After installing our clustered w08 Print Server I noticed that there was a particular Warning in the System Log. Event ID 4 "The print spooler failed to reopen an existing printer connection because it could not read the configuration information from the registry key S-1-5-18\Printers\Connections. The print spooler could not open the registry key. This can occur if the registry key is corrupt or missing, or if the registry recently became unavailable."

I looked in the registry and, no surprise, that key was missing. I hunted around a bit and found this entry. Basically to resolve this was fairly simple - add the key back in:

1. Open up the registry on the computer (I did it on all nodes individually)
2. Go to HKEY_USERS/S-1-5-18/Printers
3. Add a new Key called "Connections" (no quotes)
4. Right-click and select Permissions on the new key to verify that System has "Full Control"

Now I am not sure if this is an issue with Clustering or just a strange whacky event that happened. But if it happens to you, hopefully this will resolve it.

<rules>
  <rule>
    <path>http://YourFormsAuthSite/*</path>
    <type>FORM</type>
    <error_pages>
      <error_page>Logon.aspx</error_page>
    </error_pages>
    <auth_url>Logon.aspx</auth_url>
    <login_type>POST</login_type>
    <parameters>
      <param name="__VIEWSTATE">dDw0OTQzMjI0MjQ7O2w8UGVyc2lzdDs%2BPvhWhKKTnHpM3RIvgkgC9jJVpN%2Bg</param>
      <param name="Login1%24UserName">FormsAuthUserName</param>
      <param name="Login1%24LoginButton">FormsAuthPassword</param>
      <param name="Login1%24LoginButton">Log+In</param>
    </parameters>
  </rule>
</rules>

1. They have incorrect fieldnames (missing the Placeholders) and note the repetition of LoginButton for the Password portion of the XML demo. Obviously someone simply cut and pasted sections rather than pasting a complete, correct, XML
2. They do not have __EventValidation. Testing this with Fiddler emphasized the need for that
3. They do not have public="true" in their params which can be useful - public: If this value is not present, the parameter specified will be encrypted and stored in the search system. For encrypted parameters, the size limit is 1,024 characters. If you specify public = "true", the parameter will not be encrypted before storing in the search system. Also, the parameters size limit increases to 4,096 characters. 

• Make sure the entries on both sides of the params are htmlencoded - use http://meyerweb.com/eric/tools/dencoder/ if you need help...
• Remember that this DOES NOT do security trimming. You still need to setup an FBA security trimmer - see http://technet.microsoft.com/en-us/library/bb852172.aspx#moss12_SearchFormsAuth_ImplementSecurityTrimmer for info on that...

http://technet.microsoft.com/en-us/library/aa996343.aspx discusses means to grant access to mailboxes. The Console can grant it to individual mailboxes, but what if you want the whole kit and kaboodle? They mention using the Add-ADPermission like this from the Shell:

Add-ADPermission -Identity "Mailbox Store" -User "Trusted User" -ExtendedRights Receive-As

This seem to be fairly straightforward. For example:

Add-ADPermission -Identity "myServer\mySG\myDB" -User "myDomain\my.name" -ExtendedRights - Receive-As

But if you do that you get yelled at:

Add-ADPermission : myServer\mySG\myDB was not found. Please make sure you have typed it correctly.
At line:1 char:17
+ ADD-ADPermission  <<<< -Identity "myServer\mySG\myDB " -User "myDomain\my.name" -ExtendedRights Receive-As

The trick here is that in this case the "Mailbox Store" means something different than every other time I have run across that phrase. In this case it is looking for the AD Distinguished Name:

[PS] C:\Windows\System32>add-adpermission -identity "CN=InformationStore,CN=EX07ServerName,CN=Servers,CN=Exchange Administrative Group,CN=Administrative Groups,CN=Our Company,CN=Microsoft Exchange,CN=Services,CN=Configuration,DC=myDomain,DC=com" -User "myDomain\my.name" -ExtendedRights Receive-As

That works. Kind of intuitive, no? No? Well here is a way to find that beast:

1 - Install ADSI Edit (if you have not already) http://technet.microsoft.com/en-us/library/cc773354%28WS.10%29.aspx

2 - Open up "Configuration (NOT Domain) by selecting it in the "Select a well known Naming Context

3 - Drill down to (ready, take a breath)

• Configuration
• Your domain
• CN=Services
• CN=Microsoft Exchange
• CN=%Organization Name as stored in Exchange%
• CN=Servers
• CN=%Server Name that has the database%
• CN=%Mailbox Storage Name%
• CN="Database% (optional)
• Right Click and select 'Properties'

4. What you need to know is stored in distinguishedName. You can double-click and it will popup a textbox (as shown below). You can copy that, just DO NOT DELETE IT!!! This will give you the information you need to supply in the -Identity entry. You can also select a particular database if you so choose.

