Slashdot Mirror
Sun Unveils RAID-Less Storage Appliance
pisadinho writes "eWEEK's Chris Preimesberger explains how Sun Microsystems has completely discarded RAID volume management in its new Amber Road storage boxes, released today. Because it uses the Zettabyte File System, the Amber Road has eliminated the use of RAID arrays, RAID controllers and volume management software — meaning that it's very fast and easy to use."
Correct me if I'm wrong, but doesn't charging enterprise prices for simplified hardware that relies on commodity software solutions, kind of defeat the point?
Unless I'm misunderstanding this hardware, the entire idea is to move data safety away from hardware redundancy toward software-driven duplication. In that way, the data is safe from failure in the same way that GoogleFS protects against individual machine failures. The only difference is that Google probably doesn't pay $11,000 for 2TB of storage. :-/
One of these days, I really will understand why Sun regularly shoots themselves in the foot. Until then, I suppose I must trust them to somehow find a customer who's willing to pay exorbitant prices for an otherwise good idea. (i.e. I'd really love to see Sun bring Google-style reliability from unreliability to the market.)
BTW, here's the link to Sun's marketing on this:
http://www.sun.com/storage/disk_systems/unified_storage/index.jsp
It's actually pretty cool tech. Sun could own the market if they just understood how the market views pricing and features.
Javascript + Nintendo DSi = DSiCade
What a stupid and misleading title. You can, and I suspect most people will, use RAID with these boxes. RAID-Z more than likely, though other types of RAID are possible too. It is not a RAID-less box, it's a box without a dedicated RAID controller.
Chernobyl 'not a wildlife haven' - BBC News
I remember Sun's 52xx NAS storage line was a non-starter for many because it didn't have a lot of the competition's (NTAP) features that made it Just Work with Active Directory, CIFS, etc. I wonder if this is still the case?
Loading...
"All of the new unified storage systems include comprehensive data services at no extra cost, Fowler said. These include snapshots/cloning, restores, mirroring, RAID-5, RAID-6, replication, active-active clustering, compression, thin provisioning, CIFS (Common Internet File System), NFS (Network File System), iSCSI, HTTP/FTP and WebDAV (Web-based Distributed Authoring and Versioning)."
Note that this system includes "RAID".
Considering that they've purchased MySQL, StorageTec and Cluster File Systems (of Lustre fame), developed ZFS, implemented CIFS in OpenSolaris from scratch (not Samba based), participated in NFSv4 and constructed the thumper, these machines hardly come as a surprise.
For the last two years, almost all their moves are targeted towards one goal: Enter the storage market from a non-conventional angle. They want to do it unconventionally, because they know that storage more than anything else is becoming The commodity and today's toys won't cut it. Plus, at this point, all the mainstream storage vendors have difficulty tapping the low end. They may be able to sell their expensive products to clients with deep pockets, but for small businesses it's a different story. No to mention that they are unwilling to reinvent themselves. OTOH with all these inventions Sun may be trying to do what it did with workstations when it started in the 80s, start low and increase. Remains to be seen whether they can pull it.
This system will intelligently move the data around to put frequently accessed bits on the SSDs. This is a lot more than a 2u server with a few TB drives in a raid 10.
You don't make the poor richer by making the rich poorer. - Winston Churchill
While that may be true, you're hardly comparing apples to apples. The entry level 2TB model has 14 146GB 10K RPM SAS drives.
You'd still be able to whitebox it for a lot cheaper, but not 4x1TB SATA cheap.
It was also mentioned on the pre-announcement discussion that some people at Sun wanted to price it lower, but internally the powers that be didn't want their hardware to look "cheap". As such, prices went up. The good news is that supposedly the VARs will have some room to play with on the pricing. Not the most straight forward way to go on Sun's part, but the actual prices may be a fair bit lower at the end of the day. Hell, you can get it at 20% off from Sun directly through their try and buy program.
ZFS doesn't stand for zettabyte anything. "The name originally stood for "Zettabyte File System", but is now an orphan acronym." from wikipedia, sourced from http://blogs.sun.com/bonwick/entry/you_say_zeta_i_say .
and of course "RAID Array" is lovelily redundant phrasing.
