• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

[OFFICIAL RAID STICKY] Must read article on the history/uses/definitions of RAID

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

PYROMANIAC

Member
Joined
Dec 11, 2001
Location
Houston, Texas
RAID?? Striping? ever wonder what RAID is all about?....read....

I was out to find out about RAID and stumbled across this article explaining everything, imo, about what the basics of raid are and data striping and what RAID0,1,2,3,4,5 are and the uses of these for normal/server use. It is a very good read and I hope you get as much out of it as I did.

http://www.uni-mainz.de/~neuffer/scsi/what_is_raid.html

This also explains the history of RAID and how it originated

Edit by David, to add:

thegreek said:
Complete description and an easy-to-understand diagram of RAID levels. It also lists the advantages and disadvantages of each RAID level.

http://www.acnc.com/04_01_00.html
 
Last edited:
Great post, and very informative. Thanks for sharing! I've bookmarked these links myself and I recommend everybody else does the same.

Good find right here, and great thinking sharing it with us!
 
donny_paycheck said:
I see...not enough info for me to be able to tell what it does for level 50 though.

For 600 bucks and 3 U160 channels though, that isn't all too expensive - especially if it can do these weird RAID types.

U160 RAID.... *droooooooooool*
 
Nice links, I wanna get 3 18GB SCSI's running raid 0, but don't have the $$ so I think I'll add another 40GB IDE and make a rade setup....
 
scsi is faster which means more heat which isn't good for ocing. Just get ide 8mb cache hds or at133.
 
Royal said:
scsi is faster which means more heat which isn't good for ocing. Just get ide 8mb cache hds or at133.

A simple fan on the HDDs, or watercooling the HDDs will do the trick. However, another disadvantage to SCSI is the noise of the drives, most ppl say its a high pitched whine (mine is still in the box-havent built my system yet)
 
can someone explain or point me to an article that explains the difference between mirroring and parity??

i know that mirroring actually writes the same data twice, as in RAID 1, for instance.

and i think that with parity it does NOT actually write/store the data twice, but it can be used to reconstruct the data?? how?

I would tend to think RAID levels with parity are faster than those using mirroring? correct me if i am wrong on this (basically just a guess)
 
ok, read over that article (for the second time, infact)

from what i gather, parity is more for checking to see if the transmitted data is "correct" or not, whereas mirroring is actual copying of data to another drive or array.

so in the event that a parity check is bad, the disk must resend the information.

I'm still not clear on something though - in the article on some RAID levels with parity it says things like "rebuilding the array is difficult..."
but if a hard disk actually failed, how would it be rebuilt with just parity check data? I can see that mirroring the backup data would be an easy fix, but I am still not quite understanding how parity data can "rebuild an array" after a hard drive failure. maybe i'm not fully clear on the exact definitions of "failure" and "rebuild" ???

thanks for the link to the article, very informative.
 
Albigger said:
ok, read over that article (for the second time, infact)

from what i gather, parity is more for checking to see if the transmitted data is "correct" or not, whereas mirroring is actual copying of data to another drive or array.

so in the event that a parity check is bad, the disk must resend the information.

Exactly.

I'm still not clear on something though - in the article on some RAID levels with parity it says things like "rebuilding the array is difficult..."
but if a hard disk actually failed, how would it be rebuilt with just parity check data? I can see that mirroring the backup data would be an easy fix, but I am still not quite understanding how parity data can "rebuild an array" after a hard drive failure. maybe i'm not fully clear on the exact definitions of "failure" and "rebuild" ???
Parity striping basically sprays the image of a disk across all the other disks in the array in little pieces. When that disk fails, it can be rebuilt with the bits and pieces collectively gathered from all the other disks. That's how RAID 5 works. Here's my explanation from this thread:
RAID 5 is called "striping with parity". It offers redundancy, meaning if a disk in the array fails you can swap it and lose no data, but it uses the space much more effeciently than RAID 1 or RAID 0+1 (sometimes called RAID 10). My RAID 5 is 120gb of total disk space, but 90gb of it is usable. The remaining 30gb is reserved for parity information. For example, each drive stores parity data, or a "skeleton" of the other three drives on it. This way if one drive fails, when the new one is installed the controller can reconstruct the entire disk using combinations of the parity data on the three disks that did not fail. It's really sweet how it works. It isn't used nearly as much as RAID 0 because it requires a dedicated controller to process the striping algorithms because they are so complicated. It also doesn't offer a performance increase, although in my case I noticed one because I was moving from a CPU-reliant RAID card to one that used no CPU time. It is most often used in servers where uptime is critical because you can even hot-swap and rebuild a bad disk without taking down the array - in most cases. An example would be the server for these forums. I'd bet my money that skip runs some flavor of RAID 5.

This was an example for a 4 drive RAID 5. It's pretty advanced in its workings and I don't understand the mathematics behind it but the end result is a bit easier to grasp. Hopefully this helps.
thanks for the link to the article, very informative.
Not a problem. :)
Here are the several links I have stored here for explaining RAID to myself and others:
ARS Technica: Article written for the layman; great for an intro to RAID 0 especially - the kind that's good for us.
SearchStorage.com: This has concise explanations of each level, all on one page. Greater detail in following pages.
Raidweb.com: This has graphical and text explanations like the link I showed you before. You might prefer it to the one I posted earlier.

Those in addition to the ACNC.com link I posted above should have at least one method of explanation that suits anybody reading this. There are tons of RAID types out there but 0, 1 and 5 are the only ones I've ever seen implemented as real-world solutions. The rest seem to be created for academic purposes.

edit- I just realized I posted this earlier in the thread...no need to post it again. Oh well. :eek:
 
Last edited:
Thanks to all of u who posted links.
It really helped me, and i will see if i can find u some of my own.
 
Which is better? I would like to stripe and mirror. Is RAID 0+1 good for that application? I'd like to use it for gaming and storing files. Appreciate the answers Ladies and Gents.
 
Back