RUSSELL MCCLOY

ASP.NET Membership System.Web.Providers. Can't log in to my siteon Azure but can locally.

noreply@blogger.com (russell mccloy) — Thu, 15 Mar 2012 05:22:00 +0000

Hi there,

I had an interesting issue last night. I deployed my local Azure emulated MVC 3 role and worker role to my Windows Azure instance. All worked perfectly locally and in particular the logon page I built allows me to log onto the secure area of my MVC 3 site using System.WebProviders (ASP.Membership).

I found an issue when I tried to log onto my remote Azure site using my deployed logon view.

The logon on the server failed and told me my user name or password was wrong.

Ok, so I then pointed my local instance ASP.NET configuration (the built in membership, role etc.. tool for setting up members and roles) at my Azure SQL database and tried to manage my members from there and it worked.

Even when I changed my local connection string to point to my Azure SQL database and tried logging on, it worked. WEIRD!!! (ie.. local browser, remote Azure database).

So then today I fould this blog:

http://stackoverflow.com/questions/8415822/asp-net-membership-login-works-locally-fails-on-azure

and then this one:

http://geekswithblogs.net/DavidHoerster/archive/2010/06/15/asp.net-membership-password-hash----.net-3.5-to-.net-4.aspx

The part of the post in particular that caught my attention was this:

I tracked it down, thanks to some info in this article by David Hoerster. The problem is that the default password hashing algorithm on Azure is different from the .NET 4.0 defaults. It is set to SHA1 on Azure, and HMACSHA256 is the new standard setting on 4.0.

The rest of the post is here:

This can be fixed by specifying the hash type explicitly in web.config. If you decide to use a method like HMACSHA256, make sure you also specify a machine key - otherwise you will run into similar problems as the autogenerated machine key will differ from server to server.

The configuration element you need to change is under :

<machineKey decryptionKey="PUT_DECRYPTION_KEY_HERE"

            validationKey="PUT_VALIDATION_KEY_HERE"

decryption="AES"

validation="HMACSHA256" />

You can use this machine key generator to generate random keys in the proper format.

I will go home tonight and try this and update this post in the morning.......time passes.......ok, so went home last night and tried a few things. My password hashing algorithm was different between my local and azure. I just used the Azure machine config setting on my local and it all works fine now. Good times!!!

RuSs

A look at ASP.NET MVC 4

noreply@blogger.com (russell mccloy) — Wed, 07 Mar 2012 23:07:00 +0000

Morning,

I am currently getting up to speed on MVC 4. Thought i'd post Scot Guthrie's video on the subject.

Here is the real link as I didn't want to embed the video as the width was to large:
http://channel9.msdn.com/Events/TechDays/Techdays-2012-the-Netherlands/2364

RuSs

My weekend (I'm a geek) of deploying my new Kite Boarding site to Windows Azure with an Azure SQL database, membership and blob storage

noreply@blogger.com (russell mccloy) — Sun, 19 Feb 2012 23:59:00 +0000

Good Morning,

I just thought I'd post to my blog after a long period of being to lazy to do it and also being too busy at work to get around to it.

I spent part of the weekend deploying my latest side project to Azure. I won't tell you what it's all about until I have it all polished and shiny as right now it's a definite work in progress.

What I have deployed:

MVC 3, Html5, Razor website - I'm thinking about moving to MVC 4 as I think the beta just came out.
The site runs off Entity Framework 4.3 and I used code first poco classes with DBContext.
Users and Roles are managed by http://nuget.org/packages/System.Web.Providers
My custom data is stored an an Azure Sql relational database

I posted the following questions on MSDN and have now got the answers I need so i'll update the questhions with answers. As I was an Azure newbie a 2 weeks ago, these questions now seem a bit lame to me!!

Questions:

I have used an Azure SQL relational database for my database. They talk a lot about table storage. I assume I have done the right thing in using a relational database?
Ok, so this one was easy. Whilst I could probably use table storage to store me data (xxx.table.core.windows.net) I am not ready to ditch my relational database way of thinking. My site will have many tables with relationships so for now I am playing it safe and using SQL Azure. I have, however, used blob storage (xxx.blob.core.windows.net) to store the gallery images I need for my site. So, and I may be wrong here, if I had a list of people whom each had a picture attached to their relational database record, I think I could store the blob Url in a field in their record and, when rendering, iterate the records and retrieve the blob image from my storage account.

I have used the http://nuget.org/packages/System.Web.Providers to manage users and roles. Once I deployed to Azure using the same database in my web.release.config for my custom data AND the membership tables it all works great. I can register a new user using the standard MVC 3 register user view and I can see this new user in my Users table inside the asp membership table.
My question here then: and I have Googled this but cant find it, is how do I manage my users and roles on the azure server as I would when I use the asp.netwebadminfiles/default.aspx in Visual Studio? Do I, and I plan to this anyway down the line, write my own custom membership management code to do this or can I use an instance of asp.netwebadminfiles/default.aspx on the server somewhere?
I figured this one out too. Although there are ways to use the netwebadminfiles tool to manage users and there is also third party code that can do this for you, ultimately I want to manage these users myself. I will build all my user and role code into the site.

When I deployed for the first time it created my web svc role on Azure AND my Sql database mentioned above. Lets say I have the site live for a few months and want to redeploy, will my membership and custom data get wiped out? What if a deploy schema changes to my database, will this be taken care of without affecting my data. I know I can use a one way data sync to get the data onto local but im sure I would not have to worry about this?Not 100% sure on this one but I now know that firstly, I don't have to redeploy / publish the whole site, storage and database every time I make a change. I can use Visual Studio WebDeploy to push new changes to my site. If I needed to sync data from my local I think I could use the Data Sync in the Azure portal and I assume I could manage my users via SQL mambership stored procs. until I get all my membership screens written.

thanks

RuSs

Saving the internet

noreply@blogger.com (russell mccloy) — Mon, 05 Dec 2011 21:11:00 +0000

All,

Quote from: http://www.avaaz.org/en/save_the_internet/?mzla

"As concerned global citizens, we call on you to stand for a free and open Internet and vote against both the Protect IP Act and the Stop Online Piracy Act. The Internet is a crucial tool for people around the world to exchange ideas and work collectively to build the world we all want. We urge you to show true global leadership and do all you can to protect this basic pillar of our democracies worldwide."

