Stupid RAID question

[D. Hugh Redelmeier]

| From: Lennart Sorensen <lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org>

| On Tue, Oct 04, 2011 at 04:26:28PM -0400, D. Hugh Redelmeier wrote:

| > Note that most current consumer hard drives are not suitable for RAID.
| > The can take a large amount of time doing internal error recover,
| > enough that the RAID system decides the whole drive is dead.  That is
| > a practical disaster because rebuilding an array on a 2TB disk takes a
| > large portion of a day.  See for example:
| >   <http://en.wikipedia.org/wiki/Time-Limited_Error_Recovery>
| > This is a scam as far as I'm concerned.  Perpetrated by the last two
| > surviving drive manufacturers.
| 
| No I don't consider that a scam.

"Scam" is a subjective term.  So let me explain how I see this.

They are using market segmentation to make more money.  That's fair.
But as a techie, I like to think of commodity products being sold at
close to cost, not value.  The value of a RAID-capable disk is higher
than the value of one that is not RAID-capable.  The cost of providing
it is essentially identical (or actually lower).

The sign of a monopoly market is a vendor able to capture the value of a 
product to the customer, not just its cost of production.  Market 
segmentation is one tool used by vendors to accomplish this more 
effectively.

>From http://en.wikipedia.org/wiki/Market_segmentation

    The term is also used when consumers with identical product and/or
    service needs are divided up into groups so they can be charged
    different amounts for the services.

The cost of having a TLER option in the firmware is essentially $0.
The cost of having different firmware for identical devices, including
service cost is more than $0 (perhaps not a lot).  Eliminating this
option segments the market.

|  Most consumers don't use raid, so if
| the drive can recover from a failing sector by trying multiple reads to
| see if it can get lucky, then that makes it a better drive.

I use the same drives in different contexts during my ownership.  I'd
sure like to be able to use a drive for RAID or not whenever I chose
to do so.

WD used to allow this, but no longer.  Hitachi and Samsung still allow
this (but are going to disappear).  Seagate does not allow this now
but may have done so in the past.

I'm particularly annoyed at Seagate since they advertised my drive (a
7200.11 series) as suitable for "desktop RAID" but provided no way of
bounding the error recovery time.  When pressed by any particular user
encountering problems, if they cose to reply at all, they claimed any
particular configuration was not desktop RAID.  Notice this
327-message thread that I started:
<http://forums.seagate.com/t5/Barracuda-XT-Barracuda-Barracuda/RAID-problems-with-7200-11-drives-unified-thread/td-p/26989>

This is on top of the 7200.11 BSY firmware bug which was handled very very
badly.

| The fact raid systems have a time limit on a drive responding makes the
| above behaviour undesirable when using it in a raid.
| 
| So the two uses have mutually exclusive needs.

Right, but it is only a tiny aspect of behaviour, quite within the
realm of tweaking.

| To some extent I think WD sells raid edition drives for more money than
| non raid drives just to make sure consumers don't buy them thinking
| "raid drives must be better or faster so I will use that" and end up
| with a much less reliable drive in their single drive system.

No, they do it to make more money.

The fact that they can get away with this clearly demonstrates that
the market isn't free.  Only a monopoly/oligopoly can get away with
market segmentation like this.

Consider, for another example, how CPU virtualization technology was in essentially
all AMD chips but only some Intel chips.  My analysis:

- Intel felt it could force users requiring virtualization to pay more
  (because they dominated the market).  A sale of a cheap chip, if it
  had supported virtualization, could be cannibalizing a sale of a
  more expensive chip

- AMD was happy to sell any chips they could.  A cheap chip with
  virtualization might be displacing an Intel chip.

(So even my cheap-as-dirt AMD-based netbook supports virtualization
but Intel netbooks and most Intel notebooks don't.)

Ordinary customers might well want virtualization.  If I
remember/understand correctly, Win7 supports backwards compatibility
using virtualization.

|  People that
| need them will know what they need and get the right one.  Now I would
| prefer the old method where a software tool or a jumper could change
| the behaviour, but unfortunately we don't get that option anymore.

If that were the case, why would Seagate advertise the 7200.11 as
suitable for "desktop RAID"?
<http://www.seagate.com/www/en-us/products/desktops/barracuda_hard_drives/barracuda_7200.11>

With the 7200.12, they don't mention "desktop RAID" but they do
suggest, without reservations or qualifications:

    PERFECT WHEN YOU NEED TO:

	Build robust workstations
	Equip home servers
	Create PC-based gaming systems
	Implement a desktop RAID
	Outfit direct attached external storage devices (DAS)
	Build network attached storage devices (NAS)

No, they are certainly not in it to give a clear understanding of this
issue to the rubes.
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists