A little while back I wrote a blog about auto-elevating batch-files. Recently I found that Keith Hill has written something similar for PowerShell, which is nowadays the thing to use.....
Anyways, here's a repeat of Keith Hill's work, with all credits to Keith of cource.
See http://rkeithhill.wordpress.com/2013/04/05/powershell-script-that-relaunches-as-admin/ for more information regarding the code below.
function IsAdministrator
{
$Identity = [System.Security.Principal.WindowsIdentity]::GetCurrent()
$Principal = New-Object System.Security.Principal.WindowsPrincipal($Identity)
$Principal.IsInRole([System.Security.Principal.WindowsBuiltInRole]::Administrator)
}
function IsUacEnabled
{
(Get-ItemProperty HKLM:\Software\Microsoft\Windows\CurrentVersion\Policies\System).EnableLua -ne 0
}
#
# Main script
#
if (!(IsAdministrator))
{
if (IsUacEnabled)
{
[string[]]$argList = @('-NoProfile', '-NoExit', '-File', $MyInvocation.MyCommand.Path)
$argList += $MyInvocation.BoundParameters.GetEnumerator() | Foreach {"-$($_.Key)", "$($_.Value)"}
$argList += $MyInvocation.UnboundArguments
Start-Process PowerShell.exe -Verb Runas -WorkingDirectory $pwd -ArgumentList $argList
return
}
else
{
throw "You must be administrator to run this script"
}
}
Tuesday, September 24, 2013
Thursday, December 15, 2011
Dynamically creating lambda expressions in .Net (C#)
Dynamically creating lambda expressions in .Net (C#)
Passing lambda expressions as strings
In a situation where you want to allow a
user to create his own lambda expressions at runtime, there are not a lot of
options. Whether you read the expression from a file, or an input box, you end
up with a string that you wish to convert. There is no method in .Net 4.0 that
lets you do just that (easily).
But as for most things, workarounds exist. Below is the one I’m using. Better solutions are very welcome.
Suppose you want to call Regex.Replace with
a lambda expression for the MatchEveluator. At design time this is easy:
Regex.Replace(“MyInputString”,
@”\B[A-Z]”, m => “ “ + m.ToString().ToLower());
This results in: "My input string"
But when the lambda expression is a string, that won’t work:
String input = “MyInputString”;
String expr = @”\B[A-Z]”;
String lambda = “m => \” \” + m.ToString().ToLower();
Regex.Replace(input, expr, lambda);
String expr = @”\B[A-Z]”;
String lambda = “m => \” \” + m.ToString().ToLower();
Regex.Replace(input, expr, lambda);
This results in: "Mym => \" \" + m.ToString().ToLower()nputm =>
\" \" + m.ToString().ToLower()tring"
Logical, but not quite what’s intended.
Logical, but not quite what’s intended.
In order to convert the string to a true
lambda expression would take considerable effort. However, creating a regular ‘method’
at runtime is a lot easier.
The question then becomes: how can we
leverage the normal compiler routines in compiling lambda ‘strings’. Here’s an
example. It requires 3 steps:
Step 1: create a method that builds of the
body of a class and method
private string createRegexClassInMemory(string input,
string expr,
string lambda,
RegexOptions regexOptions = RegexOptions.None) {
StringBuilder classbuilder = new StringBuilder("using System;”);
classbuilder.Append(“using System.Text;");
classbuilder.Append(“using System.Text.RegularExpressions;");
// Note that we need to double escape 'escapes' and double accolade 'accolades'
classbuilder.AppendFormat("namespace {0}{{", GetType().Namespace);
classbuilder.Append(@"
public class RegexLambdaClass{
public static string RegexLambda(){");
// Note that we need to double escape 'escapes', except for the lambda expression, as it gets handled differently
input = input.Replace("\\", "\\\\");
expr = expr.Replace("\\", "\\\\");
// Note the lack of double-quotes around the third parameter !!!!!!
classbuilder.AppendFormat("return Regex.Replace(\"{0}\", \"{1}\", {2}, RegexOptions.{3});",
input,
expr,
lambda,
regexOptions);
classbuilder.Append("}}}");
return classbuilder.ToString();
}
Step 2: create a module to compile and run this code
using System;
using System.Text;
using Microsoft.CSharp;
using System.CodeDom.Compiler;
using System.Reflection;
using System.Text.RegularExpressions;
namespace Invocation {
/// Example
/// /x myNamespace MyClass MyFunction False param1,param2,param3,etc
/// The False states that MyFunction is not static.
