What would cause an internal HD to simply disappear in Windows?

[mscotgrove]

Thats also a classic with memory leak problems, they dont
happen that often, only after quite a while after the reboot.- Hide quoted text -

- Show quoted text -

Have you eliminated any drive mapping problems. I have noted drives
disappearing when there is a conflict with a network mapped drive, or
a drive SUBST command. I am not sure if this will happen with an
internal drive.

Michael

 

[Rod Speed]

Have you eliminated any drive mapping problems.

Unlikely to affect two drives.

I have noted drives disappearing when there is a conflict
with a network mapped drive, or a drive SUBST command.
I am not sure if this will happen with an internal drive.

Yes it can.

 

[calypso]

It's a Zalman. The same people who make those fancy overclocker heat
sinks for processors.

It's OK, very good PSU, so I'd rule it out...

It's a quad-rail PSU. Sorry, it's not a 650W, it's a 600W PSU, my
mistake. No matter, that's just Zalman's conservative rating, another
manufacturer, OCZ, repackages the same PSU and rates it at 700W.

I did have a problem with low-quality PSU's in the past, I used to go
for cheap 400W units based on price mostly, until I started noticing
that certain instabilities were being caused. So I decided to properly
invest in this one, this time. However, even still, a single 12V rail
can only supply upto 18.5A, which is 222W. I think two of those rails
are dedicated to the motherboard, and one for a high-end video card
(don't have one of those), and the rest for anything else, which
probably means the hard drives. Is 222W enough? I would've thought so.

Yup... It's not the problem then... HDD uses cca. 30W during spinup and when
working mostly 10-15W of power... It's not an issue here...

It just happened once, it isn't a recurring thing (yet). But once is
enough in this case. It's not supposed to do that at all, not unless
the drive itself dies which didn't happen.

Well, then, it's a fully qualified ghost...

--
"Cokoladans li kokoso tuce ?" upita kamion gnjeci Samurajog gnjecija.
"Nisam ja nikog bombardiro !" rece studenta pjeva "Ja samo Mercedesu bacu suncokretovm !"
By runf

Damir Lukic, calypso@_MAKNIOVO_fly.srk.fer.hr
http://inovator.blog.hr
http://calypso-innovations.blogspot.com/

 

[Yousuf Khan]

Have you eliminated any drive mapping problems. I have noted drives
disappearing when there is a conflict with a network mapped drive, or
a drive SUBST command. I am not sure if this will happen with an
internal drive.

Yeah, that's not going to be the problem. I'd have had to have actually
tried to map a network drive at the time this internal drive
disappeared. The internal drive just disappeared by itself, and as I
said before there are entries in the Windows System log showing when it
happened. It showed controller problems at the time, with no other
details beyond that.

Yousuf Khan

 

[Jonathan de Boyne Pollard]

Well, considering that the error showed up with both an IDE and a SATA
drive, it may have something to do with the motherboard's disk
controller subsystem.

There is no disc controller on the motherboard with IDE drives. That's
what Integrated Drive Electronics means, after all. Turn your disc unit
over. See the chips and circuitry? That's where the disc controller
is. The motherboard contains merely a PCI-to-ATA bridge.

If there's a problem with a PCI-to-ATA bridge, it will by and large
affect both disc units (master and slave) on both channels (primary and
secondary) handled by the bridge. Device Manager, in "devices by
connection" view, will of course tell you which ATA devices are
connected to which ATA buses, and what the two devices that you saw
named in the error log have in common.

 

[Arno]

There is no disc controller on the motherboard with IDE drives. That's
what Integrated Drive Electronics means, after all. Turn your disc unit
over. See the chips and circuitry? That's where the disc controller
is. The motherboard contains merely a PCI-to-ATA bridge.

There is a disk controller, it just has the form of a communications
controller, not a disk hardware controller. It still does things
like DMA transfer and PCI interfacing.

If there's a problem with a PCI-to-ATA bridge,

Such a thing does not exist. PCI and ATA are fundamentally
different, hence there is a controller. With a classical ISA
bus on the mainboard, it would indeed only take some sort
of bridge.

Arno

 

[Rod Speed]

There is no disc controller on the motherboard with IDE drives.

Yes there is.

That's what Integrated Drive Electronics means, after all.

There is still a controller on the motherboard as well as that.

You can see that in the device manager.

Turn your disc unit over. See the chips and circuitry? That's where the disc controller is. The motherboard
contains merely a PCI-to-ATA bridge.

Its rather more than just a bridge, particularly when its RAID capable.

