You need to build one of these:
http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/
http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I need some help. I have 20 TBs worth of storage; and I need more. Need help!!
- Thread starter WiiDSmoker
- Start date
- Sort by reaction score
Transferring 50TB at 10Mb/s would take roughly 463 days
its a suggestion, maybe the OP has a fast fibre connection? brings it down to.. ohh.. a month!
in this case, thats a VERY good suggestion! nice.
At the low end in premades, expect to pay double the raw drive cost for your system.
Atom is seriously not worth your while.
At the low end, you will want to run your RAID in software.
If you get one of very slightly tricked-out motherboards with a dual-core intel, you should have at least 6 and maybe up to 10 SATA ports in the motherboard, and it should be possible reasonably cheaply to get a board supporting up to 4x pcie 4x-16x slots.
Supermicro AOC-SAT2-MV8 is a PCI-X 64/66 card (will work in a PCI 32/33 slot) with 8 SATA ports. ~ 10-15 per port.
Supermicro AOC-SASLP-MV8 is a PCIe 4x (will work in PCIe 16x) card with 2 4x SAS ports. Including the breakout cables to 4x SATA, 8 ports will be slightly more expensive than the SAT2, but the increased controller bandwidth will make up for that. ~15-20 per port.
Sharkoon Rebel12 Value. With very minor modding (bending some steel out of the way at the top) this case will hold 12 5.25 bays, cheaply.
Coolermaster 4in3 -- using 3 5.25 bays, this holds 4 drives, with a 12 cm fan for airflow. Better airflow/cooling/noise/price performance than the option below. ~ 4-5 per HD.
Cheap 5-in-3 hotswap SATA bay -- in the same space, put 5 drives, and they're hotswappable. ~ 20-25 per HD.
Putting that all together, you get 16 or 20 drives in each Sharkoon case, and you can bolt them together with a single motherboard plus one or two Supermicro cards.
You can also get two Sharkoons, bolt them side by side, and route all the drives to a single motherboard. That would be up to 40 drives or at current capacities 80 to 120 tera, requiring 4 supermicro cards plus the motherboard.
Pricing it out (very roughly, high end estimates):
Small option, 32 TB in 16 drives:
PC parts including an 8x SATA motherboard: 400
SSD boot drive: 100
Power Supply, comma, hefty (but one will do): 100
Sharkoon case: 70
4x Coolermaster 4in3: 60
Supermicro card plus cables: 150
16 drives: 2000
32 TB NAS for 2780 + minor effort.
Bigger option, 80 TB in 40 hotswappable drives:
PC Parts including an 8x SATA, 4x pcie16 motherboard: 500
SSD Boot drives for mirroring: 200
Power supply, comma, hefty, comma, two of them: 200
2x Sharkoon case: 140
8x hotswap 5x SATA cage: 1000
4x Supermicro card plus cables: 600
40 drives: 5000
80 TB NAS for 7640 plus a bit more effort.
Mind you, these are raw capacities, I would recommend you make raid6 sets of 8 drives, losing about a quarter raw capacity, and maybe raid0 stripe those together, so 24/60 TB formatted capacity.
Atom is seriously not worth your while.
At the low end, you will want to run your RAID in software.
If you get one of very slightly tricked-out motherboards with a dual-core intel, you should have at least 6 and maybe up to 10 SATA ports in the motherboard, and it should be possible reasonably cheaply to get a board supporting up to 4x pcie 4x-16x slots.
Supermicro AOC-SAT2-MV8 is a PCI-X 64/66 card (will work in a PCI 32/33 slot) with 8 SATA ports. ~ 10-15 per port.
Supermicro AOC-SASLP-MV8 is a PCIe 4x (will work in PCIe 16x) card with 2 4x SAS ports. Including the breakout cables to 4x SATA, 8 ports will be slightly more expensive than the SAT2, but the increased controller bandwidth will make up for that. ~15-20 per port.
Sharkoon Rebel12 Value. With very minor modding (bending some steel out of the way at the top) this case will hold 12 5.25 bays, cheaply.
Coolermaster 4in3 -- using 3 5.25 bays, this holds 4 drives, with a 12 cm fan for airflow. Better airflow/cooling/noise/price performance than the option below. ~ 4-5 per HD.
