Newbie Question re hardware vs software RAID

[kony]

Drives reliability is a lot more complicated than the bearings. FDB
drives tend to be more forgiving of bumps and vibrations, and that can
affect longevity. Also the 15k FDB cheetahs have a great & proven
track record.

Yes it's more complicated than bearings, and yet there isn't
some unique failure point or "lower quality" part on ATA
drives that makes them more susceptible to failure.

Why? Why is engineering/reverse engineering knowledge so superior to
experiential knowledge?

Because you don't have a large enough sample size to make
that determination and there are many variables involved,
particularly with systems using expensive SCSI drives in
contrast to PATA.

For the most part? what about the other part?

The other part is the factors beyond the drive itself, like
the typical buyer, typical system, typical usage.

Than what DO you want? You ask me to cite specific failure points but
then say that it would not clarify a SCSI vs ATA bias?

I wrote "unless", not "not".

To determine that the failure rate is due to drive type (ATA
or SCSI) you would need to demonstrate that either type of
drive has a unique failure mode that is lower in the other
type due to drive construction, interface, or other factors
necessarily unique to that drive type,not due to other
factors not dependant on whether the drive is SCSI or ATA.
For example, claiming higher reliability when the average
SCSI drive is in a system with better power supply,
ventilation, or fewer spin-ups, is not an indication of SCSI
having superior reliability, rather than the usage and
environment was different.

I can regurgitate a lot of crap from manufacturers to explain the
alleged enterprise vs personal storage reliability, but we both know
this type of information tends to be a lot of marketing hype and tends
to be something you should take with a grain of salt. But if SCSI
only offers higher price, why do people buy them?

Because SCSI supports more drives and they're tpically
faster, the types of systems they're used in came with SCSI.
I did not claim "only" higher price, also the superior bus
and speed.

Why are they the
standard device interface for serious/mission-critical use?

Because systems designed to fill this role have been
designed for more than just reliability.

In fact
the FDB 15k Cheetahs have a great reliability reputation. They also
place at the top of storagereview.com's reliability database (if you
place any stock in that). True many ATA drives place higher than many
SCSI ones in that same survey, but I did not recommend _ANY_ SCSI
drive over _ANY_ ATA one.

Storage "groupies" will gravitate towards SCSI and higher
RPM drives, and put care and irrational love towards such
products. Do you deny that a system built with 3X the
budget (as typical in a system using SCSI drives) has
differences beyond which interface the drive uses? You
would believe that a cheap and nasty Tiger Direct Barebones
special (system build) with a SCSI card and drive in it will
be more reliable storage than a good system running a PATA
drive? It would seem you feel this way, and I disagree.
The specific drives you mentioned are usually not the only
variable in systems seeing higher drive reliability.

I can really only share my anecdotal experience which is much higher
satisfaction with SCSI disk subsystems. I've had a lot of ATA drives
that just sometimes do weird things or tend to suffer more hard errors
/corruption as they reach the end of a much shorter 'realistic'
service life. I've also seen a lot of scsi drives function perfectly
for longer periods during heavier use, and seen them deal with
problems better (without data loss/corruption). Of course not EVERY
model scsi device is going to be GREAT, but neither is everything of
anything else. Sure this is anecdotal, but it the way it's _supposed_
to be as there is _supposed_ to be a difference in reliability of
enterprise storage vs personal storage.

The way it's "supposed" to be is a vague assumption, you
would very much like to think the higher cost also brings
higher reliability because with some/other/non-drive
products it is true.

An assumption. If it were true then why is RAID common?

Is it a HUGE difference in
ALL cases, well no but with a SCSI RAID solution you are more likely
to get a full fledged, cohesive product that is more geared to serious
use.

You've been assimilated by marketing droids.

That's not ATA vs SCSI trolling, it's just how the product lines
run. Better management software, Better drivers, better features,
better error handling, better handling of configuration and
recoverability, better support, better compatibility, better warranty,
more scalable, available for faster busses, etc.

It is likely that products costing more have better
management software and drivers, features, support, and
warranty, etc, etc. This is because the higher price allows
more development of these things, and that the target
customer is more likely to need many of these features.
They are built into the price of the equipement but it bears
no relation to reliability.

