Talk:RAID/Archive 4

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Archive 3 Archive 4 Archive 5


Clarification on "proprietary".

Just my opinion, but perhaps a better term would be "non-standard"? Proprietary typically implies that the specification isn't readily available, but several of the mentioned alternatives are openly available. -Matt 20:23, 5 August 2006 (UTC)

Moved some sections, removed disclaimers

I especially think that a section which explicitly "assumes knowledge of RAID configurations" and directs to another site for the "basic explanations" is contrary to the purpose of a general-knowledge encyclopedia. I removed that disclaimer and moved that section down below the explanations of RAID configurations to hopefully give newbies some background, so they can understand Wikipedia content w/o having to reference other sites. Hope this edit didn't offend anyone, but we can certainly discuss it here and I won't be offended if someone thinks a revert is in order. Icewolf34 19:46, 8 August 2006 (UTC)

Something missing in RAID 5

Text from the RAID 5 article:

"The parity blocks are not read on data reads, since this would be unnecessary overhead and would diminish performance. The parity blocks are read, however, when a read of a data sector results in a cyclic redundancy check (CRC) error."

I understand that a CRC error is when the parity information extracted from the data blocks is different form the parity block. Am I wrong?

If I am not wrong, how is it possible to know that a CRC exists without reading the partity block?


modern hard drives have inbuilt CRC and other data integrity checks and on a read are able to either give back the data or report an error. The chances of a modern drive giving back 'wrong' data (i.e. different from that written) are very small. 21:03, 24 September 2006 (UTC)
While I certainly hope Ali is correct, I think this is an important enough fact that it ought to be mentioned in this RAID or perhaps the hard drive article. Can you give a reference? -- 18:11, 30 June 2007 (UTC)

Windows LVM/RAID issues

The section about JBOD implies Windows lacks "LVM" mechanism and only supports JBOD. However Windows XP supports software RAID 0 and 1 through using logical volume based abstraction. Windows 2000 Server and Windows Server 2003 versions additionally support software-based RAID 5. All versions support any kind of hardware-implemented RAID flavors.

SSG 06:09, 29 September 2006 (UTC)

LVM in Unix/Linux means being able to span disks together to create a pool of blocks called the Physical Volume. Logical Volumes (partitions) can be dynamically created, destroyed, expanded and shrunk on the fly in the Physical Volume. The pieces of the Logical Volume may not be contiguous on the disk. You can think of it like a "file system" for logical volumes. The logical volumes can be fragmented. The link to the LVM page explains this further.
On Windows the logical volumes (partitions) must be all in one piece. They can span disks but can't have other partitions in the middle. Even though Windows may have something called a "Logical Volume" it's not the same as LVM.
Avernar 06:51, 28 October 2006 (UTC)
But do they really support RAID-5 or is it actually RAID-4? The Help in Windows XP Professional describes its software RAID-5 as using one drive for parity- which is NOT RAID-5, it's RAID-4.

cleanup request

This article used to be much shorter and better. I suggest we limit the introduction to 3 or 4 sentences, and that descriptions of the various RAID types on this page be changed to summaries with a link to separate wikipedia pages for each type, where detailed analysis of algorithms and things can occur. There is lots of unexplained jargon and formulas here which are completely useless for someone looking up RAID in general. Also, nonstandard and vendor-specific RAID implementations need to get their own pages or be excised. $0.02 Perle 23:20, 12 October 2006 (UTC)

    • I have two SATA controller which can do RAID. Windows is implying that I can setup an array. I came here to find the answer, but this page appears to be written for persons who actually know what is RAID. How can this be a candidate to a "great page" or whatever honor?

The section on RAID-5E and RAID-6E has been rewritten to introduce bogus information

RAID-5E and RAID-6E refers to RAID-5 and -6 with spare when the spare is an active part of the block-rotation scheme. This spreads the I/O load out over all drives including the spare drive, and such a scheme is faster than RAID-5 or -6. Whomever wrote that not only is it not faster, but it is in fact impossible to make it faster, introduced active falsehood into the article. Hpa Wed Oct 25 00:43:46 UTC 2006

About Atomic Write Failure

On the subject of atomic write failure: Although it is rarely handled on the same level as RAID, it need not be. Modern OSes that use journalling have solved this problem already at the file system level. NTFS, EXT3, and ReiserFS all use journalling to ensure that writes are atomic. Implementing it on the same level as RAID would be redundant.

CobraA1 07:49, 19 November 2006 (UTC)

RAID 2 Bit or Byte??

Someone has changed the word bit to byte in the RAID 2 section. Other sites that haven't copied from here have it as bit. Can someone confirm this as I don't have any experience with RAID 2. Avernar 01:20, 23 November 2006 (UTC)

The original (for the last year or so at least) was bit, it was changed without comment by an anon recently to byte; and then flipped back to bit by another anon - it would appear to have been accidentally reverted back to byte so I've restored bit, which would be correct as my memory and the few references I have on hand support. Let me know if we need to cite. Kuru talk 05:24, 23 November 2006 (UTC)

I'm the one who changed it to byte on the 22nd of November believing I was fully reverting an otherwise odd edit. I'm leaving it as "bit" Poweroid 14:28, 23 November 2006 (UTC)

Can you Change a Raid Level? if so How?

Mrmojorisinca 13:59, 6 December 2006 (UTC)mrmojorisinca

RAID 10 on two disks

I've removed the paragraph about Linux Raid 10 on two disks from the nested raid levels section. Running raid 10 on two disks is exactly the same as running raid 1 on two disks except that the blocks are in a different order. This just adds more CPU overhead (raid 1 vs raid 10 processing) and will increase drive seeks if you have a couple of linear reads going on. And if one disk fails the other would have to seek like crazy which would reduce it's life expectancy. Not a good idea.

Also the Linux Raid 10 driver is covered in detail in the Proprietary RAID levels article and and briefly in the Nested RAID levels (when it's configured like standard RAID 10) so it doesn't need to be mentioned in the general article.

