bizarre hard drive failure

Hi all,
I bought a Dell Inspiron 1501 laptop computer (Turion 64 X2-based, ATI
chipset) and it arrived a couple days ago. In no time at all, I had
Ubuntu Feisty up on it alongside Windows Vista, and everything seemed
to work great. I used "smartctl -a" to verify that the hard drive
parameters were within the expected good range for a brand new drive,
and everything checked out. The drive is a Hitachi HTS541680J9SA00,
80gb, 5400rpm, SATA.

Today, I was playing around with hdparm to see if I could reduce the
power consumption of the disk drive. I noticed that this drive
supported the "power-on in standby" feature, and enabled it with
"hdparm -s1". This didn't seem to be a dangerous thing to do at all.
I kept using the computer for an hour or so, then put it in standby
mode, and tried to take it out a minute later.

To my surprise, when the computer booted back up, BIOS informed me
that the hard drive had failed. I couldn't believe it. I tried
booting off a USB stick with a small Linux system on it, and the boot
messages informed me again that Linux was unable to communicate with
the first SATA hard disk. I then removed the drive from the laptop
and put it into my desktop computer, where the BIOS again failed to
recognize it. I tried holding the drive in my hand as I powered-on
the desktop, and I could not feel the drive begin to spin up at all.

Is it possible that I've ruined this drive by enabling "power-on in
standby"???? Why would the drive have this option if it doesn't
work??? Is there any way to get it out of this mode? Or could this
be a freak coincidental drive failure? Any advice will be greatly
appreciated!

Dan Lenski
University of Maryland

Dan Lenski <dlenski@gmail.com> writes:[color=blue]
>Is it possible that I've ruined this drive by enabling "power-on in
>standby"????[/color]

Yes, but coincidence seems more likely. After all, the drive is
apparently not in standby mode (where you can talk to it), but it's
dead.
[color=blue]
> Is there any way to get it out of this mode?[/color]

hdparm -s0

Of course, if you can't talk to the drive, this won't work.

BTW, I don't see a point for using the -s1 mode on a laptop. When you
turn it on, you usually want to access the disk (for booting). This
feature is more for servers with huge disk arrays; there you want to
spin up the drives in a staggered fashion, otherwise you would need to
use an even bigger power supply unit.

Some -S setting in combination with laptop mode is useful for saving
power on laptops.

- anton
--
M. Anton Ertl Some things have to be seen to be believed
[email]anton@mips.complang.tuwien.ac.at[/email] Most things have to be believed to be seen
[url]http://www.complang.tuwien.ac.at/anton/home.html[/url]

On Jun 1, 3:42 am, a...@mips.complang.tuwien.ac.at (Anton Ertl) wrote:[color=blue]
> Dan Lenski <dlen...@gmail.com> writes:[color=green]
> >Is it possible that I've ruined this drive by enabling "power-on in
> >standby"????[/color]
>
> Yes, but coincidence seems more likely. After all, the drive is
> apparently not in standby mode (where you can talk to it), but it's
> dead.[/color]

Indeed, that was my thinking...
[color=blue]
> BTW, I don't see a point for using the -s1 mode on a laptop. When you
> turn it on, you usually want to access the disk (for booting). This
> feature is more for servers with huge disk arrays; there you want to
> spin up the drives in a staggered fashion, otherwise you would need to
> use an even bigger power supply unit.
>
> Some -S setting in combination with laptop mode is useful for saving
> power on laptops.[/color]

Right, I was pretty much just playing around with the settings. I
figured enabling -s would be more or less harmless. As you point out,
a drive in standby mode can normally still identify itself to the
BIOS. I've got this sneaking suspicion that my fiddling did SOMETHING
to harm the drive, but I do feel justified in sending this in under
the warranty. Two-day-old drives aren't supposed to just go into
standby mode and die.

Dan

In article <1180706166.304476.84040@p47g2000hsd.googlegroups.com>, Dan Lenski wrote:[color=blue]
> to harm the drive, but I do feel justified in sending this in under
> the warranty. Two-day-old drives aren't supposed to just go into
> standby mode and die.[/color]
I dont know much about IDE/ATA/SATA but I do kknow SCSI and if
your controller does not send the spinup command the drive will sit
there reporting Not Ready errors as the motor never gets "Hey! Wake up
you lazy piece of hardware! its a New Day!" I can do the same with
SCSI and tell my Bios not to send the Spin up command and I get
similar errors.

--

From the Desk of the Sysop of:
Planet Maca's Opus, a Free open BBS system. [url]telnet://pinkrose.dhis.org[/url]
Web Site: [url]http://pinkrose.dhis.org[/url], Dialup 860-618-3091 300-33600 bps
The New Cnews maintainer
B'ichela

On Jun 1, 3:42 pm, B'ichela <mdal...@pinkrose.dhis.org> wrote:[color=blue]
> I dont know much about IDE/ATA/SATA but I do kknow SCSI and if
> your controller does not send the spinup command the drive will sit
> there reporting Not Ready errors as the motor never gets "Hey! Wake up
> you lazy piece of hardware! its a New Day!" I can do the same with
> SCSI and tell my Bios not to send the Spin up command and I get
> similar errors.[/color]

Yes, you were right! This turned out to be an INCREDIBLY obscure BIOS
bug: most BIOSes (as well as the Linux kernel) don't know what to do
with a disk that doesn't spin up automatically, so they declare it to
be dead.