Please sign up to save the internet. This is important for all of us to continue to have an open and uncensored internet. Imagine if the police monitored all your phone calls, you wouldn't like that would you?

http://www.avaaz.org/en/save_the_internet/?mzla

thanks
RuSs

Explanation of Ninject Bindings and where we used them in our MVC Website (Part 1 - MVC Website)

noreply@blogger.com (russell mccloy) — Mon, 07 Nov 2011 06:05:00 +0000

Afternoon,

Im going to summarize what we have done with our Ninject bindings and injection which allows us to follow the "Inversion of Control (IOC Container)" principle in a project we are working on.

As mentioned we are using Ninject as our IOC container: http://ninject.org/

The project architecture looks much like the following diagram:

Architecture Diagram

I will work my way from the top to the bottom of the architecture. Starting with the MVC web site all the way through to the Microsoft Crm layer.(This post (Part 1 - MVC Website) just refers to the web site.)

One of the 2 BIG reasons we are using an IOC container (Ninject) is to allow us to separate all our layers (separation of concerns) so that they are Unit testable and also so that we can swap components out at a later date. We may want to replace our Crm later with a Sql server layer, doubtful but we are catering for this anyway as you never know what will be a requirement in the future. As long as our Sql server repository layer implements the same interface as our existing Crm repository layer, we can more or less swap out the Crm Repository with a new Sql Repository.

Mvc Website:

On the website layer we will be using Ninject to inject the following:

A GatewayAgent layer into our controller - the GatewayAgent layer separates the concerns of the controller from the concerns of our WCF SOA layer.
An AutoMapper implimentation into our GatewayAgent - the mapper will map our presentation entities to our Domain entities.
Our presentation entity validation classes into our presentation entities - as we are using Fluent validation, we may want to ditch this later and inject a different type of validation.

The following diagram show part of the web site's global.asax file. This code uses Ninject to inject the above 3 items.

The Ninject Binding in the Global.asax class

Obviously, the web site has a reference to the Ninject.dll.

Most of this is set in stone for our site but one thing to note. As more development is done we will need to add more validators as we add more presentation entities to our site. So when we add an Invoice presentation entity to the site a related Invoice Validator class will need to be added to the ConfigureKernel method, shown above, in our global.asax file.

Note: the line:

- kernel.Bind<IValidator<Contact>>().To<ContactValidator>();

means, when we call validate on a Contact presentation entity,

we would like to use the ContactValidator class to carry out the validation

Sample Contact Validator

You can see in the ConfigureKernel method above that we also inject a ContactGateway.
The following line shows this:

kernel.Bind<IContactGateway>().To<ContactGateway>();

This means that when we refer to IContactGateway we would like to use the concrete class ContactGateway.
So in our controller we see the following code:


Sample controller

Why is this good? Because you can then mock up a ContactGateway in a unit test to test your controller.

A unit test for the controller

As mentioned above, we also inject a mapper implementation into our web site:

kernel.Bind<IMapper>().To<AutoMapperMapper>();

This is used to map our domain entities to our presentation entities on the way up and our presentation entities to our domain entities on the way down.

Why is this good? We may way to use a different mapping tool other than AutoMapper in the future. Injection will enable us to swap AutoMapper out for a different mapper tool.

This is the end of Part 1 - Explanation of Ninject Bindings and where we used them in our MVC Website

I will do another couple of posts on the other layers mentioned in the architecture diagram at the start of this post.

RuSs

How Jimmy Bogard does MVC

noreply@blogger.com (russell mccloy) — Mon, 03 Oct 2011 05:26:00 +0000

Bonjour,

Following is a link to a very good post called "How we do MVC" on some DOs and DON'Ts of MVC. We are about to start a fresh enterprise level project and this article is great for getting a few things done correctly from the start.

http://lostechies.com/jimmybogard/2009/04/24/how-we-do-mvc/

There is also another article he has written on "How we do view models":

http://lostechies.com/jimmybogard/2009/06/30/how-we-do-mvc-view-models/

Hope you find this useful.

RuSs

Entity validation, unit testing and MVC 3 with Fluent - RuleSets and Predicates

noreply@blogger.com (russell mccloy) — Mon, 26 Sep 2011 00:56:00 +0000

Morning all,

I have been looking at validation for my current project. We are building an MVC 3 site that calls a wcf service later that is wrapped around Crm 2011.

Initially I was looking at data annotations out of the box. These work really well and MVC 3 works seamlessly with these with client validation and unobtrusive JQuery.

When I came to look at rulesets however it seems a lot more complicated. I them stumbled upon Fluent validation:

http://fluentvalidation.codeplex.com/

This is a very user friendly clean way of writing validation on your entities in the same way that data annotations are written. The difference came when I started to look at rule sets which you will see in my attached screen shots.

Firstly I defined 2 classes in my domain model. Thsese are called ContactForValidationPOC and AddressForValidationPOC as shown below:

----------------------------------------------------------------------------------------------------------

As you can see there are no data annotations and ContactForValidationPOC has a complex property of Addresses.

We are going to write some rulesets that will validate the contact for a STAFF ruleset and a public ruleset. (Note that there are not real world validation needs as of yet,it's just proof of concept code.)

You will notice that at the top of the ContactForValidationPOC class we have an attribute shown as follows:

[Validator(typeof(ContactValidator))]

This tells the class that there is a validator class call ContactValidator somewhere.

Also please note that I have referenced the correct Fluent  validation dlls in my project.

You can find the details of this here:

http://fluentvalidation.codeplex.com/

So back to the ContactValidator. Here it is:

You can see that we have 2 rule sets. The first is a staff rule set and the second is a public rule set.

They are quite similar and the only difference really is that the Age must be over 18 for a staff member.

Both rule sets also use a .Must predicate validator.