1] The filesystem is called ZFS not Zettabyte
2] They appear to be twice as expensive as storage solutions that sun already sells.
Compare:
Sun Storage 7210, Option 3, $117,995, 44Tb
with
Sun Fire X4500 Server, Config 4, $61,995, 48Tb
People, please stop trying to compare a couple of drives from Newegg tossed in a chassis as a similar product for thousands less, simply because you have the same storage capacity.
That's not even apples and oranges, it's more like apples and redwoods.
Last I checked Netapp was still charging $10,000 per TB! Do you really think there is no reason for this?
Experience teaches only the teachable. -AH
Will that $600 box be using 14 146 GB 10k RPM SAS disks?
These boxes aren't about providing stupid storage, their about providing massive I/O throughput. The larger boxes scale to 44TB and 576TB respectively. This also automatically moves frequently accessed data to flash drives (and RAM) for even faster I/O.
These are absolutely monstrous compared to anything you could build for $600. There seems to be quite a bit of custom hardware to power this setup.
You don't make the poor richer by making the rich poorer. - Winston Churchill
The third one I believe--the rest I'm skeptical about...
"Not an actor, but he plays one on TV."
With that said, linux REALLY needs ZFS , and not in userspace.
Due to deliberate licensing issues we won't have native ZFS in Linux any time soon. However, BtrFS should be merging into the mainline kernel soon enough (~2.6.29), and it includes most of ZFS's features plus a few of its own: storage pools, checksumming, mutable snapshots, built-in extent-level striping and mirroring, etc. It even supports in-place, reversible conversion from ext3 via a copy-on-write snapshot.
"The state is that great fiction by which everyone tries to live at the expense of everyone else." - Bastiat
enough whining by people who really dont know what they are talking about (save those of us who do)
Check out the simulator that Sun built for you to try it in a VMWare instance:
https://cds.sun.com/is-bin/INTERSHOP.enfinity/WFS/CDS-CDS_SMI-Site/en_US/-/USD/ViewProductDetail-Start?ProductRef=SunStorageSim-1.0-OTH-G-F@CDS-CDS_SMI
Second level ARC is standard in recent ZFS; you could just plonk some X25-M's in your X4240, attach a disk shelf to it, configure ZFS to use the SSD's as secondary ARC for it, and pretty much have something like what Sun are selling.
You know, just with less vendor support, and more effort involved in building, configuring, tuning and testing. If you come out of it with change from $10k, you probably earned it with the effort you put in.
Yes yes yes, you can do that with just $1000 and a afternoon at Fry's or browsing Newegg, right?
Everyone's missing the point here, and a lot of what is being said could be applied (just as wrongly) to NetApp... after all, those are just x86 boxes running a BSD kernel.
The special sauce here is not so much the underlying OpenSolaris OS (which does provid the IO and services such as CIFS, NFS, iSCSI, data replication, and so on) but the Fishworks software put on top of it. Built-in failover clustering, the integrated web GUI and CLI... if you weren't paying attention to the console during boot, you might not even have a clue that's it's OpenSolaris underneath... which is one of the marks of a good appliance OS... easy to manage and the idiosyncrasies of the underlying OS is sufficiently abstracted away.
You don't need to be a Solaris admin to use this, just like you don't need to know about BSD to run a NetApp. The difference here is that this takes pretty high-end x86 hardware and does better than NetApp, for cheap. Ever see a support contract for any of the NetApp filers? I guarantee you'll spin a 180 on your heals and pretend you never saw the number.
You should use a mix of SLC and MLC. MLC for the frequent read, infrequent write, SLC for the frequent write.
There is more underneath the covers than meets the eye.
FTFA:
So, these RAID-less devices all include optional RAID-5 and optional RAID-6?
Putting the RAID as part of the fully integrated hardware-software black box may be a convenience, and the particular way it is implemented may save money, configuration, reduce failure points, or provide some other benefit, but what it doesn't do is make the box "RAID-less" in any reasonable sense.
This is BS. Clearly.
You have certainly never done this yourself.
First of all, the P800 is a PoS for anything but the included RAID5 or 6 (we haven't even tested RAID6, IIRC).
It has a maximum number of logical disks it can create and you will most likely have to reboot the server and go into Array Manager to setup another "array" (single disk). You can't use the RAID on the card, because ZFS wants to control the disks themself, without a RAID-controller in between (and ideally no Cache-RAM).