For the complete, two-day saga of how I fixed the problem, check out
the thread in comp.sys.ibm.pc.hardware.storage:
[url]http://groups.google.com/group/comp.sys.ibm.pc.hardware.storage/browse_frm/thread/d4b6da0f662c1ecc/25bca2242db30700#25bca2242db30700[/url]

Dan

On 2007-06-01, Dan Lenski <dlenski@gmail.com> wrote:
[color=blue]
> I bought a Dell Inspiron 1501 laptop computer (Turion 64 X2-based, ATI
> chipset) and it arrived a couple days ago. In no time at all, I had
> Ubuntu Feisty up on it alongside Windows Vista, and everything seemed
> to work great. I used "smartctl -a" to verify that the hard drive
> parameters were within the expected good range for a brand new drive,
> and everything checked out. The drive is a Hitachi HTS541680J9SA00,
> 80gb, 5400rpm, SATA.
>
> Today, I was playing around with hdparm to see if I could reduce the
> power consumption of the disk drive. I noticed that this drive
> supported the "power-on in standby" feature, and enabled it with
> "hdparm -s1". This didn't seem to be a dangerous thing to do at all.
> I kept using the computer for an hour or so, then put it in standby
> mode, and tried to take it out a minute later.
>
> To my surprise, when the computer booted back up, BIOS informed me
> that the hard drive had failed. I couldn't believe it. I tried
> booting off a USB stick with a small Linux system on it, and the boot
> messages informed me again that Linux was unable to communicate with
> the first SATA hard disk. I then removed the drive from the laptop
> and put it into my desktop computer, where the BIOS again failed to
> recognize it. I tried holding the drive in my hand as I powered-on
> the desktop, and I could not feel the drive begin to spin up at all.
>
> Is it possible that I've ruined this drive by enabling "power-on in
> standby"???? Why would the drive have this option if it doesn't
> work??? Is there any way to get it out of this mode? Or could this
> be a freak coincidental drive failure? Any advice will be greatly
> appreciated![/color]

First things first: did you check the cable connections? If you have an
external USB drive enclosure, mount the drive in there and see if you
can talk to it using another computer.

--

John (john@os2.dhs.org)

In article <1180805122.479484.125550@m36g2000hse.googlegroups.com>, Dan Lenski wrote:[color=blue]
> On Jun 1, 3:42 pm, B'ichela <mdal...@pinkrose.dhis.org> wrote:[color=green]
>> I dont know much about IDE/ATA/SATA but I do kknow SCSI and if
>> your controller does not send the spinup command the drive will sit
>> there reporting Not Ready errors as the motor never gets "Hey! Wake up
>> you lazy piece of hardware! its a New Day!" I can do the same with
>> SCSI and tell my Bios not to send the Spin up command and I get
>> similar errors.[/color]
>
> Yes, you were right! This turned out to be an INCREDIBLY obscure BIOS
> bug: most BIOSes (as well as the Linux kernel) don't know what to do
> with a disk that doesn't spin up automatically, so they declare it to
> be dead.[/color]
And you are welcome! Glad this old SCSI lass could give you
some clues. ;) As for your Bios/Cmos if you can, flash it with the
latest updates. may or may not help but I don't think it would hurt
either. If your SATA controller is similar to my CMOS/BIOS on my
AHA2940UW card it may have its own configuration screen to allow you
to have it do the spin up for you. Might want to check your
manual/Manufactures website to see if it can. Never know if you get
another one of those "dead" drives and want to have your bios fire the
start spin command by default.
BTW it isn't a bug, It simply wasn not expected that someone
would want to do something so SCSI on a ATA drive. or for that matter
in a home/office PC.

--

From the Desk of the Sysop of:
Planet Maca's Opus, a Free open BBS system. [url]telnet://pinkrose.dhis.org[/url]
Web Site: [url]http://pinkrose.dhis.org[/url], Dialup 860-618-3091 300-33600 bps
The New Cnews maintainer
B'ichela

On Sun, 03 Jun 2007 00:57:22 +0000, B'ichela wrote:[color=blue]
> And you are welcome! Glad this old SCSI lass could give you
> some clues. ;) As for your Bios/Cmos if you can, flash it with the
> latest updates. may or may not help but I don't think it would hurt
> either. If your SATA controller is similar to my CMOS/BIOS on my
> AHA2940UW card it may have its own configuration screen to allow you
> to have it do the spin up for you. Might want to check your
> manual/Manufactures website to see if it can. Never know if you get
> another one of those "dead" drives and want to have your bios fire the
> start spin command by default.[/color]

Yeah, thanks for the clues! From the discussion in
comp.sys.ibm.pc.hardware.storage, I've found that most onboard SATA
controllers don't handle this feature at all, including the
brand-spankin-new laptop that this drive came in :-(.
[color=blue]
> BTW it isn't a bug, It simply wasn not expected that someone
> would want to do something so SCSI on a ATA drive. or for that matter
> in a home/office PC.[/color]

Well... that's true I guess. But many SATA drives *do*
support this feature, and most SATA BIOes *don't*. So that's a pretty bad
situation, since a drive put into this mode will appear quite dead... and
the only way to rescue such a drive that I've figured out is to hotplug
the drive into a PC with a patched Linux kernel. Rather obscure for what
is such a simple operation!

Dan