This is a great place to put more complex logic.


Here we need to ensure that there is at least 1 address in the Addresses collection that is marked as IsPrefferedAddress.

Ok, so that is all done, now we need to unit test or validation rules.

 
So I have a unit test written here. Firstly I set up a new ContactForValidationPOC entity and add a couple of addresses to it.

I then call Validate on it and ask it to validate using the Staff rule set. After validation we check that the ValidationResult (result) is valid

and then do a simple check for the number of validation errors. Obviously my test could be more detailed in a real world example.

As I debug through my test I can see that my entity is not valid (IsValid = false)

and it has a collection of validation errors on it.

This is good as we didn't enter a first name, our Age is under 18 and we set all AddressForValidationPOC's in the Addresses collection to IsPrefferedAddress = false.

Ok so we now know that our validation is working correctly on the server side.

We haven't even touched the MVC code yet.

I have whipped up a quick MVC controller with a view and have disabled the settings

in the web.config that make client side validation work as shown in the following picture:

Note: they should be set to false to prove that server side validation is still working correctly now that we are validating through MVC.

Here are my controller and view:

Note: I have just forced 2 addresses into the ContactForValidationPOC entity to make things easier, these are both set to IsPrefferedAddress = false

Here is the view that kicked off the validation:

You can see that, when debugging the controller we get the same validation results.

IsValid = false and a number of errors relating to our Staff rule set.

Note that in our controller we could have control over what rule set is loaded if we so desired:

We could use some logic to check what type of rule set to load, staff or public.

Now I will change the web.config setting to ensure that client side validation gets included:

We now can fire up the view in a web browser and see client side validation in action:

Well that is it, a high level explanation of domain entity server and client validation using Fluent, unit tests and MVC.

Some positive points:

All the validation for an entity is contained wholly on the entity.
Fluent validation is very easy to write and read.
More more complete business rules you can use a predicate delegate.
You can unit test your entity validation results with ease.
It works seamlessly with jquery.validate and jquery.validate.unobtrusive.
With MVC you could also use IOC to instatiate your validators:
http://fluentvalidation.codeplex.com/wikipage?title=ValidatorFactory&referringTitle=Documentation

Some possible negatives:

Only the following validators are supported on the client:
- NotNull/NotEmpty
- Matches (regex)
- InclusiveBetween (range)
- CreditCard
- Email
- EqualTo (cross-property equality comparison)
- Length
Cant think of any more at the moment.

I hope this was helpful.

RuSs

EDMX or DbContext API

noreply@blogger.com (russell mccloy) — Tue, 06 Sep 2011 00:54:00 +0000

Hi all,

I have used EF for a while. I am used to using an EDMX file for my model. I have also recently used Code First and the DbContext API and I really like this.
Last night I started building an app. and was thinking, am I starting on the right foot? Do I use EDMX or DbContext API? If the thing i'm building turns out to be big, will I have chosen the right technology.

What if I use DbContext API and get it to generate my database. Obviously I don't want this to keep generating my database down the track when i'm at a stable point in development (or do I?). Then if I want to tweak my database or domain model will all the strings come un done?

The DbContext API is a simplified version. Will it do everything I need? I dont think ill need stored procedures as I can do everything with linq.

So my real question is. If I am building something that is serious and out there is the real world and it may turn out to require some complex coding, which way do I go? EDMX or DbContext API?

Is EDMX now old school?

Comments / thoughts?

RuSs

A small update

noreply@blogger.com (russell mccloy) — Mon, 05 Sep 2011 05:35:00 +0000

Morning all,

Well it's been a while since I posted in my blog that i'm sure no one reads! I have finally got used to living in Melbourne and have started enjoying it more now that the weather is getting better.

I will hopefully be posting some interesting stuff on MVC up here is the next few weeks.

Here is a list of things I have looked at in the last few weeks.

MVC(3) - I have been trying to get around to getting up to date on MVC for years. Ever since I tried and failed to read Martin fowler's book on design patterns (http://martinfowler.com/eaaDev/uiArchs.html).
Well finally I have got around to it and I love it. Combined with the Entity Framework edmx or code first it is such a great way to work. This is the way, for most sites, the web should be built. No hidden viewstate, no nightmare page load order and code that you have total granular control over. Awesome!!
Razor view engine - initially it looks dodgy but after a little playing around it's cleaner and easier to read than traditional asp.net forms. One thing I have learned about Razor is, if it looks messy and hard to read, then you probably have started putting business logic where it should not go.
Unit testing and TDD with MVC - Combined with Ninject and Moq this is easy. You can directly test your controllers. Compare this to ASP.NET web forms where your controller was, sort of, your code behind. You needed httpContext etc. Now you can test just your controller logic without other webby things clouding the test.
JQuery - ok so this is not new but im loving it so far.
NuGet - the NuGet package manager is quite cool. NuGet is a Visual Studio extension that makes it easy to add, remove, and update libraries and tools in Visual Studio projects that use the .NET Framework. When you add a library or tool, NuGet copies files to your solution and automatically makes whatever changes are needed in your project, such as adding references and changing your app.config or web.config file. When you remove a library, NuGet removes files and reverses whatever changes it made in your project so that no clutter is left.
For example if you need to get the latest JQuery libraries into you solution you open Tools > library package manager > package manager console and type:

PM> Update-Package jquery

PM with auto update your JQuery to the correct version. Note that your references inside Views > Shared > _Layout.cshtml (for example) will not be auto updated.

Well, I know this is a bit of a lame post but it was more to get me back into the habit of writing on my blog. I hope to come up with some more interesting stuff soon.

RuSs

SQL Profiler - Being self-centered is a good thing (thanks Jeff Handley)

noreply@blogger.com (russell mccloy) — Tue, 07 Jun 2011 23:34:00 +0000

When I'm faced with app. debugging, especially procedures that take awhile to run, or require a series of steps in the application to fire off, I often use SQL Profiler to capture the SQL call. Then I'll re-run the call directly in SQL Management Studio. But SQL Profiler can bring back a lot of noise, making it tough to grab only what I need.

