4 drive raid -0 in vista

[Dattron]

I read there is a performance boost with raid -0, could this be true? I am
interested in setting up a 4 drive riad -0 in vista anythoughts? And will
VistaBusimess Disk Manager use 4 drive in raid 0 config?


Thx

Up to 38% better overall storage performance when configured with a
two-drive RAID 0 array, plus the ability to scale to even higher levels of
performance with a three or four drive array

http://www.intel.com/performance/desktop/platform_technologies/storage_performance.htm

 

[Dave]

oh boy..... here we go again..........

 

[Ken Blake, MVP]

I read there is a performance boost with raid -0, could this be true? I am
interested in setting up a 4 drive riad -0 in vista anythoughts?

Although some people report performance improvement with RAID0, the
great majority of those using it in a typical home environment see no
discernable improvement. I tried it here, and no longer use it.

Using RAID0 entails substantial extra risk. Without RAID0, if one of
your drives fails, you lose everything on that drive. With four-drive
RAID0, if any one of your four drives fails, you lose everything on
all four.

To me, it's foolhardy. It isn't worth running that extra risk for a
non-existent or tiny speed improvement.

 

[PNutts]

To me, it's foolhardy. It isn't worth running that extra risk for a
non-existent or tiny speed improvement.

I'm curious... Why is it considered extra risk? There is a certain risk that
any given hard drive will fail, but it remains the same for each drive and is
not doubled with two drives. Each drive still carries that same risk. For
example, if each drive has a 1 in 100 chance of failure, you have two drives
that each have a 1 in 100 chance of failure. If you do combine those numbers,
you get a 2 in 200 chance of failure which is still 1 in 100. I hope my
examples are clear.

I should admit I am hardly objective as I run RAID0 for my OS and the
performance increase is substantial.

 

[Ken Blake, MVP]

I'm curious... Why is it considered extra risk? There is a certain risk that
any given hard drive will fail, but it remains the same for each drive and is
not doubled with two drives.

Yes, it is. As I said "Without RAID0, if one of your drives fails, you
lose everything on that drive. With four-drive RAID0, if any one of
your four drives fails, you lose everything on all four."

With the four drive RAID0 he's talking about, that quadruples the
risk.

Each drive still carries that same risk. For
example, if each drive has a 1 in 100 chance of failure, you have two drives
that each have a 1 in 100 chance of failure. If you do combine those numbers,
you get a 2 in 200 chance of failure which is still 1 in 100. I hope my
examples are clear.

I should admit I am hardly objective as I run RAID0 for my OS and the
performance increase is substantial.

If you find it so, fine. As I said, I found *no* perceptible
performance improvement here, and most typical home users find that
there's little or no performance improvement.

 

[Canuck57]

I'm curious... Why is it considered extra risk? There is a certain risk
that
any given hard drive will fail, but it remains the same for each drive and
is
not doubled with two drives. Each drive still carries that same risk. For
example, if each drive has a 1 in 100 chance of failure, you have two
drives
that each have a 1 in 100 chance of failure. If you do combine those
numbers,
you get a 2 in 200 chance of failure which is still 1 in 100. I hope my
examples are clear.

I should admit I am hardly objective as I run RAID0 for my OS and the
performance increase is substantial.

Ken is right. It is not uncommon for say in 2 drives, 1 _will_ fail in 5
years.

So if you have 2 drives, and the OS is on one, there is a 50% chance in say
5 years it will fail. If it is the OS drive, you are dead. If it is the
alternate or back, it is OK except for the data on it but the OS will/should
boot.

Where as if your OS is spanned on 2 drives then the chance becomes 1, as
either drive is failure, probable failure if one fails in 5 years.

If using RAID, I much prefer 0+1, which really does kick read performance
considerably when used with 3 or more drives. But also reduces risk of
since drive failure being an issue to near zero. If using very large
quantities of disk such as a data center, RAID 0+1 or hybrids like RAID 50
work well. But for RAID 50, you need 10 or more disks to do correctly. For
home users, RAID is generally something you will not do. And if you do,
RAID 1.

No, RAID 0+1 with 3 drives is not a typo, have done this and it does work.
One of the questions I have been know to ask "disk RAID gurus" and most get
wrong.

 

[Dattron]

My home pc -uses two independant raid controllers, Adaptec 300GIG Riad -0 E:\
and an Intel 300GIG Raid -0 C:\. I install the Vista paging file to E: along
with all games and video related data. Drive C is used for the Vista 32
Operating System and all applications. There is a backup plan in place that
uses externel storage ....A usb2 700GIG SATA Drive, that automates backups of
the images of both raid controllers.

I have a duplicate system same mobo using a sigle drive config. The raid
box out performs the single dive system everytime. There is always risk of
loss of data and the odds that my single drive system goes down, is the same
as the raid system. Unless the raid system hosts a bogus drive, but that
bogus drive could just as easily be installed to the single drive box

 

[Chupacabra]

I'm curious... Why is it considered extra risk? There is a certain risk
that
any given hard drive will fail, but it remains the same for each drive and
is
not doubled with two drives. Each drive still carries that same risk. For
example, if each drive has a 1 in 100 chance of failure, you have two
drives
that each have a 1 in 100 chance of failure. If you do combine those
numbers,
you get a 2 in 200 chance of failure which is still 1 in 100. I hope my
examples are clear.

That's not how you calculate cumulative risk though. Running two drives,
where the failure of either one results in the loss of all data, is slightly
more risky than running each drive individually.

If the risk of one drive failing is 1%, then its risk of not failing is 99%.

For a two-drive RAID0 set, it would be 99% x 99%, or 98.01%

Instead of a predicted uptime of 99% on a single drive, you only have 98% on
a two-drive RAID 0 set.

On a 4 drive RAID 0 set, it would be

99% x 99% x 99% x 99%, or 96.05% predicted uptime.

The more drives you add to a RAID0 set, the lower your predicted uptime will
be.

 

[Dattron]

....and even a 96.05% uptime reguarding hard drive failure over a 5 year MTBF
span, would be unbelievably awesome to an end user. And I would think a
predicted uptime failure rate on hardrives would be taken from a lot of 100
or more hard drives. So that 99% would be more like 99.99999999998765 %