RogueSpear

Could imagine the system that had all of these threads running in the background like the java updater? Each of them taking upwards of 2MB a piece. I swear my desktop with 2GB of RAM would get swamped before long if I let all that crap run.

Removed.

You're starting to make me feel as though I haven't paying attention lol.. don't all 3 methods integrate the mass storage drivers in exactly the same way? While not the "inventor" of method 3, I believe RyanVM and some others had discussed the idea prior to me, I sort of brought it out into view and documented my procedure.. thus Method 3 was made. On the flip side of that, I also did extensive testing to show that method 3 was not a valid solution, due to the size the driver packs had grown to. I've since become a Method 2 holy roller. What I don't understand is this - what is the benefit to using DetachedProgram if you are already using presetup.cmd from Method 2? Want a good hybrid? On a DVD install, I placed all of the drivers on the DVD Method 1 style and call SetDevicePath.exe from DetachedProgram, eliminating the need for OemPnPDriversPath in WINNT.SIF. In fact with the addition of my own custom driver pack, this is necessary since I would otherwise overshoot the 4096 limit in WINNT.SIF.

Now if RealTek could only fix their piece of crap NICs and/or NIC drivers. A few months ago I instituted a new policy for myself. I do not purchase (for me or my clients) anything with any RealTek onboard. The epitome of substandard garbage hardware. The Walmart of computer componants.

Removed.

I had been planning to announce this at some point in the not too distant future, but some low life scumbag from Romania has accelerated the time table significantly by sucking up 800GB in two days of my bandwidth. All of the request referrals came from this thread - so this thread is now finished permanantly. Installers that had been posted here are currently available at my forums, but depending on what happens with my bandwidth they may not be available for a short period of time.

This sounds like Method 2 with the difference being that WatchDriverSigning is being executed from DetachedProgram. Or am I missing something here?

What conerns me and has me completely baffled is why the laptop does this during an install from CD also. At least that was the case earlier in this thread. I think that if you can't figure out the issue with an install failing from CD, you're not going to figure it out for RIS either. The two installation types really aren't as different as some might think, but there are a couple more things thrown in the mix with RIS. Therefore, it may be easier to solve if you use the less complex CD install method for troubleshooting.

I had given that some thought yes. I ended up modifying the .msi file, removing the custom action that prevents directly installing .msi file. But now I'm wondering if perhaps the install.exe could be used in conjunction with an AIP. By making the installer out of an AIP, you end up saving over 7MB of space. Personally I'd rather save the space and install it from RunOnceEx, than use the larger file from svcpack. But I'll try making a new installer tonight that uses Install.exe and see what happens from svcpack. For now I need to catch up on everything else I've let go while messing around with this thing.

Well after spending entirely too long on this thing, here is where I'm at: Installing from svcpack is giving me this error in the event log I've managed to work through some other errors, but this one has me stumped. For the time being, installing from RunOnceEx or GUIRunOnce is the best bet using Petya's original installer. If I manage to figure this thing out, I'll update. EDIT: Does anyone know if the Visual J# redist is included with .NET 2.0?

I actually had this work for me on a live PC. Guess I need to do a little testing here and figure out what's what. EDIT: yep, it's pooched. What has me scratching my head is this - under what circumstances did this install properly for me? I wouldn't have posted it if it didn't work for me. I'll d/l the file again and see where I went wrong. Hopefully I'll have a replacement up before too long.

V1.03 - 10/30/2005 - Add custom user .SIF files to the \Templates directory. I have included a sample .SIF file which will be copied by AutoRIS. Users can either edit this file to suit their needs or replace it entirely with one or more of their own .SIF files. The default .SIF file, RISTNDRD.SIF, is deleted by default. - Drivers for Intel Pro, Marvel Yukon, and RealTek are now included with AutoRIS. The .SYS files are not compressed in the distribution in order to achieve greater compression of the AutoRIS download file. The .SYS files are compressed during execution of AutoRIS. - I have a new Addon Pack for use with RyanVM's Update Pack. I had contemplated including it directly in the AutoRIS download, but opted against that as the file size would have been a bit much for something not everyone will use. You can read more about it and download it over at RyanVM's Site.