If there's a problem with a PCI-to-ATA bridge, it will by and large affect both disc units (master and slave) on both
channels (primary and secondary) handled by the bridge.

There is no master and slave with SATA.

Device Manager, in "devices by connection" view, will of course tell you which ATA devices are connected to which ATA
buses,

There is no ATA bus with SATA.

 

[Yousuf Khan]

There is no disc controller on the motherboard with IDE drives. That's
what Integrated Drive Electronics means, after all. Turn your disc unit
over. See the chips and circuitry? That's where the disc controller
is. The motherboard contains merely a PCI-to-ATA bridge.

Yes, yes, we all know that, but disk controller is the colloquial term
for the bridge circuitry, despite the inaccuracy. And Windows itself
calls them controllers.

If there's a problem with a PCI-to-ATA bridge, it will by and large
affect both disc units (master and slave) on both channels (primary and
secondary) handled by the bridge. Device Manager, in "devices by
connection" view, will of course tell you which ATA devices are
connected to which ATA buses, and what the two devices that you saw
named in the error log have in common.

Okay, I just tried that. The system is organized into two "Standard dual
channel PCI IDE controllers" (I'm using them all in IDE compatibility
mode). Each "Standard dual" has two "ATA channels", which themselves
each hold 2 drives, meaning 4 drives per "Dual channel", or 8 possible
drives altogether. The hard drive that failed was on the first
dual-channel, and the two CDROMs are in the second dual-channel.

So in other words, no connection between them.

Yousuf Khan

 

[Jonathan de Boyne Pollard]

There is still a controller on the motherboard as well as that. You
can see that in the device manager.

What one can see in Device Manager is always inferior to what one can
see with one's own two eyes physically on the disc unit itself, kiddo.

There is no ATA bus with SATA.

A serial bus is still a bus, kiddo, and a serial ATA bus is still an ATA
bus.

 

[Jonathan de Boyne Pollard]

There is a disk controller, it just has the form of a communications
controller, not a disk hardware controller. It still does things like
DMA transfer and PCI interfacing.

If it's controlling comms, then it's clearly not a disc controller. Think!

It isn't a comms controller really, of course. It is, as stated, a
bridge between two buses, the PCI bus and the ATA bus, with a bit of
support circuitry so that it can be a programmable busmaster on the former.

Such a thing does not exist.

Your PC begs to differ with you, kiddo. Every time that there's an I/O
R/W transaction on your PCI bus (matching the port ranges specified in
PCI configuration space for the function, of course), your PCI-to-ATA
bridge is happily translating that into a R/W cycle on your ATA bus,
using the data and address lines, chip selects, and R/W strobe lines
that exist on the (parallel) ATA bus just as they do on many other
computer buses. And this will continue to be, however much you protest
its nonexistence on Usenet.

 

[Jonathan de Boyne Pollard]

There is no disc controller on the motherboard with IDE drives.
That's what Integrated Drive Electronics means, after all. Turn your
disc unit over. See the chips and circuitry? That's where the disc
controller is. The motherboard contains merely a PCI-to-ATA bridge.

Yes, yes, we all know that, [...]

Apparently not. M. Arno and M. Speed don't, it seems. It's
disappointing in the former case.

Okay, I just tried that. The system is organized into two "Standard
dual channel PCI IDE controllers" (I'm using them all in IDE
compatibility mode). Each "Standard dual" has two "ATA channels",
which themselves each hold 2 drives, meaning 4 drives per "Dual
channel", or 8 possible drives altogether. The hard drive that failed
was on the first dual-channel, and the two CDROMs are in the second
dual-channel.

"dual channel" is attributive, and not a full name it's own right,
note. It's a description of the type. Some PCI-to-ATA bridges (like
the one in the Intel 82371) are dual-channel; some (like the one at
device 31 function 1 in the Intel 82801) are single-channel.

So in other words, no connection between them.

Then there's very probably no problem with your PCI-to-ATA bridge. Now
look at the power rails. Are the drives in question sharing the same
power rails from your PSU?

 

[Rod Speed]

Jonathan de Boyne Pollard wrote

What one can see in Device Manager is always inferior to what one can
see with one's own two eyes physically on the disc unit itself, kiddo.

You're so stupid that you cant even manage to work out that the controller
that does the DMA and RAID on the motherboard is clearly a controller, ****wit.

A serial bus is still a bus, kiddo,

Taint a bus when its just got one device on each end of the cable, ****wit.

There is no bus with RS232, ****wit child.

and a serial ATA bus is still an ATA bus.

Wrong, as always.

