RAID0 question

[Crusher]

You confuse statistics with risk, the 2 are related but different fields.
If you have a dollar to invest, and it is your only dollar, the risk is
high.
If you have a dollar to invest, yet are a millionaire, there is virtually no
risk.
Statistically the invested dollar is the same.

 

[Ken Blake, MVP]

You confuse statistics with risk, the 2 are related but different fields.
If you have a dollar to invest, and it is your only dollar, the risk is
high.
If you have a dollar to invest, yet are a millionaire, there is virtually no
risk.
Statistically the invested dollar is the same.

You may believe whatever you want. I have neither the time nor the
inclination to argue with you.

 

[Crusher]

Maybe you need a book on risk assessment.

You may believe whatever you want. I have neither the time nor the
inclination to argue with you.

 

[Crusher]

My opinion is a fact. The OS residing on 2 HD's is of no greater "risk" than
on a single HD. In either situation, the failure of a single HD would cause
failure. MVP must mean "I read it somewhere, and as I cant think for myself,
that makes it a fact."

You stated an opinion as fact. The opinion was wrong.

 

[Kerry Brown]

You don't obviously don't understand statistics. With two drives there is
greater chance of one of them failing than with a single drive. With any
RAID array the MTBF equals the MTBF of a single drive divided by the number
of drives. This means in a two drive RAID 0 your MTBF is halved. Any one who
works with RAID on a regular basis will tell that the formula is true.
That's why in real RAID with redundancy a lot of people use hot spares. This
only takes into account drive failure and not bad sectors. In a striped
array if you get one bad sector the controller may mark the drive as
unusable and there goes your array. With redundancy this isn't a problem.
You test the drive and possibly put it back into service because the array
continued working with a failed dive. With most RAID controllers built onto
motherboards if you are using RAID 0 and get a bad sector on one drive you
usually loose the whole array for good.

--
Kerry Brown
Microsoft MVP - Shell/User
http://www.vistahelp.ca

My opinion is a fact. The OS residing on 2 HD's is of no greater "risk"
than on a single HD. In either situation, the failure of a single HD would
cause failure. MVP must mean "I read it somewhere, and as I cant think for
myself, that makes it a fact."

 

[Kerry Brown]

If you have two cars you have a greater chance that at any given time one of
them will break down than if you just had one car. Fortunately cars are
independent of one another and you can still use the car that is working. In
a RAID 0 array the drives are not independent. If one breaks down you lose
the data on both. You need a basic understanding of statistics and
probability to continue this argument.

--
Kerry Brown
Microsoft MVP - Shell/User
http://www.vistahelp.ca

If you buy a 2nd car, does it increase the chance the first car will break
down? Of course not. You may have 2 HD in RAID, but millions are made. 2
HD doesn't "double" the risk. If their is any increased risk, it is
miniscule. And with regular backups, there is 0 risk. The same risk is
present with a single HD. I wasnt name calling, mook is a funny word

 

[Crusher]

You don't understand the word RISK. If you have 2 than thats twice as much
as one. But it doesn't make it any riskier...as in either system, the
failure of one drive will result in the loss of data.
I have 2 hard drives, and the failure rate is 3/1000/year. the probability
that my system will fail is 3 in a 1000.
You have 1 hard drive, and the failure rate is 3/1000/year, the probability
that your system will fail is 3 in a 1000.
My system probability of failure would be 6 in a 1000 only if you factor in
"concurrent" failure of both drives.
I understand statistics, I was a statistical process engineer for 8 years.
You should learn to read, write....the double negative in your opening
remark isn't proper grammar, nor does it make any sense.

 

[Kerry Brown]

It was a typing mistake and you're right it didn't make sense. It should
have read: "You obviously don't understand statistics". You say you do. It
sounds like you've had more education in that field than I have. Given this
I don't understand your lack of understanding of a basic concept. According
to this article around 3% of drives fail in the first three years of use.

http://blogs.zdnet.com/storage/?p=162

This means if you have two drives you have a 6% chance that one of them will
fail in the first three years. By your logic if you had 100,000 drives the
odds of one of them failing would be the same as the odds of the failure of
any given drive out of the 100,000 total. I'm sure the drive manufacturers
love your logic but reality shows us otherwise. It's a pretty basic concept.
Again, this only takes into account total failure of the drive. The way RAID
controllers work they mark a drive unusable after x number of errors. Many
things can cause an error, bad RAM, bad PSU, bad data cable, etc.. With
single drives the problem causing the errors can be fixed and unless the
file system was corrupted nothing is lost. With a RAID controller the
controller marks the drive bad and quits using it. With RAID 0 this means
you have lost your data. You can't rebuild a RAID 0 array.