Ok, Intel are better than AMD. I have read that piece you quoted before btw. My final statement (I mean it this time too): If you don't work in the field for a living and you simply get your information and base your opinions on what you read then you have no idea what you're talking about. Also I will reiterate that I do not care about benchmarks. If you're a benchmark whore, keep right on following Anandtech. Neophytes and hobbyists would get their butts handed to them on a plate if they ever had to implement server solutions in a corporate environment. I'm going to be unsubscribing from this thread immediately. So you all can continue to diddle yourselves over your AMD CPUs. I never realized until this thread, just how many people there were here that lacked so many critical thinking and reading comprehension skills. Not too mention how many people root for a particular brand as if it's a sports team. It's really pathetic. Get a job.

What's even more annoying is people talking about things as though they're an authority when they don't even work in the field. You apparantly didn't catch my posts where I laud AMD for their technical achievements. No, that wouldn't fit your agenda. Oh and the avatar is meant to show you what you'd see if you were to poke a hole in your belly button and look in a mirror. Glad you noticed

Yes I am telling you that (paragraph 1). What logic did you present? Typical. Your opinions seem to materialize from nowhere. I actually work with this stuff everyday, not just read about it.

My analysis is: 1. You're talking out of your a** again.

Hey if I could get the same rock solid reliability from a HP-Compaq ProLiant DL server with an AMD processor instead of Intel, fine by me. Just not willing to take that risk yet. I'll wait for some more reports from the field first. My days of risk taking on hardware are over.

AMD may be fine for your gaming rig or a home machine, but anyone worth their salt wouldn't even consider that for a database server. It would be malpractice. I will admit that AMD has some pretty compelling technology and yes they do win benchmarks, but benchmarks do not equal reliability, stability, compatability, etc. that enterprises are most concerned with. I'd be the first one to jump on the AMD train if they would do something about those issues. AMD focused almost exclusively on speed and to that end they have succeeded. Now they need to take a look at some of the other factors that make up a long term reliable CPU. I've been burned enough and personally witnessed enough other people suffer through disasters at the hands of system with an AMD CPU to make me think very very hard before ever trusting them again.

I barely contain my laughter here.. a high end database server running on an AMD rig. That's kind of like George W. Bush being the president of the United States. We've seen how well that runs.

If you're familiar with the old protocol NetBUI (sp?) you can accomplish this. It should be within the TOOLS directory on your Windows XP CD as it's not included in an XP install. Microsoft no longer supports the protocol and really there's very little use for it anymore. It was extremely chatty, heavy on the broadcasts, etc, however it is a protocol that does not route so the security implications are somewhat limited.

As long as you delete the registry entry for error reporting (HKLM\Software\Microsoft\Windows\PC Health) from the .msi file, the resultant switchless installer will install from svcpack.inf just fine. This is the same action I had to take in order to get Office 2003, and the latest refresh of Windows Messenger to install. I took the package offered by Petya V4sechkin (excellent work by the way), decompressed it, made my mods to .msi file, recompressed, etc. As a side note, I've switched from using StartX.exe to using RunHiddenConsole.exe in my switchless installers and it works great, less filling. Much less in fact. dotNET20.exe

If you have so many computers that you need a script to do this, then that means you really should be running a DHCP server and using reservations. I know that reservations defeat half the purpose of DHCP, but still it's a hell of a lot easier to change a DNS server assignment on 100 workstations via DHCP than to physically visit each one (or even to script that out too).

Once you get the RIS basics down, and it won't take you long believe me, check out my RIS guide for how to better optimize your RIS deployments. What everyone in here is saying about RIS is true too, it's really fantastic. It's like a religion, once you find it, you never go back. I think that tools like RISult and AutoRIS should only be used after you have a solid grasp on what's going on. As good as RIS is, the underpinnings can be complex. You'll find this out when something doesn't quite work right. And you'll need the know how to be able to troubleshoot and fix the problem. Once you're a RIS master though, both of these tools can save you countless hours.

I just started my experimenting around with Nero 7 last night so I don't have everything figured out yet, but so far I found that you can strip out a lot of stuff using InstallShield AdminStudio. Then when you make a AIP, all the stripped out stuff is gone. It doesn't save nearly the amount of space I had thought that it would, but it did save some. I was able to yank out all the languages I don't use and all of the Windows 9x support files. In the end though I think the best way may end up being a complete repack using the old two step snapshot. This is how I had always done it with Nero 6, and with great success.

I wish I had been as diligent as orei in writing down the filename and the directory in which I found the file, but it is definately the same filename that was missing for me as for him. And the symptoms are identical to his. It was certainly one of the stangest things I had encountered, but at least it all worked in the end.