/// <summary>
/// This class can be used to execute dynamic uncompiled code at runtime
/// This class is not exception safe, all function calls should be exception handled.
/// </summary>
public class CodeCompiler {
private CompilerParameters compilerparams = new CompilerParameters();
/// <summary>
/// Default Constructor.
/// It is necessary to have an instance of the class so that the reflection
/// can get the namespace of this class dynamically.
/// </summary>
/// <remarks>
/// This class could be static, but I wanted to make it easy for developers
/// to use this class by allowing them to change the namespace without
/// harming the code.
/// </remarks>
public CodeCompiler() {
compilerparams.GenerateExecutable = false;
compilerparams.GenerateInMemory = true;
// Add System reference
this.addReference(typeof(System.String).Assembly.Location);
}
public void clearReferences() {
compilerparams.ReferencedAssemblies.Clear();
}
public void addReference(string reference) {
compilerparams.ReferencedAssemblies.Add(reference);
}
/// <summary>
/// This is the main code execution function.
/// It allows for any compilable c# code to be executed.
/// </summary>
/// <param name="code">the code to be compiled then executed</param>
/// <param name="namespacename">the name of the namespace to be executed</param>
/// <param name="classname">the name of the class of the function in the code that you will execute</param>
/// <param name="functionname">the name of the function that you will execute</param>
/// <param name="isstatic">True if the function you will execute is static, otherwise false</param>
/// <param name="args">any parameters that must be passed to the function</param>
/// <returns>what ever the executed function returns, in object form </returns>
public object ExecuteCode(string code, string namespacename, string classname,
string functionname, bool isstatic, params object[] args) {
object returnval = null;
Assembly asm = BuildAssembly(code);
object instance = null;
Type type = null;
if (isstatic) {
type = asm.GetType(namespacename + "." + classname);
}
else {
instance = asm.CreateInstance(namespacename + "." + classname);
type = instance.GetType();
}
MethodInfo method = type.GetMethod(functionname);
returnval = method.Invoke(instance, args);
return returnval;
}
/// <summary>
/// This private function builds the assembly file into memory after compiling the code
/// </summary>
/// <param name="code">C# code to be compiled</param>
/// <returns>the compiled assembly object</returns>
private Assembly BuildAssembly(string code) {
Microsoft.CSharp.CSharpCodeProvider provider = new CSharpCodeProvider();
CompilerResults results = provider.CompileAssemblyFromSource(compilerparams, code);
if (results.Errors.HasErrors)
{
StringBuilder errors = new StringBuilder("Compiler Errors :\r\n");
foreach (CompilerError error in results.Errors)
{
errors.AppendFormat("Line {0},{1}\t: {2}\n", error.Line, error.Column, error.ErrorText);
}
throw new Exception(errors.ToString());
}
else
{
return results.CompiledAssembly;
}
}
}
}
Step 3: call the code above
using System.Text;
using Microsoft.CSharp;
using System.CodeDom.Compiler;
using System.Reflection;
using System.Text.RegularExpressions;
namespace Invocation {
/// Example
/// /x myNamespace MyClass MyFunction False param1,param2,param3,etc
/// The False states that MyFunction is not static.
/// <summary>
/// This class can be used to execute dynamic uncompiled code at runtime
/// This class is not exception safe, all function calls should be exception handled.
/// </summary>
public class CodeCompiler {
private CompilerParameters compilerparams = new CompilerParameters();
/// <summary>
/// Default Constructor.
/// It is necessary to have an instance of the class so that the reflection
/// can get the namespace of this class dynamically.
/// </summary>
/// <remarks>
/// This class could be static, but I wanted to make it easy for developers
/// to use this class by allowing them to change the namespace without
/// harming the code.
/// </remarks>
public CodeCompiler() {
compilerparams.GenerateExecutable = false;
compilerparams.GenerateInMemory = true;
// Add System reference
this.addReference(typeof(System.String).Assembly.Location);
}
public void clearReferences() {
compilerparams.ReferencedAssemblies.Clear();
}
public void addReference(string reference) {
compilerparams.ReferencedAssemblies.Add(reference);
}
/// <summary>
/// This is the main code execution function.
/// It allows for any compilable c# code to be executed.