If you have been observant you will note that the DistinguishedName (which is what is passed into the -Identity variable) matches the path you drilled down. So theoretically, you do not need to go through this. Your entry should be something like:

CN=InformationStore,CN=%Exchange Server That Has Databases%,CN=Servers,CN=Exchange Administrative Group,CN=Administrative Groups,CN=%Your Exchange Organizational Name,CN=Microsoft Exchange,CN=Services,CN=Configuration,DC=%Your Domain%,DC=%your DomainExtension"

Good luck!

NOTE: This is NOT an issue with clustering, but appears to be an issue with w08 (and w08r2) regardless of whether the drive is clustered or local. For more info look here -> http://www.myfriedmind.com/techBlog/2009/10/14/UACAndDomainAdminsPermissionsIssueOnWindows2008.aspx

============ The information below is misleading - see the above link for correction

Another addition to w08 that might trip you up is the use of the User Account Control (or UAC) to prevent Administrator accounts (other than the default created one) from doing anything useful (unless prompted). Connect that with the fact that you can only sign onto a machine once per account (see this) and you have a case where you have to log on as the non-default Administrator but are hampered in doing your work.

Put aside the annoying popups (are you SURE you want to see the security permissions? Really? Really?) there is are more serious issues. Case in point - we have a Cluster server with the Role of File Services. Logged on as a lowly Domain Admin I can not get to the actual drive that it is sharing. Let me state that again clearly

1. I am working on a Clustered w08 server with the Role of File Services
2. I am logged on with a Domain Admin account (but not with the default Administrator account)
3. UAC is turned on
4. I can NOT access the drive(s) (much less the shares) that the Cluster uses

1. Always use the default Administrator account when logging on to a Clustered w08 account. This always gives you access.
   
2. Turn off UAC ON ALL CLUSTERED SERVERS (since if it is not turned off on the host server, whichever one that is, you are going to run into the same problem).
   

I funny thing happened to me this morning. I was remotely connected to one of our w08 servers with our standard Administrator account when suddently my session came to a sudden end. I knew what must have happened, and sure enough one of my cohorts had signed on remotely to that server.

At first I thought it might be a limitation in w08 that you could now only have a single Remote Access connection, but I quickly realized that that was not the case. Instead the new tweak in w08 is that you can only sign on ONCE per account. So, when my cohort signed on with the same username it booted me off and handed my session over to him. This is new to w08, in w03 you could have sign on more than once with the same account and run different sessions.

The takeaway seems to be that you are going to have to have multiple Domain Admin accounts, probably assigning one per administrator. This will mean that you can have better security auditing (hopefully) but it also means that you will have more accounts that can do more damage.

Note that with the addition of User Account Control turned on by default this may restrict some critical tasks (see here).

Not a bad thing to have added to w08, just something to be aware of...

Part 1 - Introduction
Part 2 - Basic Example
Part 3 - Complex Example
Part 4 - Final Notes
Bonus - Ajax History and the Memento Pattern

There is some clean up I need to do in regards to my posts for Ajax History.

Where can I see an example?

You can download the zipped files here -> AjaxExamples.zip or go to http://www.myfriedmind.com/AjaxExamples and poke around

What version of .net do I need?

I want to reiterate that you must be using .net 3.5 (or higher) for this to function. The methods and properties that are used are packaged into .net 3.5 (as is Ajax itself).

How long does my history last (ie going back/forward)?

Your history lasts only as long as you are on that particular page. You can go forward and backward over your Ajax history all you want, but once you got to a different page, either preceding or following, you lose your pathway. This does not mean that if you get one of the intermediate URIs either through bookmarks, links, manually typing, etc that it will not return that particular page. It will. But the tracking of the entries in your browser history will be lost.

