I'll use this and add to it my comments in bold..
Originally Posted by trueg50
RAID 0 is great for games, etc.. but no redundancy, though most just get another drive and use that. Capacity is the number of drives x the capacity of the drives.
True. You will get virtually double, if not slightly higher than 2 x performance by utilizing two disks. This will not only double your transfer rates but also cut your seek times down and the advantage of having a raid controller to do your calculations. Better performance gains (albeit, very minimal either way) for using an external controller VS onboard, as external controllers have a dedicated CPU to do the calculations. RAID-0 requires virtually zero processing.
RAID 1 basically the two drives will be identical, one dies, just pull it and run the other one. Capacity is 1/2 x the number of drives.
The advantage of this is a live backup. However, this comes at a performance loss when WRITING data, as the data has to be written twice and you will notice some impact on your controller, even though technically it's writing data at the same time. You will notice a performance gain READING as you have two disks to read off and the RAID controller is able to utlilize both.
The capacity would be 320GB if you were running 2x320GB @ RAID-1. You basically lose one if your disks.. just to clarify
RAID 5, minimum of 3 drive this time. Tough to explain, but capacity is the number of drives you have - 1 x capacity
The advantage is the fact you again have great performance, and redundancy. You will notice similar performance with 3-RAID-5 disks, as 2xRAID-0.
If a hard drive fails, you can simply replace this with a fresh disk. The 1st and 2nd (good) disks have 1/2 of disk #3 on them each (Parity). They are then able to re-write the information back to the fresh disk. This is great for file servers.
The advantage is you get great speed writing, reading, and the parity is done by the controller. You notice a slight degradation in write speed vs 3xRAID-0 due to the fact it takes time to write the parity.
RAID-5 is my personal favourite for it's 3 factors: Performance, Redundancy and it's Economical (Vs say, a RAID-1 which is UNeconomical)
RAID 0 + 1
You have 4 drives, in 2 pairs, each pair is in RAID 0, and then the two pairs come together in RAID 1. Capacity is still number of drives x 1/2 x capacity.
Possibly the most uneconomic, however would be the choice if you had one. This yeilds the advantages of the RAID-0 speed, and also, the advantage of having this mirror onto a 2nd RAID-0 volume, effectively creating two independant RAID-0 (striped) volumes but maintaining a live backup.
You would notice a slight performance decrease from standard RAID-0 due to the backup being written but the live backup advantage is an awesome selling point. If you're going to go this option, it's best to look into RAID-5 first.
A lot of your 'live backup' performance issues can be resolved by an external raid controller HOWEVER, for standard RAID-0 it should not make a difference.
Now for another question, in terms of patritioning up a RAID volume.
Now, you will be introduced to something called the RAID-BIOS when you wish to configure your raid. This basically is where you say which disks is in what array - assuming you want to run multiple arrays on one controller (yes, it's possible to say have 6 disks, four @ raid-5 then 2 X raid-0 creating an OS/ System volume and a redundant storage volume, which is exactly the setup I run..)
Once this is configured, basically, Windows (I'm not sure about other OS's..) sees it as one volume and really doesn't know much better. The raid BIOS tells windows, hey, here is the hard disk, it's 1TB and called Volume0 (or something else if you rename it)
You may then partition the drive up once in windows setup.
If you are running a 120 GB RAID-0 set, you can just set your windows disk to 60gb, then re-allocate the other 60gb in Disk Management as it was a standard hard disk.
In essence, windows THINKS it's a single hard disk.. and can treat it accordingly.
Your second volume would appear as 'Unallocated' black space, and you could also sub-partition this drive up.
DO NOTE: If you delete your C: partition, and reinstall windows, I have experienced an issue where the MFT is moved to the D: partition, then C: becomes Partition1, D: being Partition0 (and this cannot be changed back) it's getting pretty advanced, in essense, you may end up with Windows thinking the system drive is D: .. its hard to explain, and there's no real way I saw to fix it other than blowing the entire RAID-SET away (losing all the data with it) and re-creating it, and reinstalling as partition-0.
You will become familiar with the terms Partition and Volume which are really important when referring to RAID.
I hope this helps. RAID can be a doozy if you arent familiar with it, but it's the most dead simple concept in computing I've come across. It baffles most people, but takes only minutes to learn .. once you setup and configure an array yourself, you will basically be a RAID expert, as there's not much more to it!
PhireX! (now server admin..)
PS: RAID-6 = RAID-5 WITH TWO LIVE BACKUP DISKS!!
THERES SOME RAID TRIVIA FOR YOU, IMPRESS YOUR NERDY FRIENDS!