Cheap 5-in-3 hotswap SATA bay -- in the same space, put 5 drives, and they're hotswappable. ~ 20-25 per HD.
Putting that all together, you get 16 or 20 drives in each Sharkoon case, and you can bolt them together with a single motherboard plus one or two Supermicro cards.
You can also get two Sharkoons, bolt them side by side, and route all the drives to a single motherboard. That would be up to 40 drives or at current capacities 80 to 120 tera, requiring 4 supermicro cards plus the motherboard.
Pricing it out (very roughly, high end estimates):
Small option, 32 TB in 16 drives:
PC parts including an 8x SATA motherboard: 400
SSD boot drive: 100
Power Supply, comma, hefty (but one will do): 100
Sharkoon case: 70
4x Coolermaster 4in3: 60
Supermicro card plus cables: 150
16 drives: 2000
32 TB NAS for 2780 + minor effort.
Bigger option, 80 TB in 40 hotswappable drives:
PC Parts including an 8x SATA, 4x pcie16 motherboard: 500
SSD Boot drives for mirroring: 200
Power supply, comma, hefty, comma, two of them: 200
2x Sharkoon case: 140
8x hotswap 5x SATA cage: 1000
4x Supermicro card plus cables: 600
40 drives: 5000
80 TB NAS for 7640 plus a bit more effort.
Mind you, these are raw capacities, I would recommend you make raid6 sets of 8 drives, losing about a quarter raw capacity, and maybe raid0 stripe those together, so 24/60 TB formatted capacity.
HobeSoundDarryl
macrumors G5
For media storage where speed should not be that important, why not unraid: http://lime-technology.com/wiki/index.php?title=Overview and http://lime-technology.com/
Macworld says this: http://www.macworld.com/article/146120/2010/02/unraid_server.html
He could use that 20-bay Norco for the box. Maybe 2 of them?
I keep seeing all these recommendations for various forms of RAID, but why would he want to link multiple drives if speed wouldn't matter that much (after the initial loading of all that data onto them)?
I also argue that if he actually owns the DVDs/BDs for all that media, why not let them be his master copies for backup and not put the ISOs on there? That would dramatically downgrade the storage needs while still holding the digital copy he would use most of the time on a much smaller server.
There's also the 67TB option for less than $8K: http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/ Macworld says this: http://www.macworld.com/article/142653/2009/09/backblazestoragepod.html I wonder if you used 2TB drives instead of 1.5 like they show, could this thing yield 90TB?
Macworld says this: http://www.macworld.com/article/146120/2010/02/unraid_server.html
He could use that 20-bay Norco for the box. Maybe 2 of them?
I keep seeing all these recommendations for various forms of RAID, but why would he want to link multiple drives if speed wouldn't matter that much (after the initial loading of all that data onto them)?
I also argue that if he actually owns the DVDs/BDs for all that media, why not let them be his master copies for backup and not put the ISOs on there? That would dramatically downgrade the storage needs while still holding the digital copy he would use most of the time on a much smaller server.
There's also the 67TB option for less than $8K: http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/ Macworld says this: http://www.macworld.com/article/142653/2009/09/backblazestoragepod.html I wonder if you used 2TB drives instead of 1.5 like they show, could this thing yield 90TB?
Everyone who has never done this thinks it's easy. It is at first and the managment of the storage starts to take hours out of your week.
The Reason for RAID is reliability is easy management. Lets say a SATA disk has an average life of four years. This means that if you own four disks, on average one will fail about every 12 months. If you own 40 one wil fail every 8 weeks. This guy, with his primary and backup system will have 40 drives. Every 6 or 8 weeks on average he
will have to deal with loosing a tarabyte of data. (assuming 1TB drives)
Maybe I should say that again to let it sink it. The drives will fail routinly at last one a month and each will take with it 1TB of data. The more drives the higher the number of failed drives per unit of time.
