What am I doing wrong ??? Or is Adaptec 21610SA just a crappy RAID card ?

In article ,
Malcolm Weir wrote:

Well, *somebody* thinks it worthwhile putting it in commodity
chipsets, which means it *is* happening, despite your ignorance.

"Is" sounds very recent to me.

Are you not familiar with the word "support"? Unlike you, I'm not
claiming absolutes.

1999 is not an absolute?

Once again, 1999.

(Actually, it predated 1999. But that was when the standard was
done).

Does "Is" not mean 2004?

So you don't *need* anything more than a 2400 baud phone line to
create a workload that involves more IO than you can get out of a SATA
drive. Which was what you seemed to think required gigabit ethernet!

Is this a magic trick? Can you tell us how that actually works?

But 1999, kid, isn't recent. It's a (drive) lifetime ago.

1999 = 2004?

OK. Go to www.dell.com. Look. Be enlightened.
Go to www.google.com. Enter "1000BaseT 1999". Look. Be enlightened.

Logic?

Indeed. However, unlike you, I do this for a living.

Does this mean all people everywhere who do their jobs for a living must
all be competent at what they do?

In article ,
"J. Clarke" wrote:

You have an odd definition of "recent". In the computer industry 5 years is
an eternity.

Now = 5 years ago?

As to the "evidence" that it is "happening at all", Fry's has
gigabit NICs for $7.99 and five-port gigabit switches for $19.99. That
didn't happen because nobody was buying the stuff, that happened because
the production volume is enormous.

"Happening" to me means that is being fully utilized. Are you suggesting
that people are actually filling these pipes in real office workloads
(and that real office use $19.99 switches?)?

Why is it so important to you that we believe that a common networking
technology, readily available, and becoming dirt cheap, is not common?

Not the technology, but it's utilization.

The time they introduced might.

The time who introduced what?

Gigabit devices. But just because it became available doesn't
necessarily mean there is widespread, full utilization.

(And I'd note you probably still haven't grasped that a tiny query can
trigger huge amounts of IO!).

So?

Try to follow the argument.

I don't see one.

How is five years old "recent" in the computer industry?

Now = 5 years ago?

So the computer industry says.

Is there anything more specific?

People who put together and run systems like
that get paid quite a lot more than your average PC technician.

Relevance?

flux wrote:

In article ,
Malcolm Weir wrote:

Well, *somebody* thinks it worthwhile putting it in commodity
chipsets, which means it *is* happening, despite your ignorance.

"Is" sounds very recent to me.

Are you not familiar with the word "support"? Unlike you, I'm not
claiming absolutes.

1999 is not an absolute?

No, it's (a) a year and (b) a matter of historical record.

Once again, 1999.

(Actually, it predated 1999. But that was when the standard was
done).

Does "Is" not mean 2004?

Huh? The fact that something is happening now does not mean that it just
started happening. It won't be "was" until 1 gig _stops_ happening.

So you don't *need* anything more than a 2400 baud phone line to
create a workload that involves more IO than you can get out of a SATA
drive. Which was what you seemed to think required gigabit ethernet!

Is this a magic trick? Can you tell us how that actually works?

He already did. Please try to pay attention in the future.

But 1999, kid, isn't recent. It's a (drive) lifetime ago.

1999 = 2004?

Huh? If you have a point you might want to try combing your hair over it.

OK. Go to www.dell.com. Look. Be enlightened.
Go to www.google.com. Enter "1000BaseT 1999". Look. Be enlightened.

Logic?

Something with which you are clearly unfamiliar.

Indeed. However, unlike you, I do this for a living.

Does this mean all people everywhere who do their jobs for a living must
all be competent at what they do?

He'd have to go some to be less competent than you.