 

[Rod Speed]

That mangles the story considerably. Its still a hardware controller,
controlling DMA to the drives, and optionally RAID etc as well.

If it's controlling comms, then it's clearly not a disc controller.

Wrong when its controlling communications with disks, ****wit child.

Think!

Not something you can manage, ear to ear dog shit cant do anything like that.

It isn't a comms controller really, of course. It is, as stated, a bridge between two buses, the PCI bus and the ATA
bus,

Its a hell of a lot more than just a bridge, most obviously
with the DMA and RAID functionality and the selection of
master and slave with PATA drives.

In spades at boot time when its deciding what drives are present etc.

with a bit of support circuitry so that it can be a programmable busmaster on the former.

It isnt even necessarily on the PCI bus at all, ****wit child.

Your PC begs to differ with you, kiddo.

Like hell it does, ****wit child.

Every time that there's an I/O R/W transaction on your PCI bus (matching the port ranges specified in PCI
configuration space for the function, of course), your PCI-to-ATA bridge is happily translating that into a R/W cycle
on your ATA bus, using the data and address lines, chip selects, and R/W strobe lines
that exist on the (parallel) ATA bus just as they do on many other computer buses.

Utterly mangled all over again.

And this will continue to be, however much you protest its nonexistence on Usenet.

Keep desperately digging that hole, child, you'll be out in china any day now.

 

[Rod Speed]

There is no disc controller on the motherboard with IDE drives.
That's what Integrated Drive Electronics means, after all. Turn
your disc unit over. See the chips and circuitry? That's where
the disc controller is. The motherboard contains merely a
PCI-to-ATA bridge.

Yes, yes, we all know that, [...]

Apparently not. M. Arno and M. Speed don't, it seems. It's
disappointing in the former case.

Okay, I just tried that. The system is organized into two "Standard
dual channel PCI IDE controllers" (I'm using them all in IDE
compatibility mode). Each "Standard dual" has two "ATA channels",
which themselves each hold 2 drives, meaning 4 drives per "Dual
channel", or 8 possible drives altogether. The hard drive that failed
was on the first dual-channel, and the two CDROMs are in the second
dual-channel.

"dual channel" is attributive, and not a full name it's own right, note.

Meaningless waffle.

It's a description of the type. Some PCI-to-ATA bridges (like
the one in the Intel 82371) are dual-channel; some (like the one at device 31 function 1 in the Intel 82801) are
single-channel.

More irrelevant waffle. Pity they are clearly controllers and called that too.

Then there's very probably no problem with your PCI-to-ATA bridge. Now look at the power rails. Are the drives in
question sharing the
same power rails from your PSU?

Corse they are.

 

[Arno]

If it's controlling comms, then it's clearly not a disc controller. Think!

It isn't a comms controller really, of course. It is, as stated, a
bridge between two buses, the PCI bus and the ATA bus, with a bit of
support circuitry so that it can be a programmable busmaster on the former.
Your PC begs to differ with you, kiddo. Every time that there's an I/O
R/W transaction on your PCI bus (matching the port ranges specified in
PCI configuration space for the function, of course), your PCI-to-ATA
bridge is happily translating that into a R/W cycle on your ATA bus,
using the data and address lines, chip selects, and R/W strobe lines
that exist on the (parallel) ATA bus just as they do on many other
computer buses. And this will continue to be, however much you protest
its nonexistence on Usenet.

Condensation will get you nowhere. Also check my sig if you think
I am a kid.

Arno

 

[Arno]

There is no disc controller on the motherboard with IDE drives.
That's what Integrated Drive Electronics means, after all. Turn your
disc unit over. See the chips and circuitry? That's where the disc
controller is. The motherboard contains merely a PCI-to-ATA bridge.

Yes, yes, we all know that, [...]

Apparently not. M. Arno and M. Speed don't, it seems. It's
disappointing in the former case.

I just happen to know what a bridge is and what a controller is.
That it is a "disk communications controller" does not really matter
here. But a bridge would never need a driver for anything, as it is
invisible on the software side (whith the exception of timing effects).
Also addressing on the PCI bus is different from ATA addressing, which
also means it cannot be only a bridge.

Incidentially, this question is besides the point.

Arno

 

[Jonathan de Boyne Pollard]

There is no disc controller on the motherboard with IDE drives.
That's what Integrated Drive Electronics means, after all. Turn
your disc unit over. See the chips and circuitry? That's where the
disc controller is. The motherboard contains merely a PCI-to-ATA
bridge.

Yes, yes, we all know that, [...]