Why RAID? Because RAID can recover from failure automatically without the admin having to be there and without need to stop the system and repair it. You simply load it up with hot spare drives that are empty and let the software do the re-build and when the number of hot spares is low you swap bad for good drives. When I was testing storage system for a dot com I worked for I'd simulate a disk failure by opening a chassis removing the power connector from a drive. I'd first load up the system with a big symulated work load from a database and web server and then break stuff. The system could continue just fine with multiple disk drives being killed. I even removed one CPU chip from a running server. The crashed a server but not the web site. When you have 100 of anythig you are going to have a few of those broken at any one time and if you design the system crrectly a broken disk or broken server is not an emergency.
Te article was wrong about the CPU requirements. Rebuilding a RAID is lots of work and takes hours and you'd like to continue streaming data just like normal while the repair continues.
So why do most people gladly pay $10K for a storage system? Because their time is valuable AND because their data is worth 20 or 1000 times the $10K
All that said does anyone over the age of 8 watch the same movie twice? Who has time to watch that much TV? This project seem to be something you'd want if you were a cable company or y a library
For one thing, unraid is ludicrously expensive -- $150 for up to 16 drives is the biggest package mentioned, which means that this particular guy would have to negotiate his special needs for an even more expensive package.
There are several reasons for raid. One is speed, but frankly, this is irrelevant in this case. The most important one is consolidation of multiple drives into a single filesystem. You do *not* want to manually distribute data over 40 drives. If the storage is pure archival, I suppose you *could* reasonably easily stuff the drives manually and hit something like 80-95 percent capacity, but that would be at the complete expense of any type of organisation whatsoever. A single filesystem is *much* easier to deal with.
The other reason is fault tolerance. It's not *entirely* the case that failures are distributed evenly over time. Consumer-grade HDDs, when they have specs for this, tend to be specced at between .8 and 1.2 million hours MTBF. That would mean that with 40 drives, you should expect the first fault at 20.000 hours, give or take, which is 2.3 years. However, with 400 drives, you would expect a drive failure every 2000 hours, ie every 2-3 months.
Once the drives hit 3-5 years you will start seeing actual lifetime related failure rather than the type of failure measured by MTBF.
Now, that's pure math -- with 40 drives, which is big enough for statistics to take effect, I would expect at least a 10 percent mortality rate in the first year. So you *will* be dealing with drive failures regularly. If your data is unimportant enough to be lost without recovery at that rate, you might as well just chuck it now. If you're willing to re-rip 2 tera worth of data every 3 months.. Well, you might as well just chuck the ISOs now and re-rip when you need them instead.
The question isn't "Why RAID?", it's "what kind of RAID?"
Something to get out of the way first: JBOD/RAID0, just say no. Either one would allow you to consolidate the disks into one file system, but any drive failure would trash the entire system, which Would Be Bad.
RAID1: Well, it's secure and easy, but losing fully half your capacity smarts. Even if you do do it, you will probably want to run two twenty-disk raid5/6 +hotspare arrays and mirror those, rather than mirroring individual drives and then striping or concatenating them, or mirroring two giant stripes/concats.
Which leaves raid5/6 and its variants.
My gut feeling is that 8-drive raid6 arrays still sounds decent, for the traditional raid setup, but out of forty drives you should probably count on at least one if not two hot spares simply so the rebuild will start immediately and there is less (but not remotely zero) chance of multiple drive failures taking out the array.
The backblaze approach, with 45 drives off 9 5-port port multipliers, is a pretty darn good one for this type of system. I've looked at PMs in the past, but have never managed to find them available anywhere for consumers to purchase, backblaze seems to have solved that problem for the world.
FreeNAS with ZFS is probably the best easy, non-customized way to manage a system like this. If you have 45 devices, I'd say 3 hotspares and 3 pools each with 14 devices in z2 or 3 sounds right, put together into one tank. That leaves 36 or 33 devices worth of effective storage.
You're still going to have to do backups, though, because if something unforeseen happens that causes some of the wrong disks to go out simultaneously, your data is toast.
HobeSoundDarryl
macrumors G5
I appreciate both responses. Thanks. I'm not sure though that the pro-raid arguments reflect an understanding of the unraid solution.