My co-worker's been there, done that, got the t-shirt.
So, you've got to buy one of these: http://www.sun.com/storagetek/storage_networking/hba/sas/specs.xml and connect a good external JBOD chassis to it (MSA70 e.g.). For SATA, we use Promise VtrakJ610s (which are not good, but cheap...).
Then it works.
But you will miss some features, like the red or yellow light when a disk is dead (so you'll have to count...).
And of course, you also don't get all the integration-work SUN has done with their new filers, all the statistics, all the health-checks, the GUIs.
In the end, you end-up a bit cheaper, but with a lot more labour and no support and no warrantee from anybody (best-effort only support from HP and SUN for Solaris on the DLxxx,)
Windows 2000 - from the guys who brought us edlin
Fortune 500 companies typically standardize hardware, so people who say they can buy this from here, that from there, one more thing from eBay are rediculous.
Also, to those who say small businesses can't afford this, its really an option. Some places like open source hodgepodges of hardware and some do not because their small business generates enough money that investing in enterprise class hardware with gold 4 hour response from a solid company with a history of UNIX experience and integration with Solaris.
Also, said Fortune 500 companies get massive discounts, as what you're seeing is retail price.
I've done with the DL320s, P400 and Linux, as Solaris did not work on the DL320s at the moment. It does certainly work.
Linux does not have ZFS.
Windows 2000 - from the guys who brought us edlin
You have certainly never done this yourself.
Or maybe he's still on his 1st or 2nd home-built RAID. In my experience techs tend to build maybe 3 or 4, and be around a while to see the problems, before they say "fuck this" and buy from a vendor.
Home built RAID is great for home use and as a learning experience. Personally, I'd go with a Enhance UltraStor RS16-JS SAS and an Areca ARC-1680ix, but that's just me and because I use macs at home. And I don't really mind it taking 16 hours to rebuild. Oh no, I can't stream my movies at full speed today. Try telling your boss that your ecommerce site will be 10x slower and timing out all over the place for the next day or two at christmastime but that's cool because you saved a few grand last year building the storage array yourself and see the delighted look on his face.
You would have to be nuts to try and pull the "save money by building it myself!" trick in a heavy production environment. You might get away with it for a while, maybe with one or two units. What happens when you have 12? 24? 48? Practically a full time job just babysitting the damn things. And what is your strategy for when, not if, but when the power fails?
I'm actually pretty interested in these Suns. The price isn't *that* bad when you consider that ZFS, in theory, doesn't have to rebuild and should get itself into an inconsistent state less. Those two features alone are pretty much worth the markup IMO. And furthermore, everyone knows Sun RRPs are just there to make the discounts they then offer you look better. No-one pays more than 50-80% of that if you buy more than a couple of things at a time and you have a decent VAR.
Let my new 7-digit UID be a lesson to all - write down your passwords.
Want to see something cool, check out IBM XIV. I will grant you that unless your in Israel, there is no customer visits in your future... The "E" guy from EMC went off on his own and built it (IBM bought him). He's been credited for EMC cache algorithms which pretty much put the DMX in front of everyone else. Obligitory droids comment. This is not new anymore than Microsoft inovating.
When you lose a disk ZFS has to rebuild and unless your system was over-engineered to begin with, there will be performance degradation. But now there's a pretty graph showing you how big the degradation is.
I did a benchmark comparison of the HP P400 in a DL185 vs a simple LSI 1068E SAS jbod controller. (HP branded of course) HP said it wouldn't work with the 1068E in the 12x DL185 expander setup, but of course it only took me about 10min of looking at the LSI website to find the Initiator Target firmware and get it fixed.
Basically my benchmark showed that in all cases the JBOD setup with linux software raid was 10-50% faster than the P400 controller.
They didn't ship a BBU for the P400 so it wouldn't do any write caching in the RAID5 test. The write speed was 2% the speed of linux software RAID, so I basically ignored that test.
Eventually I gave up on the DL158 and bought a Supermicro based box (SC836 chassis, great box) for a bit less, and a lot less HP bullshit.
From the preso I saw - it sounds like the goal is to have varying speed storage backing and some rather sophisticated "caching strategies" in front of it - i.e.
