Maximum PC

Hitman's RAID Guide

Questions about RAID come up all the time in this forum. I would like to give some general information about RAID, and some links where you can obtain more in depth information about RAID. I would also like to clear up some common misconceptions about RAID and a brief overview of the different RAID Levels available.

What is RAID?

Short for Redundant Array of Independent (or Inexpensive) Disks, a category of disk drives that employ two or more drives in combination for fault tolerance and performance. RAID disk drives are used frequently on servers but aren't generally necessary for personal computers.

There are number of different RAID levels:

Level 0 -- Striped Disk Array without Fault Tolerance: Provides data striping (spreading out blocks of each file across multiple disk drives) but no redundancy. This improves performance but does not deliver fault tolerance. If one drive fails then all data in the array is lost.
Level 1 -- Mirroring and Duplexing: Provides disk mirroring. Level 1 provides twice the read transaction rate of single disks and the same write transaction rate as single disks.
Level 2 -- Error-Correcting Coding: Not a typical implementation and rarely used, Level 2 stripes data at the bit level rather than the block level.
Level 3 -- Bit-Interleaved Parity: Provides byte-level striping with a dedicated parity disk. Level 3, which cannot service simultaneous multiple requests, also is rarely used.
Level 4 -- Dedicated Parity Drive: A commonly used implementation of RAID, Level 4 provides block-level striping (like Level 0) with a parity disk. If a data disk fails, the parity data is used to create a replacement disk. A disadvantage to Level 4 is that the parity disk can create write bottlenecks.
Level 5 -- Block Interleaved Distributed Parity: Provides data striping at the byte level and also stripe error correction information. This results in excellent performance and good fault tolerance. Level 5 is one of the most popular implementations of RAID.
Level 6 -- Independent Data Disks with Double Parity: Provides block-level striping with parity data distributed across all disks.
Level 0+1 – A Mirror of Stripes: Not one of the original RAID levels, two RAID 0 stripes are created, and a RAID 1 mirror is created over them. Used for both replicating and sharing data among disks.
Level 10 – A Stripe of Mirrors: Not one of the original RAID levels, multiple RAID 1 mirrors are created, and a RAID 0 stripe is created over these.
Level 7: A trademark of Storage Computer Corporation that adds caching to Levels 3 or 4.
RAID S: EMC Corporation's proprietary striped parity RAID system used in its Symmetrix storage systems.

http://en.wikipedia.org/wiki/Redundant_ ... dent_disks
http://www.webopedia.com/TERM/R/RAID.html
http://www.acnc.com/raid.html

History of RAID

RAID has been around for a long time on the server/workstation side of the house. You can read a more detailed history of RAID in the links below. In 2000, and 2001 IDE Raid controllers started to be seen in Home Computers. Company's like ABIT and MSI started to put IDE RAID controllers on their motherboards. IDE RAID quickly became popular with Enthusiast and Gamers. RAID in Home PC's became even more popular with the introduction of SATA and the WD Raptors. Home users were able to have the benefits of Fast 10K RPM Hard Drives, without having to invest in expensive and complex SCSI drives and controllers.

http://www.staff.uni-mainz.de/neuffer/s ... _raid.html
http://searchstorage.techtarget.com/sDe ... 32,00.html

RAID 0 vs Single Drive

The argument is that RAID 0 is not faster then a Single Drive. In most real world applications the speed increase from RAID 0 is so minimal that most people wont even notice a difference. RAID 0, biggest fall back is that it does not have fault tolerance, so if one drive fails all data is lost. The more Hard Drives you add the higher your chance of drive failure. The difference between RAID 0 and a Single drive depends on the Hard drives you use. For example two 36GB WD Raptors in RAID 0 are not faster then a Single 74GB Raptor, but 2 36GB WD Raptors in RAID 0 are faster then a single 36GB Raptor. More information about RAID 0 Vs Single Drive can be found in the links below.

http://www.techreport.com/reviews/2004q ... ex.x?pg=27
http://www.anandtech.com/storage/showdo ... =2101&p=11
http://www.devhardware.com/c/a/Storage- ... r-Home-PC/
http://faq.storagereview.com/SingleDriveVsRaid0

Hardware RAID vs Software RAID

Software RAID is less expensive then Hardware RAID. Windows XP for example offers RAID 0, RAID 1, and RAID 0+1 in the PRO version and with a hack you can enable RAID 5 in XP Pro. Software RAID does not require a RAID controller, the drives can be connected to any IDE or SATA controller. Software RAID can be transferred from one computer to another. If you take your Hard Drives that are in a RAID configuration and put them in another computer, XP will recognize that they are in RAID, and you can import them with out loss of data. The major drawback to Software RAID is performance. Software RAID is dependent on CPU resources, so performance take a major hit. Its also less stable then Hardware RAID.

