nmX.Memnoch

Windows User State Migration Tool 3.0.1 (Documentation)

*SHOCK* I didn't say "inexpensive". I said "cheap", in this case meaning "of lesser quality". I certainly wouldn't trust my high-dollar equipment to a cheap power supply. Keep in mind that if a PSU goes it has the potential of taking everything else with it. I don't know why people skimp on arguably the most important piece of hardware in the computer.BTW, I knew we'd already had this discussion before... http://www.msfn.org/board/index.php?showtopic=98192

You want my honest opinion? Rosewill power supplies are cheap for a reason. It may be able to handle up to 600W draw, but how many amps can it output and how many 12V rails does it have? You need a better power supply anyway. Oh...and your computer "teacher" is wrong as well. Having a faster video card isn't going to damage your CPU, RAM, motherboard or anything else. I do have some serious concerns about that power supply being able to supply enough juice to run everything (not that 600W isn't enough, but that it's not a quality power supply).

I'm not sure if this will work but you might be able to copy the "Win9x" directory from the WinMe CD to your XP Unattended media. It may not bypass the message, but it might bypass having to swap the disk. I think all it does is checks for certain files in that directory.

I missed that. For that matter, I still don't see it mentioned before the last post. *shrug*Anyway, RIS/WDS can be used in that exact method. You put the disk it, the machine then boots to the network and pulls the image from the RIS/WDS server.

That why you have a support package with the hoster. They are your "hands and eyes" on-site. Let me ask this question...is he looking for hosting or co-location? There's a difference. "Hosting" would be where he rents a dedicated server from the hoster. They own the server and are responsible for repairing it. "Co-location" would be where he owns the box and is putting it in their rack on their network. In that case he's responsible for the hardware because he owns it. This is also usually much more expensive. When I was using the planet I wasn't using RAID, but the server did have two drives in it. I had custom scripts that ran four times a day and backed up everything from drive 0 to drive 1, and everything from drive 1 to drive 0. That way if one of them failed I was assured that I had something that was at least remotely recent. It worked out very well (we had to have the OS drive reinstalled once...and didn't lose anything).

Doing a RIS/WDS setup won't cost you anything but time...providing you already have a 2000/2003 server (and a decent network). RIS is built into 2000/2003/2003SP1 and WDS is the upgrade to RIS that comes with SP2 for 2003.

I used to use The Planet. Their service was really good and they're connected to all of the major backbones. The only reason I don't use them anymore is I don't have the need now. I'm not sure if they still are, but HardOCP used to be hosted at The Planet. Bjorn3D is hosted there as well. The Inquirer is (or was), too. If you're using dynamic pages (PHP, ASP, ColdFusion, etc) you will most definitely see a difference when using a robust disk sub-system. Besides, given the option in a production server you should always use RAID (but never RAID0).