This also means that if the user modifies the URI manuallyto change what is in the has, it will lose that history since it considers you having gone to a new page.

What if I need to keep a perfect snapshot of the page itself?

If that is the case then you are going to need to look at a different way to store/retrieve the data. You will probably need to store each page, as it appears, into the database and then recall it from there. Do NOT store it in the History Points Remember, what is stored in the History Points must be tiny!

Why does Opera NOT WORK???

http://www.myfriedmind.com/techBlog/2009/09/21/Opera9x10xFailingOnAjaxHistoryAndTheHackToFixIt.aspx (thanks to Tomi for discovering this issue)

Part 1 - Introduction
Part 2 - Basic Example
Part 3 - Complex Example
Part 4 - Final Notes
Bonus - Ajax History and the Memento Pattern
Extra Bonus - Issues with Opera

This is the third in a series looking at using Ajax History in .net 3.5. I would recommend you look at Part 1 and Part 2 if you have not already done so.

Ajax out of the box is wonderful in that it reduces page reloads, but it is not wonderful in that it does not treat each Ajax call as a unique URI. As a result all Ajax modified changes are lost when the Back or refresh button are clicked. In addition there is no way to store the URL of the page in such a way that it will display the results of Ajax changes. I explain much of this in Part 1. In Part 2 I provide (very) basic example of storing/retrieving Ajax changes. However in the real world we do not live with simple pages. Most of what we want to reconstruct are complex systems. This post will attempt to give a real-life example of such a functionality.

Many people use mapping pages as an example of how to do Ajax history functionality. That is very nice because, especially for retrieving the pages, because the world generally changes very little geographically. Longitude and Latitude have been fairly constant since the early explorers. GPS does not change from year to year. So it is the perfect example of how you can use a small portion of data (GPS coords for example) to return a large complex data (a map). This is similar to the Memento pattern (take a look at http://www.myfriedmind.com/techBlog/2009/05/20/AjaxHistoryAndTheMementoPattern.aspx for more on my thoughts on that).

Using a Memento pattern is critical because if you were to try to store all that complex data in the Ajax history it would quickly fall to pieces.

So, rule #1 - do not store complex data, store simple data that can be used to reconstruct complex data

While racking my brain for what to use as a non-map example, I came across a post by Senthilkumar Moorthy on forums.asp.net in which he lays out an interaction using Ajax that I thought would be a useful example since it uses multiple history points

The page layout

For simplicity sake I am keeping their names fairly generic and not adding a lot of extra (fun) code...

• There are two UpdatePanels (upPanel1 and upPanel2).
• upPanel1 contains a TextBox (textBox1), a Button (button1) and a DropDownList (ddl1)
• when the user enters in phrase in textBox1 and clicks button1, ddl1 is populated with subchoices
• when the user clicks on ddl1 to select a choice, upPanel2 displays the details of the selected item in a handful of labels

It should be noted that I am using the Adventure Works (light) database as my source of data.

1. I will use textbox1 to search through ProductDescription and return a list of possible matching items into ddl1.
2. Selecting an item from ddl1 will then display information about that particular item in the details pane.

I have also included a class in the page called ProductInformation.

public class ProductInformation
{
      public int ProductId { get; set; }
      public string Name { get; set; }
      public string Model { get; set; }
      public string Description { get; set; }
      public DateTime LastModified { get; set; }
}

The history point events

The first thing to consider is where we want to put our history points. Ie, where do we want, if the user clicks 'Back' to roll them back to. Also, what do we want, if they post the URI, to display. It seems to me that there are clearly two logical places and these all take place when events occur (which is a good way to start)

1. when the user clicks button1 to populate ddl1
2. when the user selects an item from ddl1 and the detail is displayed

First off, note that we are storing three connected, but different points of data, as opposed to the single point in the simple example of Part 2. In addition we are having to store the Page Title (more on that later) because of an apparent issue with the way that .net handles that.

The caveats

The biggest issue here is that data changes. Here are a few possibilities

1. the entry of the textbox1 is no longer valid and returns no matching items
2. the selected item of ddl1 is no longer valid and returns no matching product
3. the details in upPanel2 have changed since the last update

Again, if this were simply mapping, #1 and #2 (changes in basic data) would not be an issue. And since most people want the most up-to-date maps, #3 does not matter. However, we are going, for the sake of study, assume that we care about all 3 of these potential issues. This means we are going to have to handle that in a graceful way when we unpack it to recreate the data.