It's clear that you're just a stupid kid who thinks that by playing word
games he proves that he's clever, when in fact you're just coming across as
one of the dumbest SOBs in Christendom. I see no point in wasting further
time on you.

plonk

(now, watch him come back with a sock puppet)

--
--John
Reply to jclarke at ae tee tee global dot net
(was jclarke at eye bee em dot net)

"Peter" wrote in
:

There is a difference between warm to the touch and 22 degrees.
Someone would need to measure temperature inside of the EMC box.
Temperature kills MTBF. Again, you will loose additional 30% of MTBF
for temperature going from 25 to 34 degC.
Seagate did not say what temperature was used for MTBF testing. They
only said that recommended case temperature from 5 to 55 degC was not
exceeded.

I'm sure EMC is keeping an eye on the temperature of their box.

About a year and a half now, but this has been the same since we
turned on the things.

And that is interesting. Despite of a "bathtub" curve for a hard drive
failure rate, putting all drives in 1.5 year from manufacturing date,
would mean the best reliability period for them.

These drives are tested before shipping to the customer. They should be
past the lead end of the bathtub curve when they arrive at our place.


--
/Jesper Monsted

"J. Clarke" wrote in
:

Which is a red flag. If it's not warm then it's probably not cooling
anything.

I'd go for the "there's so much air coming through those cases, the drives
can't make enough heat" explaination.

There's so much noise and airflow, i'm thinking they've put a row of jet
engines on that thing.


--
/Jesper Monsted

flux wrote in news:support-7DF9EF.01031908122004
@news.verizon.net:

So you don't *need* anything more than a 2400 baud phone line to
create a workload that involves more IO than you can get out of a SATA
drive. Which was what you seemed to think required gigabit ethernet!

Is this a magic trick? Can you tell us how that actually works?

Load up SQLnet, connect to your oracle server over the 2400 baud link,
enter sql statement that reads sequentially through several tables, hit
enter.

It's not too hard to generate massive loads on DBMS's (or even a plain
filesystem) with a single query - the tricky part is avoiding it.


--
/Jesper Monsted

There is a difference between warm to the touch and 22 degrees.
Someone would need to measure temperature inside of the EMC box.
Temperature kills MTBF. Again, you will loose additional 30% of MTBF
for temperature going from 25 to 34 degC.
Seagate did not say what temperature was used for MTBF testing. They
only said that recommended case temperature from 5 to 55 degC was not
exceeded.

I'm sure EMC is keeping an eye on the temperature of their box.

I'm sure they do if that is their resposibility. But while they rate their
storage unit MTBF at, let say, 400,000hr at 25degC, they allow its operation
in ambient temperature up to 50degC. Do you think that MTBF is going to stay
at 400,000hr if teperature goes substantially above 25degC? Would they do
anything if temperature inside the box will be constantly at 35degC (well
below maximum operational limit)?

And that is interesting. Despite of a "bathtub" curve for a hard drive
failure rate, putting all drives in 1.5 year from manufacturing date,
would mean the best reliability period for them.

These drives are tested before shipping to the customer. They should be
past the lead end of the bathtub curve when they arrive at our place.

Yes they should have them properly burned-in (time, temperature, load). Were
they, for sure?

In article ,
"J. Clarke" wrote:

1999 is not an absolute?

No, it's (a) a year and (b) a matter of historical record.

What is?

Does "Is" not mean 2004?

Huh? The fact that something is happening now does not mean that it just
started happening. It won't be "was" until 1 gig _stops_ happening.

But it did just start happening.


So you don't *need* anything more than a 2400 baud phone line to
create a workload that involves more IO than you can get out of a SATA
drive. Which was what you seemed to think required gigabit ethernet!

Is this a magic trick? Can you tell us how that actually works?

He already did. Please try to pay attention in the future.

No answer?

But 1999, kid, isn't recent. It's a (drive) lifetime ago.

1999 = 2004?

Huh? If you have a point you might want to try combing your hair over it.

So 1999 does equal 2004?

