Re: BUGFIX: Update Installation Compatibility Issues

Tech-Archive recommends: Repair Windows Errors & Optimize Windows Performance

I've now come up with a solution for people having problems
with Intel-chipset motherboards. Please see the bottom of this
message - as the response is posted in-line to show the entire
history of this thread's development.


Bill Drake wrote:
mabster wrote:
Bill Drake wrote:
Recently, there have been multiple reports of instability when
attempting to install the latest Windows Genuine Advantage
Update Tool (KB892130), along with similar problems with
some other updates as well.

Many of these problems can be traced back to Registry
Corruption - which prevents the updates from being properly
applied. There are also reports of updates not applying
because the files are not being properly written to the user's
hard disk - and consequently the update(s) cannot be
properly installed.

All the above is consistent with the consequences of an unstable
Hard Disk subsystem. Furthermore, because the Registry is by
far the most-commonly-accessed-and-written file on a Windows
machine - the Registry is one of the most-commonly-corrupted
files on *any* machine with unstable disk-drivers or unstable
disk-hardware.


Things to check:

1. Users with AMD-chipset motherboards - especially those
with nVidia nForce4 chipsets - are directed to download
and install updated motherboard BIOS and nVidia chipset
drivers *before* installing the latest updates.

Note: There is a problem with disk-corruption which occurs
when using older SATA/IDE drivers. This is a problem
regardless of the make-and-model of your Hard Disk.


There are also known-and-admitted problems with
some Maxtor SATA Hard Disks. Some Hitachi SATA
Hard Disks are also under suspicion.

Please check with Maxtor and/or Hitachi Technical
Support to determine if your particular hard drive's
firmware version is in need of update.


2. Users with SATA RAID0 or RAID1 arrays - on ALL chipsets
(even Intel Chipset Motherboards) are also at risk of disk-stalls
that can create corruption during Windows Updates. This is
especially a problem if the disk-stall occurs during the shutdown
process.

Again, Registry corruption is a common consequence of a
disk-stall during a software update - as the Registry is
always written to and updated as part of the Windows
Update procedure.

Similarly to Item1 above - a check for updated motherboard
drivers, RAID BIOS and hard-disk-firmware for your system
is highly recommended before performing the latest Windows
Updates.

Note: RAID0/1 stability is an ongoing problem that is not
thoroughly and completely researched. You will find
that in some circumstances - your equipment suppliers
have not yet got a handle on this situation. As a result,
they are going to try to make their screwup your fault.

DO NOT TOLERATE THIS.

Complain long and loud - especially in public places such
as these news groups - until this problem is properly
researched and fully resolved.



Hi Bill,

What about those of us who have already installed KB892130? Will
updating BIOS firmware/drivers fix the freezing that's occurring when
we attempt to visit a page requiring WGA validation?

See this thread: http://tinyurl.com/l2sd8

Matt

Hi, Matt. Here are the things I've found so far:

Observations:

1. Installation of Windows XP on an Asus P4P800SE with the latest
motherboard BIOS (1011), the latest Intel Chipset Drivers
(7.2.2.1006) and the latest Intel Matrix Storage Drivers
(5.5.0.1035) will *appear* to work properly when used with a pair
of Maxtor 6L250S0 SATA drives in RAID1 mode (Mirrored) installed
using WXP-SP2 with an F6-Floppy. My initial installs used the
factory-installed BANC1G10 firmware.

2. After the Installation of Norton Internet Security 2006 and Norton
SystemWorks 2006 Premier (includes Ghost), the array will not
work with Ghost if Norton System Doctor is running. Also, the
array will not Speed Disk - the array will hang at the start of the
"sorting" phase of a Speed Disk run. Both these problems are
reliably repeatable over multiple WXP-SP2 installations.

3. After I talked to Maxtor, They provided the BANC1G20 firmware
update. This was installed successfully - but it did NOT fix
either of the problems mentioned in Item2 on a retroactive basis.

4. Interestingly, doing a complete reformat and reinstall-from-scratch
with the BANC1G20 firmware in place allowed Speed Disk to work
properly. Obviously, something in the older firmware allowed a
very subtle disk-corruption to occur which manifested as the Speed
Disk hang.

And therefore, to answer your original question:

Yes, a complete reinstall-from-scratch *may* be required after
all the necessary driver/firmware updates are in place, in order
to clear out whatever subtle already-existing-disk-corruption is
causing the updates to fail to install properly.

5. I now have a set of Samsung drives to replace the Maxtors, and I
am experimenting with these to see if a simple change of drives
to a brand known for better compatibility will solve both problems
or not.

For more details on the loooooong history of my investigations of
this problem, see:

http://www.boardfish.com/forum/forum_posts.asp?TID=122&PN=1



Conclusions:

1. The Intel 875P and 865PE Chipsets, when used with Northwood P4
CPUs, are known to be the standard-of-reference for stability when
dealing with hyperthreaded CPUs and Dual-Channel memory.
If this hardware is cranky, the problem definitely rates
investigation and resolution. There is no excuse for problems
with hardware this mature.

2. For users with nVidia nF4-chipset motherboards, there is either a
bug in the nF4 chipset or a bug in earlier drivers that *also*
manifests as subtle disk-corruption. This is *above and beyond*
the obvious problem with startup-stalls that the Maxtor firmware
updates are designed to handle.