Avernar 10:27, 2 January 2007 (UTC)

Copyright violation

Ok, User:Matt0401 added a copy of the guide from found here (and added to WP here) and asked for permission to use on wikipedia here (I assume the forum user Matt Welde is User:Matt0401). I haven't found the permission response from the author but I have left a note on Matt's talk page asking to provide it with a request to reply here. I look forward to the response. Cburnett 18:01, 24 January 2007 (UTC)

  • I see where someone asked for permission, but did they ever actually get permission, and did the original poster specifically say the text is under a free license? Those are the important things. Even being told that "you can use this content on Wikipedia" (which it doesn't look like he was told so) is not enough to allow it here. — BRIAN0918 • 2007-01-24 18:54Z
    • Why is permission from the author insufficient? Cburnett 20:03, 24 January 2007 (UTC)

On a separate side note, I would definitely not be opposed for a rewrite with sources... :) Cburnett 20:07, 24 January 2007 (UTC)


I just moved the contents of IDE RAID to ATA RAID, because IDE is an incorrect term to refer to ATA. However, I don't really think the ATA RAID article needs to exist at all, since this information is pretty much contained in this article. Unless anyone has a good reason it should stay, I'm going to change the ATA RAID page back to a redirect and integrate any relevant info into this article. Timbatron 07:16, 21 February 2007 (UTC)

Merge Discussion ("Standard RAID levels" into "RAID")

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.

The result of the discussion was don't merge. --Richmeistertalk 16:37, 4 July 2007 (UTC)

Okay, since I noticed we've got a merge template on this page and Standard RAID levels and there was no topic to discuss this on the chat page, here's the discussion.

I vote against the merge, because Standard RAID levels is a complete technical article on specific RAID implementations, whereas RAID is probably better suited to being a general article about the concept and its origins. Having the two articles merged would make for an unreasonably long article with too varied a scope of information. — KieferSkunk (talk) — 23:09, 2 March 2007 (UTC)

Merge and split raid 0, 1, 5, & 6 into their own articles as they are definitely sufficiently long. Raid 2, 3 & 4 can be merged back into RAID as they aren't long enough. Additionally, write a short blurb about 0, 1, 5, & 6 in RAID and {{main}} link these separate articles.
Another additionally: split JBOD into its own article as its not even RAID. There's absolutely zero redundancy to it. And I'd be happy to do all that I have suggested here. Cburnett 23:45, 2 March 2007 (UTC)
I guess I can go with that. I don't see why the individual RAID levels need separate articles, though - the Standard RAID Levels article is nice because it has internal indexing for each of the major RAID levels, making it a handy reference. And if someone wants to look up any of the numbered levels, they don't need to know which levels have their own articles and which ones don't. Agreed on JBOD, though. — KieferSkunk (talk) — 19:23, 6 March 2007 (UTC)
For the record, I'm all about keeping Standard RAID Levels out of the Main article because it would just be too long.Silvie rob 17:20, 13 March 2007 (UTC)
In my oppinion, the Standard RAID levels is a very consistent and stringent article of a very apropriate length. Splitting it up would severely harm its stringent nature, with the exception of JBOD, that should rather be shortly mentioned and linked to in "See Also" in the RAID article. If the suggestion inherrent in the merger proposal is that the introduction to the different RAID levels in the RAID article is too long, then I suggest that part be condensed a bit! - Erik Madsen, Denmark 03:56, 23 April 2007 (UTC)
Against I agree with Erik Madsen, Denmark as the Standard RAID levels article is a very comprehensive and quite lengthy article. I would suggest that RAID needs some cleaning up as an article. Davidkinnen 13:25, 27 May 2007 (UTC)
Totally agree with Erik. --Richmeistertalk 11:45, 5 June 2007 (UTC)
Against. I also agree with Erik Madsen, Denmark. About JBOD: the current redirection from JBOD to RAID is awful, as JBOD is by definition not RAID. JBOD deserves it's own article. However it would be logical to briefly mention JBOD "concatenation" idea before introducing striping. --Kubanczyk 06:45, 25 June 2007 (UTC)
Strongly Against RAID configuration is a highly complex subject and extremely deserving of its own article, as it now stands. Stuffing the different RAID levels in different articles and places would make access more difficult rather than having a single coherent discussion for those looking for it. The proposed merge would do more harm and not make any improvements. Nodekeeper 12:43, 4 July 2007 (UTC)

The above discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.

RAID and Controller Failures

With typical RAID systems (which are set up for increased reliabilty (RAID 1, for example)), not throughput), controller failures are significantly more often the cause for failure than disk failures.

So, for example, two persons set up RAID systems. The first guy's RAID conroller fails after 6 months, the other guy experiences a RAID controller failure after 3 years. Who is the lucky guy? The first one, obviously, since he'll be able to run to the next computer store and buy an identical controller as a replacement. ;-)

The problem is that different RAID controllers may organize the data differently on the disk in the array. Data which was written to a disk array with one controller may not be readable by a different controller.

Note that this may apply to software RAID as well: I once experienced a failure of a Windows NT 4 server, which used an external SCSI box with a set of disks via software RAID. As the main server box failed (not the external SCSI box), the SCSI box was simply moved over to the next NT 4 server machine. The RAID setup was not recognized there; resulting in data loss. (In theory, the data about disk configurations should be exported to an "emergency disk" right after any configuration change, so it can be reused after a failure. Unfortunately, Microsoft required the use of a 1.44MB floppy disk as a "safe" storage for this data. Which makes the prodeure of generating such an "emergency disk" impossible when this data exceeds 1.44 MB. Go figure). Fortunately, the srorage space on the external SCSI box was only used for temporary files at the time. =8-O

Anyway. The above mentioned issue is a significant (and often overlooked!) problem with RAID systems. It should be addressed in the artcile as well...otherwise, I might to begin to argue that it's not NPOV ;-)

Just joking. About the NPOV thing, the controller failure issue is real.

--Klaws 08:27, 6 March 2007 (UTC)

Unfortunately, I have to agree - I myself had a controller failure and had to use special software to recover my data. I thought RAID was going to be great, but now I'm just going to use a second harddrive as a backup and not use RAID at all. It just isn't worth it if the controller easily fails. CobraA1 09:47, 8 March 2007 (UTC)

Restore to version of Sept 2006?

It was suggested that I read the Wikipedia article on RAID. The person who suggested it was thinking of an earlier version of the article, from September 2006. For example, this is a good page:

What we really like about that page is that it contains descriptions of RAID0+1, RAID1+0, etc., that are apparently very useful.

Maybe if I have time I will get around to actually reading both of them and editing them. —The preceding unsigned comment was added by Neurogeek (talkcontribs) 21:36, 15 March 2007 (UTC).

Why do they have batteries?

