is just cost effective.

That said, that's not Sun's customer base; those kids are usually using BSD or Loonix instead. Sun's customer base are corporations who need the datacenter today or they lose $12k/day. Sun's customer base are people who don't have time to screw with making the server work, because

Re:Indeed, Sun's list prices are way too high

[this great guy]

The disks would go in the chassis (see my itemized list). You may not know it but Sun is not the 1st company to use a chassis with vertical bays. Here is one example among many. The price would be more likely around 2 or 3 grands by the way, instead of 1 grand. But anyway this doesn't change the fact that this Sun box is way overpriced, even with a good 40% discount.

Regarding the mobo, just pick one with two AMD 8131 or 8132 PCI-X bridges. This will give you 4 independent PCI-X busses. The two PCI-X bridges would have to be on 2 different HT links in order to not dangerously approach the theoretical one-way data throughput limit of 3.2 GB/s of one 1600 MT/s 16 bits HT link. The two PCI-X bridges could be either connected to different CPUs or to the same CPU because the Opteron XBAR _can_ easily handle the ~3 GB/s you speak about, it has been designed to support 19.2 GB/s of HT traffic and even more with the recent upgrade to 2000 MT/s ccHT links. Now with the 4 independent PCI-X busses, you could put 4 SATA HBAs on the 1st and 2nd busses, and 2 HBAs on the 3rd and 4th busses. This way the first 2 busses will run at 100 MHz and the 2 others will run at 133 MHz, giving a practical throughput of 3.4 GB/s (2 * (100 MHz * 64 bits / 8) + 2 * (133 MHz * 64 bits / 8), and assuming a 90% efficiency as found on most PCI/PCI-X busses), this is enough to handle the 3 GB/s you are speaking about. There are plenty of single AMD 8131 mobo on the market right now starting at $250. I am sure you can find one with two AMD 813x for $500 max.

Now when I think about it you could even use SATA port multipliers in order to reduce the number of HBAs, allowing all busses to run at 133 MHz. I am aware of 12-port and 24-port SATA HBAs (Areca comes to mind) but those are outrageously expensive and are not necessary to handle all that throughput. My experience and those of my friends playing with high-end enterprise gear prove that _very_ simple and inexpensive PCI-X SATA chips such as the SII3124 or Marvell 88SXxxxx are way sufficient to handle the max combined read throughput of any number of disks attached to their SATA ports. The reason being that the designers of such chips have come up with a simple and performant hardware interface optimized to reduce the CPU load. I know for a fact that the SII3124 design is somewhat close to the AHCI spec which is the best example of a performant SATA hw interface.

So I _do_ believe that it is possible to build a $13-14k server with 48 SATA disks in 4U offering ~3 GB/s of raw read throughput. I don't understand why so many people refuse to believe that, especially since other posters in this thread have pointed out that some vendors are already selling similarly priced servers !

Re:Indeed, Sun's list prices are way too high

[XeroDegrees]

this configuration==$0.94cents per GB or $1.10 inc vat

Re:Indeed, Sun's list prices are way too high

[XeroDegrees]

oops that should be $1.73/gig ex vat and $2.03 inc vat. Mix up between DNUK, XE, Slashdot tabs and a calculator window, also that config is for 18TB not 12TB

Re:Indeed, Sun's list prices are way too high

[htd2]

Try buying a similar set of components off the shelf from a major systems vendor such as HP or IBM.

IBM and HP want ~14-15K for a 2U 2x 2.6Ghz dual core Opteron 285 server fully loaded with memory and buying the 500 GB SATA II drives bare from them will cost you close to $700 per drive. The HP DL585 which is a similar size to Thumper though it cannot hold more than 4 drives comes in at $20K with 2 x dual core Opteron 285's 16 GB of RAM and 2 internal drives.

So for ~$47K you would have yourself a 2 CPU 2U server and 46 disks sitting on a table or for $52K you could have a 2 CPU 4U server with 44 drives sitting on the table.

Sun's prices would seem to be competitive when compared with other major systems vendors and that is probably what matters to the bulk of their customers.

Re:Indeed, Sun's list prices are way too high

[_damnit_]

Well, I enjoyed this conversation. I don't believe I will ever build a box as you have described, but should you choose to more power to ya.