- Some amount of RAM for frequently active/accessed files
- Some amount of SSD drives for Level 2 data access, faster, but not an entire array (too prohibitively expensive)
- A bunch of spindle drives behind the whole thing for "slow" data.
It still follows the principal of more spindles for faster seek times, but a layered approach makes sense and I'm surprised I haven't seen more companies go towards it.
The techs added in (dtrace, zfs, etc) are there for better tracking of hot files, snapshots online, etc. All wrapped in a GUI front-end for the single-Admin company.
As far as the market - I guess it's who or how you define a "small" business. I'd say probably one large enough to run Oracle for some record keeping would fit small in their playbook, as this seems clearly aimed at a market that needs faster access storage. "Small" at that level seems to become semantics - someone with 50 employees but tons of warehousing comes to mind (think Dunder Mifflin if that helps)
But SUN is FAR from being the inventor of charging people $50k for something they could just as well get for free...
Name ANY big IT vendor, they all do it. My father can tell some amazing stories on that subject. Not a new phenomenon either.
Now, if you are the GOVERNMENT, they'll give you the special bonus public sector price, $150k!!!
"Malo periculosam, libertatem quam quietam servitutem." -- Jefferson
I'm not sure which part of ZFS is considered "old". Glad RAID is working out for you. Be happy that you have not hit any of the issues that RAID has.
But we're in the terabyte size drive age now. If all you can do is raid, your data is going to go bye bye. You need ZFS. Go google for the intro paper they wrote on why use ZFS.
It does, using Fuse. If OP used ZFS via Fuse for production boxes, he's a bigger moron than I thought possible.
And furthermore, everyone knows Sun RRPs are just there to make the discounts they then offer you look better. No-one pays more than 50-80% of that if you buy more than a couple of things at a time and you have a decent VAR.
Oh man, you haven't seen IBM work that game. There's retail. Then there's preferred which is about 1/2 X. Then there's the "just because we love you discount" bringing things down to about 2/5 X. Still above reasonable, but, hey, senior management never got fired for buying IBM right?
Then, next year, you find out what the annual support and maintenance cost. 20%. Of the freaking *RETAIL* price.
to me. Coming from high performance transaction processing land where an operation means 'the data is ON the platter' you can't do that more often than the platter rotates to the point where the head is over the sector where the write operation starts. Basic math, 15k RPM spindle = roughly 300 times/sec. Multiply by however many spindles you got, that's what you're max throughput is.
This is one reason why IN THEORY at least an SSD would be so great, that latency is much less. So basically I'm thinking they just aren't talking about what you're talking about, and maybe that makes sense, if you're running a trading operation say, you just DO NOT CARE what is buffered someplace, if it isn't physically on the drive, it doesn't exist.
"Malo periculosam, libertatem quam quietam servitutem." -- Jefferson
Guh. Sorry. I'm tired, and re-reading my comment the english is well-formed but the concepts are jumbled nonsense. Let me try again, by your leave...
Yes, it's unavoidable to rebuild when you lose a disk, and there will be a performance hit unless you go for full on 100% redundancy, and not many companies can afford to do that with a lot of data.
ZFS offers a number of benefits, though, in the event of drive failure-triggered rebuild, in that it basically knows where the data is and only bothers with that. A hardware controller has no idea what's data and what is blank space and so just redoes everything. In theory, assuming the MB/s of rebuild is the same, a ZFS rebuild of a half-full array should take half the time of a traditional controller.
It is also much more intelligent about *what* it rebuilds, starting at the top and then descending down the FS tree, marking it as known good along the way. This means that if a second drive fails halfway through the resync, instead of a catastrophic failure you still have the data up to the point of failure.
I can't remember where I read that; maybe here: http://blogs.sun.com/bonwick/entry/smokin_mirrors
But I didn't even want to talk about drive-failure rebuilding, what I actually wanted to say that ZFS is, in theory, less likely to get itself into an inconsistent state in the case of power fluctuations, controller RAM failures, drive failures w/ pending writes, that kind of thing. That's the kind of rebuild I meant - after some kind of catastrophic failure. I should probably have said "integrity checking" though.