I fully acknowledge I'm just offering an opinion. I thought my
perspective bared some clarification for the OP and group but I don't
really care what you use or think is better. If you think PATA is
better please use them. I'm simply not interested in getting further
involved in a potential "SCSI vs xATA" flame/troll. (this smells like
a Folkert)

There are valid reasons to use SCSI, but claiming higher
reliability without being able to cite differences between
SCSI and PATA hardware that specifically cause higher
reliability isn't evidence, more like superstition or a
failure to consider the other variables.

One reason for higher reliability of SCSI might be that SCSI
controllers are more likely to avoid electrolytic
capacitors. On the other hand it's usually not the
controller that fails on PATA.

 

[The professor]

There are several sites offering this kind of service on the web but I
don't know about their reliability. I have tested some of them without
problem but it is not sufficient to say they are reliable (all of them
claim they are).

 

[kony]

In message <[email protected]> kony


As a general rule, if you buy the cheapest SCSI drives on the market,
you'll get exactly the same hardware, just a different controller with a
SCSI interface rather then an IDE interface.

However, if you're comparing a higher end SCSI drive to a low end IDE
drive, you'll definitely see better quality parts on the higher end
drive.

Substantiate "better quality".

 

[Curious George]

Substantiate "better quality".

That is part of the problem. "better quality" is a qualitative
judgment and not a quantitative one. I think you are oversimplifying
the issue. Just because identifying a difference in specific
attributes as it relates to "quality" over multiple, quickly changing
product lines is not easy doesn't mean there necessarily isn't a
difference. Furthermore the SCSI protocol is more robust and it
wouldn't surprise me if that is the reason it wasn't passing on errors
in similar situations like I have observed with ATA.

Finally, you are expanding the scope of issue too broadly. As I said
before I am not endorsing every scsi drive every scenario as being
vastly more reliable than every ATA drive in every situation. The OP
is concerned about reliability and has been burnt by unreliable backup
solutions and is seeking a "full-fledged" & reliable hardware raid
solution. I am replying that a better SCSI RAID solution will tend to
be a safer bet toward that end. Even if ATA drives are just as
reliable as SCSI ones, an ATA raid solution won't be as reliable if it
has poorly written drivers, bad management software, or features which
obstruct certain attempts to reconcile some types of failures or if
support is lacking. The OP really wants to know about viable and
reliable RAID so your experiences with SATA or PATA RAID are really
more appropriate than whether you believe ATA and SCSI drive build
quality to be identical.

 

[kony]

That is part of the problem. "better quality" is a qualitative
judgment and not a quantitative one. I think you are oversimplifying
the issue. Just because identifying a difference in specific
attributes as it relates to "quality" over multiple, quickly changing
product lines is not easy doesn't mean there necessarily isn't a
difference. Furthermore the SCSI protocol is more robust and it
wouldn't surprise me if that is the reason it wasn't passing on errors
in similar situations like I have observed with ATA.

Finally, you are expanding the scope of issue too broadly. As I said
before I am not endorsing every scsi drive every scenario as being
vastly more reliable than every ATA drive in every situation. The OP
is concerned about reliability and has been burnt by unreliable backup
solutions and is seeking a "full-fledged" & reliable hardware raid
solution. I am replying that a better SCSI RAID solution will tend to
be a safer bet toward that end. Even if ATA drives are just as
reliable as SCSI ones, an ATA raid solution won't be as reliable if it
has poorly written drivers, bad management software, or features which
obstruct certain attempts to reconcile some types of failures or if
support is lacking. The OP really wants to know about viable and
reliable RAID so your experiences with SATA or PATA RAID are really
more appropriate than whether you believe ATA and SCSI drive build
quality to be identical.

This seems terribly complex to you perhaps, but it's not
that difficult. OP was concerned about drive failure.
There is no evidence that the much more expensive solution
you suggest will make any difference in that regard. What
it will do is take substantially more of a budget that would
be best put towards a 2nd, alternate backup in addtion to
the array, which is not meant, not designed to be a backup
strategy at all regardless of whether PATA, SATA, or SCSI.