Will someone in the know please post a description of the role, size, pros/cons of batteries in hw raid controllers? MrZaiustalk 21:40, 17 March 2007 (UTC)

I'm pretty sure it's to keep the cache from being lost which could yield corrupt/inconsistent data on the drives if lost. But I wouldn't say I'm "in the know." Cburnett 23:31, 17 March 2007 (UTC)
I've been outta the game for a while, but here's my take. Disk controllers are often equipped with Write-Back Cache (WBC). WBC is a chunk of memory that stores 'writes' to the disk. When an OS sends a write, it hits the cache, and the controller then signals 'write complete'. The benefit is that memory works in Nanoseconds while disk-drives work in Milliseconds. The cache is later 'flushed' down to the disk. The battery does exactly what CBurnett says. It keeps the memory 'alive', should the power fail. Once the power is restored, the cache is 'flushed' to the disks, and everyone is happy.
This is true. By the way the cache is a popular but optional add-on to RAID controllers. Generally cache has nothing to do with basic idea of RAID. --Kubanczyk 06:06, 25 June 2007 (UTC)

Explaining RAID to Mgmt

Explaining anything to management is a chore. Trying to get them to understand the cost/benes of different Raid levels is like pulling teeth! I learned a nice gimmick to do this. RAID is a 'triangle'. The three sides are: 1) COST, 2) PERFORMANCE and 3) AVAILABILITY. You can get any two, but never three. RAID-0 is good cost and performance but bad availability. RAID-1 is good performance and availability but bad cost. RAID-5 is good cost and availability but bad performance.

I'm not sure the addition to Basic Function is clear in its meaning.

The edit in question, states:

At the very simplest level, RAID combines multiple (and now even on X-large single) hard disk drives into a single logical unit.

I'm not sure I understand the meaning of "(and now even on X-large single)" and I'm guessing it should be removed. Any insights? I don't want to edit it out just because _I_ don't understand it. Dr. Zed 15:04, 22 April 2007 (UTC)

I rewrote the whole section. The "X-large single" didn't make sense at all. Cburnett 16:19, 22 April 2007 (UTC)

Merge with RAID controllers

Since the two topics are so closely related, does anyone think these two topics should be merged? Royallywasted 07:10, 6 May 2007 (UTC)

They should not be merged. I added {{main|RAID#Hardware RAID}} to Disk controller#Hardware RAID so each half could be better written. Cburnett 15:17, 6 May 2007 (UTC)
Shouldn't be merged. There should be RAID about ideas and the other separate article RAID controller about specific (i.e. hardware) implementation of these ideas. --Kubanczyk 06:02, 25 June 2007 (UTC)

Slightly incorrect

The phrase "a multi-threaded operating system... can perform overlapped I/O, allowing multiple read or write requests to be initiated without waiting for completion on each request" is slightly incorrect. A system with asynchronous I/O can also allow that. (that is not theoretical. The following would go to far to make it to this page, but An example is the original Mac OS, which even in 1984, when it only ran one application at a time, theoretically could reorder asynchronous I/O requests.

"throughput" replaced with "I/O performance"

In the lead section i can read They offer, depending on the scheme, increased data reliability and/or throughput.. I think the term throughput is misleading, since it would suggest the popular belief that RAID schemes do only affect sequential performance and do not improve non-sequential performance (such as random 2KB reads or writes), while in fact RAID is able to improve both sequential and non-sequential performance. Therefore, i have changed 'throughput' into 'I/O performance'. Anyone disagrees? --FluffleS 15:10, 20 June 2007 (UTC)


In a multi-threaded operating system (such as Linux, FreeBSD, Mac OS X, Windows NT/2000/XP/Vista and Novell NetWare) the operating system can perform overlapped I/O, allowing multiple read or write requests to be initiated without waiting for completion on each request. This is the capability that makes RAID 0/1 possible in an operating system.

I don't believe it. There is nothing about RAID which says that it needs to happen on a multi-threaded operating system, not need a request be overlapped. These may be useful things for performance (they can also be nasty things for integrity) but they are not essential. Indeed, I thought it was often an option as to whether the application you wrote wished to use asynchronous or synchronous I/O. I've been bold - give a cite if you want to put it back in. Spenny 13:19, 4 September 2007 (UTC)


We need to have some discussion on RAID 2, either here or in the 'non-standard RAID' article. There is almost nothing said about RAID 2, as to what it is, or how it can be used. I ran across an article in the December 2002 IEEE Computer Magazine that states "Raid-2 requires the use of nonstandard disk drives and is therefore not commercially viable". But that is all that it says about it. We need more information on RAID 2, even if it not used, and why it is not being used. 17:56, 27 September 2007 (UTC)

Sector failure

the meaning

Does a sector failure destroy just data from corresponding sectors, or from the entire array? It's not entirely clear at the moment. --Tom Edwards 15:02, 20 October 2007 (UTC)


"Provides fault tolerance from disk errors and single disk failure..." Correct me if I'm wrong, but if more than 1 disk fails in a RAID 1 of more than 2 disks, can't it still recover from the failure since they are all mirrored? Viper5030 (talk) 18:21, 5 December 2007 (UTC)Viper5030

You mean if disk 1 mirrors disk 2 and disk 3 mirrors disk 4. In that case yes, unless two in a pair fails. But it's arguable that they're two independent raid systems, and then no.- (User) WolfKeeper (Talk) 19:06, 5 December 2007 (UTC)
No, I think you misunderstand. It's possible to have a single RAID 1 array with more than 2 drives, where all the drives are mirrors of each other. What Viper is saying, and I agree with, is that to account for the case that there are more than 2 drives the sentence should read; "Provides fault tolerance from disk errors and failure of all but one of the drives". - Concentric (20th May 08 22:04 GMT) —Preceding unsigned comment added by (talk) 21:05, 20 May 2008 (UTC)
Please source this. Bulbous (talk) 15:59, 21 May 2008 (UTC)
Speaking theoretically it is a case of a replication-based fault-tolerant system in which a majority quorum is used to determine "which value is correct?" RAID has absolutely zero guarantees on data integrity so it is possible for both drives in a 2-disk RAID 0 to have the exact same error and you will not see an error produced by the RAID system: the data matches. (That's assuming it checked for equality to being with, which I think is a false assumption).
So if you had a 5-disk RAID 0 you could assume that if 3 or more drives read the same value then that is the correct value. If you have 3+ drives online still then you could assume that they hold the right value if they all match. If you have a 4-disk system and 2 fail then you have no majority of disks to "vote" on the correct value.
Of course, this assumes that the RAID controller verifies correctness across drives on read. With RAID 0 you can always recover from any number of lost disks provided you have at least one drive standing (just re-mirror it!). Just hope that that one disk hasn't been corrupted in any way. Cburnett (talk) 20:50, 5 December 2007 (UTC)