3. However, because of the issue mentioned in Item4 in the
Observation Section, I am now focusing on the interaction
between the SATA drivers and the SATA firmware in the
drives -- because I think Item4 above indicates there is a
problem with the SATA flow-control of Data I/O during regular
disk use *over and above* the problem with the startup-stalls
that Maxtor has already admitted.

4. I already have a case open with Microsoft about this issue, as
there is a KB article (906678) which may also have a bearing
on this problem. Since Ghost 9/10 is a DotNet 1.1 application,
this is definitely a possibility and warrants further investigation
with the proper hardware protocol analyzers. This is required
in order to ensure bus activity on the ICH5R is occurring reliably,
both between the ICH5R and the SATA drives *as well as*
between the ICH5R and the dedicated CPU I/O bus-channel
serving the ICH5R when run in enhanced SATA mode.

Note: Once the ICH5R has been verified-stable, it will be
an easy exercise to compare the bus traces of the
mature hardware with the newer nVidia nF4 and
Intel 900-series chipsets - to ensure the other
chipset-driver sets are working correctly as well.


Comments:

1. As usual when dealing with multi-vendor compatibility issues,
I'm getting *deathly silence* from all parties when I bring my
observations and conclusions to their attention.

2. Since it is obvious there is something seriously wrong, (and
this newsgroup is ample evidence of that), and there has
been no movement on this problem for several months, I am
no longer willing to hear lies, evasion and specious BS about
this issue.

3. I want this problem investigated and resolved. With the
demonstrated lack of responsibility inherent in the lack of
a coherent response to this problem - the only way to deal
with this issue is to publicize the problem widely and get
lots of people royally annoyed at the stupidity and
intransigence of the people responsible for yet another
debacle.



SUCCESS! I've localized the problem with the ICH5R Hard
Disk Controller chipset interaction and come up with a workable
configuration for Intel-chipset motherboards using the ICH5R
Hard Disk interface.

History:

1. There are two different version-sets of drivers available for
the ICH5R. The original set is Version 3.x - of which the
last available is Version 3.5.3.2847.

2. There is a later set of drivers which is *supposed* to
work with both the ICH5R and the newer ICH6R chipsets.
The driver is made available by Intel as usable for either
chipset. This is the Version 5.x set - of which the latest
available is Version 5.5.0.1035.


Observations:

As per my forum post at Boardfish, I tried both the latest
and one-previous versions of the 5.x-series driver. Both
driver versions exhibited identical problems, so I originally
focused in other areas.

After exhausting other avenues of exploration, I finally went
back and tried the Version 3.5.3.2847 drivers. This solved
the problem - both Ghost 9/10 and Speed Disk now work
stably and reliably.


Conclusions:

1. There are two steps to resolving this problem. The first
(and completely counter-intuitive) item is to use the *older*
Hard Disk Controller drivers - not the latest 'n greatest.

Note: This is *not* applicable to ICH6-chipset motherboards,
since the 3.x-series drivers are not documented as
compatible with the ICH6R chipset. At the moment,
this solution is only applicable to the ICH5R chipset.

2. Regardless of the application of the correct Hard Disk
Controller Driver set, it is still necessary to check for Hard
Disk firmware updates as well - and to update if newer
firmware is available from your Hard Disk manufacturer.

Note: The *firmware* update was the original solution to
the Speed Disk problem. Consequently, the evidence
shows that getting the flow-control working properly
for Disk-Subsystem I/O requires fixes in the drive itself.

3. Over and above the fix which occurs by regressing to the
Version 3.5.3.2847 driver - the ultimate fix for this problem
will require an update to the 5.x-series drivers to address
the improper support in the 5.x-series drivers for Volume
Shadow Copy services when run on ICH5R hardware.


Hope this helps others having grief with Hard Disk subsystem
instability.


Best I can do for now. <tm>


Bill



.

Relevant Pages

  • Re: Motherboard you recently bought for a new DAW? CPU and CPU fan?
    ... Maybe I don't know where to shop, but I find that motherboards are somewhat like mattresses - there are many, many choices, and it's rare that you can find the same one in two places. ... difficulty finding drivers unless it's new stock and you get the original installation disk. ... And if there have been updates since (motherboard hardware driver updates aren't nearly as frequent as outboard hardware updates) they may no longer be available. ...
    (rec.audio.pro)
  • Re: device drivers
    ... > does any one know how i can find out what drivers are in my system, ... Microsoft has these suggestions for Protecting your computer from the ... I'll mainly work around Windows XP, as that is what the bulk of this ... How to configure and use Automatic Updates in Windows XP ...
    (microsoft.public.windowsxp.general)
  • Re: 800703E7 error message
    ... I get this message when trying to install service pack 2. ... > Check for hardware driver updates? ... > Patches and Updates! ... > drivers for your hardware/operating system. ...
    (microsoft.public.windowsupdate)
  • Re: hard disk activity light flashes at regular intervals
    ... Do I have a software or hardware problem? ... - A need to optimize your system, cleanup - update drivers, patch it - etc. ... Microsoft has these suggestions for Protecting your computer from the ... How to configure and use Automatic Updates in Windows XP ...
    (microsoft.public.windowsxp.general)
  • Re: Service pack 2
    ... think of it this way: "I should have done the updates and cleaned ... What to Know Before You Download and Install Windows XP Service Pack 2 ... application, Disk copier, etc.) You'll be glad to know that if you have ... You also have hardware on your machine that requires drivers to interface ...
    (microsoft.public.windowsxp.newusers)