By design, ZFS never holds critical data in memory only and so at least in theory should always be consistent on-disk. Basically it shouldn't need to fsck. That is a giant advantage to me, if it turns out to be as good in reality as it sounds on paper. Of course, that also has a lot to do with the capabilities of the FS proper, but removing the evil, evil HW controllers from the picture can only be a plus.
I don't know why, but RAID controllers are the most unreliable pieces of hardware I have ever known, besides the drives themselves (but at least they are consistent and expected to fail). Get a few of them together and something WILL go wrong, more often than not in a horrible and unexpected way. When some RAM goes bad in a HW RAID controller you are in for a whole lot of subtle, silent-error-prone fun. Anything that gets the HW controllers out of the picture is a win for me.
And don't even mention the batteries in HW raid controllers. They are the wrong solution to the power failure problem, especially since it's always after a failure that a disk will decide it's had enough of spinning and would just like to sit still for a while, thank you very much. Drive failure with pending writes! Exactly the words every administrator wants to hear. Almost as good as power failure with pending writes. Combine the two (highly likely!) for maximum LULZ. Ok, this is turning into a rant, I better stop.
Anyway, thanks for the corrections. My original comment (and probably this one) came across as a confused mess upon re-reading .. sorry .. will sleep now : )
Let my new 7-digit UID be a lesson to all - write down your passwords.
The goal of this product is to compete with Netapp. If you've ever experienced Netapp licensing/pricing, this Sun solution is a bargain. People seem to be forgetting that this is a storage appliance.
Perhaps I wasn't completely clear. I am fully aware that ZFS supports all these features; I meant only that these ZFS features are also in BtrFS. BtrFS has some additional features not in ZFS, and visa-versa; there is a more extensive list on the BtrFS Wiki.
Among all the features of BtrFS the most significant is native Linux compatibility, which ZFS does not have and is not likely to acquire.
"The state is that great fiction by which everyone tries to live at the expense of everyone else." - Bastiat
This is meant to be 100x faster than the storage you're talking about:
First: This uses Hybrid Storage Pool:
The Hybrid Storage Pool combines DRAM, SSDs, and HDDs in the same system, dramatically reducing bottlenecks and providing breakthrough speed.
Second: The system's hybrid architecture gives you the speed and performance you need to shatter the I/O bottlenecks with no administrator intervention. In fact, Hybrid Storage Pools with SSDs can improve I/O performance by 100x compared to mechanical disk drives.
All of the features I mentioned are in ZFS. The GP wanted a filesystem with the functionality of ZFS, so those are the points I emphasized. BtrFS has other features which can be found on its homepage.
"The state is that great fiction by which everyone tries to live at the expense of everyone else." - Bastiat
Small Businesses are businesses that make under $25M/year by definition.
the company I work at (multinational semiconductor business) has just been degraded to "small business" by you.
You insensitive clod.
(AC, for obvious reasons)
You know your "x% of the Sun price" comment makes me think that no storage vendor offers anything between the $120/TB and $750/TB range. Disks sell for $120/TB raw; DIY solutions (counting the PC hardware) can be doable around $200-250/TB; and the less expensive offerings from storage vendors always sell above the $750/TB mark. Something tells me a storage startup targetting the $300-500/TB range would be very, very successful...
It's funny how this viewpoint is always the one promoted on slashdot. One could argue that the Linux GPL is the problem. FreeBSD and Mac OS X had no problem integrating ZFS into their code precisely because the ZFS license (CDDL) allowed it.
so for "small" SAN/NAS usage with replication/clustering to tolerate 1 completely dead appliance you need deep pockets ... or go back to 2 Linux boxes with drbd. Correct me if I'm wrong.
"I love my job, but I hate talking to people like you" (Freddie Mercury)
It's not "unconventional" to mimic everything NetApp does five or ten years later.
Am I part of the core demographic for Swedish Fish?
Grr, when I visit sites like eweek.com, I really wish I had an "anti-bookmark" system so that if I ever attempt to go back there a dialog will pop up which says "These idiots screwed with the functionality of your back button last time you visited -- are you *sure* you really want to go there again?", and which defaults to "Heck, No".
Wankers.
ZFS does what I consider raid 10, 5 and 6.
