Turn complex and repetitive tasks into simple operations
As a systems administrator, you're well aware of how busy you are. If you're not putting out four-alarm fires, you're playing catch-up on last month's and maybe even last year's projects. The idea that you can squeeze anything else into your schedule seems as preposterous as Microsoft Bob 5.0. Yet there's one technology that's well worth making time for—Windows PowerShell, an interactive scripting and command-shell environment that lets you automate administrative tasks and access a wide range of information.
With PowerShell, you can run commands directly at the command prompt or run scripts that contain those commands. PowerShell supports its own scripting language, which leverages the Microsoft .NET object model to combine the rich features of object-oriented programming with the ease of command-shell scripting. What that means for you is a powerful environment that can turn complex and repetitive tasks into simple operations. Through PowerShell, you can access a variety of systems and technologies, such as Active Directory (AD) and Windows Management Instrumentation (WMI) to perform such tasks as retrieving event log entries, disabling user accounts in AD, and retrieving a computer's user-defined shares.
PowerShell runs on Windows Vista, Windows Server 2003 SP1, Windows Server 2003 Release Candidate 2 (R2), and Windows XP SP2. It will also run on Windows Server 2008 (formerly code-named Longhorn Server). You can install PowerShell on x86, x64, and IA64 processor architectures. However, before you install PowerShell, you must first install Microsoft .NET Framework 2.0. You can download the .NET Framework at http://msdn2.microsoft.com/en-us/netframework/aa569263.aspx and PowerShell at http://www.microsoft.com/technet/scriptcenter/hubs/msh.mspx. To install either product, simply run the setup program and follow the steps in the installation wizard.
After you've installed PowerShell, you're ready to go. Click Start, All Programs, Windows PowerShell 1.0, then Windows PowerShell. In the PowerShell window, you can run commands or PowerShell scripts (.ps1) files by entering the command or filename at the command prompt. To test your installation, type
at the command prompt and press Enter. This displays information about getting help in PowerShell—a handy command to be sure. (For more cmdlets that are helpful when learning PowerShell, see the sidebar "PowerShell Pointers.")
You're now ready to run commands and scripts. All you need to do is to learn a little about the PowerShell language. To help you with that, I'll review three sample scripts—RetrieveAppEvents.ps1, DisableUser.ps1, and FindShares.ps1—that demonstrate many of the basic concepts in the language and show you how easy it is to get started with PowerShell.
RetrieveAppEvents.ps1 in Listing 1 retrieves entries from the local application event log and saves them to a text file. As callout A shows, I begin the script by defining the $date variable. A dollar sign always precedes parameter and variable names. The variable uses the Get-Date cmdlet to retrieve the current date and time (aka datetime). A cmdlet, which is similar to a function, performs a specific action and usually takes the form of verb-noun. I then use the AddDays method to obtain the datetime exactly 24 hours (i.e., 1 day) prior to the current datetime and assign that value to the $date variable.
Next, I create the FormatEntryType function, as callout B shows. A function is a named block of code that performs a specific action. After you create the function, you can reference it anywhere in your script and the block of code will run. In this case, the FormatEntry function retrieves the content of a text file, modifies that content, and saves it to a second text file. The function takes the $file parameter, which passes the pathname of the target text file into the function.
The first command in the function's statement block (enclosed in curly brackets) uses the Get-Content cmdlet to retrieve content from the text file in $file. Notice that a pipe (|) follows the cmdlet. This indicates that the content should be passed down the pipeline to the next cmdlet. One feature that makes PowerShell so useful is the ease with which you can create pipelines to pass information from one statement to the next.
In this function, I pass the data retrieved by Get-Content down the pipeline to a ForEachObject cmdlet, for which you can use the alias ForEach or %. The ForEach cmdlet lets you iterate through objects within a collection. In this case, the collection is made up of the content of the text file. By default, the objects in a file collection are delineated by line breaks, which means the collection contains one object per line. (You can override the default behavior, but for the purposes of this example, line breaks work well.)
The ForEach cmdlet uses an expression, enclosed in curly brackets, to process each object in the collection. The expression begins with the $_ symbol, which refers to the current input object from the collection. The expression then uses the -replace operator to replace any error object with an *** ERROR *** object. In other words, any line that contains only the word error is replaced with *** ERROR ***. A second ForEach cmdlet performs a similar operation on warning objects.
The second ForEach cmdlet pipes the content to the Out-File cmdlet, which sends the content to the AppEvent_EntryTypes.txt file. Each time you run the function within a script, the content will be inserted into that file.