He shot down his own method with that statement. Which would you rather do...update 1 image or update 20-30 images? I'm pretty sure the answer is a no brainer. Have y'all thought about the possibility of using RIS or WDS? Short of that sysprep is best option. Besides, for some of the hardware you can write custom scripts to detect if that hardware is installed (because the basic driver would already be there) and then silently install the "required" applications for those pieces of hardware. I had a script around here somewhere that I used to detect if the machine had an Intel NIC...if it did it used an Intel command line utility to set the link speed of the NIC. EDIT: I found the script. This is the section of the script that we used to detect what kind of NIC was installed. Note the lines that install Intel PROSet if it detects certain Intel NICs. The script language used is KiXtart (my fav but requires a 3rd party parser...which can be included in your media). Also note that this code is a few years old...my coding style has changed quite a bit since then, but there's not too many different ways of doing what's done in this script. $objWMIService = GetObject("winmgmts:\\@WKSTA\root\cimv2") $colItems = $objWMIService.ExecQuery("Select * from Win32_NetworkAdapterConfiguration",,48) For Each $objItem in $colItems Select Case InStr($objItem.Caption,"Intel") And InStr($objItem.Caption,"1000") ? "Intel Gigabit Network Adapter Detected" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b If Exist ("C:\Program Files\Intel\PROSetWired\NCS\WMIProv\") = 1 ? "Intel PROSet already installed" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b Else ? "Installing Intel PROSet for Wired Connections" Shell "%COMSPEC% /C START /WAIT C:\Install\IntelPROSet\PROSet.msi /QB" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b EndIf Sleep 2 ? "Setting network connection to 100Mbs Full Duplex" ? Shell "%COMSPEC% /C cscript C:\Install\IntelPROSet\fsdset.vbs 4" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b Case InStr($objItem.Caption,"Intel") And Not InStr($objItem.Caption,"Wireless") ? "Intel Network Adapter Detected" If Exist ("C:\Program Files\Intel\PROSetWired\NCS\WMIProv\") = 1 ? "Intel PROSet already installed" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b Else ? "Installing Intel PROSet for Wired Connections" Shell "%COMSPEC% /C START /WAIT C:\Install\IntelPROSet\PROSet.msi /QB" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b EndIf Sleep 2 ? "Setting network connection to 100Mbs Full Duplex" ? Shell "%COMSPEC% /C cscript C:\Install\IntelPROSet\fsdset.vbs 4" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b Case InStr($objItem.Caption,"3Com") ? "3Com Network Adapter Detected" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b Case InStr($objItem.Caption,"Broadcom") And InStr($objItem.Caption,"Gigabit") ? "Broadcom Gigabit Network Adapter Detected" Color y+/b ? "-------------------------------------------------------------------------------" Color w+/b EndSelect Next

You could write a script and use Log Parser 2.2 to pull the data from System Event Log. The main Event ID you're looking for is 10.

Are you sure he's not running XP x64? Also, the second quote in my post above isn't correct regarding physical memory...well, it is but that quote specifically talks about page address space (virtual memory) and not physical memory.

Some parts are still made stateside. Both Intel and AMD (both American companies) have fabs in the states, but the final product is assembled in other countries.

He had to have done something. XP Pro 32-bit doesn't natively support more than 4GB RAM. http://www.microsoft.com/whdc/system/platf...PAE/PAEmem.mspx

Not sure where you read that at but I've never heard that before. Again, I would say the opposite is true. A drive formatted as a single partition with a good defrag program will always perform better than a drive split into multiple partitions.

I addressed that part in my previous reply. See the quote below. Well, to make it mostly transpart will take some work on your part. As far as I know there's no way to make it 100% transparent. There will be some minor changes for them, but it shouldn't be anything complex or complicated. For instance, you'll have to create their domain account, set a temporary password, and flag the account to "Change password at next logon". Give them the temporary password and when they login the first time they'll be required to set a new password. Other than that, the USMT is what you'll use to copy their user settings into the new user profile (and by that, I mean the structure under the "Documents and Settings" folder on each machine).As for adding the share to each user's logon, you can do that several ways. You can either setup a basic logon script to map the share to a drive letter each time they logon, or you can manually map the share one time with a persistent map. Personally, I recommend just using a logon script so you can automatically add new shares as necessary. One step at a time though.

You are correct in that XP 32-bit can only "see/use" up to 4GB of RAM. The system properties will show anywhere from 3.25GB to 3.75GB depending on the number of PCI cards and the amount of RAM on the video card (if it's a dedicated game server I would hope that you're just using some low-end card). To see/use all 8GB you'd have to use one of the following: Windows XP Professional x64 Windows 2000 Advanced Server with PAE enabled Windows 2000 Datacenter Server with PAE enabled Windows Server 2003, Enterprise Edition with PAE enabled Windows Server 2003, Datacenter Edition with PAE enabled Windows Server 2003, Standard x64 Edition Windows Server 2003, Enterprise x64 Edition Windows Server 2003, Datacenter x64 Edition Windows Server 2003 R2, Enterprise Edition with PAE enabled Windows Server 2003 R2, Datacenter Edition with PAE enabled Windows Server 2003 R2, Standard x64 Edition Windows Server 2003 R2, Enterprise x64 Edition Windows Server 2003 R2, Datacenter x64 Edition The Service Pack level doesn't matter (although you should always use the latest and install all hotfixes). See this page for information on Page Address Extensions (PAE). It's basically a band-aid fix for 32-bit operating systems. To effectively use more than 4GB you need to make sure to go with one of the x64 versions. Not true if you're hosting multiple games at the same time, especially with some of the newer games. As I understand it, Crysis can be pretty intense on the server side. Besides, I've said it before and I'll say it again, I would much rather have more than enough power (extra CPU cycles in reserve) than not have enough to do the task.