Where is RAID 2 (parity via hamming codes)? —Preceding unsigned comment added by (talk) 01:46, 24 October 2007 (UTC)


The comment in the description of RAID 3 doesn't make sense from a raw I/O perspective: with both RAID 3 and RAID 5 all disks are involved in the write -- there isn't any more load on the parity disk with RAID 3 than the other drives in either situation. However it would appear there is a performance advantage on reads with RAID 5 vs. RAID 3. Does anyone know if this is stated incorrectly? Are there any statistics that could show that it is actually superior read speeds with RAID 5 rather than worse writes speeds with RAID 3? (talk) 17:59, 18 April 2008 (UTC)

RAID 4 & 5 minimum disk amount wrong

While useless, you can run RAID 4 & 5 on two disks. This is, in effect, RAID 1. -- RichiH 16:43, 30 October 2007 (UTC)

...and I think for that reason it is not appropriate to describe 4 & 5 as being possible on 2 disks. Put another way, it is not helpful to describe these logical niceties, it simply confuses. Spenny 17:14, 30 October 2007 (UTC)


RAID 2 should be included in this page Although RAID 2 doesn't really have any commercial implementation, I think the article should give the reader a good overview of the different RAID implementations, adding RAID 2 will make that overview better.

As said below: RAID 2 uses Hamming codes (wiki article available) to do error checking. Other than that, RAID 2 is very similar to RAID 3. The reason why RAID 2 hasn't been a commercial success is because of the fact that using Hamming codes as error checking mechanism requires a great deal of disk space (also the raid controller has to perform complicated calculations to do the error checking).

I'm certainly not an expert on this issue, so somebody with more expertise should verify this and maybe add some good structured text about this to the article. —Preceding unsigned comment added by (talk) 14:47, 5 November 2007 (UTC)

[RAID0] Number of disks

Is possible use 3 disks to make a RAID 0? Or is necessary couple numbers (as 2, 4, etc)?

18:39, 4 December 2007 (UTC)Renato S. Yamane

Any number would work. For 3 disks: blocks 0,3,6,9 would go on disk 1; 1,4,7,10 on disk 2; and 2,5,8,11 on disk 3. Cburnett 19:08, 4 December 2007 (UTC)

[RAID0+1, 1+0] Number of disks

The article says that minimum 4 disks are required, however it is possible to use in a two-disk scenario. This is used for example in HP Proliant DL320 G4, which can only house two disks. —Preceding unsigned comment added by (talk) 13:14, 8 January 2008 (UTC)

This link also says 4:
I would reason that it is 3, 1 'real disk' and 1 disk as a raid 0. For example, 1 'real disk' of 240 GB is mirrored on a raid 0 of 2 120GB disks. But apparenly they are counting that a RAID 0 must be implemented on each one of the mirrors. I can't figure how can you call a 2 disk arrangement 0+1 as you either mirror or create one :bigger disk but you can't do both. PuercoPop (talk) 08:18, 8 November 2008 (UTC)

RAID n+1 Alternate Meaning

The '+1' is used in this article to suggest a nested raid system using 'raid 1'. However this term is also used to denote a hot-swap spare disk, probably incorrectly, in some publications. For instance, a RAID 5+1 is taken to mean a RAID 5 with a hot-swap disk which can be brought into the array in the event of a disk failure.

This may not be correct, but it may avoid confusion to readers if this was pointed out, and possibly lead to the correction of this misnomer. —Preceding unsigned comment added by (talk) 13:46, 4 March 2008 (UTC)

The first mention of anything "X+Y" is in RAID#Nested levels where it says:
Which to me spells it out well enough. Not sure where this confusion could/should be addressed. Cburnett (talk) 01:08, 5 March 2008 (UTC)

Issues with RAID

Concerning Atomicity: Are those issues also resolved with simple data journaling, or is that not enough? Esspecially are they resolved by Reiser and by ext3 with full journaling?

Concerning Unrecoverable data: Most drives remap bad sectors on write, don't they? So, if a sector is unreadable on one disk, but there is enough redundancy, the controller could simply _write_ the reconstructed sector to the disk, and it's the same as if it would use an own remapping table, wouldn't it?

--JensMueller (talk) 01:01, 10 February 2008 (UTC)

RAID5 algorithm

somewhere I found this one:

( (Drive 1) XOR (Drive 2) ) XOR (Drive 3) = (Drive 4)

I do not get this: why 3d+1p?

I think is better 2d+1p in a single xor operation

actually I thought raid5 algo was like this:

1d' 1d" 1p 2d'

2p 2d" 3p 3d'

3d" 4d' 4d" 4p

5d' 5p 5d" 6p

6d' 6d" .. ..

.. .. .. ..

...only a single xor computed on d' d" pair.

this is k complex algo, instead the above one is more and more complex (and slower) for bigger arrays.

ok, xor is fast... but N*xor is slower than only one. —Preceding unsigned comment added by (talk) 15:35, 7 September 2009 (UTC)

Beyond RAID6

RAID5 has single parity, RAID6 has dual parity.

I'm sure there are error correcting codes that can go beyond that.

What issues would occur with e.g. a (hypothetical) 6+3 RAID 6+? Are there experimental implementations of such stuff? --JensMueller (talk) 01:08, 10 February 2008 (UTC)

Well, both of those use parity. RAID 2 uses a Hamming code which I would consider "beyond" simple parity. What exactly do you mean by "beyond"? Cburnett (talk) 03:40, 10 February 2008 (UTC)
triple parity, i.e. that three devices can fail without using data. --JensMueller (talk) 09:55, 10 February 2008 (UTC)
Oh, I just saw that this is addressed in Standard RAID levels ... -- (talk) 00:05, 23 March 2008 (UTC)


I have semi-protected the page because of CONSTANT change from "Inexpensive" to "Independent". The original article by Patterson, Gibson, & Katz is titled "A Case for Redundant Arrays of Inexpensive Disks (RAID)". They coined the term and that's what it should remain unless someone comes up with a really compelling reason to override those who coined it.

