Disk Enclosures
cancel
Showing results for 
Search instead for 
Did you mean: 

Raid 1+0 Raid 10

Bob_233
Occasional Contributor

Raid 1+0 Raid 10

I have a question about RAID 10. Is there a place on the HP website that explaines Raid Levels? From searching on the internet it looks like RAID 10 is also refered t0 as RAID 1+0 or 0+1 but in other places raid 0+1 is different to Raid 1+0. Does anyone have more information on this?
12 REPLIES
Michael Schulte zur Sur
Honored Contributor

Re: Raid 1+0 Raid 10

Uwe Zessin
Honored Contributor

Re: Raid 1+0 Raid 10

Bob,

I have never seen an 'official' definition for RAID10, RAID01, RAID1+0 or RAID0+1. What does the first number define? The lowest RAID-level or the one above? There can be arguments for either way.

To avoid that confusion I call them 'striped mirrorsets' (a RAID-0 that includes multiple RAID-1) or 'mirrored stripesets'. It depends on the ability of the storage controller what combinations it supports.

I have also seen customers who run one RAID-level (often RAID-0) on the storage system and then mirror (use RAID-1) the data to another storage system using host-based software - clearly 'mirrored stripesets'.
.
Michael Schulte zur Sur
Honored Contributor

Re: Raid 1+0 Raid 10

Hi again,

on a hsg80 you have mirrorsets, which you add to a stripeset. In the end it should make no difference, just a different hierarchy.

Michael
Uwe Zessin
Honored Contributor

Re: Raid 1+0 Raid 10

Michael, it can make a _big_ difference.

Let us take a look at the differences:

- mirrored stripesets:

You bind four disks with 36 GBytes together to form a 144 GByte stripeset. Next you create a second stripeset that way. On top of that you create a mirrorset to keep the contents of both stripesets equal. The resulting mirrorset and what the hosts sees is, of course, 144 GBytes.

What happens when a single physical disk fails? On all implementations that I know this hierarchy is a strictly layered one - the upper layer (your mirrorset) does not know how the lower ones (your stripesets) are implemented. The result is a complete failure of the stripeset.

Now, the HSG does not permit this type of layering, but on host-based solutions I have seen, this results in the defective stripeset being 'split off' from the mirrorset.

You have to create a new stripeset. You cannot simply replace the failed disk with a new one, because this is missing the metadata. You need to create a new set which updates the metadata on all physical disks. Only then you can add the 'repaired' stripeset back to the mirrorset.

Now, the mirroring mechanism does not know anything about the underlying object - it could be a 144 GByte disk as well. It will start mirroring ALL 144 GBytes again, just because a single 36 GByte disk has failed.

- striped mirrorsets:

As you correctly say, this is what the HSG can do. You take two 36 GByte disks and create a mirrorset from them. You do this four times and get four mirrorsets, each with a size of 36 GBytes. Next you go and create a single stripeset that uses those four mirrorsets. That stripeset is, surprise, 144 GBytes in size. So, for the host this makes no difference.

Now, what happens when a single physical disk from a mirrorset fails? Again, the host can continue to work. Even better: the stripeset continues to work, too. As long as there is an appropriate spare-disk the HSG will go and add it to the mirrorset.

And here is the big difference:

You loose a single physical disk with 36 GBytes of data and you only need to copy that 36 GBytes over to the replacement disk.

Even better - if you loose only one disk in both stripesets of the first example you have lost the entire LUN. For the second example you would have to loose both disks of a single mirrorset.

Perhaps someone with a stronger mathematical background than me can calculate the propabilities(sp?) for a total failure for each example, please?
.
Karthik S S
Honored Contributor

Re: Raid 1+0 Raid 10

There is doc on veritas site explaining this (relevent to VxVM),

http://seer.support.veritas.com/docs/231471.htm

-Karthik S S
For a list of all the ways technology has failed to improve the quality of life, please press three. - Alice Kahn
Darius Rudzika
Occasional Advisor

Re: Raid 1+0 Raid 10

Hello,
on HP website dont know, but try this one:

http://www.acnc.com/04_01_00.html

Best regards,
Darius
"In pure practice - everything woks, but nothing clear. In theory - everything clear, but nothing works. In most favorable case when theory meets practice - nothing works and nothing clear"
Michael Schulte zur Sur
Honored Contributor

Re: Raid 1+0 Raid 10

Uwe,

at first it seemed logical, what you said. But on second thought, I would see no necessity to disable the stripeset because of one failed disk? This would mean, that a failure in the other stripeset would take the mirrorset down. This would be not reasonable and stupid in my eyes. The rest of the impaired stripeset could be updated until the defect disk is replaced. So, why the risk?