/// </summary>
/// <param name="code">the code to be compiled then executed</param>
/// <param name="namespacename">the name of the namespace to be executed</param>
/// <param name="classname">the name of the class of the function in the code that you will execute</param>
/// <param name="functionname">the name of the function that you will execute</param>
/// <param name="isstatic">True if the function you will execute is static, otherwise false</param>
/// <param name="args">any parameters that must be passed to the function</param>
/// <returns>
public object ExecuteCode(string code, string namespacename, string classname,
string functionname, bool isstatic, params object[] args) {
object returnval = null;
Assembly asm = BuildAssembly(code);
object instance = null;
Type type = null;
if (isstatic) {
type = asm.GetType(namespacename + "." + classname);
}
else {
instance = asm.CreateInstance(namespacename + "." + classname);
type = instance.GetType();
}
MethodInfo method = type.GetMethod(functionname);
returnval = method.Invoke(instance, args);
return returnval;
}
/// <summary>
/// This private function builds the assembly file into memory after compiling the code
/// </summary>
/// <param name="code">C# code to be compiled</param>
/// <returns>the compiled assembly object</returns>
private Assembly BuildAssembly(string code) {
Microsoft.CSharp.CSharpCodeProvider provider = new CSharpCodeProvider();
CompilerResults results = provider.CompileAssemblyFromSource(compilerparams, code);
if (results.Errors.HasErrors)
{
StringBuilder errors = new StringBuilder("Compiler Errors :\r\n");
foreach (CompilerError error in results.Errors)
{
errors.AppendFormat("Line {0},{1}\t: {2}\n", error.Line, error.Column, error.ErrorText);
}
throw new Exception(errors.ToString());
}
else
{
return results.CompiledAssembly;
}
}
}
}
Step 3: call the code above
// We need to add the necessary references !!!!!
cc.addReference(typeof(System.Text.RegularExpressions.Regex).Assembly.Location);
string result = (string)cc.ExecuteCode(
createRegexClassInMemory(
input,
expr,
lambda,
regexOptions: RegexOptions.IgnoreCase),
GetType().Namespace,
"RegexLambdaClass",
"RegexLambda", true);
Note that if you don’t need lambda expressions, you could add the parameters as proper ‘method’ parameters. The build-up and execution code could then be as follows:
private string createRegexClassInMemory() {
StringBuilder classbuilder = new StringBuilder("using System;”);
classbuilder.Append(“using System.Text;");
classbuilder.Append(“using System.Text.RegularExpressions;");
// Note that we need to double escape 'escapes' and double accolade 'accolades'
classbuilder.AppendFormat("namespace {0}{{", GetType().Namespace);
classbuilder.Append(@"
public class RegexLambdaClass{
public static string RegexNoLambda(string input, string expr, MatchEvaluator me, RegexOptions regexOptions = RegexOptions.None){");
classbuilder.Append("return Regex.Replace(input, expr, me, regexOptions);"); classbuilder.Append("}}}");
return classbuilder.ToString();
}
Invocation.CodeCompiler cc = new Invocation.CodeCompiler();
// We need to add the necessary references !!!!!
cc.addReference(typeof(System.Text.RegularExpressions.Regex).Assembly.Location);
string result = (string)cc.ExecuteCode(
createRegexClassInMemory(),
GetType().Namespace,
"RegexLambdaClass",
"RegexLambda", true,
input, expr, new MatchEvaluator(delegate), RegexOptions.IgnoreCase);
Wednesday, December 7, 2011
Metro
you can check out any time you like, but you can never leave
Taking the metro
Embarking on new travels; mostly by car, sometimes by bus or train, this time by metro.
Living in the Netherlands I've given up on the train- and bus-thing, as it is simply to cumbersome, error prone and riddled with cost-incurring 'facilities'.
You can read all about it on the internet.
So the launch of a new metro was right up my alley, and I was privileged to be allowed a test-run. Which I did.
Before continuing, I want you to know that from buses and trains you can easily leave (note that you pay extra for not checking out; very much so in the case of trains).
By the way, leaving while it's still running falls in the category "don't try this at home dude"; which you couldn't anyway by virtue of not being at home, seeing as you're in the bus or train.
Well, it's all different in the metro. Oh yes. Checking out or not checking out makes hardly any difference (except for clotting up part of the brain). BUT YOU CAN NEVER LEAVE.
Here's my story
So I got on this brand new metro. It's not quite finished, and thus I need to stay flexible and open-minded. I thought I was.....
Drove around, tried some of the new features, got stuck in the features (eventually throwing them on a big heap and leaving them to rot), got expelled back to the good ol' black-top, been dragged along, managed to hop on the metro again, but couldn't find the exit. Oh bummer.