For example, with unraid you use user shares to consolidate physical drives into single (virtual) volumes. It presents itself as if they were set up as raid0 when, in reality, there are 2+ drives each set up independently of each other (the independence is invisible, as 2+ drives consolidated into a user share look like 1 bigger drive). When a drive fails, the other volume(s) are still "whole", meaning they can be accessed because their contents are not spread over multiple discs (like raid) such that if one drive goes down, some of their content in each & every file goes down with it.
Those raid options with built-in backup (like raid 5) are great because you can swap in a new drive and the raid can rebuild itself (resurrecting the lost drive onto the new drive). Unraid uses a dedicated parity drive to do the same. Still, the difference is that in Unraid you store one of the movies or ISO on one physical volume, not spreading it over multiple volumes. In a catastrophic event where more than one drive is lost before you notice, RAID will lose EVERYTHING, while unraid will lose only the movies stored on the drives that failed- all the other drives will still be accessible (and the content they hold "whole).
In this case, the guy already has a dedicated physical backup. He claims to own all these movies on disc. Assuming that's true, I would take that to mean that raid-like backup is not a major priority. He can always use the physical discs to resurrect the system in a catastrophic loss. However, it doesn't really matter. Both raid and unraid have backup options to restore a dead hard drive in the array. Raid however loses everything if 2 drives manage to go down. Unraid will still have access to all but what was stored on those 2 drives.
Unraid volumes don't sacrifice some space on every drive to support the backup. 2TB is 2TB in unraid. One disc gets dedicated as a Parity drive (a backup for all discs in the array), but otherwise, all hard drives have all of their space fully available for storage. Raid has to give up a chunk of space on each drive to make it's backup model work. This may be a "6 of 1, half dozen..." argument, but it might matter- more so in this case, where backup might be covered by physical movie discs on hand.
The unraid tradeoff is speed. Unraid works with single volumes at a time, so read/write speeds are limited to individual hard drive read/write speeds, with writes slowed a bit more for the parity drive to adjust it's setting to backup new files that are added. But this particular application is great for that, as speed of read/writes is not crucial; most of the time, he would be reading- not writing- movie files. Once the parity drive is set up and all his movies are in place, writes would be limited to when he adds a new movie. Most of the time, he should be reading from the array.
The implication that RAID can recover via swapping in a replacement drive and unraid can't is wrong. Unraid will rebuild a lost drive too when you swap in a new one.
The implication that unraid would show every drive as a separate volume is wrong. It lets you decide how many drives to bundle together into a single "virtual" volume (called a user share). In this case, the OP could bundle all the drives into a single volume, though it might make more sense to bundle them as M4V and ISO volumes, or maybe even categorize them in other ways (movies, TV shows, photos, etc or maybe by genre or popularity).
Another big advantage to unraid is that you can mix and match drive sizes. If the OP starts with 2TB drives today but then wants to add some >2TB drives (when they are released) tomorrow, no problem. If he has some <2TB drives laying around with nothing to do, no problem. If he has some old IDE drives gathering dust, no problem. unraid will take, mix and match sizes (and drive speeds) and add that to the array.
The implication that unraid is expensive compared to Raid solutions is most mysterious of all. I have been looking for a big storage solution myself for a very long time. One of unraids best points is the very low cost of it. I can't hardly find big storage raid-based solutions that don't seem to always start at fairly big bucks for just the driveless hardware to start. Since the cost of the drives should cancel themselves out in either solution, I'd love to see- say- a 30 drive raid box beat 2 15-drive unraid boxes on driveless box cost. While you can do better than the pricing on the unraid website for 15-drive boxes- especially if you build your own- 2 of their turnkey, ready-to-load (with up to 30 drives) boxes can cost as low as $1199 each. Could someone point me to a good 30+bay Raid 5/6 driveless box available for less than $2500? Note, this is not asking for just a case... I'm talking about something more apples-to-apples, as in plug in some drives and you have a storage server (not plug in some drives, and cards, and memory, and power supplies, etc).
And there would be no custom cost negotiations, the OP could just use a couple of 15-bay boxes to meet his needs. Trying to do this in one bigger box (all the bays he needs) via raid is going to be a lot more expensive. But if anyone can show me that that is not true, I'd appreciate finally seeing a raid solution big enough for this kind of need that will come in at lower cost than a couple of Unraids. I've been looking for that a very long time myself.