Apparently not. M. Arno and M. Speed don't, it seems. It's
disappointing in the former case.

I just happen to know what a bridge is and what a controller is.

No, kiddo, you really don't in this case. And as I said, that's
disappointing in your case. Think! If ever you find yourself with the
same (mis-)understanding of hardware as Rod Speed has, alarm bells
should be going off.

That it is a "disk communications controller" does not really matter
here. But a bridge would never need a driver for anything, as it is
invisible on the software side (whith the exception of timing effects).

I've news for you. Not only do PCI-to-ATA bridges have drivers, but
even PCI-to-PCI bridges have drivers. On Windows NT, for example, the
latter are driven by the PCI.sys driver. The idea that bridges don't
need drivers is not in accordance with what you'll find actually
happening on your PC. And the consequent idea that if something has a
driver it cannot be a bridge is, accordingly, nonsense.

Also addressing on the PCI bus is different from ATA addressing, which
also means it cannot be only a bridge.

Two different buses bridged together aren't required to have the same
addressing, kiddo. That's often the point of their being two different
buses that require a bridge connecting them.

 

[Yousuf Khan]

There is no disc controller on the motherboard with IDE drives.
That's what Integrated Drive Electronics means, after all. Turn your
disc unit over. See the chips and circuitry? That's where the disc
controller is. The motherboard contains merely a PCI-to-ATA bridge.

Yes, yes, we all know that, [...]

Apparently not. M. Arno and M. Speed don't, it seems. It's
disappointing in the former case.

If you're going to be completely pedantic about it, and continue on with
the networking analogy, then "bridge" isn't necessarily the right term
either. In the case of a RAID arrangement, then it's acting more like
router. In the case of an ATAPI device, it's acting more like a gateway.

Yousuf Khan

 

[Jonathan de Boyne Pollard]

There is no disc controller on the motherboard with IDE drives.
That's what Integrated Drive Electronics means, after all. Turn
your disc unit over. See the chips and circuitry? That's where
the disc controller is. The motherboard contains merely a
PCI-to-ATA bridge.

Yes, yes, we all know that, [...]

Apparently not. M. Arno and M. Speed don't, it seems. It's
disappointing in the former case.

If you're going to be completely pedantic about it, and continue on
with the networking analogy, [...]

It's not a networking analogy. Things that connect computer buses are
known as bridges, bus bridges if there's scope for confusion with any
other sorts of bridges. There are many sorts, because many computer
buses have been connected together over the years, from Unibus to the
VESA Local Bus. In the world of PCI we have PCI-to-ISA bridges,
PCI-to-ATA bridges, PCI-to-PCI bridges, and of course the bridge between
PCI bus 0 and the processor bus.

In the case of an ATAPI device, [...]

In the case of an ATAPI device, the actual bus is still an ATA bus. The
bus itself isn't any different. (If it were, after all, one wouldn't be
able to correctly connect the device to the cable.) Nor is the
PCI-to-ATA bridge connecting it to the PCI bus any different. The
difference between ATA and ATAPI is in the disc unit itself, at the disc
unit command level and (to a lesser extent) the disc unit device
register level, not at the bus level. The bus is the same, and the
bridge is the same.

 

[Arno]

There is no disc controller on the motherboard with IDE drives.
That's what Integrated Drive Electronics means, after all. Turn your
disc unit over. See the chips and circuitry? That's where the disc
controller is. The motherboard contains merely a PCI-to-ATA bridge.

Yes, yes, we all know that, [...]

Apparently not. M. Arno and M. Speed don't, it seems. It's
disappointing in the former case.

If you're going to be completely pedantic about it, and continue on with
the networking analogy, then "bridge" isn't necessarily the right term
either. In the case of a RAID arrangement, then it's acting more like
router. In the case of an ATAPI device, it's acting more like a gateway.

Indeed. A bridge is ISO/OSI layer 2, which means same addressing
on both sides, but potentially different speeds, collision domains
or bus access arbitration. An important factor is here that both
connected networks can initiate transfers and are generally considered
to be on the same hierachical level. I agree on the router and gateway
analogy.

As addressing is different on PCI and (S)ATA, and the disks are
certainly not on the same hierachical level as the computer side,
the term "bridge" goes out the window.

In a hierachical situation the device is generally called a
"controller", since it has power over what is attached to it,
i.e. "controls" it.

With regard to ISO/OSI layers, "bridge" would also not qualify,
as due to DMA, NCQ, etc., the PC side disk controller is at
least on Layer 3, possibly higher.

Arno