Heck, even in 'expelled' state there was no way to truly leave....
Who was it wrote: you can check out any time you like, but you can never leave.
But hey, that's okay, I didn't really wanna leave yet, fully trusting in finding the exit later (as I am writing this from home, I clearly did, so don't worry your pants off - it would leave me with a mental image I am not too keen on).
Played the piano (cool eh, a piano in the metro), threw it on the heap...
Played sudoku, threw it on the heap...
Tried the ink-pad...
Hey, hold on a minute.... sudoku I can tear up and throw in the bin.... ah, got that sorted neatly (notice that in this brave new world I can conjure up a new sudoku in no time flat)
Hmmm.... a piano lid can be slammed shut.
Chess boards can be smashed to smithereens....
Seems like every feature can be destructed if one truly wishes to do so....
Ahh.... thus one worry less.
Would this work for the whole metro? Nope. It will not allow itself to be destroyed so easily. Leaving metro is a whole different story. I eventually manged by crawling in a far corner, prying open a panel, clicking a button, and finally leaving...
It was a trip I'm not likely going to forget any time soon... Will I be sufficiently disturbed to try it again some other time? No doubt.....
Microsoft Windows 8 with the new Metro UI
If you like to try Windows 8 with the new Metro UI yourself, then please do so in a Virtual Machine. The latest VirtualBox (by Oracle) has knowledge of Windows 8 making the installation a breeze.
Some Metro-apps throw you right back to the old desktop when they're started, which is rather confusing.
Most (if not all) 'metro-apps' can be closed by using Alt-F4 (yes, the very old Windows Alt-F4), but you'll probably loose any unsaved data. (Alt-F4 ends up in a blank screen - green in my case - on which you can click to get back to Metro).
Shutting down Windows is possible, but it requires some searching (left-bottom corner hover, followed by settings, followed by power, followed by shutdown).
Still, Microsoft is working on the environment as we speak, hopefully ironing out most of the issues.
I'll give it another test-drive in 2012.
Be brave
Taking the metro
Embarking on new travels; mostly by car, sometimes by bus or train, this time by metro.
Living in the Netherlands I've given up on the train- and bus-thing, as it is simply to cumbersome, error prone and riddled with cost-incurring 'facilities'.
You can read all about it on the internet.
So the launch of a new metro was right up my alley, and I was privileged to be allowed a test-run. Which I did.
Before continuing, I want you to know that from buses and trains you can easily leave (note that you pay extra for not checking out; very much so in the case of trains).
By the way, leaving while it's still running falls in the category "don't try this at home dude"; which you couldn't anyway by virtue of not being at home, seeing as you're in the bus or train.
Well, it's all different in the metro. Oh yes. Checking out or not checking out makes hardly any difference (except for clotting up part of the brain). BUT YOU CAN NEVER LEAVE.
Here's my story
So I got on this brand new metro. It's not quite finished, and thus I need to stay flexible and open-minded. I thought I was.....
Drove around, tried some of the new features, got stuck in the features (eventually throwing them on a big heap and leaving them to rot), got expelled back to the good ol' black-top, been dragged along, managed to hop on the metro again, but couldn't find the exit. Oh bummer.
Heck, even in 'expelled' state there was no way to truly leave....
Who was it wrote: you can check out any time you like, but you can never leave.
But hey, that's okay, I didn't really wanna leave yet, fully trusting in finding the exit later (as I am writing this from home, I clearly did, so don't worry your pants off - it would leave me with a mental image I am not too keen on).
Played the piano (cool eh, a piano in the metro), threw it on the heap...
Played sudoku, threw it on the heap...
Tried the ink-pad...
Hey, hold on a minute.... sudoku I can tear up and throw in the bin.... ah, got that sorted neatly (notice that in this brave new world I can conjure up a new sudoku in no time flat)
Hmmm.... a piano lid can be slammed shut.
Chess boards can be smashed to smithereens....
Seems like every feature can be destructed if one truly wishes to do so....
Ahh.... thus one worry less.
Would this work for the whole metro? Nope. It will not allow itself to be destroyed so easily. Leaving metro is a whole different story. I eventually manged by crawling in a far corner, prying open a panel, clicking a button, and finally leaving...
It was a trip I'm not likely going to forget any time soon... Will I be sufficiently disturbed to try it again some other time? No doubt.....
