Mr Snrub

That is why I specifically stated a browser (HTTP) proxy, and only for that specific name by using a PAC or equivalent - this would not affect the domain functions of the clients, but is ultimately a hack rather than a solution.

I would check that the SATA disk is visible natively by XP x64 first, otherwise it might fail to boot due to a missing driver.Boot from the XP x64 CD when you get the SATA disk connected, and see if it appears in the list of disks at the point where you choose to install the OS - if it can then you should be good to go, clone the disk and make the BIOS change to alter the boot sequence to check the SATA disk before the PATA one. The first time it boots up there might be some switching around as the SATA controller and its devices are re-evaluated, and disk signatures are updated in the registry.

Definitely a problem created by nLite, or a bad burn. This is the reason I don't use 3rd party tools to make slipstreamed or unattended install discs, and prefer to do it by hand (and test using virtual machines and ISOs rather than physical media).

The biggest challenge with home networks for XP users is usually getting to see anything in the first place ("simple file sharing" being enabled by default but people wanting to expressly share specific folders rather than a "Shared Folders" entry in their profile). Given that you can see the shared folders and access them without an authentication prompt (or error), then I don't think your issue is with the guest account or network share permissions. How many objects are present in any given folder? If more than 1,000 then you can run into prolbems with partial lists due to a legacy issue with 8.3 filenames. If you map a drive letter to a share, then open a command prompt and switch to that drive and enter "DIR", does it list all of the objects, or is this list also truncated?

This should have been taken into account when designing the AD and DNS infrastructure, unfortunately the DNS servers have to be authoritative for the domains you have defined - this appears to be an inconsistency in DNS if you check multiple DCs they report themselves as the authority, but it is meant to be like that. This is why many people recommend using corp.mycompany.com for internal (AD) use and not overlapping public DNS spaces. I guess one workaround would be to define a proxy server for your browsers which is only used for http://mycompany.com, and direct connectivity is made for other requests. If the proxy server is a member of the domain as well, then you might be able to hack a HOSTS entry for mycompany.com to point to the public IP, but this could make domain membership weird for this server - easier to not have it in the domain at all and resolve names via public DNS. But this is a hack to workaround a flaw in your AD design unfortunately.

Other than the ability to uninstall SP2 and the associated disk space for the backup of the SP1 versions of the files replaced, there should be no difference in functionality. In the brief time your system is at SP1, the firewall won't be activated by default and the Messenger service will still be enabled, so there is a difference before you install SP2.

http://www.msfn.org/board/index.php?showtopic=51542

Thermal compound I've heard only a couple of people have issues with, from applying too much paste - but this would imply a heat issue anyway so something that Motherboard Monitor Lite should show you. Did you use the fan that came with the CPU, or the ICE heatpipe from the Shuttle? I would take the following troubleshooting approach: - verify the fans in the system are operational - install MBM Lite and configure it to show: > fan speeds > temperatures > voltages If the problem does not appear to be rising temperatures, failing fans, or fluctuating power on the rails then I would start stripping the system to the bare minimum and see if the problem is still present. Take out the wifi card, 1 of the DIMMs and the DVD-RW drive - if the problem persists still, swap the DIMM with the one you removed. This will let you rule out (or identify) one of the components if the system is stable or not.

Sure there are built-in ones, but how useable is the "Standard VGA Controller" driver which Windows falls back to for compatibility?It's not fair to compare Vista 32-bit with Aero Glass in all its glory with Vista 64-bit having to use the unoptimized and poorly-performing compatible driver on the same hardware. I have a system with an Radeon x800 and ATI have at least provided drivers that work with both 32-bit and 64-bit, but I've not used this other system much for testing.