The code at callout C retrieves the application event entries and assigns the results to the $events variable. To retrieve data from the application events log, I use the Get-Eventlog cmdlet and specify Application as a parameter. I then send the event data down the pipeline to the Where-Object cmdlet. The backtick (`) at the end of the line indicates that the statement continues to the next line. However, you don't have to use a backtick when a line breaks at a pipe.
The Where-Object cmdlet filters the data based on the expression defined in the curly brackets. As with ForEach, you use $_ to reference the current object within the collection. In this case, the collection is made up of the event entries. You can also use $_ to reference specific properties within the object. For example, the expression uses $_ to reference the TimeGenerated property. You reference an object's properties by adding a period followed by the property name. The expression then uses the greater than (-gt) operator to compare the TimeGenerated property's value to the value in the $date variable. As a result, the $events variable includes only events generated within the last 24 hours.
Next, I use the $events variable to access the events. The statement in callout D passes the content in $events down the pipeline to a ForEach cmdlet. The ForEach expression consists of a Out-File cmdlet that outputs the event content to AppEvents.txt. The Out-File cmdlet includes the -Append option to ensure that each event is added to the file without overwriting any events. In addition, the cmdlet includes the -InputObject option, which uses the $_ symbol and property names to specify the types of data to save to the file. As a result, the output file includes only the timestamp, entry type, source, and message associated with each event.
Finally, the code in callout E calls the FormatEntryType function and passes the AppEvents.txt file's pathname to the function. As you saw earlier, this function retrieves the data from the first text file, updates the data, and adds it to the AppEvents_EntryTypes.txt file. Figure 1 shows a sample event from AppEvents_EntryTypes.txt.
To run RetrieveAppEvents.ps1 from the PowerShell command prompt, you simply need to type the script's pathname and press Enter. So, the command might look like
However, PowerShell prevents you from running scripts by default. To modify the default behavior, you must change the Power-Shell security settings. To get started, you can change the settings by entering the command
Thereafter, you can run scripts that you create, but any other script must be digitally signed.
Now that you've seen your first script, you should be familiar with many of the basic PowerShell concepts. As you'll see, a lot of these concepts apply to other scripts.
DisableUser.ps1 in Listing 2 disables user accounts in AD. As callout A shows, I begin this script by using the Param keyword to define a parameter ($sam) that passes in the user account that's entered on the command line when the script is run. When using this keyword, you have several options:
You can use only the Param keyword and the name of the parameter in parentheses.
You can define a default value in case no value is specified when running the script.
You can use the Throw keyword (as I've done here) to return an error message when no value is specified. When returning an error message, you must enclose the Throw keyword and message in parentheses and precede the parentheses with a dollar sign.
The code at callout B locates the user account in AD. This code begins by creating an object that searches the directory. To create the object, I use the New-Object cmdlet with the DirectorySearcher class in the DirectoryServices namespace. (For information about the DirectoryServices namespace, go to http://msdn2.microsoft.com/en-us/library/ system.directoryservices.aspx.) I then assign the object to the $ds variable.
Next, I create a filter on the $ds object by setting the object's Filter property ($ds.filter). The filter is based on the Active Directory Service Interfaces (ADSI) attributes defined after the equal sign (=). The filter removes all values except those that conform to the attribute definitions. As a result, the filter returns a user account that is part of the person object category and the user object class and that has a SAM name that matches the one specified in the $sam parameter.
After creating the filter, I use the FindOne method of the $ds object ($ds.findOne()) to retrieve the user account. I assign the account to the $dn variable.
The code in callout C defines two variables. The first variable, $desc, stores the user account's description. To retrieve the description, I use the $dn variable to call the account's properties, then call the Description property ($dn.properties.description). The second variable, $date, stores the current datetime, which I obtain with the Get-Date cmdlet. Both these variables are used later in the script to update the user account's description.
The code in callout D disables the user account. The section is encased in an If statement block that runs when the If condition ($dn.path.length –gt 0) evaluates to true. The condition compares the length of the $dn object's Path property to 0. The Path property contains the LDAP location of the user account in AD. When the Path property contains a value, the If statement block runs.
The first command in the If statement block uses the Path property to create an ADSI object for the user account. The ADSI object, which is assigned to the $user variable, is used to access the object's properties and methods, including the AccountDisabled property.
In the next statement, I set the AccountDisabled property to true. Because AccountDisabled is stored in a binary collection in AD, I use the InvokeSet method on the PowerShell base object (psBase) to update the property. The InvokeSet method takes two arguments: the property name (AccountDisabled) and the new value. To set the AccountDisabled property to true, I use the built-in $true variable.