Microsoft Windows 8 with the new Metro UI
If you like to try Windows 8 with the new Metro UI yourself, then please do so in a Virtual Machine. The latest VirtualBox (by Oracle) has knowledge of Windows 8 making the installation a breeze.
Some Metro-apps throw you right back to the old desktop when they're started, which is rather confusing.
Most (if not all) 'metro-apps' can be closed by using Alt-F4 (yes, the very old Windows Alt-F4), but you'll probably loose any unsaved data. (Alt-F4 ends up in a blank screen - green in my case - on which you can click to get back to Metro).
Shutting down Windows is possible, but it requires some searching (left-bottom corner hover, followed by settings, followed by power, followed by shutdown).
Still, Microsoft is working on the environment as we speak, hopefully ironing out most of the issues.
I'll give it another test-drive in 2012.
Be brave
Thursday, September 8, 2011
Ferrari: a Point to Share
What's the point of having a great system if you're not allowed to use it. Here's my take on it.
What ?!??!
A Ferrari going 10 miles an hour on 3 wheels, while listening to your favorite music on the latest iPod-Nano, plugged into universal media connector. What ?!?!!???
My birthday
I was going to be 18, and was ecstatic to
learn that my mom and dad had bought me a Ferrari. No, not a matchbox type
(that would really ruin this story), but a real one (yellow if you care to
know). Didn’t believe it at first, thought I was dreaming. But I wasn’t…. So I called all my friends to take her for a
spin….
feRRaRi
The motor puRRed……Nice!!!!!! A gentle
start…. 2 Mph, 5 Mph, 7 Mph….. really getting the hang of it now…. 10 Mph… 10
Mph… 10 Mph…. terrible scraping noise…. and sitting crooked to boot.
Ok, so there’s clearly something wrong here.
I already thought something fishy was going on when I spotted only 3 wheels;
but not being able to go faster than 10Mph in a Ferrari?!??!? Surely the
machine must be broken. Techno-savvy as I am I assumed a reboot was needed, so
I stopped the car, got out (that’s important, as otherwise the sensors might
think you’re still driving, and thus failing on the reboot) , got back in
(equally important, as otherwise driving gets really difficult, esp. seeing as
my dad had removed the remote control for safety reasons), started the car
again, and of we went… at a whopping 10 Mph.
My dad; Gotta love this guy
So what happened? Being conscientious as he
is, he had decided to limit to speed of my brand-new-all-out-honest-to-god
Ferrari. For two reasons: environment and safety. Furthermore, he had removed
one wheel so that he could save on the cost of this fantastic machine.
My Ferrari being a Share Point
Having gotten to grips with the situation,
we decided to talk about all this in my new car (or what’s left of it). I mean,
what else could I use it for, other than having a cup of coffee and collaborate
on possible changes that would render the machine somewhat useful (or at least
a bit less useless). Funny though, my dad had a UMC specially fitted. One in
the front and one in the back. If was perfect for my Nano, giving us some
really nice background music. Not to mention the coffee-maker stuck on the
ceiling…
Back to business
If you believed the story above to be true,
then this story really is for you. You see, the above actually does happen on a
regular basis. Not with Ferrari’s I’d have to admit, but with the Ferrari
amongst Web-Portal-Collaboration-Software called SharePoint.
Companies go all out buying this rather
expensive piece of tooling, and then strip it down, because they don’t trust
the users to behave. So where the tooling is designed to allow end-users to
interact with, and setup new web-sites and pages, companies try to remove the
options for fear of “web-site proliferation”. Unwarranted fear in my mind.
True, when given the rights, end-users will experiment and create new sites.
But isn’t that the whole idea? Isn’t that why you bought SharePoint… so that
new sites can be easily created. And who knows better what they want from a new
site then the actual user?
Instead, a small group of individuals has
been tasked with deciding what users want, and create the sites for them
(including some fancy shmancy functionality no one really needs); invariably
resulting in sites that are painfully underused. An embarrassment even. Is the
truth hurting already? Are you feeling the pain I deliver?
And why the fear? Take the following
scenario:
Project teams have a need to collaborate.
The team-lead or project-lead gets accredited with the rights to create project
sites.
Now how many project-sites is he going to create? One, five, one hundred? I dare bet that (s)he will create a maximum of five project sites. No more…
Now how many project-sites is he going to create? One, five, one hundred? I dare bet that (s)he will create a maximum of five project sites. No more…
“But” you say “what if he creates 25; and
all his friends as well”. Here’s the parry: have administrators monitor the
amount of sites per user. Should this fall above a threshold, then issue a
warning. Does the culprit not listen? Then remove the rights to create sites. Now
this is time and money well spent; and proliferation will soon die out, I
assure you.