I'm not sure why that mindset change would come in, specifically from the perspective of an external data store.

It doesn't matter to a buyer if the LVM is hidden from the NFS hosts. Someone still has to maintain that box. They want to know what's involved and if they have to train their sysadmins on another product.

Really, the hardest part about selling something as cool as ZFS is that it sounds like bullshit. Sun comes off sounding like a snakeoil salesman with a panacea for your storage problems. Plus, Sun doesn't have a great track record of getting filesystems to stick in the realworld over the last 8 years. They bought SAM-FS and QFS which both are merely obscure technical knowledge now for me.

Re:Indeed, Sun's list prices are way too high

[1lus10n]

no enterprise warranty. Certainly not on site 2 year. Which is included in the price of the system.

Next, just find me a freaking chasis that will hold 48 hdd in 4U of space for 1k.

Re:Indeed, Sun's list prices are way too high

[this great guy]

If you think that an on-site 2-year warranty on a 4U box and a top-of-the-line chassis are worth Sun's price ($33k, or $20k assuming a 40% discount), then you are gullible and uninformed of the real value of such things. Proof ? This guy bought a similar server from a smaller vendor, with twice the storage capacity at half the price of Sun. Guess that you could throw in any expensive warranty and still be well below Sun's prices.

Re:Indeed, Sun's list prices are way too high

[Anonymous Coward]

Well, yes and no. 4U goes for $50/mo these days from a place like ColoPronto (it may be cheaper elsewhere - I haven't looked in a long time.) So, sure, you're looking at an extra $100/mo. On the other hand, the Thumper fully built out is $48k, whereas the equivalent machine stitched together is about $11k. So, you're looking at about a $37k difference, which at $100/mo is almost 30 years of datacenter rental to break positive.

The company I work for has recently built four large datacenters. For true enterprise data centers, it cost us about $5k per usable RU to design and build the facilities. So, if I can save 8U that saves me about $40k.

Re:Indeed, Sun's list prices are way too high

[1lus10n]

an expensive warranty from WHO ? I doubt the smaller vendor can afford to have 24x7 next day onsite support. You think you can find a reputible third party that provides such coverage for under 20k your nuts. A chasis such as that is going to run you several grand (5k +) EASY. If you can even find one.

Next lets factor in the comparison. First it doesnt hold as much data (from the vendor: "2 + 2 HDD bays, supporting up to 16TB capacity."). Second it takes up (almost) twice as much space. Space is money, cooling is money, power is money and density of data or cpu's is money. And it doesnt even mention warranty or provided OS. You want enterprise OS's on something like this, thats some more cash right there.

Re:Indeed, Sun's list prices are way too high

[this great guy]

An expensive warranty that smaller vendors COULD provide (but most DON'T, my point was to show the price difference). And they don't because clients who buy hardware from small vendors know that even the most advanced and expensive vendor warranties and technical support contracts cannot match the efficiency, cost and speed of interventions performed by technicians hired by a company to maintain its own infrastructure.

Real world examples: UPS don't use external services to fix their vehicles, they hire mechanics and run their own repair shops. Google don't buy support contracts from Sun/HP/Dell, they hire technicians and engineers and maintain their own servers.

Imaginary example: company A buys 10 of those Sun servers with a 2-year warranty (so $20k each), company B buys 5 cheap servers with twice the storage capacity ($16k each) and hires a technician for 2 years ($40k/year). Over 2 years company B has saved $40k (do the math), plus it has the advantage of having the technician on-site (faster interventions than Sun), and he would be able to maintain other servers at no additional cost for you (a full-time technician maintaining only 5 servers would have a lot of spare time).

Is it so hard to understand ? Profitable companies prove that it is the right way to do it (UPS, Google...).

Re:Indeed, Sun's list prices are way too high

[1lus10n]

Yeah the problem with that scenario is that both UPS and Google take advantage of warranties, and have the size to not have to worry about downtime. They can afford to buy backups of the backups. Most companies cannot.

What your analysis also fails to take into account is any sort of failure. How much money do you think it would cost a mid-sized insurance or loan company if they had the backplane fail on a system like this ? What technicians do you know who can re-work a motherboard ? I have worked for two of the 5 largest computer companies in the world and even they have only a scarce few who can do work like that.