I have the sample code included below for you to download and I am using the AdventureWorksLT for my data. You can snag it from CodePlex at http://www.codeplex.com/MSFTDBProdSamples/Release/ProjectReleases.aspx?ReleaseId=4004 . The code itself is fairly rough in patches since I really wanted to hit the main points about Ajax History but let me know if I need to clarify any section of it just so you can understand what I am attempting to show.

Creating the basic methods to create the page display

My methods for populating the DDL are as follows:

    private void PopulateDDL(string text)
    {
        ddl1.Enabled = true;
        StringDictionary _results = GetMatchingDescriptions(text);
        if (_results != null)
        {
            ddl1.DataSource = _results;
            ddl1.DataTextField = "Value";
            ddl1.DataValueField = "Key";
            ddl1.DataBind();
        }
    }
    private void PrefixDDL(string text, string value)
    {
        ddl1.Items.Insert(0, new ListItem(text, value));
        if (ddl1.SelectedIndex != -1)
        {
            ddl1.SelectedItem.Selected = false;
        }
        ddl1.Items[0].Selected = true;
    }

My method for the Display Info (upPanel2) are as follows:

    private void DisplayInfo(string selectedValue, DateTime requestedOn)
    {
        if (String.IsNullOrEmpty(selectedValue))
        {
            DisplayError("That selection is no longer valid. Please retry");
        }
        else
        {
            int _productId;
            if (Int32.TryParse(selectedValue, out _productId))
            {
                ProductInformation _product = GetProduct(_productId);
                if (_product == null)
                {
                    DisplayError("That selection could not be found. Please retry");
                }
                else
                {
                    lblLastModified.Text = _product.LastModified.ToString("F");
                    lblName.Text = Server.HtmlEncode(_product.Name);
                    lblModel.Text = Server.HtmlEncode(_product.Model);
                    lblDescription.Text = Server.HtmlEncode(_product.Description);
                // what if the data is old? notify the user
                    if (requestedOn < _product.LastModified)
                    {
                        lblRequestedOn.Text = String.Format("This item has been modifed since request at {0}", requestedOn.ToString("F"));
                    }
                    else
                    {
                        lblRequestedOn.Text = requestedOn.ToString("F");
                    }
                }
            }
            else
            {
                DisplayError("That was an invalid selection");
            }
        }
    }
    private void ResetDisplayInfo()
    {
        lblRequestedOn.Text = "";
        lblLastModified.Text = "";
        lblName.Text = "";
        lblModel.Text = "";
        lblDescription.Text = "";
    }
    private void DisplayError(string errorText)
    {
        ResetDisplayInfo();
        lblName.Text = errorText;
    }

I will call these methods not merely when the page is initially created but also when it is REcreated during the ScriptManager's Navigate command.

Creating the HistoryPoint methods/event handler

Apart from the modifications that are made similar to what was done on Part 2, the main changes deal with the methods to store the History Points and to recreate the page after the History Points have been reloaded. In other words what happens when a History Point is set and what happens when a History Point is retrieved. Because we do not want to store the ENTIRE page in the browser URI the fundamental question is "what can I get away with?" This is a good time to practice your YAGNI (You ain't gonna need it) to eliminate all but the most essential and simplest pieces of information. In this case I have three pieces that can not be reduced any more

1. the text in the textbox
2. the selected value of the dropdown list
3. the DateTime that the page was loaded (to see if the data has been changed)
4. (bonus) In addition, due to a 'feature' of Ajax History in .net 3.5, I am also including the Page Title. I will discuss more about this later, but just know that it is included.