Educate
The point I’m trying to make is that it
seems much wiser to not cripple the tooling you spent so much money on (and the
action of crippling typically costs way more than the initial outlay), but
instead educate people in the proper usage of your great piece of software. Use
it to its full potential, and give end-users what they really want (not what
you believe they should want), by giving them freedom and flexibility. Sure,
provide them with extras, shortcuts, easy to use modules, and perhaps fix the
branding. By all means, help them along by setting up an initial site for them. But do not
limit the potential.
Don’t give your employees a 3-wheeled
Ferrari doing a max of 10Mph, out of pure fear. Instead give them the Ferrari
you can afford, and teach them how to drive and use it safely and properly.
Keep an eye on them, and let them know when they cross the line. Revoke their
license if they don’t listen. But please don’t punish the good citizens by
limiting functionality. Punish the culprits instead by revoking their
license…..
Don't be like my dad.
Don't be like my dad.
Wednesday, March 24, 2010
Running Windows 2008 (runtime) off a USB Drive Without VHD
This post describes installing MS-Windows 2008 on a USB Drive without the help of a Virtual Hard Drive (VHD). In an earlier post I described how to do this WITH a VHD. |
Installing windows onto a USB device requires quite some steps. In what situations this is possible I cannot say, but it obviously works in some.
Below is a description of a Windows 7 installation setting up a new USB HardDisk with "Hyper-V Windows Server 2008 R2".
Let's take every single step into account.
(Intermezzo's are used to show errors you would get when skipping steps. One of them is important.)
Assumptions
Your USB HDD drive letter is going to be "U:".
Your DVDDrive has drive letter "V:".
You have administrator rights.
Tools needed
Your Windows Installation Disk.
BootSect.exe (can be found in the "Boot" directory of your windows install disk).
Windows Automated Installation Kit (WAIK) (Needed for ImageX).
Steps
1 - You will need access to the BIOS setup.
Make sure that your BIOS supports booting into a USB device.
2 - If you've got a brand new drive, you will need to initialize it. You can do this in Window's Disk Management tool, which is available in all windows versions.
Make sure you choose for NTFS. Give it a drive letter (we need it later). For the rest of this tutorial, I will assume drive letter "U:".
INTERMEZZO:
Trying to boot from the USB Drive will now give the following error.
Non-System disk or disk error.
Replace and strike any key when ready.
Trying to boot from the USB Drive will now give the following error.
Non-System disk or disk error.
Replace and strike any key when ready.
3 - Make sure the partition on the USB HDD that you wish to install in is Active. You can do this in de Windows Disk Management.
(This step got me going for a while……)
Note that the option to mark the partition as active will be greyed out if the partition is active by default.
4. Install WAIK. You should now have three "ImageX" executables (one for AMD64, one for Intel64, and one for 32 bits.) These can be found in C:\Program Files\Windows AIK\Tools.
NOTE: For me the only one working was the AMD version, although I have an Intel T5500 processor !!?!??!?!
5 - Open an elevated command prompt.
Go to the correct WAIK directory for your machine.
Type the following command (DO NOT FORGET THE DRIVELETTER:)
BootSect /nt60 U:
Note the colon.
(Bootsect.exe updates the master boot code for hard disk partitions to switch between BOOTMGR and NTLDR.)
Leave the CMD open. We'll need it again later.
IMPORTANT
We need to make sure that your machine can actually boot into the USB Drive before we continue. Make sure the USB Drive is the first to boot so that a automatic restart will go back to the USB Drive.
Try it now, and make sure you get the following error when rebooting your machine. After that, boot back into windows so you can continue with this tutorial.
INTERMEZZO:
Trying to boot from the USB Drive will now give the following error.
BOOTMGR is missing.
Press Ctrl+Alt+Del to restart.
INTERMEZZO:
If you were now trying to install Windows onto the USB Drive, you would receive the following error:
We need to make sure that your machine can actually boot into the USB Drive before we continue. Make sure the USB Drive is the first to boot so that a automatic restart will go back to the USB Drive.
Try it now, and make sure you get the following error when rebooting your machine. After that, boot back into windows so you can continue with this tutorial.
INTERMEZZO:
Trying to boot from the USB Drive will now give the following error.
BOOTMGR is missing.
Press Ctrl+Alt+Del to restart.