--dave
davecb@spamcop.net
I'm using ext3, of course. ZFS on FUSE is nowhere near production quality. The original idea was to use Zumastor but it is not production-ready, either.
There's merit to criticizing Sun's prices, but this box is aimed towards smaller shops that with high throughput and large volumes of critical data. There are (for now, anyway) many small shops that offer specialized services to financial companies. And even with all of 10-30 employees, they easily wade through several gigs of data per day. Not only that, but in order to get contracts they need to demonstrate that they have top-to-bottom support on hand; something Sun does well (albeit at an exhorbitant price). Otherwise, they will face stiff lawsuits if their downtime costs their customers, for less-than-due diligence. This is why comparing to Google doesn't quite cut it and why monolithic companies like IBM still offer consulting services (at obnoxious prices). No doubt, Sun really needs some serious introspection if they want to remain a player but there's still a need.
Guh. Sorry. I'm tired, and re-reading my comment the english is well-formed but the concepts are jumbled nonsense. Let me try again, by your leave...
Yes, it's unavoidable to rebuild when you lose a disk, and there will be a performance hit unless you go for full on 100% redundancy, and not many companies can afford to do that with a lot of data.
Actually, ALL the companies that I've worked for (from startup to Fortune 100) do full 100% redundancy (RAID 0+1) for all their filesystems (on Sun Solaris). Not some, ALL. OS, straight mirrored. Application space, straight mirrored. Database, straight mirrored. Disk is cheap compared to software cost. Have you priced Oracle Enterprise lately?
The only thing worse than a Democrat is a Republican.
Because it uses the Zettabyte File System, the Amber Road has eliminated the use of RAID arrays, RAID controllers and volume management software -- meaning that it's very fast and easy to use."
A new spin on fakeraid.
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.
The Linux kernel has been GPLv2 for a long time. If Sun wanted ZFS support in Linux they could have made it work easily. The license they chose, and the patents they hold, allow them to call it "open source" while avoiding the possibility of native support in the most popular open-source alternative to their own Solaris operating system. I don't think that's a coincidence.
A much more detailed debate on the LKML is summarized on KernelTrap in support of my position. About a third of the way down the page you'll find this quote: "there are senior Sun programmers who have on record stated that one of the reasons why Sun picked the CDDL was precisely because it was incompatible with GPL and Sun fears Linux."
"The state is that great fiction by which everyone tries to live at the expense of everyone else." - Bastiat
ext3 in RAID-6, BTW. But with the configuration I posted in the first post of this thread, you can build a very cheap SAN/NAS. It does work. You can get support from Sun for the software (Sun sells Solaris support contracts) and HP for the hardware. Obviously, you won't have the nice GUI, but it's TEN times cheaper per TB.
RAID-Z (as well as it's other flavors, e.g., RAID-Z2) is not just a way to arrange disks to be more reliable and/or provide better throughput, but is an advanced file system. This means that RAID-Z offers features like compression, privileges, quotas, etc. that are above and beyond the RAID-5 specification. This is a good thing or a bad thing, depending on your requirements (probably a good thing, 'though).
One of the advantages of having a ZFS-based RAID in this type of configuration is the ZFS file system is transaction-based and so performs better in a network configuration, where local caching of data can corrupt file systems without like features (see http://blogs.sun.com/roch/entry/nfs_and_zfs_a_fine).
RAID-Z will require more write I/O and CPU than a hardware-based RAID solution, but it is possible that Sun has create such hardware in their (quite expensive) solution. At the same time, if you are creating a dedicated network storage device rather than sharing the hardware with other activities, you'll never notice the extra overhead. And if you don't want to buy Sun's hardware, they have given the RAID software away, so you can build your own at no (additional) cost.
Ah, yes, I was mainly thinking about non-mission-critical data, for example vast amounts of user-uploaded data for web sites.
You would have to be utterly crazy not to guarantee full redundancy on, say, a user database or business documents. However, it's quite a different matter to guarantee full 100% redundancy for, say, a few hundred TBs of user photos and videos. When you are offering a free service, it's difficult to make a business case for an incredibly expensive full-redundancy setup just to rule out an unlikely event which would maybe annoy a tiny subset of your non-paying users.