I finally created a custom trace template and set it to be my default. It will only show me calls made from my local workstation to the development database. (we have a shared development database*)

This article will help you set up a personalised trace. For expample you can filter by Application Name or Host Name to reduce the noise in your trace.

Link to the full article by Jeff Handley

Thanks

RuSs

Finding all stored procedures updated in the last (n) days

noreply@blogger.com (russell mccloy) — Wed, 23 Mar 2011 01:23:00 +0000

This was quite helpful:
SELECT name
FROM sys.objects
WHERE type = 'P'
AND DATEDIFF(D,modify_date, GETDATE()) < 30

My new life in Melbourne

noreply@blogger.com (russell mccloy) — Sat, 15 Jan 2011 03:33:00 +0000

Morning,

After 8 years in London I have decided enough is enough. I'm on the way to Melbourne to see what the future holds.
Im currently getting my CV ready and if anyone out there has any leads I can follow in Melbourne that would be great.
I'm currently in Sydney with my family and will head south in the next couple of weeks.

RuSs

Re-map an SQL server login to a database login

noreply@blogger.com (russell mccloy) — Wed, 10 Nov 2010 16:41:00 +0000

Hello,

Sometimes users get un-mapped from the server to database login after a database restore.

Here is the command to fix it:
EXEC sp_change_users_login 'Auto_Fix', 'SRSUser'

RuSs

Workflow 3.5 and MOSS 2007 - This task is currently locked by a running workflow and cannot be edited.

noreply@blogger.com (russell mccloy) — Wed, 10 Nov 2010 11:52:00 +0000

Morning,

Today I had an interesting issue with my WF3.5 SharePoint workflow.
A user was attempting to action a SharePoint workflow task and was given the error:

This task is currently locked by a running workflow and cannot be edited.

I have seen this error a small number of times (5-10) amongst the 1000's of workflow instances that have passed through my production system.

I am always very careful with quiescing and always release a new workflow version the correct was using workflow.xml, feature.xml and stsadm.

There was a time a few weeks ago when we had to do an emergency bug fix as some of my business logic was not functioning as expected.

I know that I can never change an activity in my workflow without quiescing otherwise the serialization and deserialization (dyhydrate / re-hydrate) from the persistence store gets confused and hence an error like this occurs.

The funny thing is, I only changed code in a utility class and only the code inside an existing method (ie...method definition was not changed) so I thought that there was no need for quiescing. (no core workflow code was altered).

Anyway,just writing here incase anyone has an opinion on this.

From now on I will quiesce no matter what the emergency is!!

RuSs

Workflow 4.0 Activity Designer

noreply@blogger.com (russell mccloy) — Tue, 19 Oct 2010 08:54:00 +0000

In this post i'm going to explain how to make a custom activity better visually.
First of all I have a custom code activity called RequestManagerAction. It inherits from CodeActivity and overrides its Execute method. (I may do a post on code activities later).

The visual representation of code activity, by default, looks like this:

And by the end of this post it will look like this:

The first step is to create a new peoject for your activity designer. As we want to "design" RequestManagerAction I have created a new project called: RequestManagerActionDesignerLibrary. And inside this I add a new Activity Designer Item from the Workflow - installed templates called RequestManagerActionDesigner.xaml.

the initial view after opening RequestManagerActionDesigner.xaml is a split screen with the Xaml that represents the design in the bottom part of the screen and the actualy visual design in the top part.

In the bottom pane we can see the Xaml. First of all I want to set the Collapsible property to false as I don't want my designer to be collapsible.

The 2nd taks is to add an ArgumentToExpressionConverter to my Xaml. This defines the mapping between arguments and expressions and provides the functionality to convert between these objects. Typically, this converter is used declaratively when using an ExpressionTextBox in a custom activity designer.

The next thing I want to do it sort out my layout. As my custom activity has 3 properties:

The main aim of what we are doing here is to show ExpressionTextBoxes for all 3 of these.
I will add a grid of 2 columns and 3 rows for our properties.

You can see that inside this grid we have some (3) collapsed TextBlock sections. I will explain these now.

Each TextBlock is for a property.
The first TextBlock is for the ExpenseRequest property of our custom activity:

You can see the obvious items marked out in the above picture.Note that the position inside the grid defined above is set by the Grid.Row and Grid.Column indexes which are zero(0) based.
You can also see that our expression text box is bound to ModelItem.SubmitExpenseRequest.

SubmitExpenseRequest is a class in my Contracts project that is part of my domain model:

The ModelItem represents the activity that that the designer is currently editing. This object sits between the visual elements that present an activity to the developer and the in-memory representation of the activity.

You can now see that we are well on the way to building our custom design:

I will now go ahead and add the other 2 ExpressionTextBoxes and the relevant properties.

You will notice the little smiley icon in the top left of the designer. To add this, I added the following xaml:

Now I am ready to associate my designer with the actual code activity that is already part of my workflow. If you remember it currently looks like this:

First I will add a reference to the project that my code activity is part of:

Once its referenced we can add an attribute to the top of our codeActivity class (Don't forget the System.ComponentModel import at the top!):

Once you have built the projects you will see the "RequestManagerAction" code activity in the toolbox. Drag this onto your workflow surface if it's not already there and you will see that the look of the activity is not just a small yellow box any more. Nice!!

This is obviously more useful as we don't need to use the properties pane to set the properties for our code activity, we can just type them straight into the workflow.

RuSs

Workflow 4.0 - Adding annotations to your workflow tracking profile

noreply@blogger.com (russell mccloy) — Fri, 08 Oct 2010 09:54:00 +0000

Lets say you have a workflow that is running on a DEV / TEST and LIVE server. Lets say, also for the sake of it, that we are using the same AppFabric database for all 3. (yes, this is not the best idea but to explain annotations it helps.)

So what we want to acheive is the show, in AppFabric monitoring, which server we are using. So the monitoring with show either:

ServerThatEmittedThis: DEV_Server
ServerThatEmittedThis: TEST_Server
ServerThatEmittedThis: LIVE_Server

Again, this example is useless in the real world but i'm sure by the end of this post you will be able to imagine some real world uses for this functionality.