If anyone has an idea how to avoid this constant change so semi-protection can be removed then I'm all ears! Cburnett (talk) 21:23, 22 February 2008 (UTC)

The problem is that we need a reliable source to indicate its current meaning. Just because it meant something in the 1980s, does not mean that it still means the same thing.
However, if no such source shows up, why not keep "inexpensive" and make a note (later, under "Meaning of Acronym", for example) to the effect that ""Inexpensive" is sometimes replaced with "Independent", but the former term is the one that was used when the term "RAID" was first coined by at Berkeley"--Ernstk (talk) 16:51, 1 March 2008 (UTC)
Second sentence: RAID is also sometimes referred to as "Redundant Arrays of Inexpensive Drives" or "Redundant Arrays of Independent Disks/Drives". That doesn't stop drive-by changing of Inexpensive to Independent. Cburnett (talk) 17:02, 1 March 2008 (UTC)
Can't the introductory sentence simply be replaced with the following?
"RAID (Redundant Array of Independent Disks or Redundant Array of Inexpensive Disks) is......."
This includes both terms immediately so any argument over the correct one is quashed. The terms also appear in alphabetical order, if anyone should argue over which appears first (!) Both terms were used recently in the final of the respected British quiz show University Challenge - if they couldn't decide on a correct answer, no-one can! --- Soulhunter123 (talk) 21:24, 2 March 2008 (UTC)
That is one solution, yes. Though I could pose the argument that Inexpensive should appear first since it was the first used term (chronological over alphabetical since alphabetical is wholly arbitrary and dumb-luck in ordering).
The whole thing is an interesting deal. The authors used "Redundant Array of Inexpensive Disks" and first introduce the acronym RAID in section 6 as "our RAID." Since then the acronym has been accepted over Redundant Array of Inexpensive Disks and then changed to mean something different (probably out of confusion). That considered, I reject that Inexpensive and Independent are equal in terms of how you propose they should be presented.
I'd propose not defining RAID and leaving that to the first section on History but I guarantee...GUARANTEE...someone will add it to the introductory sentence. The problem is people changing without bothering to read. The rest of the introductory sentence starts with "as named by the inventors" which Independent is not what they named it. So rationally explaining terms won't work either because people aren't thinking before they change it. Cburnett (talk) 22:01, 2 March 2008 (UTC)
The disambiguation page for "Raid" reads "Redundant Array of Independent/Inexpensive Disks", and has done for over a year. Can't this just go in the opening sentence, and let's be done with it? The second sentence can simply read "Originally dubbed Inexpensive Disks by the creators, the different naming convention (Independent) has since arisen within the industry." And finally, I'm totally impartial to the subject of RAID, but I think chronological order is silly for a neutral, encyclopedic point of view. Alphabetic is the way to go. --- Soulhunter123 (talk) 01:02, 3 March 2008 (UTC)
Please don't bastardize the meaning of "neutral" as it means in WP:NPOV. Cburnett (talk) 01:09, 5 March 2008 (UTC)

I have edited the introduction to encompass both names, entirely seperately, in chronological order (due to the added descriptions in parenthesis). Hopefully this will put an end to the whole issue. Personally, I still think the article is messy, particularly the introduction which drones on for about three paragraphs longer than it could be. If you could unprotect the page now, that would be great. --- Soulhunter123 (talk) 17:17, 7 March 2008 (UTC)

Agreed--KelvinHOWiknerd(talk) 07:40, 29 March 2008 (UTC)
Do we have a cite for "Independant"? Bulbous (talk) 14:36, 29 March 2008 (UTC)

There seems to be a (trivial) error in the semi-protected section "RAID — which stands for Redundant Array of Inexpensive Disks (as named by the inventor[1])" Reference 1 seems to show 3 inventors, not 1, so that should read " the inventors[1]" —Preceding unsigned comment added by (talkcontribs) 13:10, 23 September 2008 (UTC)

Done. Note that the page is no longer protected, so you are welcome to make such changes yourself in the future. --Zvika (talk) 17:32, 23 September 2008 (UTC)

Umm, if all of the disks are required in order for a RAID system to work at full capacity, then the disks are not independent. Additionally, all disks in a RAID system are working in tandem, NOT independently. Aditionally, the technology of RAID was introduced to combat the necessity of the SLED (Single Large Expensive Disk) so please change it back to Inexpensive. —Preceding unsigned comment added by (talk) 00:13, 12 January 2009 (UTC)

It's perhaps a semantic argument, but the disks can be seen as being accessed independently, at least by the controller itself. Actually reading from a Volume on those disks in the OS requires more more than one disk, but unless someone actually proposing to changing the name to Redundant Array of Interdependent Disks, I don't think it much matters either way. I'd like to re-write the blurb on the marketing reason for changing the name to at least include the obvious, which is that often, the disks are not inexpensive, it was simply the original intention. Ar-wiki (talk) 16:23, 9 February 2009 (UTC)

rewrite so that dumb people can understand it?

this article is a little bit hard for non-technical people to understand. (talk) 13:14, 13 March 2008 (UTC)

I quite agree, and I have re-written/re-structured the introduction and first section to better explain RAID in a more simplistic, readable manner. Hope this helps you and others! --- Soulhunter123 (talk) 04:27, 14 March 2008 (UTC)

Dumb people rewrite - Purpose and basics

I agree that the article lacked a simple explanation (not for dumb people, but for people who didn't know it all beforehand), and added a new "Purpose and basics" section. It's been extensively edited. I didn't entirely agree with the edit, and was trying to improve the new version, but came to the conclusion that it was basically less clear and simple than my original; I'd like others, particularly the non-technical, to consider the versions and decide what text to use. For now I've reinstated my version, extensively revised, for consideration, and would suggest that neither Soulhunter123 nor I touch the section for a while.

Basically, this section should be simple but never simplistic, and should convey:

  • RAID looks to you just like one disk.
  • Different RAIDs give you a combination of faster performance and data safety against disk failure.
  • RAID 0 uses at least 2 disks and is faster but unsafe (I originally said 2 disks, unduly restrictive)
  • RAID 1 uses 2 disks, protects data, and loses 50% of capacity
  • RAID 5 uses any number of disks, protects data, and gives more capacity that RAID 1.
  • It is possible to protect data against more than one disk failing; read the article for details.

In fact, maybe the whole section should be replaced by just the above? It's actually undesirable to give too much detail and consider all possibilities (mirroring more than 2 disks); after all, it's followed by a whopping great discussion of just about everything. This section must be simple, but I do think it should avoid being simplistic.

Details I wasn't happy about:

'extra "summary" data' seems confusing

"is written alongside the main data on a disk" it's distributed over the array, not on the same disk as the data in question - this is confusing

"data on it is reconstructed from the summary data on the other disks" it's reconstructed from the remaining good data, corrected with the redundant data. I think I made the same implication

"[redundant RAIDs] requiring, on average, roughly half the size of the main data" no

"For increased performance, there are various combinations of configurations" I think for the user unfamiliar with RAID, RAID 0 is enough. The full article details the others.

