Help Understanding RAID

This is a pretty good place to start:

http://www.bit-tech.net/hardware/2007/06/11/understanding_raid/

Gives you an overview, and then provides specific answers.

 

The article linked by Macman provides a good overview of RAID.

A couple of points I'd add:

• You seem to be confusing NAS and RAID. NAS is Network Addressed Storage (ie, storage that exists on the network, not inside of just one computer), RAID is the method by which data is stored redundantly on a number of disks. Most multi disk NAS boxes support RAID, but you don't have to use it - you can just use regular disks. Similarly you can get RAID boxes that connect to the computer directly, not over the network.
• RAID isn't a backup. If you use RAID 5 to spread your data across a number of disks, then it'll preserve that data if one of the disks dies. That's great; but disk crashes aren't the only bad thing that can happen to data. Fire, theft, virus, accidental deletion, corruption, water damage, physical damage are all (likely) events that RAID doesn't protect against. All that RAID does is to reduce the amount of time you're down after a crash. You still absolutely need a backup (in fact, more so - because when RAID goes really wrong the proprietary formats can sometimes make recovery much more difficult).
• Given that you can't get away from taking backups, often the cost of RAID (cost of hardware, cost of additional parity drives, cost to performance) starts to dull it's appeal. If you don't currently have proper backups (2 sets of backup disks, one set kept offsite) then this should be the first thing you spend your money on.

Look laterally at your problem and don't be too over-impressed by 'magic bullet' solutions that seems to protect your data through disk crashes and other problems. If you need the assurance of batter backups, then get some disks and make better backups.

If you need to access your data kept inside your MacPro in the event that the MacPro fails - within 2 minutes you can open the side, remove your data disk, drop it in a dock (pictured) and get back to work:

 

There's no "should". It's all a matter of what measures you're willing to take to ensure the safety of the data. The more you're willing to do the better your odds.

You could certainly use just a large, single external drive but you're out of luck if anything happens to the drive.

With RAID you at least have some redundancy if a drive fails. I'd recommend keeping a spare on hand to minimize your down time if you go this way.

You can use Time Machine or another software backup solution but you don't have to. You can use the RAID as an external drive.

You can add internet or other offsite backup as well.

Again, it all depends on what lengths you're willing to go to to protect your data.

 

I have owned several NAS RAID 0/1/5/10 boxes... and I have decommissioned all of them. I also own a Windows Home Server w/ 8TB of data... which is currently sitting with the power turned off.

The problem for me... these NAS solutions do not fit into my backup strategy. I use both Time Machine for local backup... combined with Crashplan+ for cloud backup. Neither of those works well (or at all) backing up NAS data. This leads me to find a replacement technology for both the NAS boxes & the WHS.

My recommendation:

1. First, backup your data twice... locally and cloud. Once you have your data backed up twice... then you are covered. This should be your #1 priority.
   
2. Do not rely on any backup strategy that is not automated, with zero human involvement. Don't get me wrong... if you want to make a backup occasionally and keep a copy away from the house... go ahead and do it. Just don't fool yourself into thinking that it is a reliable backup strategy. For example... I religiously make an ANNUAL "extra safe" backup copy of my >100K photos... an keep it locked up at work. I do it EVERY YEAR WITHOUT FAILURE... except there is just one small problem. I think if I was to pull that drive out of my office... I would find that it has been 3 years since I really did it. It is just human nature.
   
3. Next... do not add any storage that does not fit into that fully automated dual backup methodology. This means that you either need to add local storage... or add "intelligent networked storage" that can have its own automated dual backup.

I *THINK* that eventually I will buy a Mac Mini and use it with Lion Server. That way I can set the Mac Mini to create local TM backups... and also backup to the cloud. It can serve files throughout the house... including to ATVs and other devices. I will just add direct attached drives to it via FW800, or TB as I need more storage. I'll also be able to use it for other "server functions" I may desire using Lion Server.

/Jim

 

Without going in the tech jargon!
Here are the high light of RAID
RAID 0 (Stripe) result FAST data divide among drives, lost 1 drives, all data is GONE - usable 100% RAW space
RAID 1 (mirror) real-time back up 1:1, able to rebuilt data when lost a drive. Data is critical during rebuilt and drive failed, control with hot-spare feature will minimize the critical time - usable 50% of RAW space
3Way RAID1 (Mirror 3 drive) - Three drives mirror, data in critical ONLY when lost TWO drives, otherwise it is the SAFEST - usable 33% of RAW space
RAID10 (Mirror and stripe) - FAST and SAFE, lost TWO drives (4 drives raid set) or more (> than 4 drive raid set) are OK
RAID5 - (Parity) speed depend on controller, hardware is fast, where software raid is generally slow, when lost a drives, then data is critical, including during rebuilt - usable N-1 RAW space
RAID6 - (2x Parity drive) speed depend on controller, hardware is fast, where software raid is slow, when lost TWO drives, then data is critical, including during rebuilt - usable N-2 RAW space

As far as backup concern, the ideal is use 1-2-3 rules (this is not what I came up, but I read it some where)
3 - Copies - have a multiple copies - recommend 3 copies
2 - different format - internal vs external; DAS vs NAS.
1 - Off side - Must have a copy remotely (out side the facility)

Here is what I have:

My company data is back up to a NAS (eBOX-N) at boss's house use RSYN to remotely sync via VPN to the Server at WORK schedule at 8:30pm (Mon-Fri)

 

Just want to point out that hard drive docking is not suggested for long term use.
No anti-vibration, no heat dissipation.

 

Keep it simple and straight forward.
RAID 0 for high performance. More drives= faster speed until reach interface transfer rate bottleneck. Lose one drive you lose everything.

RAID 1 = mirror= you lose half of the capacity. Speed is also depending on the RAID engine and how many drives in the RAID set.
**Side note** Currently mot of the RAID card (at least what I've been doing, Areca, ATTO, QSAN, LSI) supports up to 16 physical drives per RAId group.
You can lose half of the drive but it is costly to use RAID 1.

RAID 5, using one drive capacity to create parity. Lost more than one drive, you are done. Avoid using this unless have good RAID card using a dedicated CPU. Something like High Point 2314, Silicon Image 3132, 3124 chipset solution, please avoid at all cost.

RAID 6 same idea as RAID 5 but two parity data across all drives; thus, took 2 drive space as parity and you can lose no more than 2 drives.

The best back up solution is using LTO or a separate RAID set and swap out the drives. Something like
2 Bay Removable RAID

But again, for high performance I will use RAID 0 or RAID 6.
For long term backup using LTO or RAID 1.

 

There is a lot of marketing smoke around RAID and equivalent like Drobo for example: put your disk in our box, we'll do our magic on them and you'll never have to worry about losing your data anymore. As firestarter said, the reality is a bit more complex.

IMHO, if you don't run a server or provides services where you really need a minimal downtime, you should avoid RAID.

You could get a NAS though (you are not forced to configure your disks in RAID in a NAS) that will act as your external drive (gigabit ethernet is faster than USB 2.0 and comparable to FW800) plus USB drives attached to the NAS for the local backup.

For offsite backup, you could use a cloud solution or a second NAS located in another place.