The system is bugchecking and auto-restarting - this was apparent from the SaveDump event reported earlier.There are also many mindumps being created and they are not consistent with their STOP error codes, which usually indicates a hardware problem, here are the last few dump analysis summaries: Driver faults are usually consistent in the bugchecks, creating 1 or 2 different STOP codes - but even the 2 0xD1 bugchecks above are not consistent in their stack traces. Integrated GPU, NIC & audio and high quality PSUs in Shuttles usually mean there isn't a problem providing enough juice - I ran an SN41G2 system with a 9800 Pro AGP card, a 1Gbps NIC, a DVD-RW and 2x HDDs without a problem on the stock PSU.From what I understand, the problem with cheap PSUs isn't their maximum wattage but their ability to provide a "clean" input without loads of fluctuation. Heat, however, can be an issue with Shuttles - though poorly mounted CPU heatsinks usually mean the system refuses to even present the POST screen (I have had this a couple of times). @IsLNdbOi: What other hardware is installed inside the Shuttle? What do you have for RAM (amount, number of DIMMs & type), CPU, HDD(s) & CD/DVD? Anything plugged in the PCI slot? I suspect faulty RAM, CPU or mainboard - but I would install and run Motherboard Monitor Lite and log all temperature sensors and fan speeds - it could be a chipset fan failure or something. Check the fans are also clear of dust and check the fans all spin up correctly - chipset(s) & ICE.

It isn't achievable, reliable virus detection has to work off signatures - known patterns of strings that occur in sequences within a file.This makes it a reactive process: the virus has to appear, be noticed, have a signature developed and available for download, then the users download the updates - this is the window where the virus is out in the wild and potentially dangerous. There are "heuristic" ways of scanning files to see what "might" be a virus, but these often don't even successfully spot variants of existing viruses so I don't rate those highly. There's no "silver bullet" for malware - the method of deployment, infection and function is down to the whim of the guy that programmed it, as well as the target for their malicious code. A common argument is that the more competition that exists in a market, the higher the quality of the products as the companies have to outshine one another - this has been thrown around aimed at Microsoft as people perceive monopolistic practises, but it has to swing both ways. I do agree with you that the quality of the offering from Microsoft is likely to be better than 3rd party ones (in terms of UI and engine) as their programmers have a lot more experience with kernel mode programming, writing for multiple platforms, making code efficient and interacting with the OS in supported ways. The AV vendors have years of experience specifically with virus detection and signatures, so may have a head start in that area - however with technologies like BITS, Microsoft may have an update process that is more lean, efficient and has less of an impact.

The checks from even 100 such apps done daily would not have an impact on general system performance - the information exchange required is so small it is over very quickly.The only time the user is likely to notice any result of an update is when the app actually did have something to download and install - and then it's disk access with a little CPU time thrown in, once the download is complete. Separate update agents, such as the Java Update Scheduler jusched.exe, insert themselves into the user logon process and so can make that one-time process a little longer than normal, but have zero impact on actual performance during the logon session. Most apps tend to do update checks on being launched, or with every N launches, rather than have a separate agent.

STOP 0xc0000221 KB article (officially listed as XP, but I would imagine the same applied for 2000)

I think only Administrators can run defrag, the documented workaround is to use runas: http://support.microsoft.com/?id=231176

Unfortunately nVidia haven't produced 64-bit Vista drivers suitable for the 7800GT so I can't test it in depth, but I did have the 32-bit and 64-bit versions on my system a few builds ago.They performed pretty much equally, but we are still a long way from any performance enhancing yet so I don't expect there to be a distinction between them. The bugcheck screen in Vista is still blue, by the way. Once Vista is launched I'll most likely be ready to replace my XP x64 installation with 64-bit Vista on my 2 main rigs, with a virtual machine running 32-bit XP for any legacy apps or IE plugins when I need them.

Check your PMs, alternatively just zip the last 3 dumps and attach them.

This is unlikely in most cases - network connections are rarely a resource in contention, and the frequency & impact of update checks for applications is invariably tuned to be low.It doesn't do software vendors any good to have their products check for updates every minute as this would just overload their servers with pointless requests - a balance is struck as to how often updates are likely to come out, and how critical it is to get clients updated in a timely fashion. Doing a weekly check for AV updates would be silly, as would doing hourly updates for Adobe Acrobat Reader or something similar. If an application is launched which does an update check, it should not have any impact on the system as a whole, possibly not even on the application itself if it is a separate thread. It takes a lot for a system to have a performance hit due to extreme network traffic, and servers are much more likely to notice a hit before clients and they would be receiving and trying to handle requests rather than issuing them.

The read-only flag set on a folder is normal, it's actually used by the system for internal purposes - as you have pointed out the folder itself cannot be read-only.The GUI provides the ability to set or clear the flag so that you can can change the flag on the child objects under it in one action, instead of having to recurse down the folder structure manually. Test: If you try to clear the read-only flag on a folder, but select "Apply changes to this folder only", do you get the access denied message? If YES, then go to the Security tab in the folder properties and click Advanced, then click the Owner tab on the property sheet which appears - who is listed as the "Current owner of this item"? If NO, then check the owner of each of the files inside the folder in the same way - can you also change the read-only flag on a per-file basis without an error?