"lost or corrupted" for all practical purposes, lost. Data which is corrupted but not lost makes you think of perhaps a document with a bad paragraph

"Another approach, 'RAID 1', stores the same data on each disk in the array so that the failure of one disk causes no loss. This configuration allows the user to consume only half the total capacity of the array's disks." In practice, it's very rare to mirror more than 2 discs. And if we do, we can lose all disks except one, and we lose the capacity of all disks except one.

I'm sure what I wrote can be dissected in this way, and improved in the process.

Best wishes, Pol098 (talk) 17:07, 14 March 2008 (UTC)

The vote of a "Techy" on this subject may not be worth much but I feel this section makes the content more confusing instead of less. It is my opinion that this section adds nothing unique to the article as a whole and much of what it does describe is done out of any logical order. As a result I would suggest this section be removed from the article entirely. The very next section is mostly a repeat of this anyway. Furthermore I would suggest that RAID levels not be referenced until after the section on RAID levels, as knowing the any part of how one paticular set-up works in not integral to the understanding of the basic idea. Neverlucky 13 (talk) 22:34, 27 January 2009 (UTC)


I've removed Sixth and Tenth from the section about the original paper because they weren't discussed in the original paper. They are discussed elsewhere in this article. There are still a few internal inconsistencies in this article. Richard Manion 13:50, 22 April 2008 (UTC)

Reliability terms unclear

Failure rate "The mean time to failure (MTTF) or the mean time between failure (MTBF) of a given RAID is the same as those of its constituent hard drives, regardless of what type of RAID is employed." This does not clarify whether this is the "sticker" MTTF on the drives, or a number calculated from the average or lowest of these drives. Failure rate is not a synonym for MTBF, though they are related. Possibly both should be defined and linked to the appropriate articles. —Preceding unsigned comment added by (talk) 13:23, 26 May 2008 (UTC)

RAID Principles

A mention somewhere of information theory is possibly appropriate, given that RAID is a communications system where you're essentially sending data to yourself, and the channel is the RAID array. I'm pretty sure RAID codes are close to other communications codes, also. —Preceding unsigned comment added by (talk) 13:23, 26 May 2008 (UTC)

RAID 50 not mentioned?

I note that RAID 50 (5+0) is not explained in this article, surely it is common enough to do so? I have personally not come across RAID 53 before now! I do not feel qualified to make the changes, in general I congratulate the article. (talk) 20:25, 26 January 2009 (UTC)QuagSwag

Space efficiency

What is the source for the space efficiency figures in the article. They don't seem to be the theoretical maximums. Plugwash (talk) 22:58, 26 January 2009 (UTC)

Shouldn't RAID 1 space efficiency be n/2 ? —Preceding unsigned comment added by (talk) 16:56, 19 February 2009 (UTC)

RAID 3 Error Correction

In the section Standard Levels, the part on RAID 3 is misleading or incorrect. While the article currently states that RAID 3 can withstand a failure in the parity drive, in actuality it can both detect and correct errors in the case of disk failure of any drive. Structured Computer Organization 5th Ed. by A. Tanenbaum (note: this is my first attempt at a contribution to a wiki please excuse any deviations from standard practices) states, in a section on RAID that

     "At first thought, it might appear that single parity bit gives only error detection, not error correction.  For the case
   of random undetected errors, this observation is true.  However, for the case of a drive crashing, it provides full 
   1-bit error correction since the position of the bad bit is known. If a drive crashes, the controller just pretends that
   all its bits are 0s.  If a word has a parity error, the bit from the dead drive must have been a 1, so it is corrected."

So while the error in the article is technically small I feel it overlooks a rather large feature of this RAID level. (talk) 19:15, 27 January 2009 (UTC)

Discussion of Write Penalties associated with parity-based forms of RAID

I was really surprised to see there's no information on the various write penalties associated with RAID-5 and RAID-6.

Has there been a previous discussion on this and perhaps there were reasons for excluding the topic? (talk) 18:48, 29 January 2009 (UTC)

Not just RAID-5 and RAID-6, or even all parity based forms of RAID. RAID-1 has a write penalty as well, and it's not parity based. —Preceding unsigned comment added by (talk) 12:36, 20 March 2009 (UTC)

RAID 5 not good enough

I've noticed a growing sentiment against RAID 5 among administrators of large data centers. There was a great talk "Trends from the Trenches" at the Harvard Biomedical HPC Summit last year, which detailed the problems of RAID 5 masking serious errors until catastrophic loss occurred. They recommend RAID 6 now, and they reject any storage/controller products that are not proactive about disk scrubbing and consistency checking. See: (talk) 20:02, 31 March 2009 (UTC)

FreeBSD loader

I'm reverting the addition of "the" from the reference to loader for FreeBSD within the "Operating system based ("software RAID")" section. 'loader' is actually the name of the program

Taken from the manpage at

The program called loader is the final stage of FreeBSD's kernel boot- strapping process.

Legios (talk) 06:48, 9 April 2009 (UTC)

RAID bug killer

what does any of this have to do with the bug spray? shouldn't there be a disambiguation? (talk) 19:43, 5 June 2009 (UTC)

There is a disambiguation page: Raid
-Garrett W. (Talk / Contribs / PM) 03:41, 6 June 2009 (UTC)
That is interesting because if you google search for Raid the first result is the Wikipedia page RAID which is identical to the Raid except for capitalization. This seems quite confusing but thanks. (talk) 06:01, 6 June 2009 (UTC)
Google doesn't make a difference between capitalized and normal letters. Rchard2scout (talk) 18:07, 4 September 2009 (UTC)

Striping versus mirroring

I came here to learn, among other things, what these terms mean. I go away uneducated. Mirroring is defined in the introduction, but striping is not defined but just "used." Please add a definition for the term "striping". —Preceding unsigned comment added by (talk) 18:59, 4 July 2009 (UTC)

OS X Server raid 5

The source referred to in the page [7] is dead, this is the new page: [1] BUT it has no mention of raid 5 anymore, so I think it's canceled for snow leopard server. —Preceding unsigned comment added by (talk) 11:58, 7 July 2009 (UTC)

"Hot spare"

In the section "How parity is calculated" is the phrase, "replacement ("hot spare") drive". I do not believe "hot" is germane to the subject. "Hot" refers to the fact that the computer does not need to be powered down for drive replacement. It has little to do with the fact that the RAID array can rebuild a failed drive onto a new drive. (talk) 13:17, 29 September 2009 (UTC)

Vertical striping

We should mention somewhere vertical striping ... I'll add more to this later, but just a reminder! —Preceding unsigned comment added by Fishsponge (talkcontribs) 16:04, 23 October 2009 (UTC)

Fake RAID: hardware or software?