My method to store the History Point looks like this:

    protected void StoreHistoryPoints(string textBoxText, string ddlValue,
        DateTime requestedOn, string pageTitle)
    {
        ScriptManager _mySM = ScriptManager.GetCurrent(this.Page);
        NameValueCollection _historyPointEntries = new NameValueCollection();
        _historyPointEntries.Add("whatTextBoxContains", textBoxText);
        _historyPointEntries.Add("whatWasSelectedInDDL", ddlValue);
        _historyPointEntries.Add("requestedOn", requestedOn.Ticks.ToString()); // ticks are shorter in length
        _historyPointEntries.Add("pageTitle", String.Format("AV - {0}", pageTitle));
        _mySM.AddHistoryPoint(_historyPointEntries, String.Format("AV - {0}", pageTitle));
    }

It is very straightforward. I create a NameValueCollection to hold my minimal data and then call the AddHistoryPoint.

Before I continue I should mention that there are three overloaded methods for AddHistoryPoint

• ScriptManager.AddHistoryPoint(string, string)
• ScriptManager.AddHistoryPoint(string, string, string)
• ScriptManager.AddHistoryPoint(NameValueCollection, string)

The second two have an additional string parameter which is used to set the Page Title when the History Point is stored. In this example I use it to set the Page Title to whatever bike model that they are looking at. Thus if they save it in their favorites it gives a user friendly name.

HOWEVER!!! While this works going forward that information is lost when the Navigate method is called and the HistoryPoint information is unpackaged. Ie - going forward the ScriptManager changes the Page Title, but when the page is loaded using the "Back" button or from a manually typed URI, it does not set the Page Title, nor is it accessible through the HistoryEventArgs. As a workaround I have included the Page Title as one of the KeyValue pairs in my history point so that I can recreate along with the rest of the page. Of course, since this is an issue I could simply have set the Page Title manually (for example in the portion of the code that determines the bike to display) but I wanted to show you this overload.

Also note that I could as easily have done this in my method:

    protected void StoreHistoryPoints(string textBoxText, string ddlValue,
        DateTime requestedOn, string pageTitle)
    {
        ScriptManager _mySM = ScriptManager.GetCurrent(this.Page);
        _mySM.AddHistoryPoint("whatTextBoxContains", textBoxText);
        _mySM.AddHistoryPoint("whatWasSelectedInDDL", ddlValue);
        _mySM.AddHistoryPoint("requestedOn", requestedOn.Ticks.ToString()); // ticks are shorter in length
        _mySM.AddHistoryPoint("pageTitle", String.Format("AV - {0}", pageTitle));
    }

There is actually no discernable difference except for the Page Title being handled by the Script Manager going forward. The NameValueCollection that the HistoryEventArgs of the ScriptManager's Navigate method returns are the same in both methods. The code that I display below for the Navigate event handles them exactly the same because both return the same NameValueCollection.

Quick note also that all these entries must be string values (they are stored in the URI after all). For this reason I have changed the DateTime to use ticks instead of writing it out, but you could have easily have chosen a format which would return a simpler string. The key is to approach this with a minimalist attitude. Not being too obsessive, but the shorting the string, the shorter the URI, the better for the system.

Recreating the page

Now I have the process in place for the storing of History Points. I can call this method as my final command from both the button1_Click event handler and the ddl1_SelectedIndexChanged event handler and it will store the information I pass it into the history. The Page does not need to know what this data represents. It is the Caretaker (see Ajax and Memento). All it knows is that it needs to store this information and return it when requested.

The second portion is the unpackaging when the History Points have been returned. This is handled through the ScriptManager's Navigate event. One quick thing to note - this event is called when the "Back" or "Refresh" commands are given to the browser, but it also is called when the page is loaded with the specific History Point URI, for example from a "Bookmark" or from a link.

My event handler looks like this:

   protected void sm1_Navigate(object sender, HistoryEventArgs e)
    {
        long _requestedOnAsTicks;
        bool _hasHistory = long.TryParse(e.State["requestedOn"], out _requestedOnAsTicks);
        if (_hasHistory)
        {
            DateTime _requestedOn = new DateTime(_requestedOnAsTicks);

            string _textStored = e.State["whatTextBoxContains"];
            if (!String.IsNullOrEmpty(_textStored))
            {
                // restore entries on TextBox and DDL
                textbox1.Text = _textStored;
                PopulateDDL(_textStored);
            }
            string _ddlValueStored = e.State["whatWasSelectedInDDL"];
            if (!String.IsNullOrEmpty(_ddlValueStored))
            {
                ListItem _matchingItem = ddl1.Items.FindByValue(_ddlValueStored);
                if (_matchingItem != null)
                {
                   // we have victory, display the item...
                    _matchingItem.Selected = true;
                    DisplayInfo(_ddlValueStored, _requestedOn);
                }
                else
                {
                    // something is wrong, the item no longer shows. Display error...
                    ListItem _errorItem = new ListItem(_ddlValueStored, "");
                    DisplayInfo("", _requestedOn);
                }
            }
            else
            {
                // since no DDL value was selected, add prompt at beginning of DDL
                PrefixDDL("Select One", "");
            }