Profitable companies eh ? I could list off HUNDREDS of companies I have dealt with on multiple levels that not only use those warranties but have site wide support contracts because even though it costs millions it could save many many times that. Sun for example has* platinum contracts that guarantee the system will be back up and running within 3 hours regardless of the problem.

Not to mention, but why would they buy more sun boxes when they hold more data than the smaller vendors (16TB vs 24TB source: vendors), take up less space, use less power per GB and so on. The sun boxes are more efficient (a fucking HUGE reason for an enterprise that would buy dozens or even hundreds of these).

Google and UPS are horrid analogies. UPS is automobiles, hardly an effective example since there is no "enterprise" car company. Google is equally has bad if for no other reason than the scale. They have by some estimates in excess of 50k computers, cheap crap that fails constantly ... but it doesnt matter because they are building purely for redundancy and scalability. A luxury that most companies cant afford and few shareholders would approve of.

Here is a list of companies that use enterprise storage solutions because its the safest way to spend there money: Every financial institution in the world. Every government lab in the country. A few large energy companies, a few large animation studios, a few car companies, a few large insurance companies. Thats just off the top of my head, I am not going to discuss specifics. Mostly for the privacy of those locations.

*= at least they used to, not sure if they still do.

Re:Indeed, Sun's list prices are way too high

[this great guy]

They can afford to buy backups of the backups.

Exactly ! This is what I wanted you to say. They can afford it in part *because* they maintain their infrastructure internally, don't buy expensive hardware/parts and save money. Look at my company B example, they have $40k left to buy backup servers, spare parts, etc.

What your analysis also fails to take into account is any sort of failure.

I agree that correctly handling failure is a very important point in order to build a reliable infrastructure out of cheap hardware. It requires time & effort. But IMHO there are plenty of cases where people say it is very hard where in fact it's perfectly doable, and would be cheaper in the long run. I think this way of designing complex IT architectures is going to explode in the near future (10 years). This is also why the debate is so intense in this thread. This is a new way of doing things, this scares people who were used to buy trusted/expensive servers up to now. Google is the clear pioneer in this field.

What technicians do you know who can re-work a motherboard ?

Not many. But look at what Google would do: trash it and buy a new one. This is the advantage of commodity parts, just identify the failing component (mobo, memory, cpu, NIC, switch...) and replace it.

Not to mention, but why would they buy more sun boxes when they hold more data than the smaller vendors (16TB vs 24TB source: vendors), take up less space, use less power per GB and so on. The sun boxes are more efficient (a fucking HUGE reason for an enterprise that would buy dozens or even hundreds of these).

No. All you need is a 48-bay 4U chassis similar to Sun's to match their storage density. So technically speaking a high-end Sun server would have the same storage density and watt/GB consumption than a DIY server.

[Google has] by some estimates in excess of 50k computers.

Actually the estimates are between 100k and 500k. But my point still holds: when they started, with only 100 machines, 1k, 10k, etc they always kept using commodity parts, no matter how large their datacenters were.

Re:Indeed, Sun's list prices are way too high

[1lus10n]

Okay. Find me a 4U 48 bay chasis. Make sure it has hot swap power supplies, 1500W each. You know why they kept using commodity parts ? Because Sun declined to build what they wanted.Google has over 1000 technicians, they hire large amounts of people just to handle hardware failures and expansion. They spend more on infrastructure than half the fortune five hundred.

It would be awesome if most companies would spend the money on the tech people, but they dont. They spend it on sales and marketing.

Another unit is called Bambi because

[Tandoori Haggis]

its deer.

(er sorry folks..)

I second the motion

[_damnit_]

Now let's hope Berman doesn't screw up the "STC" (Star Trek Collections) unit of measurement too by adding to it!! Leave it alone!

Bad idea from a storage management point of view

[HockeyPuck]

If you liked the concept of the e450, you'll like this box.

If you are interested in storage consolidation and increasing utilization while reducing storage islands. This isn't for you.

With 48disks, you'll want protection... all implemented in software raid. So you do raid-5, probably create raid groups of 12 disks? 8 disks? as the number of disks in the raid group goes down, the amount of disk you waste on parity, and the amount of CPU cycles done on calculating parity goes up.

