RAID 0 is usually a foolish choice for desktops

[Milleron]

If more proof of this old contention is needed, there's a cutting-edge
review by Anand Shimpi on AnandTech.com:
http://www.anandtech.com/storage/showdoc.html?i=2101

As always in these tests, the real-world improvement achieved with
RAID 0 varies between 0 and 4%, which is simply imperceptible. The
price that's paid is two-fold: (1) the difference between the RAID
and a single drive of the same capacity, and (2) the DOUBLING of the
chance of a hard-drive failure.

For the life of me, I can't understand why so many users decide to
install RAID 0 on desktops.

RAID 1 is another matter entirely, but, as Anand says, "If you haven't
gotten the hint by now, we'll spell it out for you: there is no place,
and no need for a RAID-0 array on a desktop computer. The real world
performance increases are negligible at best and the reduction in
reliability, thanks to a halving of the mean time between failure,
makes RAID-0 far from worth it on the desktop."


Ron

 

[Leythos]

For the life of me, I can't understand why so many users decide to
install RAID 0 on desktops.

I strongly agree with you as long as we further qualify it to mean home
user desktops or systems that don't contain high-performance RAID
controllers.

In real-world testing of quality RAID systems, RAID-0 can have up to 20%
performance increase, but there isn't a single motherboard with built-in
RAID controller, and not one cheap (non-scsi) RAID controller that will
let users see this performance.

But, as I said above, with the qualification of quality, I completely
agree with you and have maintained the same stance for years.

 

[Doug Ramage]

I strongly agree with you as long as we further qualify it to mean home
user desktops or systems that don't contain high-performance RAID
controllers.

In real-world testing of quality RAID systems, RAID-0 can have up to 20%
performance increase, but there isn't a single motherboard with built-in
RAID controller, and not one cheap (non-scsi) RAID controller that will
let users see this performance.

But, as I said above, with the qualification of quality, I completely
agree with you and have maintained the same stance for years.

I would broadly agree. I have been using 2 x 74Gb Raptors in a RAID0 array
for several weeks. There is a noticeable improvement in performance,
especially video editing.

I was aware of the potential drive failure aspect - but I am not too
bothered about that, as my critical data is not the RAID drives, and I have
backups of my OS and apps on other drives.

As already said, probably not worth the money for your average desktop.

 

[Tim]

Agree 100%. It's a good article. SCSI is still smoother esp with a dual CPU
system. Perhaps the NCQ drives will help iron things out, then dual core
opterons will bring a smile to everones faces.

- Tim

 

[Doug Ramage]

Yes, I have a dual Pentium III system with 2 x 36Gb SCSI drives - pretty
quick in its day.

 

[_P_e_ar_lALegend]

Il Thu, 01 Jul 2004 23:45:07 +0000, Milleron ha scritto:

If more proof of this old contention is needed, there's a cutting-edge
review by Anand Shimpi on AnandTech.com:
http://www.anandtech.com/storage/showdoc.html?i=2101

As always in these tests, the real-world improvement achieved with
RAID 0 varies between 0 and 4%, which is simply imperceptible. The
price that's paid is two-fold: (1) the difference between the RAID
and a single drive of the same capacity, and (2) the DOUBLING of the
chance of a hard-drive failure.

It's pretty much three years I run Raid 0 now with various chipsets and
configurations (3ware, promise, high point and now SI) without any failure
problem.

BTW, become I do care about my data, I do backups often. For now I never
had to use my backups.

For the life of me, I can't understand why so many users decide to
install RAID 0 on desktops.

Becouse of improvements in performance?

RAID 1 is another matter entirely, but, as Anand says, "If you haven't
gotten the hint by now, we'll spell it out for you: there is no place,
and no need for a RAID-0 array on a desktop computer.

Anand need to do some real life computer: tests are not enough.

 

[Winey]

Agree 100%. It's a good article. SCSI is still smoother esp with a dual CPU
system. Perhaps the NCQ drives will help iron things out, then dual core
opterons will bring a smile to everones faces.

What are NCQ drives?

Glad you still like SCSI. If you look at the pricing for some of
the high-perf 73 GB ATA drives, you're going to pay about what the
same drives cost in SCSI-land.

--W--

 

[JTS]

Good points. The only thing I use RAID for is to kill bugs.

 

[Milleron]

Il Thu, 01 Jul 2004 23:45:07 +0000, Milleron ha scritto:


It's pretty much three years I run Raid 0 now with various chipsets and
configurations (3ware, promise, high point and now SI) without any failure
problem.

And every minute of every hour of every day that you operate this
RAID, you STILL have TWICE the chance of a HD failure that you would
have without it. Have you heard about the guy who jumped off the top
of a fifty-story building? As he passed the third floor, someone in a
window yelled "how you doing." The jumper yelled back "OK, so far."

BTW, become I do care about my data, I do backups often. For now I never
had to use my backups. Keep making those backups!


Becouse of improvements in performance?

Didn't you read the article??? There are NO significant improvements
in performance.

Anand need to do some real life computer: tests are not enough.

I disagree strongly on two counts:
1) Your so-called "real-life computer tests" are not tests at all.
They're nothing more than SUBJECTIVE impressions of performance,
strongly biased by the money and time you spent on the RAID, and they
mean absolutely nothing in the face of true OBJECTIVE tests.
2) AnandTech's benchmarks DID include "real-world" content-creation
benchmarks, and the results speak for themselves. There is NO
PERCEPTIBLE improvement in real-world applications with RAID 0.

Ron

 

[Lorenzo Sandini]

For the life of me, I can't understand why so many users decide to

install RAID 0 on desktops.

You're overlooking the key element here Ron.