Kernel dump will be way to big to attach here - attach the minidumps that are already produced.If I need a kernel dump later we can sort it out separately.

IRQL_NOT_LESS_OR_EQUAL - most likely a buggy 3rd party driver.The system is only configured to produce minidumps so there's not much info, but if you zip & attach the dump file(s) mentioned I'll take a look. You might want to consider setting the machine to produce kernel memory dumps: - Right-click My Computer, click Properties - Click Advanced tab - Click Settings button under "Startup and Recovery" - Under "Write debugging information", select "Kernel memory dump" This will produce a single MEMORY.DMP file in your %systemroot% folder - C:\Windows by default on XP - instead of a new Minixxxxxx-xx.dmp file every time it restarts. Make sure your page file is at least as big as the physical memory installed in the system too.

"Incorporate some of the OS architecture" is going a bit far - anyone that understands how device and filter drivers work can make a firewall or AV product (and at the same time realise there is always a way to circumvent it if a user with administrative privileges launches a malicious process).Consumption of memory is not as critical as having efficient and stable code - people get a bit hung up on trying to reduce the memory footprint of processes when it can have little or no effect in system performance. Any "intelligent" software will eventually make a mistake in a decision it makes, changing the users' perception of it being a fantastic feature to a liability - I for one would not want a program to determine for itself what is safe and unsafe, especially regarding trusted sites/zones. This is what kicked off the anti-trust case against Microsoft - the alleged "bundling" of software that is not essential to the OS which puts vendors of similar products at a disadvantage.(Only after proving IE provides key rendering services to the OS did it get left incorporated in Windows - the EU forced Microsoft to produce a version of Windows without Media Player so people had the choice to not have it - though I know of noone who opted for this choice.) AV (with free updates), a fully-featured personal firewall and a word processing package bundled with Windows can never happen due to legal & business practice issues. Hence why Microsoft can make separate products and sell them so as to not have an unfair advantage over the vendors of competing products. Even incentives are frowned upon - e.g. "Buy Product X and get Product Y for $1!", so you can't argue that MS Office should be cheaper for people buying Windows at the same time - it's not allowed by law.

You don't know what multi-homed means, but you have installed VMWare on this machine and the host is trying to obtain DHCP settings from one of its guests? Okay, uninstall IPv6 as you will not need it. You can't route anywhere as you don't have a default gateway on the network adapter, and did you join the host to a domain running on a server in a guest OS or something? (DNS suffix is "domain.actdsltmp") Your VMWare (possibly NAT?) adapter appears to have a conflicting subnet with your physical network. 192.168.0.1 must be your router, connected to the LAN that "NETWORK 2" is connect to, but when VMWare is installed it is screwing up the configuration by putting the IP of your router on the virtual interface. Multi-homed means the machine has multiple network interfaces (physical or virtual). Your machine is multi-homed with 2 physical network adapters (192.168.0.0/24 and 10.10.10.0/24) and with 2 virtual network adapters provided by VMWare (192.168.0.0/24 and 192.168.40.1). The strange-looking IP addresses are IPv6 which you will not need. A gateway is the device which network traffic is sent to when the destination is not on your local networks. Having no default gateway means you can't route. Having more than one gateway can cause routing problems or delays as the computer will just the use first it "sees" even if it isn't right (unless you set up static routes). Leave VMWare out of the equation until you understand more about networks & IP routing, then reconfigure "VMware Virtual Ethernet Adapter for VMnet8" so it doesn't try to use the same subnet as your physical network adapter.

Is this machine multi-homed? If that is output from "route print" on the XP x64 machine then it seems to imply that it has 2 interfaces (or at least 2 bound IP addresses) on the 192.168.0.0/24 network... I agree with Gouki that we need either a diagram of your network to understand the layout, and ideally the complete output from: ipconfig /all route print

Did you stop/disable the "DNS Client" service?

Title is slightly out of context, the article states: I consider their test flawed in that they are purposely using unpatched versions of the browsers, and I wonder if they are also logged on with a user with admin rights... what was it meant to "prove"? I would guess that the most significant proportion of malware infections comes from visiting warez, hacking/cracking and "free" porn sites.