Hardware RAID is more expensive, and requires a RAID controller. RAID is setup in the controllers BIOS, the RAID Levels available depend on the controller that you are using. Hardware RAID is typically more stable, and uses minimal CPU resources. Most newer motherboards have onboard RAID controllers. The higher end motherboards onboard RAID controllers, offer more RAID Levels. Hardware RAID is not dependent on software. You may be required to install drivers (F6) for your Controller during XP setup for all your RAID array to be seen by windows. Hardware RAID is the preferred method.

http://searchstorage.techtarget.com/exp ... bucket=ETA
http://www.gamepc.com/labs/view_content ... 005&page=1
http://www.tomshardware.com/storage/20041119/index.html
http://www.software-raid.tk/

Is RAID Right for You?

Well that depends. It's not right for everyone. There are risks, and frustrations associated with RAID. If RAID fits what you need it for then its right for you. Keep in mind that RAID is not a substitution for Backing Up. Even if you are using RAID 1, you should still do regular backups. Fault tolerance is meant for quick recovery in case of drive failure, it does not protect you from catastrophic disasters.

If you have more questions or comments about RAID feel free to post them in this thread.


Hitman

I'm going to expound on the last thing you said there, hitman

The major downfall of RAID1 is that the data is synchronized between the two hard drives, so if you delete your wedding photos on one, you also deleted it on the other (you may be able to recover it, of course, but that's not what we're talking about here )

RAID1 is only helpful if one drive fails. This is, of course, not common among new hard drives, so it's debatable whether RAID1 is actually helpful for most applications, as long as you do regular backups.

Two put it shortly, RAID1 protects your hard drive from a hardware failure, but NOT software.

Software raid is on par of hardware raid with reliability and performance on raid controllers under like $200...

Heck, OS X has a the software to put thumb drives in RAID. I'm not sure why anybody would want that, but...

Not true, Software RAID performance isn't on par with a $20 RAID controller.

I don't see any way the software could be as good as hardware. It'd take more of your CPU cycles - CPU cycles that would be better off being given to F@H

$20 raid controllers are software based, they have the computer processor do the calculations fyi. Hence why i said around $200, that is where hardware controllers are used.

The same reason people didn't use onboard sound cards 3 years ago and got a creative card.

what if you have two 40 gig hard drive.
which will be faster? setup as Raid0 or just 2 separate drives?
those link is confusing...
so... hopefully you guys can give me a simple answer and some explanation

For pure WinXP overall performance, I have long been a major proponent of using 2 drives independantly. By doing so, you can place the page file on the drive that does not contain your windows / gaming / pagefile heavy apps and get a more noticeable improvement in actual performance then any simple RAID 0 solution will grant you.

Would JBOD be appropriate for this thread? Or should it be considered an array different and separate from RAID?

I am trying to think of which RAID is best for me. Level 5 sounds the best but what number and type of drives do you need? Like if I used a few 250 GB drives how many would need what storage would I get, and could I use an existing 250GB drive I have?

2 40gb HDs in RAID 0 will perform faster then a single 40gb HD. However the speed increase won't be very noticeable. You also increase your chance of data loss. Because if one drive fails you will loose all your data. A better reason to put 2 40gb HD in RAID 0 is that they will appear as one 80gb HD.



JBOD (Spanning) comes in handy if you have several Hard Drives of varied sizes, that you want to put in an array so that they appear as one drive. The nice thing about JBOD is the drives don't have to be the same size, so if you have one 80gb drive, one 100gb drive, and one 20gb drive they will show up as a 200gb drive. The problem with JBOD is there is no redundancy. So you need to make sure you have your data backed up. Also there is no performance increase.



RAID 5 requires you to have at least 3 hard drives of approximately the same size. 1/3 of the total drive space will be used for redundancy. So 3 250GB HDs in Raid 5 would show up as just under 500GB. In my main system I have 4X250gb drives in RAID 5. Windows XP sees that array as one 700GB drive.

what is the real world difference between 0+1 and 10? say i had 4 500Gb drives, both 0+1 and 10 would give me 1000Gb of useable space and 100% redundency (as far as my understanding goes) so what is the odds in terms of tranfer rate and latency?

If I had 4 hard drive in RAID5 and 2 failed, did I lose everything?

With RAID 0+1, you can only lose a single disk.

RAID 10 (RAID 1+0), you can lose multiple disks, so long as none of the underlying RAID1 "disks" under the 0 are completely compromised.

Example:

4 RAID1 "disks", each comprised of 2 physical drives per RAID1.

RAID 0 them together.

Now, you could, conceivably, lose 4 physical drives (1 per RAID 1 "disk") without losing the array.