The GEEK factor )

Lorenzo

 

[Barry Watzman]

An interesing comment, the author notes that:

"The price that's paid is two-fold: (1) the difference between the RAID
and a single drive of the same capacity"

I detect a presumption that RAID is more expensive. In fact, it's often
LESS expensive. I've bought Western Digital 1200JB's (7200rpm, 8 meg
cache) for as low as $59. You cannot buy a 240 gig drive for $118.

 

[Leythos]

I detect a presumption that RAID is more expensive. In fact, it's often
LESS expensive. I've bought Western Digital 1200JB's (7200rpm, 8 meg
cache) for as low as $59. You cannot buy a 240 gig drive for $118.

No one ever suggested that a RAID-0 array capacity was cheaper than non-
RAID-0 capacity, the comment about about the COST paid for loss of
everything, down time, etc....

While a MIRROR could cost ($) as much as 4 times the cost of a RAID-0,
the MIRROR would provide redundancy and protection in case of drive
failure.

RAID-1 offers fast reads, but slower writes, but RAID-0 is just not
worth the loss of all data on BOTH drives.

 

[_P_e_ar_lALegend]

RAID-1 offers fast reads, but slower writes, but RAID-0 is just not

worth the loss of all data on BOTH drives.

Just do backups and u are fine.

 

[Leythos]

Just do backups and u are fine.

Most people don't even purchase something for backup, most don't even
know what a tape drive is, most people don't even know what Windows
Update is, most people don't even run the update for the AV program.

If you build an 120+GB RAID-0 Array, do you know that it would cost more
than the price of a new computer to back it up to tape?

 

[MikeSp]

Although it does not absolutely ensure that the drives will actually be more
dependable, but the WD SATA Raptors do have a 1.2 million hour MTBF which is
significantly better than any PATA drives--thus RAID is less likely to fail
when using these enterprise drives than non-RAID that uses regular PATA
drives. This is in regard to the question about one of the drives failing
when using a RAID configuration.

MikeSp
---------------------------------------------

 

[_P_e_ar_lALegend]

Most people don't even purchase something for backup, most don't even
know what a tape drive is, most people don't even know what Windows
Update is, most people don't even run the update for the AV program.

So I think those people wont go raid.

If you build an 120+GB RAID-0 Array, do you know that it would cost more
than the price of a new computer to back it up to tape?

Well, DVDs backup? In this days dvds are pretty much cheap, also double
layer. Or maybe just backup valuable data?

Considering many people reinstall OS every season change, it wont be a
real problem to reinstall OS.

BTW, I'm with the same upgrade system formatted in 1996! And I do have
Raid 0 starting in 2000.

Then, again, in normal condition, Raid 0 wont fail.

It's incredible how low level is the understanding and the misconception
of average users. :-(((

It remind of the time when people got worried about PCI and then they
demonized USB and now they do demonize raid 0 ))))

 

[Pavel VYCHODIL]

I think Anand is not true.
He used tests just for desktop performance. Minimal disk load. Much more for
graphics, processor and chipset :-(.
I dont agree, that RAID 0 for desktop is nonsense. A lot of people need
great disk subsystem performance. AV editing and many others.

Anand didnt do tests with copy from 1 partition to the second, no DVD
burning and normal office work etc or Database creation and indexing.
I think, that reliability and SMART is now much bigger than before 2-3yrs
and price is good for two smaller disks than one bigger.

Pleva

 

[Leythos]

Then, again, in normal condition, Raid 0 wont fail.

It's incredible how low level is the understanding and the misconception
of average users. :-(((

It remind of the time when people got worried about PCI and then they
demonized USB and now they do demonize raid 0 ))))

Well, I can assure you that I completely understand all of the
electronics and methods behind raid, raid controllers, and MTBF as it
relates to real-world installations.

RAID-0 as long as there are no glitches in the system, the controller
card, no glitches in power, no failures in the drives, is a nice way to
get a little performance boost in drive bandwidth, but, when you take
into consideration all of the things that can render a RAID-O array
dead/unreadable, then it's hardly worth it.

Yea, I know, you've had it working without problem - none of us have
said it doesn't work, in fact, it works quite well. The problem is that
when there is a "glitch" and you expect to recover from it, you have
twice the chance that you won't recover without loss as you do with a
single drive or a MIRROR.

I don't know anyone that didn't like PCI, but, as a IT professional, I
can't stand USB, at least when you have more than one device per USB
port. I've seen to many of those little 12" screen laptops where
everything has to be chained to a single USB port and then they start
adding hubs... It's no end of problems. Don't take this wrong either,
USB for single chain devices is great - much better than Serial!

Now, in case you missed the point of RAID-0 in this discussion, it's
that RAID-0 is not fault tolerant and offers more opportunity for
Disaster than a single drive system - is a basic statistic - two drive
needed to make the system work means a higher probability that one drive
will fail more frequently than a single drive in another system.

 

[Ron Reaugh]

There is one major flaw in his article. He claims that the biggest possible
stripe size provides the best performance. In fact usually a stripe size of
16K or 32k for two drives is usually optimal. Also Anand doesn't seem to
understand the definition of what a stripe is.

A stripe is the set of data than spans the entire RAID 0 array until it
starts back on the first drive again. A stripe unit is the amount/unit of
data on a single drive in a RAID 0 set until the linear stream starts on the
next drive.

 

[Ron Reaugh]

Most people don't even purchase something for backup, most don't even
know what a tape drive is, most people don't even know what Windows
Update is, most people don't even run the update for the AV program.

If you build an 120+GB RAID-0 Array, do you know that it would cost more
than the price of a new computer to back it up to tape?

Tape, god no. Use a SATA drive in a removeable tray ~$220 for 250GB.