For example, I am not privy to Google's internal workings, but I very much doubt they have guaranteed full redundancy for every single video that has ever been uploaded to YouTube. Admittedly, they don't use RAID, they use a custom FS, but the principle is the same. The cost of absolutely guaranteeing so much (mostly low-value) data would be incredible, and I can't believe they would do it.
I've studied the systems of high-load social networks like Mixi and LiveJournal, and unless something has changed, they do not do it. I can't imagine Wikipedia has full redundancy on its images, or RapidShare on its user files, or Flickr, etc etc. Hell, there was an "incident" earlier this year when darling-of-the-blogosphere VC-funded Joyent, ironically using ZFS, were forced to admit they did not have redundant storage for data uploaded into not one but two of their *paid* online storage products. Something went wrong, the service was down for a week while they sorted it out, and they then decided to pull the product from the market rather than move to full redundancy since it would be too expensive. And that's when the customers were paying them!
http://www.joyeur.com/2008/01/22/bingodisk-and-strongspace-what-happened
So, it's not uncommon at all. I would actually be pretty surprised if any large percentage of the huge amount of bulk data uploaded to free services around the web was stored with the "enterprise grade" 100% redundancy you're talking about.
Databases and business documents, though, hell yes : )
Let my new 7-digit UID be a lesson to all - write down your passwords.
(emphesis mine)
I agree with what you're saying, but remember two things:
1) Its the license they chose. The developed the code, and they can choose how they license it.
2) I object to enclosing open source in quotes. The CDDL IS in fact Open Source.
The fact that something is incompatible with GPL (which is an aspect of the CDDL inherited from the MPL it was based on), does not mean its not Open Source, anymore than the fact that you can't freely use GPL code with proprietary code without opening up the proprietary code within certain guidelines means that GPL code is less open than BSD code.
http://en.wikipedia.org/wiki/CDDL
There are different licenses with different goals.
This space for rent. All reasonable inquiries will be entertained at proprietors discretion.
I think it should be pointed out that the definition isn't that simple. Here's a table (PDF) of how the U.S. Government defines a "small business." It varies by sector and sub-sector (a 'small' peanut farm is defined differently from a 'small' aerospace parts manufacturer, since the latter is significantly more capital intensive than the former).
Most IT-related activities are at the higher end of the spectrum (see p.30 in the PDF) which tops out at $25 million/year ("services") or 150 employees ("value added reseller"), but there are some odd special cases in there. "Technical consulting," for instance, is $7M, and "Engineering services" is only $4.5M, but "Custom Computer Programming Services" is $25M. Makes you wonder who that was gerrymandered for...
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
Nobody is saying it doesn't work. What you are building, and what Sun built are two entirely different classes of equipment. If you don't understand that, and still believe your Yugo can outperform a Lexus, more power to you.
This seems to be a day for misinterpretation. I did not mean to imply that the CDDL is not open source; I was simply setting the term off from the rest of the sentence, e.g. "someone says, 'ZFS is open source'" or, equivalently, "someone says that this project is 'open source.'" In other words, what quote marks used to mean before they were co-opted to indicate sarcasm.
Regardless of Sun's intentions, the effect is that they get the good PR associated with releasing code under an open-source license while keeping their most popular open-source competitor, Linux, from simply porting over the interesting bits. Most open-source project aim for technical and legal compatibility; thus the GPL and similar standard licenses. That Sun apparently does not share these goals bodes ill for those inclined to treat them as just another open-source contributor.
"The state is that great fiction by which everyone tries to live at the expense of everyone else." - Bastiat
For example, I am not privy to Google's internal workings, but I very much doubt they have guaranteed full redundancy for every single video that has ever been uploaded to YouTube. Admittedly, they don't use RAID, they use a custom FS, but the principle is the same.
You're right... Google actually uses triple redundancy, and that's cheaper than "enterprise-grade" anything.
Misinterpret what you wrote. Its difficult to track the nuances of writing, especially in an internet age where multiple different conventions abound.
Technically they are still in competition, and if they see Linux as a competitor, so much the better. I am sure the GPL community will come up with an alternative to ZFS as soon as they can. The fact that BSD and OS X can incorporate ZFS though, does imply a certain amount of "legal compatibility" on the part of CDDL though. OS X is even a commercial product, that could arguably be considered a competitor against Sun's workstations.