You're headed in the right direction with some of your thoughts about AD but you are missing a few steps. First you'd have to make the Server 2003 box a domain controller. That's a fairly simple process and since you only have the one server and handful of workstations there shouldn't be anything overly complicated in the network design. Definitely do a bit more reading before you jump into setting up AD though. While it isn't difficult, it can be a bit daunting the first time or two. Be prepared to spend a little time configuring the server. You want to make absolutely sure that's right before you start adding computer and user accounts to the domain. As for adding the computers to the domain, you're sort of correct. What you'd do is go to My Computer Properties (right click My Computer and choose Properties), then go to the Computer Name tab. From there you should see a Change button. Click that button. At the bottom you'll see two options under Member of; Domain and Workgroup. Right now you're on Workgroup. But after you setup AD you'll want to change that to Domain. In the Domain box, type either then NetBIOS name (i.e. domain) or the full domain name (i.e. domain.local). It'll ask you for a username/password that has authentication to join computers to the domain (the Administrator account or another account that has been given Domain Admin privileges). After you've done that, you'll need to create user accounts in Active Directory using the Active Directory Users and Computers snapin. The User State Migration Tool will come into play when you get ready to have your users start logging onto the domain instead of with their current local accounts. It should allow you to copy their local account profiles to the new domain profile (on the workstation) so that it looks fairly transparent to them. Honestly, I know all of the details on USMT since I don't have much first-hand experience with it...everything I've done since it was released has been straight to a domain model instead of a workgroup model. But...definitely keep reading and try to plan out as much as you can. There are plenty of smart people here who can help with recommendations on how you should or shouldn't setup your domain. Personally, I am of the opinion that you should definitely be looking at a domain model since you already have what's necessary to start.

Definitely give it a look. And yes, they have trial versions available. You shouldn't need anything higher than the Professional version. Once you get a server going at the cafe you may also want to look into purchase a license for the Admin Edition. It'll let you control settings on all of the machines remotely, plus it has reports so you can monitor the fragmentation status of each machine. A lot of people tend to get confused with NTFS permissions. It's easy to do at first. The main thing you need to make sure of is that the folder isn't inheriting permissions from the parent folder. Once you turn that setting off you can then customize your permissions and enforce them down the tree on all sub-folders below the restricted one (any new folders will automatically inherit). CACLS is very limited so you definitely want to look at either XCACLS or SubInACL. Careful with SubInACL though...it's got some crazy syntax rules. I know for a fact that you don't need a server just to run SiteKiosk. Now if you get into using the actual SiteCafe features then you might. It's definitely made to help manage and enforce a customer's usage time on the cafe machines though. Glad we got that one straightened out.

Hehe...welcome to XP SP2 and Server 2003 have the Messenger service disabled by default. Not to mention that it would be blocked by a cable/DSL router anyway.

For "Everyone" to work you'd have to enable Guest access...which is something that's (rightly) disabled by default in Server 2003. Using "Everyone" is a big "no no" and it's what used to be people into security trouble back in the day (when it was the default NTFS permission under NT 3.51/4.0). Why not just finish the setup and start yourself an Active Directory? It's not that difficult...you already have the server running and it probably wouldn't take much more planning to configure AD, join all the machines and create user accounts for everyone. You can use the User State Migration Tool to migrate their local profile to their domain profile. Well...before we get too much into that...how many machines and users are we talking about?