 

[Curious George]

Yes it's more complicated than bearings, and yet there isn't
some unique failure point or "lower quality" part on ATA
drives that makes them more susceptible to failure.

Well I don't know about that, and I doubt you do too. A
generalization like that is highly problematic when dealing with
multiple manufactures & product lines. Firmware is a good example of
one of the items on a drive whose quality is difficult to measure by
the end user and is critically important to its function. Describing
overall quality in a piece by piece manner is highly problematic.

Lets say we inventory all the component parts of every SCSI and ATA
drive. So what? What are you going to do with that information to
make a judgment about reliability of the fully assembled operational
units? The individual components are simply not the only variables
that affect operational success and longevity.

What you need to look at is the operational success of the fully
assembled devices in real world use for their promised service life.
You need to look at things like operational MTBF (not usually
disclosed by manufacturers) or hard error rates over time (impossible
to do over a substantial population esp if you want to get this from
ATA users). You might also want to do synthetic lab testing of a
large quantity of drives for the full length of their projected
service life (as opposed to the way synthetic MTBF is conducted) and
examine various error rates as well as outright failure. If these
were possible, at best examinations as thorough as these would be
little more than of historical interest & would not be valid to base
current purchase decisions on.

Therefore I don't see how your conclusion re reliability is any more
of a gut feeling from anecdotal experience than mine. Please show us
your concrete data which supports your position in a new thread like
"ATA vs SCSI reliability" (or maybe something better worded to avoid a
flame war).

 

[Curious George]

This seems terribly complex to you perhaps, but it's not
that difficult. OP was concerned about drive failure.

I guess you didn't actually read his post or pay attention to the
subject. He's concerned about overall reliability as well as
compatibility. He did not discuss drive failure specifically. Also
there is a lot more to reliability than just outright and complete
drive failure.

There is no evidence that the much more expensive solution
you suggest will make any difference in that regard.

Absence of evidence is not evidence of absence.

What
it will do is take substantially more of a budget that would
be best put towards a 2nd, alternate backup in addtion to
the array, which is not meant, not designed to be a backup
strategy at all regardless of whether PATA, SATA, or SCSI.

He needs backup & raid both carefully planned and catered to his
needs. How he uses his budget is for him to decide and involves
information not yet shared.

 

[kony]

Well I don't know about that, and I doubt you do too. A
generalization like that is highly problematic when dealing with
multiple manufactures & product lines. Firmware is a good example of
one of the items on a drive whose quality is difficult to measure by
the end user and is critically important to its function. Describing
overall quality in a piece by piece manner is highly problematic.

Lets say we inventory all the component parts of every SCSI and ATA
drive. So what? What are you going to do with that information to
make a judgment about reliability of the fully assembled operational
units? The individual components are simply not the only variables
that affect operational success and longevity.

You are the one making the assertion that there is higher
"reliability", it is not up to me to prove it.

What you need to look at is the operational success of the fully
assembled devices in real world use for their promised service life.
You need to look at things like operational MTBF (not usually
disclosed by manufacturers) or hard error rates over time (impossible
to do over a substantial population esp if you want to get this from
ATA users). You might also want to do synthetic lab testing of a
large quantity of drives for the full length of their projected
service life (as opposed to the way synthetic MTBF is conducted) and
examine various error rates as well as outright failure. If these
were possible, at best examinations as thorough as these would be
little more than of historical interest & would not be valid to base
current purchase decisions on.

Therefore I don't see how your conclusion re reliability is any more
of a gut feeling from anecdotal experience than mine. Please show us
your concrete data which supports your position in a new thread like
"ATA vs SCSI reliability" (or maybe something better worded to avoid a
flame war).

Again, you made the claim.

 

[Curious George]

You are the one making the assertion that there is higher
"reliability"

And you are the one asserting reliability is the same. So?

it is not up to me to prove it.

Ahh I see. Your plan is to simply contradict what I say without
substantiative evidence.

Hmm.

I formerly apologize to the group for having fed the troll. Boy do I
feel silly.