So you can see that in my current tracking profile I have added an annotation of my DEV server as follows:

You can see in the pic. "ServerThatEmittedThis: DEV_Server" which will now show up in AppFabric monitoring.

So I will now run the workflow which is hosted in IIS / AppFabric.

My workflow runs and is persisted into my persistence store. So I click on the Active or Idle Intance in my AppFabric dashboard:

Note I could also navigate to the annotated info if my workflow wasnt using a persistence store by Looking on the dashboard for the WCF Call History section and click on the link in the Completed Calls section.

Once I can see the instance I click to select view tracked events:

Once here, I can see all the events that my workflow has passed through.

Note that I am using a custom tracking profile that emits less than some of the default profiles. My custom tracking profile named: ExpenseServiceTrackingProfile is shown in the first screen shot of a part of my web.config in this post.

As i placed my annotation in my web.config as a child node to the activityStateQuery named Screen Expense I know to expect to see my annotation in AppFabric at the point where my screen expense activity emitting its tracking record.

As you can see from the above pic. I have clicked on the Screen expense item and at the bottom you can see a tab called "tracked variables" shown below:

So I know that the workflow (server) emitting this tracking record is from my dev server.

Rubbish example, but I hope you get the point.

AppFabric Hosting PowerShell Cmdlets

noreply@blogger.com (russell mccloy) — Tue, 05 Oct 2010 09:22:00 +0000

This seems to be the place for the AppFabric Hosting PowerShell Cmdlets:
http://msdn.microsoft.com/en-us/library/ee767662(v=WS.10).aspx

The following shows the current state of my AppFabric. It has nothing persisted.

I start a workflow that I know will persist to my persistence store:

I can then look back at my AppFabric dashboard and see that 1 workflow is persisted and active.

I can then open Windows PowerShell Modules. It calls this from the server menu behind the scenes:
%SystemRoot%\system32\WindowsPowerShell\v1.0\powershell.exe -NoExit -ImportSystemModules

Ok so it's loaded. Now, just as a simple example we can query for all instances in the default persistence store.

We call "Get-ASAppServiceInstance".

It returns all the info about my persisted workflow instance.

This is interesting as I have just noticed that I have an extra unnecessary service contract name, a spelling mistake and hence an extra bookmark.

Ok, so I went in, altered the service contract name to get rid of the extra IService ref as it should only have an IExpenseService ref. I also renamed my bookmark from ManagerActionComplete to ManagerActionCompleted. Then I re-ran my workflow thinking all was ok.

An exception was thrown so obviously I just did something incorrectly.

I re-ran my "Get-ASAppServiceInstance". and saw the following information which of course is very obvious:

The exception is:

ExceptionMessage: A bookmark with the name 'ManagerActionCompleted|{http://tempuri.org/}IExpenseService' already exists.

Obviously I just named a bookmark in my workflow to the same name as another one.

I renamed ManagerActionCompleted to ManagerActionCompletedReminder as this bookmark gets resumed after a delay activity expiration time has elapsed.

So "Get-ASAppServiceInstance" now shows:

We then wait for the delay time to expire and run "Get-ASAppServiceInstance" again and we then see:

Ok so I set off to write myself an investigative blog into Windows PowerShell Modules for AppFabric and this technology already showed me a couple of issues with my workflow. Nice!!

Again, look at this link for loads of other useful commands:

http://msdn.microsoft.com/en-us/library/ee767624(v=WS.10).aspx

thanks

RuSs

Why does my workflow die when I add a TransactedReceiveActivity?

noreply@blogger.com (russell mccloy) — Tue, 05 Oct 2010 08:17:00 +0000

dzien dobry,

I dont know the answer to this one but hopefully I will be able to update this post when I find out.

The issue is this:
When I add a TransactedReceiveScope to a fully functional, persisted workflow monitored by AppFabric the workflow throws an exception.

I have looked at the exception and my trace / log files and can't find any useful information.

My Workflow looks like this:

And the error is this:

The error text is:

The execution of the InstancePersistenceCommand named {urn:schemas-microsoft-com:System.Activities.Persistence/command}LoadWorkflow was interrupted by an error.

And the stack trace is:

Server stack trace:

at System.ServiceModel.Channels.ServiceChannel.HandleReply(ProxyOperationRuntime operation, ProxyRpc& rpc)

at System.ServiceModel.Channels.ServiceChannel.EndCall(String action, Object[] outs, IAsyncResult result)

at System.ServiceModel.Channels.ServiceChannelProxy.InvokeEndService(IMethodCallMessage methodCall, ProxyOperationRuntime operation)

at System.ServiceModel.Channels.ServiceChannelProxy.Invoke(IMessage message)

Exception rethrown at [0]:

at System.Runtime.Remoting.Proxies.RealProxy.HandleReturnMessage(IMessage reqMsg, IMessage retMsg)

at System.Runtime.Remoting.Proxies.RealProxy.PrivateInvoke(MessageData& msgData, Int32 type)

at WindowsClient.ExpenseServiceRef.IExpenseService.EndGetData(IAsyncResult result)

at WindowsClient.ExpenseServiceRef.ExpenseServiceClient.WindowsClient.ExpenseServiceRef.IExpenseService.EndGetData(IAsyncResult result) in C:\PROJECTS\TelerealTrillium.Expense.Workflow.Soln\TelerealTrillium.Expense.Workflow.Soln\WindowsClient\Service References\ExpenseServiceRef\Reference.cs:line 532

at WindowsClient.ExpenseServiceRef.ExpenseServiceClient.EndGetData(IAsyncResult result) in C:\PROJECTS\TelerealTrillium.Expense.Workflow.Soln\TelerealTrillium.Expense.Workflow.Soln\WindowsClient\Service References\ExpenseServiceRef\Reference.cs:line 537

at WindowsClient.Form1.OnStartCompleted(IAsyncResult asr) in C:\PROJECTS\TelerealTrillium.Expense.Workflow.Soln\TelerealTrillium.Expense.Workflow.Soln\WindowsClient\Form1.cs:line 93

Ill keep hunting for solution and will post back when I find it.