Michael
amhakassa
Honored Contributor

Re: Raid 1+0 Raid 10

Hi Bob,

May this anser this question?

RAID 10 is Striping of mirrored arrays. Or 2 RAID 1s with a RAID 0 to span them together. Data is striped across disks, as in RAID 0 and each disk has a mirror disk. RAID 10 provides 100% data redundancy and supports larger volume sizes, but at a relatively high cost. RAID 10 can have 2 drives fail (one on each array) and data is still available. RAID 10 provides the best performances for most applications where redundancy is required and cost is not a factor. RAID 10 requires a minimum of 4 disk drives, 2 mirrored sets striped via RAID 0 *** Arrays are striped together not mirrored ***

Cheers
Amha Kassa
Wolfram Eckert
Occasional Advisor

Re: Raid 1+0 Raid 10

Hi All,

some limitations on what was said:

1)RAID 0+1: If your RAID1 contains two RAID0 Stripesets (lets say A and B), then they act separately as 1 physikal Disk each in the RAID1 (Mirrored Disks).
As we all know, if 1 Disk in a RAID0 Stripeset is gone, the complete Stripeset is gone at the same moment, no matter if it contains 2 or 15 or 50 Disks!
- So, you lost (logical)Disk A of your 2 Disks in the Mirror. It makes no difference to disable or not Disk A - its gone, and you are no longer redundant.
Next: if any of the physical Disks in the other Stripeset (Logical Disk B) defects, the other Side of the Mirror is gone.
Means: You can configure a RAID with 54 Disks, if you loose 1 Disk in each Stripeset, the hole RAID 0+1 is gone!
A quite risky configuration without advantage.

2) RAID1 1+0: You take the same 54 Disks, but build 27 RAID1 Sets (Mirrors), each redundant.
Next you do a Stripe (RAID0) over this 27 (logical) Disks (A,B,C,D...).
As long as you not loose 2 Disk in the same Mirror, you are still working on your Data, even if you loose 27 Disks (an extreme Picture)!

3)Rebuild:
RAID 0+1: You have allways to rebuild your complete Stripeset - 27 Disks (A) to copy from the hopefully working Stripeset B.
RAID 1+0: You rebuild the 1 (physical) Disk that oyu lost.

4) A HP Smart Controller does this automatically for you that way, - you have no joice.

Sorry for my English I am not a native Speaker as you can see.
- Hope that helps!

Wolfram




Alicia White
Esteemed Contributor

Re: Raid 1+0 Raid 10

Here's a document I downloaded from HP some time ago: I don't remember where I got it, though. I probably got it from a NetServer manuals page since that is my main area of interest.

I also tend to rename these documents when I find them because HP usually gives it some number for a title that is indecipherable to us mere mortals. So, this isn't the original title but it is an HP document that does a thorough and pretty clear job of explaining the different RAID levels.

I hope you find it useful.

Alicia
Alicia White
Esteemed Contributor

Re: Raid 1+0 Raid 10

I found a "Raid Tutorial" PDF on HP's site. It looks to be better than the other one I posted: it explains everything (difference between logical & physical drives, good explanation of each RAID which details the advantages/disadvantages, etc.).

http://h200007.www2.hp.com/bc/docs/support/SupportManual/lpn12749/lpn12749.pdf

Alicia
Uwe Zessin
Honored Contributor

Re: Raid 1+0 Raid 10

Michael,
on mirrored stripesets the 'mirrorset handler' does not know how to 'poke' into the underlying stripesets - it just sees two large arrays of logical blocks that it sends the write (and read) requests to. It is the 'stripeset handler' that decides how the disk blocks are mapped (think about the chunk size).

Sure, it would 'make sence' if it worked the way you describe, but that would require special knowledge of the mirror and stripe handlers - you very likely won't find that in a strictly layered architecture.


Alicia,
thanks for those documents - they are indeed useful.
.