I think this edit is incorrect, because the paragraph starting with Since these controllers use proprietary disk layouts... refers to the one that precedes it, Because these controllers often try to give the impression of being hardware RAID controllers.... I think the real error is that those two paragraphs should go at the end of the Hardware-based section, not the Software-based section. --Fstanchina (talk) 16:41, 27 January 2008 (UTC)

I don't think so, because the controllers that "use proprietary disk layouts..." are hardware, while software RAID is the one that tries "to give the impression of being hardware RAID...".
-Garrett W. (Talk / Contribs / PM) 03:35, 22 November 2009 (UTC)

Two partitions to emulate Raid 1E

I added a fact tag to the suggestion an administrator could make two partitions and then mirror across disks and stripe between partitions as I think it needs to be attributed to a reliable secondary source making that suggestion. It seems a rather odd suggestion to me as it would be particularly punitive on the writes since you can't sequential write (with 1E I presume the mirrored strip [2] is written sequentially). Perhaps this config is useful in some cases, I've never set up a server or RAID array and I guess if all you care about is read it should be fine. Also the sentence needs to be improved if it's kept. As I understand it, to emulate a Raid 1E config, if you have drive A B C with each having partition 1 & 2 you need to do something like have A1 and B1; C1 & A2; B2 & C2 as a set of mirrors. You then stripe across these 3 mirrors. Alternatively, you need to make 2 stripe arrays each containing 1 partition from each disk and then you mirror across these stripes but you need to make sure the mirror isn't on the same drive. I guess if it's possible to specify the order of disks in the strip array (so for example stripe array 1 writes to disk A then B then C while strip array 2 writes to disk B then C then A) then it should be fine. The sentence doesn't really convey that, at least to me. Indeed I thought it was suggesting you mirror across each 3 disks and then stripe between partitions on a disk which gives you a non-sequential mirror array with the capacity of 1 disk. Nil Einne (talk) 13:12, 7 October 2008 (UTC)

Wow that's confusing. I don't have a clue as to exactly what you're trying to describe.
-Garrett W. (Talk / Contribs / PM) 03:35, 22 November 2009 (UTC)

Strategy comparison for RAID implementations

The aforementioned 1E is getting down to the "partition level" and seems to be more likely in a software-implemented RAID strategy, but if there is a reference that could be cited for the fact tag purposes for either hardware, or any implementation of this, I suppose that would clarify the idea. Also, there is a merge / new article proposal for a section from Vinum volume manager at the moment. --Kuzetsa (talk) 20:55, 1 November 2008 (UTC)

Remove 'backup' in Purpose and Basics?

In the Purpose and Basics section, it says that RAID 1 "could be described as a real-time backup solution." I would recommend changing it to read "RAID 1 (mirrored settings/disks) provides redundancy for saved data" or something like that. RAID 1 is not a backup because you can't revert to a previous version of a file and can't recover deleted data any easier than you could with a single disk. If data is corrupted from bad blocks on a single drive, chances are good that it can be recovered/rebuilt from the mirror (good drive), but again, that's redundancy, not backup. Werikblack (talk) 16:22, 18 March 2009 (UTC)

I agree. I'd probably go with something along the lines of "RAID 1 (mirrored settings/disks) provides redundancy for existing data in situations where up to n-1 disks can fail without loss of data" or something along those lines. I have seen some text books list RAID 1 as a "backup solution", but agreed - I don't think it should be written to seem like a backup solution.
Legios (talk) 02:11, 19 March 2009 (UTC)
I've had a stab at rewriting it. Hopefully it sounds a bit less.. backup-y.
Legios (talk) 01:05, 20 March 2009 (UTC)
Looks great! Thanks for the edit! I like the images too.
Werikblack (talk) 05:01, 1 April 2009 (UTC)
Agree Fault tolerance is certainly not the same as backup. Consider the case of a malware infection: with RAID-1, both drives would be equally infected, and neither would provide an uninfected copy to restore.
-Garrett W. (Talk / Contribs / PM) 03:35, 22 November 2009 (UTC)

Performance impacts of software RAID

From the article: "With traditional "real" RAID hardware, a separate controller does this computation. In other cases the operating system or simpler and less expensive controllers require the host computer's processor to do the computing, which reduces the computer's performance on processor-intensive tasks."

I feel that this needs some qualification. "reduces the computer's performance on processor-intensive tasks" should read "reduces the computer's performance on tasks that are both processor- and disk-intensive" -- if the RAID isn't in use, it doesn't affect performance at all. Also, for the most part reads and writes to the software RAID are I/O bound, even under RAID-5. Modern CPU's can perform the RAID computations fast enough that they are always waiting for the disk to catch up, so the real difference in performance between hardware and software RAID is often negligible.

I don't know enough about CPU pipelining to say this authoritatively, but I also think it's likely that the non-I/O computations of a given task (that is, the ones that would be slowed down by the CPU having to do software RAID computations as well) can often be run in parallel with the RAID stuff, so they're not waiting on the CPU. (If those tasks *do* need CPU registers or components that are being used for RAID computation, odds are it's because they wanted a disk access, and if that's the case the task is I/O bound anyway, not CPU bound.)