========================== Time passes =========================

Ok so I came back to this little issue after remembering about the MSDTC.
I realized that my app. updates my persistence store AND my custom expense database inside the TransactedReceiveActivity. (I update my custom expense database using Entity Framework 4.0)

Now my connection strings for these 2 things are different obviously so I am trying to wrap a transaction around some activities that span 2 databases.

So what I did first was enable MSDTC on my local server that hosts my WCF Workflow 4.0.
It was like:

Now its like:

But that still didnt work. Although I did see a one off Transaction Aborted error which I can't reproduce.

I then went to my remote SQL server that hosts my Persistence store and custom Expense database which is what I should have looked at first really.

I enabled the same open-ish security on the Sql box and it fixed everything!!

thanks

RuSs

Resetting the AppFabric dashboard whilst in development iterations

noreply@blogger.com (russell mccloy) — Tue, 05 Oct 2010 07:41:00 +0000

Bonjour,

I have found a little bit of t-sql that cleans out the required sql tables that the AppFabric dashboard uses. I have not researched this much but it seems to do the job.

It is very useful if you are doing constant development iterations and you don't want loads of instances of your workflow clogging up the AppFabric dashboard.

The t-sql is this:

delete [dbo].[ASWfInstancesTable]

delete [dbo].[ASWfEventsTable]

delete [dbo].[ASWcfEventsTable]

delete [System.Activities.DurableInstancing].[KeysTable]

delete [System.Activities.DurableInstancing].[InstancesTable]

Probably not a god idea to rely on this in anything but a development context.

Please comment if you know any more on this.

thanks

RuSs

Workflow Tracking Profile Editor

noreply@blogger.com (russell mccloy) — Fri, 01 Oct 2010 16:02:00 +0000

Here is a great link on Dennis van der Stelt's blog about info. about the workflow tracking profile editor.
http://bloggingabout.net/blogs/dennis/archive/2010/06/08/workflow-tracking-profile-editor.aspx
This tool makes it easy to create a tracking profile. It allows you to load up a WF4 XAMLX file. It then shows the editor and you’re allowed to browse through all (composite) activities. After right-clicking an activity (with a blue circle on it) you can specify what you exactly want to track. It even knows everything about arguments and (scoped) variables so you can use checkboxes to turn them on or off.

Here are some screen shots:

So following on from my other post:
http://russellmccloy.blogspot.com/2010/10/workflow-40-sample-development.html

I will create a simple tracking profile for my workflow.

Currently my workflow is using a default tracking profile that is set by my root web.config:

You can see that it is using the "HealthMonitoring tracking Profile"which seems to be the default one. there are a few others to choose from and to configure these you:

right click on your WF in IIS and select "Manage WCF and WF services" and then "Configure..."
Click on Monitoring
then use the slider to choose what you require.

And you can see that in my web.config that is local to my service that I have no tracking profile:

So I will use the Workflow Tracking Profile Editor to point to my current workflow definition and add a tracking item.

For the purposes of this exercise I have added an int32 variable to my xamlx file called:
TempTrackingExpenseId

I then use an Assign activity to set the variable. I suppose I could track the expenseId that hangs of my custon ExpenseResponse object. I tried that before and I could not get anything to appear in appFabric. So for now i'll just stick with a simple integer.

I then right click in the blue spot in the Workflow Tracking Profile Editor:

and select "track variables in this activities scope...":

And for now ill just select every event type for my variable:

I then save my new tracking profile using the Workflow Tracking Profile Editor and insert the resulting Xml into the web.config of my xamlx service.

Now that I have done this I will fire up my workflow using a windows form client as mentioned in my earlier blog:

http://russellmccloy.blogspot.com/2010/10/workflow-40-sample-development.html

Workflow 4.0 - Hosting with AppFabric

noreply@blogger.com (russell mccloy) — Fri, 01 Oct 2010 15:29:00 +0000

Following on from my last post I will attempt to share with you, the steps I took to get my workflow hosted in AppFabric with Persistence and Monitoring enabled.(I will not talk about AppFabric's caching abilities as I know nothing about these at this point!!)

Ok so from my first post which is here:
http://russellmccloy.blogspot.com/2010/10/workflow-40-sample-development.html

I outlines how I began with workflow 4.0 and a few of the issues I had. I eventually got to a point where I developed a workflow application that, in a very simple way, allowed for the approval of an expense.

what I will now try to talk about is getting that workflow and it's parts hosted in AppFabric.

To start with, I created the workflow project as follows in this pic:

Workflow project creation

When you create a new WCF Workflow Service Application the settings default to using the visual Studio 2010 development server. If we want to use AppFabric we need to change the project so that it is hosted, first of all in IIS. The following picture shows this:

Visual Studio Project Properties

If you use the "Create Virtual Directory" button, then VS 2010 will, hopefully create the IIS vertual directory for you as shown here:

IIS virtual Directory

Right so we now have a Wcf service in IIS but it's sort of useless unless we built our own monitoring and hosting tools. Remember that AppFabric gives you loads of stuff out of the box. It also manages hosting and will re-hydrate a workflow from the persistence store when a delay timer expires.

So now what we need to do is install AppFabric.

Once appFabric is installed you need to configure appFabric.

To configure AppFabric click on the following link:

You will get an explanation page:

then the connection setup page for your monitoring and persistence database:

Before you point to the database you will need to run the following scripts that are @:

C:\WINDOWS\Microsoft.NET\Framework\v4.0.30319\SQL\en:

SqlWorkflowInstanceStoreSchema.sql

and

SqlWorkflowInstanceStoreLogic.sql

the setup wont work with out these. I think they setup a basic persistence store and appFabric then adds additional functionality to it. Not sure right now.

After you run these 2 scripts successfully on the database of your choice. (You can use any SQL db with any name to store persistence and monitoring data) you can then continue with the AppFabric setup. You may get a warning about Sql agent not being detected as in a runnable state but I checked and mine was running so I think you can assume everything is still ok.