Lastly, the new version of unraid apparently will work with 20 drives- a perfect match for the OPs desire for something like that Norco box. 2 of those could yield 80TB of gross storage. Allocate 2 parity drives (one for each box) and the net will be very near 80TB of gross storage too, in just 42 drives. How many drives would it take to accomplish the same via one of the raid with backup options?
No doubt about many advantages of raid- especially speed- vs. unraid. But this guys application is more of a write once: read many times thing, where he apparently has hard copy backups for all content on hand anyway. Other than that first transfer of data onto the server, write speed should not be important to him. And serving movies is probably a 1-movie-at-a-time (or maybe a few movies at a time) proposition, so read speed is no big deal either. Since unraid read speeds are pretty much limited to hard drive read speeds, the experience should be as fast as storing a movie on a Mac hard drive now and streaming it (aka plenty fast, for this application).
For example, with unraid you use user shares to consolidate physical drives into single (virtual) volumes. It presents itself as if they were set up as raid0 when, in reality, there are 2+ drives each set up independently of each other (the independence is invisible, as 2+ drives consolidated into a user share look like 1 bigger drive). When a drive fails, the other volume(s) are still "whole", meaning they can be accessed because their contents are not spread over multiple discs (like raid) such that if one drive goes down, some of their content in each & every file goes down with it.
Those raid options with built-in backup (like raid 5) are great because you can swap in a new drive and the raid can rebuild itself (resurrecting the lost drive onto the new drive). Unraid uses a dedicated parity drive to do the same. Still, the difference is that in Unraid you store one of the movies or ISO on one physical volume, not spreading it over multiple volumes. In a catastrophic event where more than one drive is lost before you notice, RAID will lose EVERYTHING, while unraid will lose only the movies stored on the drives that failed- all the other drives will still be accessible (and the content they hold "whole).
In this case, the guy already has a dedicated physical backup. He claims to own all these movies on disc. Assuming that's true, I would take that to mean that raid-like backup is not a major priority. He can always use the physical discs to resurrect the system in a catastrophic loss. However, it doesn't really matter. Both raid and unraid have backup options to restore a dead hard drive in the array. Raid however loses everything if 2 drives manage to go down. Unraid will still have access to all but what was stored on those 2 drives.
Unraid volumes don't sacrifice some space on every drive to support the backup. 2TB is 2TB in unraid. One disc gets dedicated as a Parity drive (a backup for all discs in the array), but otherwise, all hard drives have all of their space fully available for storage. Raid has to give up a chunk of space on each drive to make it's backup model work. This may be a "6 of 1, half dozen..." argument, but it might matter- more so in this case, where backup might be covered by physical movie discs on hand.
The unraid tradeoff is speed. Unraid works with single volumes at a time, so read/write speeds are limited to individual hard drive read/write speeds, with writes slowed a bit more for the parity drive to adjust it's setting to backup new files that are added. But this particular application is great for that, as speed of read/writes is not crucial; most of the time, he would be reading- not writing- movie files. Once the parity drive is set up and all his movies are in place, writes would be limited to when he adds a new movie. Most of the time, he should be reading from the array.
The implication that RAID can recover via swapping in a replacement drive and unraid can't is wrong. Unraid will rebuild a lost drive too when you swap in a new one.
The implication that unraid would show every drive as a separate volume is wrong. It lets you decide how many drives to bundle together into a single "virtual" volume (called a user share). In this case, the OP could bundle all the drives into a single volume, though it might make more sense to bundle them as M4V and ISO volumes, or maybe even categorize them in other ways (movies, TV shows, photos, etc or maybe by genre or popularity).
Another big advantage to unraid is that you can mix and match drive sizes. If the OP starts with 2TB drives today but then wants to add some >2TB drives (when they are released) tomorrow, no problem. If he has some <2TB drives laying around with nothing to do, no problem. If he has some old IDE drives gathering dust, no problem. unraid will take, mix and match sizes (and drive speeds) and add that to the array.