            // this is a required hack because even though page title is specified in the AjaxHistory, it
            // does not work when you navigate backwards..
            string _pageTitle = e.State["pageTitle"];
            Page.Title = _pageTitle;
        }
        else
        {
            // if there is no DateTime stored than this is the original (ie no history) -> reset form
            textbox1.Text = "";
            ddl1.Items.Clear();
            ResetDisplayInfo();
            Page.Title = "AV";
        }
    }

The Navigate event uses the HistoryEventArgs. This really only has one thing in there - the State, which is the NameValueCollection that you passed the ScriptManager during the AddHistoryPoints calls.

The first thing I do is see if there is anything in there. Since I am wanted to get the date and the TryParse is a convenient way for me to verify that it is a valid long, I check it there. If there is no history than this is a non-modified Ajax Page (probably the very beginning) and I simply reset everything.

Once I know there is a history I have to go through the process of recreating the page. I extract the appropriate information and repopulate the Textbox, the DropDownList (if appropriate) and the Display (if appropriate). By doing this I am creating a complex page from simplistic data.

Because my data may have changed (say the Description has been updated) I have also included a notification to the user if this takes place in the DisplayInfo method. However, I no longer have access to the original data. That data is lost to me because all I am passing are the minimal pieces to recreate the page. What is displayed is the CURRENT information on that particular product.

Also, note the setting of the Page.Title. Remember that even though I did pass that as a parameter during my AddHistoryPoint method call, it is no longer available to me during the Navigate event. If you want to use a unique Page Title when an Ajax modification takes place, make sure to pass the Page Title as its own KeyValue pair and then unpack and recreate it.

Final Notes

If you have stuck in this long, congrats!!! I know this may still seem murky, probably more due to my explanation than to anything else, but there are a few more things I want to touch on. I will actually them in my (potentially) last blog on this - part 4 - wrap up.

Zipped Example

AjaxExamples.zip (1.37 MB)

Part 1 - Introduction
Part 2 - Basic Example
Part 3 - Complex Example
Part 4 - Final Notes
Bonus - Ajax History and the Memento Pattern
Extra Bonus - Issues with Opera

Handling Ajax history has been a critical issue since Ajax first started. With the advent of .net 3.5 the ability to easily deal with that is rolled into the code. This is part 2, here we will look at a simple example of handling Ajax history to give a basic understanding of the process. If you have not read Part 1, I suggest you do that now. Part 3 will expand on handling Ajax history when the data is more complex.
For simplicity sake we are only going to look at an example of very, very basic history. We are going to look at a page that displays changes made to the time.

A basic assumption is that you are fairly familiar with Ajax (http://ajax.asp.net). We are going to be using the ScriptManager (of course) and an UpdatePanel.
Our basic page looks like this

<%@ Page Language="C#" AutoEventWireup="true" CodeFile="AjaxHistory.aspx.cs" Inherits="AjaxHistory" %>

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">

<html xmlns="http://www.w3.org/1999/xhtml">
<head runat="server">
    <title></title>
</head>
<body>
    <form id="form1" runat="server">
    <asp:ScriptManager runat="server" ID="sm1" EnablePartialRendering="true" />
    <div>
        <asp:UpdatePanel runat="server" ID="upME">
            <ContentTemplate>
                <asp:Button runat="server" ID="uxChangeTime"
                    OnClick="uxChangeTime_Click" Text="-- change:" />
                <asp:Label runat="server" ID="lblTime" />
            </ContentTemplate>
            <Triggers>
                <asp:AsyncPostBackTrigger ControlID="uxChangeTime"
                     EventName="Click" />
            </Triggers>
        </asp:UpdatePanel>
    </div>
    </form>
</body>
</html>