 

[kony]

I guess you didn't actually read his post or pay attention to the
subject. He's concerned about overall reliability as well as
compatibility. He did not discuss drive failure specifically. Also
there is a lot more to reliability than just outright and complete
drive failure.

Anyone is concerned about reliability and compatibility.
You want to pretend these imply SCSI, but they do not.

Cut out the random BS, specifics count.

Absence of evidence is not evidence of absence.

Again, nonsense.

He needs backup & raid both carefully planned and catered to his
needs.

You're reading this off of a brochure aren't you?

How he uses his budget is for him to decide and involves
information not yet shared.

Again, nonsense... not that what you wrote is nonsense, it
is true, but that you wrote it as some kind of "news"
towards your argument, is nonsense.

Who suggested how he uses his budget is for anyone else to
decide? Apparently that's exactly what YOU are thinking
since you keep pushing SCSI instead of addressing the point
I made. Note that I never pushed PATA or SATA instead,
rather questioning your flawed logic about reliability. You
have completely missed the boat here. Someone has
brainwashed you really, really good and you're in too deep
to even make any sense out of it.

_IF_ the OP wants RAID5, so be it. It is likely that SCSI
*IS* the best solution for RAID5, but reliability has
nothing to do with it.

 

[kony]

And you are the one asserting reliability is the same. So?

So, you made the inital claim, it is your burden.

What alternate universe do you live in where it is not this
way?

If you're going to start telling a fictional story it's only
considerate to finish the story.

Note that I never argued that OP shouldn't use SCSI.

 

[Curious George]

Anyone is concerned about reliability and compatibility.
You want to pretend these imply SCSI, but they do not.

Cut out the random BS, specifics count.

That's funny

Again, nonsense.

again, funny

You're reading this off of a brochure aren't you?


Again, nonsense... not that what you wrote is nonsense, it
is true, but that you wrote it as some kind of "news"
towards your argument, is nonsense.

That's funny

Who suggested how he uses his budget is for anyone else to
decide? Apparently that's exactly what YOU are thinking
since you keep pushing SCSI instead of addressing the point
I made. Note that I never pushed PATA or SATA instead,
rather questioning your flawed logic about reliability. You
have completely missed the boat here. Someone has
brainwashed you really, really good and you're in too deep
to even make any sense out of it.

_IF_ the OP wants RAID5, so be it. It is likely that SCSI
*IS* the best solution for RAID5, but reliability has
nothing to do with it.

Allright Folkert (or student of the folkert school of troll). Take a
deep breath, count to ten, and take your happy pill. Your eyes really
jump across the lines when you try to read when mad.

 

[Curious George]

So, you made the inital claim, it is your burden.

I am responsible for my clarity of writing - you are responsible fro
yours.

There is a difference between opinions and claims. If my opinions are
wrong and your claims are right you should have substantive evidence.

What alternate universe do you live in where it is not this
way?

I seems logic is not your strong suit. I offered an opinion re scsi
raid. You are making unsupported assertions about drives which do not
directly translate to a complete raid solution.

What kind of world do you live in where "you're wrong" is all you need
for a valid argument?

If you're going to start telling a fictional story it's only
considerate to finish the story.

Then you are very rude indeed.

Note that I never argued that OP shouldn't use SCSI.

Well that's not true but if you are now going to endorse scsi raid
then your ravings are all the more absurd and a blatant troll.

 

[kony]

On Wed, 17 Nov 2004 23:55:51 GMT, Curious George

What kind of world do you live in where "you're wrong" is all you need
for a valid argument?

There are a thousand different lies and misconceptions one
could fit into a single post. Are you dense enough that
you think OTHERS have to disprove all of them? The one
making the claim is the one with the burden. It was this
way yesterday, today, and will be a week from Tuesday too.

 

[Curious George]

So, you made the inital claim, it is your burden.

What alternate universe do you live in where it is not this
way?

Nonsense. Only in a court of law does only 1 of 2 parties bear the
burden of "proof". This is a discussion forum not a court. If you
want to make this a competition the only logical corollary is a debate
where both "pro" and "con" must demonstrate their arguments.