The implication that unraid is expensive compared to Raid solutions is most mysterious of all. I have been looking for a big storage solution myself for a very long time. One of unraids best points is the very low cost of it. I can't hardly find big storage raid-based solutions that don't seem to always start at fairly big bucks for just the driveless hardware to start. Since the cost of the drives should cancel themselves out in either solution, I'd love to see- say- a 30 drive raid box beat 2 15-drive unraid boxes on driveless box cost. While you can do better than the pricing on the unraid website for 15-drive boxes- especially if you build your own- 2 of their turnkey, ready-to-load (with up to 30 drives) boxes can cost as low as $1199 each. Could someone point me to a good 30+bay Raid 5/6 driveless box available for less than $2500? Note, this is not asking for just a case... I'm talking about something more apples-to-apples, as in plug in some drives and you have a storage server (not plug in some drives, and cards, and memory, and power supplies, etc).
And there would be no custom cost negotiations, the OP could just use a couple of 15-bay boxes to meet his needs. Trying to do this in one bigger box (all the bays he needs) via raid is going to be a lot more expensive. But if anyone can show me that that is not true, I'd appreciate finally seeing a raid solution big enough for this kind of need that will come in at lower cost than a couple of Unraids. I've been looking for that a very long time myself.
Lastly, the new version of unraid apparently will work with 20 drives- a perfect match for the OPs desire for something like that Norco box. 2 of those could yield 80TB of gross storage. Allocate 2 parity drives (one for each box) and the net will be very near 80TB of gross storage too, in just 42 drives. How many drives would it take to accomplish the same via one of the raid with backup options?
No doubt about many advantages of raid- especially speed- vs. unraid. But this guys application is more of a write once: read many times thing, where he apparently has hard copy backups for all content on hand anyway. Other than that first transfer of data onto the server, write speed should not be important to him. And serving movies is probably a 1-movie-at-a-time (or maybe a few movies at a time) proposition, so read speed is no big deal either. Since unraid read speeds are pretty much limited to hard drive read speeds, the experience should be as fast as storing a movie on a Mac hard drive now and streaming it (aka plenty fast, for this application).
its a suggestion, maybe the OP has a fast fibre connection? brings it down to.. ohh.. a month!
You need a damn fast fibre then. As you know, we got some speedy fibers here and the fastest upload speed you can get is 10Mb/s while fastest download is 200Mb/s. Even if you had EXTREMELY fast upload, e.g. 100Mb/s, it would still take 46 days. IIRC uploads cost more for the ISP thus upload speed is usually lower
When would you get the time to watch all of those films? I watch maybe 2 a week at most. Many of them aren't worth re-watching
When would you get the time to watch all of those films? I watch maybe 2 a week at most. Many of them aren't worth re-watching
I'm a hoarder
I'm a hoarder
I think I am to with some things. How many Dvd's and Blu Rays do you have? I'm really curious.
I hope you're able to figure something out. What are you are currently doing now for the 10 drives that you own? There are 3 TB drives out now, that might keep the number down that you need.
HobeSoundDarryl
macrumors G5
Could you send a link pointing to a 3TB drive? Do you mean 1 drive, as in a single 3TB 3.5" internal SATA or similar? I'm not aware that anything over 2TB in one drive has been released.
Seagate has released a 3TB external drive in their Go-Flex line. It has not yet been released as a bare drive. Anandtech reviewed it recently. Bare drives will be available soon, I am sure.
HobeSoundDarryl
macrumors G5
Great, then pairing this impending advance (3TB drives) with my long rant up above about unraid vs. raid... and considering that the OP says he is full at 20TB and needs more, just one unraid box with 15 bays at $1199 or so seems like it would give him up to 45TB of storage.
He could put in maybe 8 of these new 3TB drives, adding 4 more TBs of space to his current situation, then add additional drives as needed. Again, since unraid can use various sizes of drives, if 3TB drives are a stepping stone to 4TB, maybe by the time he was close to filling the 24TBs, he could add 4 additional TBs via drive #9, and still have 6 bays left for future storage growth.
I'm still looking for any of the pro-raid guys to come back with any near-turnkey (all the parts are there- just insert the hard drives) raid box capable of the same, and available for around that $1199 range. I'm genuinely interested is such a thing actually exists.