OK. Go to www.dell.com. Look. Be enlightened.
Go to www.google.com. Enter "1000BaseT 1999". Look. Be enlightened.

Logic?

Something with which you are clearly unfamiliar.

Logic?

Indeed. However, unlike you, I do this for a living.

Does this mean all people everywhere who do their jobs for a living must
all be competent at what they do?

He'd have to go some to be less competent than you.

Huh?

On Wed, 08 Dec 2004 06:03:24 GMT, flux wrote:

In article ,
Malcolm Weir wrote:

Well, *somebody* thinks it worthwhile putting it in commodity
chipsets, which means it *is* happening, despite your ignorance.

"Is" sounds very recent to me.

That's because you are stupid.

Are you not familiar with the word "support"? Unlike you, I'm not
claiming absolutes.

1999 is not an absolute?

No, it's a date of a specific event, moron.

Once again, 1999.

(Actually, it predated 1999. But that was when the standard was
done).

Does "Is" not mean 2004?

No.

So you don't *need* anything more than a 2400 baud phone line to
create a workload that involves more IO than you can get out of a SATA
drive. Which was what you seemed to think required gigabit ethernet!

Is this a magic trick? Can you tell us how that actually works?

No magic, just some elementary knowledge, which you obviously lack.

Here's how:

1. Take a large database. I'll assume something SQL based for the
sake of argument, but that only impacts the details, not the
principle.

2. Do something like "SELECT AVG(some_numeric_field) FROM
some_large_table;"

3. That's it. The database engine will read the entire table.

You truly are ignorant, aren't you?

But 1999, kid, isn't recent. It's a (drive) lifetime ago.

1999 = 2004?

Learn to read, you hopelessly arrogant and ill-informed idiot!

OK. Go to www.dell.com. Look. Be enlightened.
Go to www.google.com. Enter "1000BaseT 1999". Look. Be enlightened.

Logic?

Obviously, you're still not enlightened.

Do what you were told, son.

Indeed. However, unlike you, I do this for a living.

Does this mean all people everywhere who do their jobs for a living must
all be competent at what they do?

No.

Now, answer the question, kid!

What do you do?

Malc.

In article ,
Malcolm Weir wrote:

On Wed, 08 Dec 2004 06:03:24 GMT, flux wrote:

In article ,
Malcolm Weir wrote:

Well, *somebody* thinks it worthwhile putting it in commodity
chipsets, which means it *is* happening, despite your ignorance.

"Is" sounds very recent to me.

That's because you are stupid.

That makes it less recent?

Are you not familiar with the word "support"? Unlike you, I'm not
claiming absolutes.

1999 is not an absolute?

No, it's a date of a specific event, moron.

So what?

Once again, 1999.

(Actually, it predated 1999. But that was when the standard was
done).

Does "Is" not mean 2004?

No.

Yes, it does.


Is this a magic trick? Can you tell us how that actually works?

No magic, just some elementary knowledge, which you obviously lack.

Here's how:

1. Take a large database. I'll assume something SQL based for the
sake of argument, but that only impacts the details, not the
principle.

2. Do something like "SELECT AVG(some_numeric_field) FROM
some_large_table;"

3. That's it. The database engine will read the entire table.

And that is supposed to transverse the 2400 baud network?

You truly are ignorant, aren't you?

But 1999, kid, isn't recent. It's a (drive) lifetime ago.

1999 = 2004?

Learn to read, you hopelessly arrogant and ill-informed idiot!

I guess that means 1999 does not equal 2004.

OK. Go to www.dell.com. Look. Be enlightened.
Go to www.google.com. Enter "1000BaseT 1999". Look. Be enlightened.

Logic?

Obviously, you're still not enlightened.

Obviously.

Do what you were told, son.

Indeed. However, unlike you, I do this for a living.

Does this mean all people everywhere who do their jobs for a living must
all be competent at what they do?

No.

Now, answer the question, kid!

What do you do?

Correct misinformation with information.