Now I'm ready to update the user account's description. To do so, I call the $user object's Put method, which takes two arguments: the property name (Description) and the value. In this case, the value is made up of a combination of the $desc and $date variables and the word disabled. The value takes the original description and appends the word disabled followed by the current datetime.
After updating AD information, you must commit the changes, so I call the $user object's SetInfo method. I then display two messages by using the Write-Host cmdlet. The first message says the account has been disabled. The second message displays the account's distinguished name (DN).
The code in callout E highlights the final part of the script, which is an Else statement block. The Else statement runs when the If condition evaluates to false. For this script, the Else statement uses the Write-Host cmdlet to display a message that says the user account wasn't found.
That's all there is to DisableUser.ps1. When you run this script, you must have access to the AD store. I tested this script on a computer running Windows 2003 Enterprise Edition that was configured as a domain controller (DC). I also tested the script on an XP machine against a Windows 2000 DC.
FindShares.ps1, which Listing 3 shows, retrieves a list of the user-defined shares on a computer. Like DisableUser.ps1, FindShares.ps1 begins by defining a parameter. As callout A shows, the $computer parameter passes the computer name to the script when you run it. However, the parameter uses a default value rather than returning an error message when a parameter isn't provided. In this case, the default value is a period, which refers to the local computer.
The code in callout B uses the Get-WmiObject cmdlet to create a WMI object. The cmdlet uses the -Class option to specify the Win32_ Share class, the -Namespace option to specify the root\CIMV2 namespace, and the -ComputerName option to specify the computer name in $computer. Of these options, the -Class option is the most important because it determines the type of information you can access through the WMI object.
After accessing the WMI class information, I pass it down the pipeline to a Where-Object cmdlet. The Where-Object expression, enclosed in curly brackets, includes three conditions. The first condition uses the not equal (-ne) operator to compare the Caption property's value to the phrase default share. For the condition to evaluate to true, the property's value can't equal the phrase. The second condition uses the -notlike operator to compare the Caption property's value to the remote* value. Notice the use of the wildcard, which can represent any characters. For the condition to evaluate to true, the Caption property's value can't begin with the word remote, but it can end with any characters. The final condition is similar to the second condition, except that the Caption property's value can't begin with the word logon.
The Where-Object expression uses the -and logical operator to link the three conditions, which means that they all must be true for a share to be included in the list. I pass the filtered list down the pipeline to a Sort-Object cmdlet, which sorts the list of shares based on the Name property (by default, in ascending, or alphabetical, order). I assign the sorted WMI information to the $shares variable.
The code in callout C is an If statement block. The If condition ($shares -ne $null) uses the -ne operator and the $null system variable to specify that the $shares variable can't contain a null value. When the $shares variable's value isn't null (i.e., the If condition evaluates to true), the If statement block runs.
The If statement block begins with the Write-Host cmdlet. Because the cmdlet specifies no content, it simply returns a blank line. This provides extra spacing to better display the information in the PowerShell window.
The next statement is a ForEach statement. The ForEach statement isn't the same as the ForEach-Object cmdlet, even though they perform the same function. Adding to the confusion is the fact that one of the ForEachObject cmdlet's aliases is ForEach. Here's how you can tell them apart: When ForEach is at the beginning of a command, it's a ForEach statement. When ForEach is within a pipeline, it's a ForEach-Object cmdlet.
The ForEach statement iterates through the objects (i.e., shares) in the $shares collection. The statement defines the $share variable, which refers to the current object. The ForEach expression uses the $share variable to take action on each object. The first command in the ForEach expression is a Write-Host cmdlet that writes the Name property's value to the PowerShell window. The script accesses the name property through the $share variable. The second WriteHost cmdlet writes the Path property's value to the PowerShell window. The final Write-Host cmdlet simply adds a line after each iteration to make it easier to read the list of shares.
When the If condition ($shares -ne $null) evaluates to false, the Else statement block in callout D runs. The Else statement block contains its own If and Else statement blocks. The nested If condition specifies that the computer name must equal a period. When the computer name is a period, Write-Host uses the COMPUTERNAME environmental variable to return the name. Note that to access an environmental variable, you must precede the variable name with $env:. When the computer name isn't a period, Write-Host returns the name stored in $computer.
When you run FindShares.ps1, you'll obtain a list of shares. Figure 2 shows sample output from this script.
Only Scratching the Surface
As you can see, with PowerShell, you have a lot of flexibility in the type of information that you can access and what you can do with that information. These three examples of how to use PowerShell only scratch the surface. The more effort you devote to learning PowerShell, the greater your payoff will be. And who knows, with these new skills, you might have time to complete last year's projects.