As the industry moves to FC boot and iSCSI boot to alleviate the need to stock disk drives from 15 different vendors, this is an interesting idea for those who don't want to have a raid array. But in most shops, huge internal storage is sooooo '90s.
How do you replicate this beast? VeritasVolume Replicator. Serverless backup? Nope.

Since when is 4-8 disks standard in a 4U

[pl1ght]

Does the author of this article even work with servers? most of my 1U's come with 4 drives now 2U's with 8 at least. I have 4 4U boxes with 32 disk configurations in them, and they were pretty standard file servers. 48 tho. Thats still hot.

Seek time?

[corychristison]

What I am wondering is about seek time... with that many drives, and that much 'virtual' space [assuming this is on some form of RAID] what would the average seek time be?

Re:Seek time?

[Anonymous Coward]

Seek time is a property of disks. It's a couple of milliseconds. It doesn't matter whether there's one disk or lots. Having lots of disks generally reduces overall response time, though, because ios can be serviced concurrently.

Re:Ah, yes...a machine labelled "Sun"... appropria

[h8sg8s]

Not really. It's listed as drawing 1500 watts - that's about what 4 medium-sized Intel servers draw. Saw one at a meeting, took the cover off and the drives were only lukewarm. Large airflow for a 4u chassis.

Re:Bad idea from a storage management point of vie

[setantae]

True, although ZFS will support raidz2, which allows for the failure of 2 devices in a raid5 type setup. This reduces parity wastage somewhat.

Re:Bad idea from a storage management point of vie

[h8sg8s]

As to backup and replication, think zfs: http://www.sun.com/2004-0914/feature/ Lots of folks are seeing this as simply a 2 socket server with lots of disk. With zfs it's more like a huge disk farm with an open, hackable interface and nice manners at the back end.

Re:Bad idea from a storage management point of vie

[HockeyPuck]

The problem is this 'island' of storage is inaccessible by any other host. You can't mount it (not using NFS) for block based backups on a dedicated media server...

Thumper -_-

[rea1l1]

Were any little bunnies hurt in the creation of this device?

A feature that doesn't strike my fancy

[Immercenary_2000]

According to TFA (the Fake article pulled out of thin air) this thing check your data for naughty bits and disallows you to store them. This works perfectly with the technology's slogan:

Thumper Data Storage: "If you can't store anything nice; don't store anything at all."

I'm hear all week. Try the veal.

Re:A feature that doesn't strike my fancy

[Immercenary_2000]

I just woke up so let me correct some of the spelling errors I just made: According to TFA (the Fake article pulled out of thin air) this thing checks your data for naughty bits and disallows you to store them. This works perfectly with the technology's slogan: Thumper Data Storage: "If you can't store anything nice; don't store anything at all." I'm here all week. Try the veal.

Seen them in operation already

[QuantumMajo]

Japan's TSUBAME (see the system at http://www.gsic.titech.ac.jp/) is made up of both x4500 and x4600 systems. I've been in the Thumper room - it's loud as a jet engine in there and cooling is an issue, but only because the room is old. It's an impressive set-up, and made to be upgraded. They've got 1.1 Petabytes of storage now.

Re:Bambi

[Anonymous Coward]

Thumper? I hope the sand worms stay away...

.

"If you can't say something nice, don't say anything at all!"

Re:Bad idea from a storage management point of vie

[Anonymous Coward]

iSCSI. :) Most OSes have a iSCSI target available now. ~400 MB/s out of the box. You could throw two Infiniband or 10Gb/s Ethernet adapters for more fun. The SPOF for me is the lack of redundancy, but there are many apps that don't require that extra level of redundancy.

As an HPC system administrator, these three boxes fill an interesting niche. The x4600 in particular is interesting. There are smaller tier-2s that make 8 socket Opterons (iWill, Supermicro, and Tyan(?) make the mobos), but these machines have a mixed reputation. Sun's support infrastructure is attractive, and I've been impressed with their hardware engineering. It's also easier to get an order for Sun, Dell, HP, or IBM through purchasing than these smaller vendors. Same story with the x4500. The 8000 makes sense when you look at is as a IO-intensive system (rather than as a high-density system) or as a VMWare Infrastructure box.