The OP could also go ahead and build his own with that 20 bay NORCO box, use the 8 or so bays now with unraid "as is", then upgrade to the new version of unraid when it is released so that he is ready to fill in all 20 bays when he gets to that point. Via all 3TB drives, that would triple his current storage in a single box, though again, if he buys what he needs now, the >3TB drives to come could actually grow that same solution into something even more than a 60TB, one-box solution.
From my perspective, this is looking more and more like the way to go for this kind of need. But I (too) would welcome any good alternative via any raid with backup box.
He could put in maybe 8 of these new 3TB drives, adding 4 more TBs of space to his current situation, then add additional drives as needed. Again, since unraid can use various sizes of drives, if 3TB drives are a stepping stone to 4TB, maybe by the time he was close to filling the 24TBs, he could add 4 additional TBs via drive #9, and still have 6 bays left for future storage growth.
I'm still looking for any of the pro-raid guys to come back with any near-turnkey (all the parts are there- just insert the hard drives) raid box capable of the same, and available for around that $1199 range. I'm genuinely interested is such a thing actually exists.
The OP could also go ahead and build his own with that 20 bay NORCO box, use the 8 or so bays now with unraid "as is", then upgrade to the new version of unraid when it is released so that he is ready to fill in all 20 bays when he gets to that point. Via all 3TB drives, that would triple his current storage in a single box, though again, if he buys what he needs now, the >3TB drives to come could actually grow that same solution into something even more than a 60TB, one-box solution.
From my perspective, this is looking more and more like the way to go for this kind of need. But I (too) would welcome any good alternative via any raid with backup box.
Currently I have 20 Terabytes worth of storage. 10 2 Terabyte drives. They are all completely full. I'm estimating that I need about 35-40 more Terabytes to complete my collection. Unfortunately I haven't even been able to begin to back things up; which I'm worried about, but not currently at this time.
I've contemplated running a bunch of 5 Bay NAS's; but I'm thinking I'd like to try something else.
Is there anyway I can get a long box that can work like a NAS; but hold 10-20 SATA drives. Or something!
I need some ideas. Please help
How about BUYING the media as DV or Bluray Disc? 2000 Blurays would still fit into a nice shelf.
Unraid is presumably organised as a filesystem-level form of concatenation/jbod internally.
I didn't see that it has the option of a parity drive -- that does narrow the gap quite a bit.
Unraid, like FreeNAS or Ubuntu w/software RAID, run on exactly the same hardware -- a PC with a bunch of hard drives attached to it. The difference is that UnRaid apparently costs $150 for up to 16 drives, and will presumably cost more for more drives. FreeNAS (So named because it's FreeBSD based, not for free-as-in-beer) is free-as-in-beer.
For a 40 drive system, unraid probably costs $300+, and there is no cost on the FreeNAS side to offset that, as the hardware is the same.
A raid-based system *can* run in N+1 mode, that's basic RAID5. Does UnRaid support hotspares? It doesn't seem to be a bad solution, I'm just not sure it scales well to the sort of volumes we're talking about. It might work reasonably well for the 16 drive setup, if you're willing to pay the cost.
It's actually a pretty good idea, as far as I'm concerned, but I'm not sure their implementation is worth what they charge for it. In this particular use case, it definitely has its good points.
The current 3 TB drive is a 5-platter, 600 GB/platter monster that slurps power, emits relatively high quantities of heat and noise, and potentially may have reliability issues (due to the above). Using it as an internal drive will also void the warranty. The 2 TB drives today are (AFAIK) all 4-platter, 500 GB/platter devices, available in 5400/5900 rpm "Green" versions, that are probably a better fit. Do take note of TLER though for raid purposes. Most of them don't have it, outside Enterprise/Raid Edition products, which can cause them to drop out of an array at the drop of a hat. Some of them have more problems than others.
What's coming next that's really interesting is 666 GB/platter, bringing 3-platter 2 TB and 4-platter 2.5 TB capacity points. That'll be next year, probably, though. For pretty close to all the other capacity points than 3 TB, 600 GB offers no real advantage over 500 GB platters -- it offers 600/1200/1800/2400/3000. *Maybe* we'll see 4x625=2500 in the intermediate, but it doesn't seem terribly likely. Even 2 TB is still coping with "what the hell do I need that much storage for", so I don't think there'll be many hard drive manufacturers chomping at the bit to put out minor capacity increments.
