Re: WD doesn`t use disks with Linux - was Re: Re:Western Digital GreenPower drives and Linux

Previous thread: Incorrect SMP topology reported on AMD K7 MP system by Bill Johnstone on Tuesday, May 13, 2008 - 1:53 pm. (1 message)

Next thread: [PATCH ?] pcmcia: ti113x: fix ENE_TEST_C9_PFENABLE definition by Roel Kluin on Tuesday, May 13, 2008 - 2:04 pm. (1 message)
From: Roland
Date: Tuesday, May 13, 2008 - 2:00 pm

I`ve got a real funky response from WD support  today, i`d like to share 
here.
This is just too weird....

After asking if they could give a comment if the problem with the high 
Load_Cycle_Count may be an issue of their drives or an issue of Linux i got:

------
Dear......
The problem is, that we actually don`t use the disks with Linux. I´m sorry 
that i cannot send further information to you.

Sincery yours
<insert support employee name here>
Western Digital Service and Support
------
(translated from german)




So, WD sells enterprise SATA disks, optimizied for 24/7 & Raid array usage, 
but they don`t seem to actually test them with Linux or at least are not 
able to "support" Linux usage.......


--

From: Arnd Hannemann
Date: Tuesday, May 13, 2008 - 2:20 pm

Well the answer was probably just from a first level support person so
nothing one should get excited about.
Anyway to the high Load_Cycle_Count; if I understood the technology
correctly the WD GreenPower drives are exactly "Green" because they
unload the heads on a plastic ramp (no landing zone!) quite often during
longer idle periods. Thus they can power off the positioning logic and
actuator of the heads. -> safe Power
This also explains why the power usage under seek is comparable with
other "non-green" drives...

my 1 TB drive has also a relative high value:

193 Load_Cycle_Count        0x0032   200   200   000    Old_age  
Always       -       2786


Just my 2 cents,
Arnd
||
--

From: Jan Engelhardt
Date: Tuesday, May 13, 2008 - 3:29 pm

I don't find it Green when I have to buy new disks because of

ramp or LZ, it's quite a force with which the head is parked.

--

From: Arnd Hannemann
Date: Wednesday, May 14, 2008 - 12:48 am

Well I'm no expert, but I would say comparing to the normal wear out due
to use (seeking, bearing wear out, vibrations, spin up, bumps) the
factor of a controlled unload/load cycle on a plastic ramp is probably
negligible.

BTW: my hitachi notebook drive, which I believe has a similar

And it didn't fail on me yet...
Of course this is just one data point.

Regards,
Arnd Hannemann
--

From: linux-os (Dick Johnson)
Date: Wednesday, May 14, 2008 - 6:20 am

Practically nobody has any on-line support that means anything
anymore. The person who wrote the response is probably in the
Philippines, makes about $US 3.00 per hour, and is allowed to
pee twice a day. Don't worry. The tiny "bump-proof" drives
cycle on an off many times to conserve power. They also
keep the heads from landing "crashing" by inserting a plastic
wedge between the swing-arms during power down. It's probably
a lot less harmful than letting the heads touch down somewhere.

It's not just the positioning logic that gets powered off. The
discs are spun down, which saves a lot of power.


Cheers,
Dick Johnson
Penguin : Linux version 2.6.22.1 on an i686 machine (5588.29 BogoMips).
My book : http://www.AbominableFirebug.com/
_

****************************************************************
The information transmitted in this message is confidential and may be privileged.  Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited.  If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to DeliveryErrors@analogic.com - and destroy all copies of this information, including any attachments, without reading or disclosing them.

Thank you.
--

From: Jan Engelhardt
Date: Wednesday, May 14, 2008 - 6:39 am

And drives latencies into the skies.
Not to mention that the permanent spinups and downs also affect life.
--

From: linux-os (Dick Johnson)
Date: Wednesday, May 14, 2008 - 9:09 am

Yes. Maybe there is a way to turn if off? That's why we
should be able to get somebody from WD who will answer
a Linux question, instead of the usual on-line "help."

I'm going to poke around and see if I can  find a real
person at WD.

Cheers,
Dick Johnson
Penguin : Linux version 2.6.22.1 on an i686 machine (5588.29 BogoMips).
My book : http://www.AbominableFirebug.com/
_


****************************************************************
The information transmitted in this message is confidential and may be privileged.  Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited.  If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to DeliveryErrors@analogic.com - and destroy all copies of this information, including any attachments, without reading or disclosing them.

Thank you.
--

From: linux-os (Dick Johnson)
Date: Wednesday, May 14, 2008 - 10:15 am

It took about 3/4 hour to access their web-page and get
a question submitted. It has been my experience that companies
that don't maintain their web-pages, don't really care about
customers. My query was, "How do I turn off Green Power so
I don't have the horrible latencies on virtual memory systems
that have active swapping?"

In the unlikely event that they answer me, I will put the
answer on this list. My guess is that if they do answer
me, the answer will be, "Duhh?" We'll see.


Cheers,
Dick Johnson
Penguin : Linux version 2.6.22.1 on an i686 machine (5588.29 BogoMips).
My book : http://www.AbominableFirebug.com/
_


****************************************************************
The information transmitted in this message is confidential and may be privileged.  Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited.  If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to DeliveryErrors@analogic.com - and destroy all copies of this information, including any attachments, without reading or disclosing them.

Thank you.
--

From: linux-os (Dick Johnson)
Date: Thursday, May 15, 2008 - 1:39 pm

The only answer I got was the "canned" answer about
contacting your hardware and software vendor. There
don't seem to be any humans employed there, anymore.
That's unfortunate because Lake Forest, California is
a nice place.

Cheers,
Dick Johnson
Penguin : Linux version 2.6.22.1 on an i686 machine (5588.29 BogoMips).
My book : http://www.AbominableFirebug.com/
_


****************************************************************
The information transmitted in this message is confidential and may be privileged.  Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited.  If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to DeliveryErrors@analogic.com - and destroy all copies of this information, including any attachments, without reading or disclosing them.

Thank you.
--

From: Pádraig Brady
Date: Friday, May 16, 2008 - 2:35 am

I wrote a few notes about load cycles here:
http://www.pixelbeat.org/docs/hard_disk_reliability/

It's bad that the WD GreenPower doesn't support
`hdparm -B` setting its parameters, but I found I
didn't need to tweak these anyway.

I just got the distro to write less often to the drive
with the (rel)atime mount options. With that I got an
acceptable load cycle rate.

Pádraig.

--

From: Jan Engelhardt
Date: Friday, May 16, 2008 - 3:21 am

The problem is not how often Linux wakes the head up due to
writes of any sort, but how often the head decides to go
down without our consent.
--

Previous thread: Incorrect SMP topology reported on AMD K7 MP system by Bill Johnstone on Tuesday, May 13, 2008 - 1:53 pm. (1 message)

Next thread: [PATCH ?] pcmcia: ti113x: fix ENE_TEST_C9_PFENABLE definition by Roel Kluin on Tuesday, May 13, 2008 - 2:04 pm. (1 message)