The codebehind sets the label text to be the current date/time on every click:

(c# code)
protected void uxChangeTime_Click(object sender, EventArgs e)
{
    string _currentDateTime = DateTime.Now.ToString("F");
    lblTime.Text = _currentDateTime;
}

(vb code)
Protected Sub uxChangeTime_Click(ByVal sender As Object, &_
  ByVal e As System.EventArgs) Handles uxChangeTime.Click
    Dim _currentDateTime As String = DateTime.Now.ToString("F")
    lblTime.Text = _currentDateTime
End Sub

As you can see this is fairly straightforward. Clicking on the button displays the current DateTime in the label. However, there is no history. Clicking back will not display the last DateTime displayed but either be grayed out (if this was the first page loaded) or take you back to the page before this one.

To enable history is very, very simple in 3.5

1. modify the ScriptManager, setting a property or two and registering an event (Navigate) for the ScriptManager
2. modify the Click handler to notify the ScriptManager to record a History Point
3. handle the event called by the ScriptManager to rollback, when required, the page

Step 1: Modify the ScriptManager

The modifications are very simple - set EnableHistory equal to true and specify what should handle the Navigate event. You can also specify if you want the HistoryState to be 'secure' by setting EnableSecurityHistoryState equal to true, but this is optional.

Your ScriptManager should now look like this:

<asp:ScriptManager runat="server" ID="sm1"
     EnableHistory="true" OnNavigate="sm1_Navigate"
     EnableSecureHistoryState="true" />

Step 2: Add an AsyncPostBackTrigger for the Navigate event of the ScriptManager

<asp:UpdatePanel ...>
    ...
    <Triggers>
         <asp:AsyncPostBackTrigger ControlID="uxChangeTime"
                     EventName="Click" />
         <asp:AsyncPostBackTrigger ControlID="sm1" EventName="Navigate" />
    </Triggers>
</asp:UpdatePanel>

Step 3: Modify the Click handler (or whatever event it is that would trigger a HistoryPoint to be recorded

What you are doing here is notifying the ScriptManager what information to store as a HistoryPoint by calling its AddHistoryPoint method. There are three possible overloads and we will look at the other two next blog, but for now we will use one of the simplest, a single key/value string pair to store the current DateTime with the key "myTime":

(c# code)
protected void uxChangeTime_Click(object sender, EventArgs e)
{
    string _currentDateTime = DateTime.Now.ToString("F");
    lblTime.Text = _currentDateTime;
    ScriptManager.GetCurrent(this).AddHistoryPoint("myTime", _currentDateTime);
}

(vb code)
Protected Sub uxChangeTime_Click(ByVal sender As Object, &_
  ByVal e As System.EventArgs) Handles uxChangeTime.Click
    Dim _currentDateTime As String = DateTime.Now.ToString("F")
    lblTime.Text = _currentDateTime
    ScriptManager.GetCurrent(Me).AddHistoryPoint("myTime", _currentDateTime)
End Sub

Step 4: handle the ScriptManager Navigate event

In this event we retrive the State property of the HistoryEventArgs and use that to set the page to be what we want it to be. In this very simple example we will use it to set the value of the label to be the DateTime that that particular page displayed:

(c# code)
protected void sm1_Navigate(object sender, HistoryEventArgs e)
{
    string _historyValue = e.State["myTime"];
    if (_historyValue != null)
    {
        lblTime.Text = _historyValue;
    }
}

(vb code)
Protected Sub sm1_Navigate(ByVal sender As Object, &_
  ByVal e As System.HistoryEventArgs) Handles sm1.Navigate
    Dim _historyValue As String = e.State["myTime"]
    if(_historyValue <> Nothing)
    {
       lblTime.Text = _historyValue
    }
End Sub

Voila! We now have a functional Ajax page with history. Try it!

What has happened is that we are using History Points. When we click the button it determines the current DateTime, displays it, and stores it as a history point. When we click the "Back" button it retrieves the previous history point (which was the current DateTime in string format) and displays that (see chart below).

Next blog: a look at more intensive use of History Points...
btw, for prep, take a looksee at my blog on the Memento pattern and Ajax history to understand my basic approach: http://www.myfriedmind.com/techBlog/2009/05/20/AjaxHistoryAndTheMementoPattern.aspx