True, I wouldn't treat them as just another contributor, but since they've based their license on MPL, I assume all the work of the Mozilla community should be similarly viewed?
This space for rent. All reasonable inquiries will be entertained at proprietors discretion.
Somehow the beginning of my message got cut off (thats what I get for not hitting "preview"). Should have started:
"Sorry to misinterpret what you wrote."
This space for rent. All reasonable inquiries will be entertained at proprietors discretion.
I'm not saying my current setup (DL320s, SA P400, RAID-6, Linux, ext3) can outperform, or even be on-par, with the RAID-less storage appliances Sun introduced. I'm saying a setup with DL180/DL185, Solaris and ZFS can.
They do ship a BBU for the P400. Part number is 383280-B21
Don't worry about the misinterpretation. In retrospect it's easy to see how you could draw that conclusion, just as it was the other two times it happened to me today. :-)
The fact that BSD and OS X can incorporate ZFS though, does imply a certain amount of "legal compatibility" on the part of CDDL though.
True. However, Linux is one of the most obvious targets for any code transplants from Solaris, and the one that presents the greatest threat to Sun. My main point is that Linux's use of the GPL isn't the problem; indications are that Sun didn't want Solaris features to end up in Linux, and chose their license accordingly.
... since they've based their license on MPL, I assume all the work of the Mozilla community should be similarly viewed?
Your example supports my argument. You'll note that the Mozilla code is all tri-licensed, MPL/GPL/LGPL, in the interests of maximizing compatibility with other open-source projects.
"The state is that great fiction by which everyone tries to live at the expense of everyone else." - Bastiat
That's a pretty extraordinary claim. I think a citation is needed.
I find it very difficult to believe that Google immediately and permanently makes a triple-redundant (not just local redundancy, separate-system redundancy) copy of every single byte ever uploaded to Youtube.
Their filesystem is highly locally redundant in itself, superficially comparable to RAID-6 or better. But you're asking me to believe that they then have another, and then another, full copy of that entire installation?
I don't think so, but I'd be interested to be proven wrong.
Let my new 7-digit UID be a lesson to all - write down your passwords.
I don't know what you mean by local redundancy or separate-system redundancy, so I can't answer it. The Google File System paper says "By default, we store three replicas". Each replica is on a different server, although they are probably all in the same datacenter. This is a higher level of redundancy than RAID 10.
Local redundancy = some protection built into a single system to protect against failing drives. Like RAID 5 or 6, or GFS storing copies of data on various servers in the cluster.
Seperate system redundancy = having a full failover redundant machine/cluster in case the first one falls over.
Not exactly standard terms I know but I suspected our difference was hinging on the definition of "redundancy" so wanted to be more specific.
The Google File System paper says "By default, we store three replicas". Each replica is on a different server, although they are probably all in the same datacenter.
Yeah, that's pretty much what I thought? I thought you were claiming they had three DCs mirroring each other or something crazy.
This is a higher level of redundancy than RAID 10.
I see the problem. We are talking about different things.
RAID10 is only locally redundant, ie, inside the server/cluster. If the RAID controller screws up or the server blows up, you lose. That is not what I mean by redundancy. I would call that "fault tolerant", not redundant.
What I mean by redundancy is having a hot replica next to the first server to failover to. Or, even better, in a different location, on different power.
Google's system is not directly comparable because although it's local redundancy they mean when they say "three replicas", it's still divided up between different hardware, and it's hardware failures of varying kinds that we are trying to guard against here. So they do have hardware redundancy, which is excellent and way better than RAID10. However, if the cluster fails - not that I can imagine that happening, knowing Google - there is no higher level redundancy to switch to. Not that they need it; works well for them.
Anyway hope I straightened that out and am making sense ...
Let my new 7-digit UID be a lesson to all - write down your passwords.
NexentaStor has been in the market for a while and seems to be doing well. They run on virtually any x86 and PogoLinux is offering an integrated solution with support. And yes prices are much, much less for much more capacity than Sun. Sun beats them on SSDs though. But some of the software Nexenta has done looks quite cool and we're using it behind ESX including the free Nexenta plug in to discover VMs and do ZFS replications. Have not messed around with the CDP yet or much else in NexentaStor.