In addition to all of that, multi-core processors would just offload RAID stuff to a different core. This would have zero performance impact on applications unless they were well designed to take advantage of multiple cores (which most applications are not), and are performing tasks that are parallelizable (many tasks are not). (talk) 22:39, 26 March 2009 (UTC)

My experience is basically consistent with your observations. In fact you may have understated the case a bit. Applications can be both CPU and I/O intensive but quickly alternate between those resource requirements rather than needing both at once. An app may grab a chunk of data off one storage pool, crunch it, write the result out to another storage pool, grab another another chunck, etc. Even if the crunching is very demanding, if it can't happen until the next chunk of data shows up, then the CPUs are free to do storage related grunt work in the mean time. So I'd say "tasks that are simultaneously CPU and I/O intensive." Unfortunately I don't have a handy source to cite other than my own experience configuring platforms for applications that work exactly that way. —Preceding unsigned comment added by (talk) 17:37, 5 December 2009 (UTC)

Raid10 and minimum drive requirement

The definition of Raid10 provided on this wiki page states it is possible to implement with 2 drives. However the external links at the bottom of the page only show support for a 4 drive implementation as the minimum. Can you clarify how it is possible to implement Raid10 with just two drives? Thanks. Rjdevine (talk) 17:19, 20 November 2009 (UTC)

I am also not aware of RAID 1+0 being possible with 2 disks. If no one shows how this is possible within a reasonable amount of time, I'll probably just take that bit out of the article.
-Garrett W. (Talk / Contribs / PM) 03:35, 22 November 2009 (UTC)
AFAIK in practical terms RAID 1+0 and 0+1 both require a minimum of four drives. The semantics of some volume management software is such that constructing nested RAID sets involves combining two or more entities each of which must consist of one or more drives, so within those microcosms it would seem theoretically possible to have 1+0 with two drives, but not really. I suspect that's the source of the confusion. Taking that bit out sounds reasonable. —Preceding unsigned comment added by (talk) 17:01, 5 December 2009 (UTC)
OOPS! Just noticed that is comment specifically addressed RAID 10. The problem with RAID 10 is that it's a non-standard level which different vendors can implement in different ways. The only thing I would expect from something called RAID 10 without knowing the details is basic behavior something vaguely like some combination of RAID 1 and RAID 0. I would not infer anything about the configuration procedure or data layout from this term. (I've seen at least one RAID 10 implementation that was unique as far as i know, but worked just fine.) —Preceding unsigned comment added by (talk) 17:46, 5 December 2009 (UTC)

Needs a more practical "configuring SATA RAID" section

The overwhelming majority of people looking at this article are trying to figure out what are the implications of lines like this in the spec for a new mainboard:

  • SATA 6Gb RAID 0
  • SATA 3Gb RAID 0, 1, 5, 10

That is exactly the specification you will see presently from Gigabyte describing the capabilities of their P55- and also X58 (Intel LGA1156) and 790FX (AMD AM3) mainboards. It's also exactly what you can expect to see in the first generation from ASUS, ASRock, MSI, Eclipse, and so on. So what this means is that the people who are choosing whether to buy this part or not, or what it might be able to do with their new/expensive SATA6Gb SSD discs and their legacy SATA3Gb discs, really need simple coverage of the tradeoffs in using two SATA6Gb in a RAID 0 (the mainboards at present only support two SATA6G connections, that's a limitation of the Marvell SE9128 controller that these boards will be using probably until 2011/2012 when this capability goes in the main chipset.

Obviously (to anyone who understands RAID), it's the OS and applications and at most a few settings files that should be on the RAID0 array which is an 800 mile an hour gokart.

User data should be on something like a SATA3 RAID5 or (if you must) RAID1 or RAID 10 or just kept on USB3 drives (all the SATA6G motherboards support 4.8Gbps USB3 as well) or (best) a file sharing router with only a local cache of data on the scary RAID0.

So the most common configuration problem is going to be setting up two SATA6G drives in a RAID0 for OS and applications and two or more SATA3G or USB3 drives in a RAID 1/10/5. Factors like how identical the drives must really be (identical models with same number of platters and so on, or just nearly identical capacities?), whether you can really do it with two and if so with what controllers, whether RAID 5 is advisable at all (see the point above about major data centers moving to RAID6 exclusively), all should be covered.

But as for the use of RAID and SATA6G, there really is no other sensible configuration (you aren't using SATA6G for reliability you're using it for speed, and you can't set up a RAID 5 or even a RAID 10 or RAID 1 with current motherboards) for the next year or so on current desktop PC technology, so this should be fully explained.

It would take no more than a paragraph to explain this and would save a lot of data. ;-D —Preceding unsigned comment added by (talk) 17:28, 4 December 2009 (UTC)

This is great stuff, but I'm not sure it belongs here. Come to think of it, I might say the same about my own comment about hardware economics elsewhere in this discussion. The topics of disk configuration, volume management, RAID level selection, etc, may be involved enough to deserve a separate article. —Preceding unsigned comment added by (talk) 17:51, 5 December 2009 (UTC)

RAID0 vs. faster bus

Another issue worth exploring in this article is whether RAID0 performance is worth the risk or whether simply moving up to a faster bus speed with a single disk is a better choice. With USB3 getting about 5.5x the burst speed of USB2 in high end systems of the type that might go SATA3G or SATA6G RAID0 instead, it's legitimate to ask whether the complexity and risk of RAID0 configuration is worth it at all. It may have been a stopgap, like earlier IDE and ATA RAID systems were, until faster and wider buses came along. At 4.8Gbps (USB3) and 6Gbps (SATA6G) the PCIe channels are saturated anyway and one starts to interfere (marginally) with graphics performance on a typical desktop PC because of use of multiple channels. One doubts if any RAID0 configuration is going to speed up disk access by 5.5x or even double.

So maybe there needs to be a big caveat that unless you have two drives doing nothing of any great value that are identical or nearly so, a RAID0 slaving them together to run an OS and applications may actually be outperformed by one new hard drive on USB3 or SATA6G. —Preceding unsigned comment added by (talk) 18:44, 4 December 2009 (UTC)

The comment re RAID 0 vs faster bus describes a specific example of a more general idea that could stand more emphasis.
There are numerous factors that can effect overall disk subsystem performance including the performance characteristics of the disks themselves, the characteristics of the disk connection, the I/0 bus, the amount of main memory in the system, size of file system buffers, write-back vs write-through caching, etc.
Covering all of these in detail would probably lead to unwarranted scope creep. Probably best to simply mention that there are a variety of considerations that affect disk performance and that performance and robustness characteristics attributed to RAID levels hold only when everything else remains the same. —Preceding unsigned comment added by (talk) 16:51, 5 December 2009 (UTC)

"Three Mirrors"

Worth noting that mirrors (RAID 1) can have two storage entities mirrored against each other but are not restricted to only two. Three or more are allowable as part of the standard and many volume managers support this.

having more than two mirrors not only increases robustness and read performance, but facilitates some interesting data management techniques like splitting off a mirror, dumping it to tape, and then re-syncing it, etc.

in order to detach and reattach mirror elements in this way they need to be able to function independently of each other. so 0+1 and 5+1 work for this but 1+0 does not. Might be worth pointing this out in discussion of 0+1 vs 1+0. —Preceding unsigned comment added by (talk) 17:09, 5 December 2009 (UTC)