INTERMEZZO:
If you were now trying to install Windows onto the USB Drive, you would receive the following error:
6 - In de Elevated CMD run the following:
imagex /info V:\sources\install.wim
This will show you all the entries (or versions) available in the WIM-file; Hyper-V Server 2008 R2 has only one entry; it will have "index 1". No changes have been made yet.
Remember that V: was assumed to be the drive letter of your DVD Drive (where the Windows Install disk lives).
With the previously found index number, run the following:
imagex /apply /check /verify V:\sources\install.wim 1 U:
This step takes a while, and will place a complete image of the Windows version you want to install onto the USB drive.
7 - Boot files
Let's copy the necessary boot file to the USB Drive (use the elevated DOS prompt again):
bcdboot U:\Windows /s U: /v
(Note: if this fails it may help to disconnect and reconnect the USB Drive.)
8 - Clean up
You can now close the CMD where you ran BcdBoot.
Remove your Windows Install disk from the drive. It is no longer needed.
Reboot the machine, and make sure the FIRST boot device is the USB Drive. Windows may want to reboot a few times, and it would be annoying if it would boot into the wrong version.
Boot into the USB Drive and see the magic happen.
Thursday, March 11, 2010
Running Windows 7 / 2008 (runtime) off a USB Drive using a VHD
This post describes installing MS-Windows 7 or MS-Windows 2008 on a USB Drive with help of a Virtual Hard Drive (VHD). In a later post I describe how to do this WITHOUT a VHD (which is easier for Windows 2008 but more complex for Windows 7). |
Installing MS-Windows 7 or MS-Windows 2008 onto a USB device requires quite some steps. In what situations this is possible I cannot say, but it obviously works in some.
Below is a description of a Windows 7 installation setting up a new USB HardDisk with "Hyper-V Windows Server 2008 R2".
Let's take every single step into account.
(Intermezzo's are used to show errors you would get when skipping steps. One of them is important.)
Assumptions
Your USB HDD drive letter is going to be "U:".
Your DVDDrive has drive letter "V:".
You have administrator rights.
Tools needed
Your Windows Installation Disk.
BootSect.exe (can be found in the "Boot" directory of your windows install disk).
Windows Automated Installation Kit (WAIK) (Needed for ImageX).
Steps
1 - You will need access to the BIOS setup.
Make sure that your BIOS supports booting into a USB device.
2 - If you've got a brand new drive, you will need to initialize it. You can do this in Window's Disk Management tool, which is available in all windows versions.
Make sure you choose for NTFS. Give it a drive letter (we need it later). For the rest of this tutorial, I will assume drive letter "U:".
INTERMEZZO:
Trying to boot from the USB Drive will now give the following error.
Non-System disk or disk error.
Replace and strike any key when ready.
Trying to boot from the USB Drive will now give the following error.
Non-System disk or disk error.
Replace and strike any key when ready.
3 - Make sure the partition on the USB HDD that you wish to install in is Active. You can do this in de Windows Disk Management.
(This step got me going for a while……)
Note that the option to mark the partition as active will be greyed out if the partition is active by default.
4. Install WAIK. You should now have three "ImageX" executables (one for AMD64, one for Intel64, and one for 32 bits.) These can be found in C:\Program Files\Windows AIK\Tools.
NOTE: For me the only one working was the AMD version, although I have an Intel T5500 processor !!?!??!?!
5 - Open an elevated command prompt.
Go to the correct WAIK directory for your machine.
Type the following command (DO NOT FORGET THE DRIVELETTER:)
BootSect /nt60 U:
Note the colon.
(Bootsect.exe updates the master boot code for hard disk partitions to switch between BOOTMGR and NTLDR.)
Leave the CMD open. We'll need it again later.
IMPORTANT
We need to make sure that your machine can actually boot into the USB Drive before we continue. Make sure the USB Drive is the first to boot so that a automatic restart will go back to the USB Drive.
Try it now, and make sure you get the following error when rebooting your machine. After that, boot back into windows so you can continue with this tutorial.
INTERMEZZO:
Trying to boot from the USB Drive will now give the following error.
BOOTMGR is missing.
Press Ctrl+Alt+Del to restart.
INTERMEZZO:
If you were now trying to install Windows onto the USB Drive, you would receive the following error:
We need to make sure that your machine can actually boot into the USB Drive before we continue. Make sure the USB Drive is the first to boot so that a automatic restart will go back to the USB Drive.