If you still don't understand what I have been consistently saying you
should be working with a reading comprehension tutor, or wait until
such things are covered in school rather than creating alternate
universes. You contest my specific recommendation for a raid
subsystem through unsubstantiated sweeping generalizations about
drives. This is both too unspecific to be accurate and doesn't
directly contest my recommendation as drives are not the only element
of raid. Your rantings have avoided my subsequent explanation of
reliability and of methodological problems in measuring reliability.
You are continuing to demonstrate nothing more than inexperience and
naiveté toward argument and disk subsystems. A troll should have more
skill.

 

[kony]

A troll should have more
skill.

In other words, you don't have any evidence at all, you were
just pulling that "reliable" part out of thin air.

 

[Curious George]

In other words, you don't have any evidence at all, you were
just pulling that "reliable" part out of thin air.

So a $15 raid controller with poorly written drivers that crash in the
middle of an OS install is reliable? What about a RAID 5
implementation which does not verify writes properly and allows
transient disk errors to creep in?

In other words you are pulling some gut feeling you have about ATA vs
SCSI drive build quality out of thin air and applying it to a problem
which is more complex than you understand. You do not appreciate what
reliability really means.

 

[kony]

So a $15 raid controller with poorly written drivers that crash in the
middle of an OS install is reliable?

Is a $400 raid controller with poorly written drivers that
crash in the middle of an OS install, MORE reliable?

You're wasting our time with vague implications rather than
specifics.

What about a RAID 5
implementation which does not verify writes properly and allows
transient disk errors to creep in?

That has nothing to do with which drive you choose, and
there is nothing inherant about SCSI that results in write
verify.... if you find your choice of RAID needs additional
write verify to ensure no errors, I'd call that anything but
a sound solution. BTW, back on topic this is a DOS Ghost
imaging we're looking at, ANY disk solution can be verified
with Ghost, IIRC.

In other words you are pulling some gut feeling you have about ATA vs
SCSI drive build quality out of thin air and applying it to a problem
which is more complex than you understand. You do not appreciate what
reliability really means.

Funny you tried to twist around the facts, as unbelievable
as it seems plenty of us have no problems using MANY
different storage solutions, making it seem like you're
incompetent if you need something specific just to do a
simple backup.

 

[Curious George]

Is a $400 raid controller with poorly written drivers that
crash in the middle of an OS install, MORE reliable?

how silly

You're wasting our time with vague implications rather than
specifics.

how silly. Your specificity is daunting.

That has nothing to do with which drive you choose,

only you are fixated on drives

plenty of us have no problems using MANY
different storage solutions, making it seem like you're
incompetent if you need something specific just to do a
simple backup.

Read before you write. the RAID is for his production machine not a
backup nas.

I'd start saving for that reading comprehension tutor if I were you.

 

[kony]

how silly

yes, it does show how pointless your arguments are. This
isn't 1993, it's quite easy to find drivers (hint, every
system running an operating system has managed this task
that seems difficult to you, or if you want to consider only
PCI add-on card, again I ask, what evidence do you have that
ANY won't work?). If you do know of some that just won't
work, then at least you have something useful to share.

how silly. Your specificity is daunting.

Vagueness benefits nobody, in the end a specific solution is
chosen, based on the secifics OF that solution. You seem to
have no grasp of anything but the one solution you mentioned
and so no way to contrast it with (anything at all).

only you are fixated on drives

Did you read the post that started this thread? I'll give
you a hint, in the very first sentence...

Read before you write. the RAID is for his production machine not a
backup nas.

I'd start saving for that reading comprehension tutor if I were you.

Where do you get "nas" from?
You are amusing if nothing else. Here's where this thread
started, a direct quote from the OP:

"I'm looking at installing a SATA RAID 5 system in my home
PC to provide a means of data recovery in case of hard disk
failure"

This is getting a bit old, there's nothing new you have to
add so might I suggest that if you, personally, have some
mental block against the myriad "means of data recovery"
that are successfully employed every day, all over the
world, then just offer tips on what you DO know instead of
what you don't.