I know drives are cheap, an 80GB will cost me about 40 US$ + TAX here, but I'm starting my business in a few weeks and cash reserves are low now. Time to restore my System drive now will be about 3 minutes vs. the complete drive, including 200-220GB of games, will take me 50 minutes and that's unacceptable when the internet/cybercafés are full. People will see an empty PC and I need to tell them I don't have space, so they walk away. Believe me, I have been there . Most games don't actually require a reinstall. For example, most id Software games can just be copied off...some of the newer ones may require importing a registry value for the CD key. In your case you wouldn't even have to do that since you're imaging the OS drive and keep the games on the second drive. Just restore the image and all application install settings would be retained. I could throw in more ram and keep the page file smaller too. in this case the page file is stored at the beginning of partition too, so only the head has to skip the first 10GB, not that much. If it were high-performance workstations or server I would do like you suggest. By the way, additional wear isn't that much bigger, a fragmented drive would be worse and you know it. More RAM isn't going to solve anything unless the machine is running out of memory. Do not confuse physical memory with virtual memory. Windows does use virtual memory regarless of how much physical memory you have (cluberti...your turn to chime in ).And of course a fragmented drive is worse. But if you aren't running any sort of defrag software all you've done with multiple partitions is compound the problem. Diskeeper 2008...install it and forget it, literally. It defrags files and folders in real time. It will even keep the MFT from fragmenting. No it's not easier at all. Folders can be locked down very easily. That's what NTFS permissions are for and there are several tools available to do that. You can either:1. Do it with the GUI options available on a folders properties under the Security tab 2. Do it using the command line using the built in CACLS command 3. Get the extended version of CACLS (XCACLS) 4. Use SubInACL, yet another Microsoft tool with even more functionality than CACLS or XCACLS How do you think server shares are secured? I certainly don't create a new partition every time I need to create a new share on one of my servers. Also, if you really want to keep the machine clean and locked down you should look into using the something like either Windows SteadyState (the latest version of teh Shared Computer Toolkit...handbook link...technical FAQ link) or something like SiteKiosk, which was designed specifically for internet/cyber cafes. I don't think it was highjacked...I think we just got further into discussions about partitions is all.

Just some side information, but the new licences that will come out for XP and Vista will have no "key", they generate there hardware key but you need to activate your windows. I could be wrong, just let me know if I am .As cluberti said, they still use a key...you just don't put it in during install. Instead you enter the key at the time of Activation. OK!!

My guess would be either, as puntoMX said, a RAM compatibility issue...or a power supply compatibility issue.

Trust me, there is no such thing as a Site License for Windows. I do work in IT...for a very large organization (the US Air Force). On the base I work on we have well over 15K machines...and that's just at one base. It's not called a "site license", it's called a "Volume License" and that is what people are calling the "Corporate Edition". Under the Volume License licensing you are still required to purchase a seperate license for every single workstation, laptop, tablet, etc that you are going to install Windows on. http://download.microsoft.com/download/f/4...99_PROC_WLG.PDF As you will note in the PDF, Volume Licenses do not cover the initial Windows license (i.e. you still have to purchase a full OEM or Retail license). Generally, companies get that initial license with the purchase of new PCs. What the Volume License provides for is upgrades and Software Assurance. It also provides for the use of a single install key that doesn't require activation. Speaking of which, I guarantee that the install you are using at work requires a product key. There is no version of Windows that doesn't require one. However, you can integrate the product key into the media (and Volume keys don't require activation). Take a look at your install media you use at work. In the i386 directory that should be a file called WINNT.SIF. In that file will be contained settings that Windows Setup uses, including the product key.

You do understand what I mean by partitions hurt performance and create unnecessary wear on the drive right? I would much rather have a machine that runs faster while I'm using it than worrying about the size of or how quickly I can restore a drive image (15 vs. 30 minutes restore?). Not to mention the fact that I much prefer clean installs over a restore any day. That's why all of my data is stored on a seperate drive from my OS/apps and I keep backups of just the stuff I would restore (Favorites mainly...and if I lose those it's no big deal). I would also much rather my drives last longer than generating unnecessary wear. punto, in your case...seriously, drives are cheap. Put a second drive in those public access system and drop the partition(s). Your drive image should only be as big as the data that's on the drive....not the size of the entire partition. If Acronis creates images that way then I'd look for another drive image app. Also, putting the page file on a second partition of the same drive will hurt performance even further. To open an application the drive has to access both the executable and the page file at the same time, which makes the drive heads move all over the place. With a second partition the drive heads have to move further, which creates both latency and additional wear.