Following this you should get a success message.

And then a final restart IIS message:

Note: I did not select to setup the caching part of AppFabric.

After this you need to configure the local details for your site / xamlx service hosting in IIS and AppFabric. The setup we have just done relates the overall global setting for our IIS / AppFabric server. We can get more granular at the site level:

Setup monitoring. Accept default as specified earlier and add some trace and message logging as shown in the following pics:

You may get an issue with net.pipe missing. You can add this in the sites advanced settings as shown below:

After this is done we can run our windows client in Visual Studio. The windows client submits expense requests to our workflow.

Before we do this please not that I have neglected to add the c:\MailboxExpenseApproval directory on my server. This will enable us to see the workflow experience and error and "suspend" its self. Using AppFabric we will be able to add the directory to the file system and "resume" the workflow.

So here is the windows client

You can see that after submission we get an error:

It looks like an Smtp error because c:\MailboxExpenseApproval directory on the server is set as the pickupDirectory for sending mail. As it's not there, we get an error which is good for us right now.

If we look in AppFabric we will see that the workflow is suspended:

We can then add the c:\MailboxExpenseApproval directory and then resume the workflow using AppFabric.Nice!!

I have then build a web site that is hosted in the same IIS / appFabric as the xamlx service. The website is used to response the the Smtp emails that managers receive asking them to approve expenses. There are links in the emails to point to the approve / reject form:

Once approved the user will see this message:

The persisted instsance of the workflow will then complete and be automatically removed from the persistence store. (Note: you can set some config to keep completed persisted workflow instances if you prefer).

That's all for now. Will add more later.

RuSs

Workflow 4.0 a brief overview

noreply@blogger.com (russell mccloy) — Fri, 01 Oct 2010 09:28:00 +0000

This is a very simple post about my recent findings on using workflow 4.0 (WF 4.0). I build a simple WF which managed expense approval.

Server Specs

Windows server 2008 R2 standard (Virtual server with 4gigs RAM)

Components:

Visual Studio 2010
SQL 2008
IIS and Appfabric

Technologies:

Visual Studio 2010
Workflow 4.0
Entity Framework 4.0
Linq
SQL 2008
IIS
AppFabric
SMTP

Initial setup steps:
Install Visual Studio 2010 Professional
Install appFabric (More on appFabric configuration in a later post hopefully)
Setup monitoring and persistence (I used the default settings - ie. root web.config in IIS holds settings for any child sites.)

Visual Studio 2010 Development:
I decided that I wanted to try this technology with a useful purpose in mind. It is not a full scale app. but most of the basic building blocks are there.

The app. I build is an expense approval app. using workflow 4.0 (WCF Workflow Service Application):

This project allows you to build a workflow and expose it as a WCF service. I ended up at this decision as I started developing workflow using System.Activities.WorkflowApplication. This worked well until I tried to use a delay timer in my WF. I was persisting my workflow successfully but after the expected time period for my timer had elapsed, the workflow was still "asleep" in my persistence store.

Following this I hunted the web for a better soln. I didn't necessarily need a delay timer for anything at this stage but, because the WF was unusable with a delay timer I thought I should keep hunting for a better solution. I came across loads of posts (some wrong, some out of date - this was very frustrating).

I decided to try and use the WorkflowServiceHost (http://msdn.microsoft.com/en-us/library/system.servicemodel.workflowservicehost.aspx) as it says on the "tin" that this will manage the re-hydration of my workflows from the persistence store.

I can't remember where I got to with the WorkflowServiceHost but in the end after I posted a question on this post:

http://social.msdn.microsoft.com/Forums/en-US/wfprerelease/thread/ebfa7571-9879-4b0c-93ff-0134764262f7

and Stuart pointed me in the right direction I moved to my current, as mentioned above, setup.

This being windows server 2008 R2 with full appFabric monitoring and persistence installed. I thought, why try and reinvent the wheel when this great infrastructure is provided more or less out of the box.

So, back to the actual workflow development. Here is an overall picture of my workflow. It needs refactoring but i'll plonk it here for now and come back to it later. (Certain parts, like the duplication in the parallel activity need refactoring).

Note: I wont be going into to much the detail with regard to my code as i'd rather explain the steps taken and a higher level.

The workflow receives an ExpenseRequest. If amount of this ExpenseRequest is under £100 the expense gets auto approved and the expense submitter is notified.

If the expense is over £100 a notification is sent to a manager for approval. If said manager does not action this request within a period of time the workflow re-hydrates from the persistence store and reminds the manager.

I tested this re-hydration (with my WF hosted bty AppFabric but submitting an expense for over £100 and letting the expense notification reach the manager (Note: the delay activity for the reminder was set to 10 minutes at this point)).

I then rebooted the server and waited the remainder of time that the delay timer required.

After the timer was due to expire the workflow came alive and sent the reminder to the manager. Good times!!

So anyway, The workflow "code" functions correctly. Some points about it that I will point out are.

The workflow (workflow as a Wcf service hosted in AppFabric (IIS)) is consumed by 3 clients:

One is a console app. that I have just for some initial setup testing.
The second is a windows forms client that is used to submit the expense request.
The third is a web app. that is used to approve or reject the expense.

Note: the manager links to the web app. from a link in the Smtp email that the workflow sends out.

If you right click on the workflow ExpenseService.xalmx file and select "view in browser" you can see the Wcf definition and Wsdl for your workflow. This is good to know as we now know we can definitely host our workflow in IIS / AppFabric. More good times!!

In my clients I added a service reference to my Workflow .xamlx service. Here are the steps to to this:
Note: that I selected to "generate asynchronous operations" as I dont want my client apps. to hang around waiting for a "long running process" to complete when all I really care about is whether my ExpenseRequest was sent correctly.