I didn't see that it has the option of a parity drive -- that does narrow the gap quite a bit.
Unraid, like FreeNAS or Ubuntu w/software RAID, run on exactly the same hardware -- a PC with a bunch of hard drives attached to it. The difference is that UnRaid apparently costs $150 for up to 16 drives, and will presumably cost more for more drives. FreeNAS (So named because it's FreeBSD based, not for free-as-in-beer) is free-as-in-beer.
For a 40 drive system, unraid probably costs $300+, and there is no cost on the FreeNAS side to offset that, as the hardware is the same.
A raid-based system *can* run in N+1 mode, that's basic RAID5. Does UnRaid support hotspares? It doesn't seem to be a bad solution, I'm just not sure it scales well to the sort of volumes we're talking about. It might work reasonably well for the 16 drive setup, if you're willing to pay the cost.
It's actually a pretty good idea, as far as I'm concerned, but I'm not sure their implementation is worth what they charge for it. In this particular use case, it definitely has its good points.
The current 3 TB drive is a 5-platter, 600 GB/platter monster that slurps power, emits relatively high quantities of heat and noise, and potentially may have reliability issues (due to the above). Using it as an internal drive will also void the warranty. The 2 TB drives today are (AFAIK) all 4-platter, 500 GB/platter devices, available in 5400/5900 rpm "Green" versions, that are probably a better fit. Do take note of TLER though for raid purposes. Most of them don't have it, outside Enterprise/Raid Edition products, which can cause them to drop out of an array at the drop of a hat. Some of them have more problems than others.
What's coming next that's really interesting is 666 GB/platter, bringing 3-platter 2 TB and 4-platter 2.5 TB capacity points. That'll be next year, probably, though. For pretty close to all the other capacity points than 3 TB, 600 GB offers no real advantage over 500 GB platters -- it offers 600/1200/1800/2400/3000. *Maybe* we'll see 4x625=2500 in the intermediate, but it doesn't seem terribly likely. Even 2 TB is still coping with "what the hell do I need that much storage for", so I don't think there'll be many hard drive manufacturers chomping at the bit to put out minor capacity increments.
HobeSoundDarryl
macrumors G5
Yes, here's good info about that: http://lime-technology.com/technology
Conceptually, I'm pretty impressed with the unraid proposition for this kind of application. It is JBOD-like, but that seems good for this kind of need... especially against future expansion needs as those larger storage capacities come online.
In any event, whether we use current 2TB drives, coming 2.5TB or 3TB drives, it is still looking like a low-cost way to go for this when the hardware box and all the parts within it are included. Everytime I go big raid box hunting (for anything much more than 8 bays), it seems the "good ones" are all starting around $4999 and going up. Just about every time I take a good look around, I seem to keep finding the (potential) "one" best suited for my own needs priced at around $7999.
On the other hand, this unraid thing looks very good in comparison, as long as the application is much like the OPs... especially where write speeds are not very important and read speeds are "good enough" at single hard drive read speeds.
He owns all of them already, he just feels the need to keep all of the original, full-sized ISO rips on his drives for some reason.
He owns all of them already, he just feels the need to keep all of the original, full-sized ISO rips on his drives for some reason.
OK, but then he already has the best backup - the physical medium itself.
OP if you want to keep it simple, convert them to a quality .mkv and stop this nonsense. Keep all your Originals in a dark and well air conditioned room and they will last you for at least 2 Decades and think over your storage problem again then!
HobeSoundDarryl
macrumors G5
By then it will all fit in his inner-ear embedded iPhone 25- a magical, whimsical, device with smellivision and portable hollodeck technology.
I know you are right 100% about this, but for my Blu-Rays I at least want to keep them in .ISO for the Boxee Box. As for my DVDs, I am encoding them all to M4V and then deleting the ISOs for them.
In the end, I will still need a significant piece of hardware for my Blu-Ray ISOs.