Try it now, and make sure you get the following error when rebooting your machine. After that, boot back into windows so you can continue with this tutorial.
INTERMEZZO:
Trying to boot from the USB Drive will now give the following error.
BOOTMGR is missing.
Press Ctrl+Alt+Del to restart.
INTERMEZZO:
If you were now trying to install Windows onto the USB Drive, you would receive the following error:
6 - Now things get complicated.
First we are going to create a VHD in which we can place an Image of the Windows version we want to install. In this tutorial I will call this Image.VHD. Note that this VHD is also the one that will eventually be used for new Windows Install. Make sure it's big enough, or make its size dynamic. (I've used fixed for my test, but I intend to change that). For permanent use I would recommend a name like "Win2K8R2_Image.VHD".
In the CMD we still have open:
Diskpart
(Starts the disk management CMD-utility)
create vdisk file=U:\Image.vhd maximum=14500 type=fixed
(This creates a VHD on the USB Drive with the size 14500 MB. As this file will later hold the actual Windows Version, it should be placed on the USB device, where it can be found by the boot manager. (This step will take a few minutes.)
select vdisk file=U:\Image.vhd
(Selects the created disk)
attach vdisk
(Attaches it so Windows can see it)
list disk
(Shows you all the disks in the system, note the number of your newly created VHD)
select disk disknumber
(Selects the newly created VHD)
create partition primary
(Creates a primary partition)
select partition 1
(Selects the primary partition)
active
(Sets it to active)
format fs=ntfs quick
(Format it with NTFS and do it quick)
assign
(Assigns a driveletter)
list volume
(Note the driveletter the new drive got!)
(Starts the disk management CMD-utility)
create vdisk file=U:\Image.vhd maximum=14500 type=fixed
(This creates a VHD on the USB Drive with the size 14500 MB. As this file will later hold the actual Windows Version, it should be placed on the USB device, where it can be found by the boot manager. (This step will take a few minutes.)
select vdisk file=U:\Image.vhd
(Selects the created disk)
attach vdisk
(Attaches it so Windows can see it)
list disk
(Shows you all the disks in the system, note the number of your newly created VHD)
select disk disknumber
(Selects the newly created VHD)
create partition primary
(Creates a primary partition)
select partition 1
(Selects the primary partition)
active
(Sets it to active)
format fs=ntfs quick
(Format it with NTFS and do it quick)
assign
(Assigns a driveletter)
list volume
(Note the driveletter the new drive got!)
Leave DiskPart open and start a new CMD with administrative rights.
Please take note of the DriveLetter assigned. This will be referred to as "vhddrive:" from now on.
7 - In de Elevated CMD run the following:
imagex /info V:\sources\install.wim
This will show you all the entries (or versions) available in the WIM-file; Hyper-V Server 2008 R2 has only one entry; it will have "index 1". No changes have been made yet.
Remember that V: was assumed to be the drive letter of your DVD Drive (where the Windows Install disk lives).
With the previously found index number, run the following:
imagex /apply V:\sources\install.wim 1 vhddrive:
This step takes a while, and will place a complete image of the Windows version you want to install into the VHD.
8 - Boot files
Let's copy the necessary boot file from the new VHD to the USB Drive:
bcdboot vhddrive:\Windows /s U: /v
9 - Clean up
You can now close the CMD where you ran BcdBoot.
Go back to the CMD with DiskPart, en run:
detach vdisk
Exit
Exit
Close the CMD Window.
Remove your Windows Install disk from the drive. It is no longer needed. The created VHD will be used for installing Windows.
Reboot the machine, and make sure the FIRST boot device is the USB Drive. Windows may want to reboot a few times, and it would be annoying if it would boot into the wrong version.
Boot into the USB Drive and see the magic happen.
Friday, March 5, 2010
Credibility
I was more than surprised to see someone disbelieving a rather very interesting post, with quality content, because he ONLY HAD 6 COMMENTS. Unbelievable but true. Credibility is rated by the number of comments on a post. This a particularly awkward for new posts, as by definition new posts have ZERO COMMENTS. Oh dear.... should we start commenting on our own posts in order to attract attention?
Would the content of the comments matter?
Great post!
by Adje
Wonderfully put
by Frank
Experienced the very same thing!
by ScreenMan
Are you kidding me?
by TheDude
Would the content of the comments matter?
Great post!
by Adje
Wonderfully put
by Frank
Experienced the very same thing!
by ScreenMan
Are you kidding me?
by TheDude
Subscribe to:
Posts (Atom)