Timing coincides with ZFS release

[saleenS281]

They just released ZFS in the Solaris10 06/06 update 2. Funny how this gets released at the same time :) Guess they're trying to pimp that new filesystem with some massive storage to boot.

awesome!

[RelliK]

Very cool! Thanks for the link. According to my calculations, that's actually closer to $21000 for 32*750GB drives (that's including 2x dual core 2.0GHz Opterons, 4x 1GB ECC DDR400, 2x 74GB WD Raptor system drives, 2x 3ware 9550SX-16ML controllers and 32x 750GB Seagate drives), but still, that's $0.85/GB with top of the line components inside! Not bad! I'll definitely keep this link in mind. That's exactly what I wanted.

Re:awesome!

[Anonymous Coward]

3ware is top of the line now?

Re:Bad idea from a storage management point of vie

[Zheng Yi Quan]

Granted that replicating the data off the x4500 is a problem, but it's going to be a problem for any system with this much disk throughput. How are you going to match 2GB/sec with any SAN or external storage? You can't. That's the price you pay for such a fat pipe between the drives and the CPUs.

Fortunately, ZFS means you won't have to worry as much about data loss. Striping all data across 48 disks with 64-bit checksumming is already solid, and you can still implement raid-5 or -6 on top of that with whichever OS. (One of the white papers for the x4500 lists the MTBF as 23 years given spare drives... surely an exaggeration, but I'd settle for 1/5 of that.)

What's old is new again.

[autophile]

We used to have a bunch of hard drives in our data center. They only stored about 35 megabytes each. We called them Thumpers, too.

--Rob

small potatoes

[dwater]

Some guy claims even a DVD can hold 50TB :

Re:small potatoes

[dwater]

oops...http://hardware.slashdot.org/article.pl?sid =06/07/11/209229

Little solo... Datadirect has bigger solutions.

[draziw]

www.datadirectnet.com Has been shipping 48 bay enclosures for a while now. SAF2248Put 5, 10, 15, or 20 behind a big RAID head (S2A9500 with 20 back end (disk) links, and 8 FC4 (Fibre Channel 4 - rated at 400 MB/sec) links on the front and it's a fairly big fast system. Lot of national labs going that route. You can snag manuals here if you want more details manuals

--
Good Karma and all I got was this silly sig line

Re:Bad idea from a storage management point of vie

[Paul Jakma]

OpenSolaris iSCSI target support is underway.

Dear Sun

[MalusCaelestis]

If you give me one of these systems, I promise to review it on my blog.

Best Wishes,
Greg

What 4U "standard"

[thenerdgod]

Yes, when standard for 4U server is four to eight hard disks

Bullpucky. Maybe on your planet. A PC 4U NAS box in my world holds 24 SATA HDDs. Oh, you mean a standard 4U Server... Which usually means a quad-CPU box with 4GB of RAM and a couple fugly FC controllers. See, your problem is that thumper is for Storage in which the 4U form-factor is for drives, and the standard is more like 12 to 24.

</flame>

I just can't wait. . .

[AcidLacedPenguiN]

. . . to defrag one of these :P

Approximate time remaining? All of it.

Re:I just can't wait. . .

[theProf]

You don't.
You run ZFS, which does not require defragging.

Re:I just can't wait. . .

[Anonymous Coward]

oh. Thanks for the information:)

though it was sort of a shot in the dark towards comedy.

SATABeast

[Ardvaark]

Or ... you could buy a Nexsan SATABeast and get 42 drives and 20 TiB of storage in 4u and not wait a year. No real story here.

But why?

[davidmcw]

I'm a big Sun bigot, always have been. But where does this sit. Seems like a solution without a problem to me...

Hasn't anyone noticed that Sun also released ...

[axle_512]

another new server called the X4600???
It can be configured with 8 Dual-core AMD Opteron chips, each running at 2.6 GHZ.
WOW!

How about $28,700 for 36 TB in 5U

[djoslin]

Using an AIC case, Supermicro motherboard, dual dual core Xeons, 4gb ram, Areca raid cards and 48 Seagate 750s I can build a system that has 36 TB raw at a cost of about $.80 per TB. We already have scads of systems based on 4u/24 disc cases. They work great.