HOWEVER, if you completely lose one of the RAID1 "disks" (both drives in the mirror), your entire RAID "goes away".





YEP! RAID5 can sustain a maximum of 1 drive loss. This is why things like "hot spares" are good ideas. That and smart controllers that notify you right away (some of the Dell controller alarms scream bloody fscking murder when a drive goes).

Is there any way to recover data on a pair of RAID 0 drives without going to a data recovery company?

The hardware RAID controller sees the drives and doesn't report that there are any problems. Can I rebuild the RAID without wiping the data?

Likely the answer to both of these questions is "no".

That is the typical type of question a person would use to justify not using RAID5 when they had a choice to use it. Chances of all your hard drives being taken out in a RAID 5 array when the tornado lifts your mobile home off the ground and smashes up your computer would be about the same result. Common, lets be realistic; Hard drives don't always completely fail nor all at the same time and most of the time give you warning that they are failing while operating in a reduced performance. Chances are slim at best that you loose more than one drive at once. If you don't trust me, use RAID 6 or some other RAID that supports multiple redundant parity drives.
More information on failures with Hard Disc Drives: http://www.overclockers.com/tips1035/index06.asp

About Data recovery.. I one day was trying to upgrade another computer's measly 20gb hard drive with a 200Gb hard drive. I was using an application called Paragon Partition Manager, copying the hard drive, when my power supply blew out. What are the chances of that! Seemed that the Paragon software modifies the boot sectors on all hard drives while it is in use, and the power interruption took out my software Spanned array, both source and destination hard drive and my C: drive. No problems with recovering my C: drive as I re-installed windows XP with the repair option (And re-install most my apps) BUT, Alas the spanned array was non-recoverable. Windows XP kept running out of memory when trying to re-attach the devices.

I wasn't about to accept this. Family pictures and the lot of personal stuff over 10 years worth LOST!

After 2 weeks of research and trying different software, I came to this conclusion, 100% data recovery from the array.
1: Re-create the array without a format
2: Run a deep scan using Runtime Software's Getdataback NTFS.

I simply connected another computer on the network and copied all the files over while in Windows XP. Took about 4 hours to scan 250Gb in four hard drives spanned and about a day to copy it.

So, sometimes you can recover data with damaged arrays.

I also found File scavenger partially usefull, as it had an option to select several physical hard drives in Raid 0, 1, 5, or jbod emulation to recover data, but it was able to recover only 75% of my data. 25% of the files were cross-linked with the contents of other files and was not a viable solution, but the option was a creative one which the other competing products did not come with.

I would mention the other software here as well, but since they didn't work completely, or at all, their names are not worth mentioning.

I since built a RAID 5 array using six 400GB drives and keeping a spare 400GB in my computer powered up and not used, so when I have another one fail, I don't loose anything and rebuilding can be done without swapping or fussing. I also since had one of those 400Gb drives fail, and I found that with software RAID 5, I was unable to use the device until it was completely rebuilt as it was not properly calculating parity until the spare drive was rebuilt in the array.

If anyone is into Hardware RAID, 3ware sells some expensive and impressive RAID cards in different form factors, (PCI-X, PCI-E) in 4, 8, 12, and more SATA ports. For the price, I am certain that performance would be impressive. But if you have a 3Ghz P4 or a 2Ghz AMD machine, you would be looking about 8Mb/s write speed with software RAID 5. Not that bad. I however have had poor customer service and poor results using Highpoint cards, but found them to be more flexible on features. I would generally suggest a buyer beware for Highpoint. Promise cards however have never given me troubles and have always been true to specifications even when they were limited in available features.

Another interesting thing I do with my RAID cards, is do a multi-nested RAID setup. I can Stripe pairs of hard drives and then Software RAID 5 them together, or Jbod (Span) pairs up for capacity and then RAID 5 them. This way I can combine drives of dissimilar capacities using Jbod in order to make up arrays with the same sizes and then Raid 5 them together. If one or more of the drives fail in a Jbod array, I simply take out all the drives in that array and replace those drives with one of equal size or rebuild the jbod with another set of dissimilar hard drives which equal or exceed the minimum hard drive capacity for that set of devices in the RAID 5 array.

Most RAID Cards also do not offer your operating system to monitor SMART status (Self-Monitoring, Analysis and Reporting Technology) of your physical discs as well, so when your SMART reports issues which could identify possible problems in the near future with your hard drive, you are simply not informed about it until it actually fatally fails. (SMART monitor software: http://www.hdsentinel.com/) For this reason, software RAID could be a feasible option. I hear however that the newer "More expensive" Raid cards come with their own monitoring software, but to what type of available information they pass onto the user I don't know.

If your machine has plenty of room for hard drives and you got a hefty power supply, playing with arrays can be fun.