Once I have my xamlx service referenced in my client apps. I need to write some code to instantiate it. first of all I need these lines:

private static ExpenseServiceClient client = new ExpenseServiceClient();

request.Expense = new Expense();

request.Expense.Amount =200;

request.Expense.Title = "";

client.BeginSubmitExpense(request, OnStartCompleted, null);

As I said, I am doing this asyc so I only care that my request was successful. Do be sure of this I subscribe to the IAsyncResult.
The code looks like this:

SubmitExpenseResponse result = (SubmitExpenseResponse)client.EndSubmitExpense(asr);

Within the method that waits for the IAsyncResult I have some try catches.

catch (CommunicationException)

catch (TimeoutException)

catch (Exception)

I can then decided what do in my client app. when something goes wrong.

You can probably see in my workflow that there are some custom activities like:
AutoScreenExpense, NotifyExpenseSubmitter, RequestManagerAction, SaveExpense, UpdateExpense. These are my own activities that inplement System.Activities.CodeActivity. AutoScreenExpense is a flowchart xaml workflow that I have built that just takes are of the £100 limit decision mentioed earlier.

After you build your custom activities they will appear in the toolbox.

AutoScreenExpense is a separate workflow that I have exposed as a custom activity inside the ExpenseService.xamlx workflow. Nice!To write a code based activity you can either implement System.Activities.CodeActivity or System.Activity.NativeActivity.As you can see I implement CodeActivity. Also note that this activity has a public InArgument which is the ExpenseRequest coming into the activity. the expenserequest is then mapped to my Entity Framework 4.0 entity and then saved and a response is sent back to the workflow.

Another thing I nearly forgot is that the way to get data into and throught a workflow is to use arguements or variables. In this case, I think because I have used a xamlx project type, I only have access to variables as follows:

you can see my ExpenseRequest variable another one for response and a boolean that is used internally by my workflow. Also, I have an ExpenseIDHandle that is used for correlation so that the workflow know which instance of a workflow to work with when a caller asks it to do something.

I mentioned that I am using Entity Framework 4.0 to manage my database updates. I have a couple of tables in an SQL 2008 database. One holds the record of the expense to be approved and the other is just a lookup table:

Obviously this is not very fancy but I plan on building on it later.

So now I have a workflow that I can host in IIS and AppFabric. In my next post I will do this.
Over and out!!
RuSs

Ps. If I have got anything wrong here please let me know

Building queries for working with Folders

noreply@blogger.com (russell mccloy) — Tue, 07 Apr 2009 08:47:00 +0000

Credit to:
http://sharepointmagazine.net/technical/development/writing-caml-queries-for-retrieving-list-items-from-a-sharepoint-list for this info.

------

A first variant I will explain is how you can work with folders and sub folders. A folder is a special list item on a list or document library. If you execute a standard CAML query you will end up with list items from the root folder.

If you want to query all folders and sub folders of a list or document library, you have to define extra query options. If you are working with the object model you have to set the ViewAttributes property of the SPQuery object as follows:

qry.ViewAttributes = "Scope='Recursive'";

If you work with GetListItems method of the Lists.asmx SharePoint web service, you have to define an extra node with the QueryOptions element:

XmlNode queryOptionsNode = camlDocument.CreateElement("QueryOptions");

queryOptionsNode.InnerXml = "";

If you want to query a specific sub folder using the SPQuery object, you have to set the Folder property:

qry.Folder = list.ParentWeb.GetFolder("Folders DocLib/2008");

When working with the web services you have to do the following:

XmlNode queryOptionsNode = camlDocument.CreateElement("QueryOptions");

queryOptionsNode.InnerXml = " Folders DocLib/2008";

-----

So my CAML looked like this:
<Query>
<Where><Eq><FieldRef Name='FileLeafRef'/><Value Type='Text'>testdoc.docx</Value></Eq></Where>
</Query>
<ViewFields>
<FieldRef Name='Title' />
</ViewFields>
<QueryOptions>
<Folder>Documents/[002643][Test%20project%20for%20Adam]/[ACPC][126573][1][writeFileAndXmlTest2]</Folder>
<OptimizeFor>FolderUrls</OptimizeFor>
</QueryOptions>

RuSs

DLL Binding Redirect

noreply@blogger.com (russell mccloy) — Tue, 10 Mar 2009 10:53:00 +0000

We had an issue where an old app was breaking because it was looking for an old version of Microsoft AJAX when we had upgraded to Microsoft AJAX 3.5.

To quickly get around this issue we did a binging redirect:

<dependentAssembly>
<assemblyIdentity name="System.Web.Extensions" publicKeyToken="31bf3856ad364e35" />

<bindingRedirect oldVersion="1.0.61025.0" newVersion="3.5.0.0"/>
<publisherPolicy apply="no" />
</dependentAssembly>

Pretty easy. Although we may move to the machine.config for a more global solution.

RuSs

Detect columns and copy a row - t-sql

noreply@blogger.com (russell mccloy) — Wed, 04 Mar 2009 15:18:00 +0000

This is a way to copy a row in a table without caring whether someone has added new columns.

USE [TRAM]
GO
/****** Object: StoredProcedure [dbo].[COPY_Subleases_2] Script Date: 03/04/2009 15:16:34 ******/
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
ALTER PROCEDURE [dbo].[COPY_Subleases_2]
(
@iSourceTable VARCHAR (150),
@lPrintIt INT,
@CurrentAssetID INT,
@NewAssetID INT,
@CurrentsubleaseID INT OUTPUT,
@executeIt int
)
AS
BEGIN

DECLARE @cName varchar(150)
DECLARE @lCnt int
DECLARE @lColumnList varchar(MAX)

SET NOCOUNT ON;

-- SET @iSourceTable = 'PropertyAnnualProperty'
-- SET @lPrintIt = 1
SET @lColumnList = ''

DECLARE _LIST
INSENSITIVE SCROLL CURSOR

FOR

SELECT column_name

FROM

information_schema.columns

WHERE

table_name = @iSourceTable AND
data_type <> 'timestamp' AND
column_name <> 'SubLeaseID' AND
column_name <> 'AssetID'

ORDER BY
ordinal_position

OPEN _LIST

IF @lPrintIt = 1 print 'Column names returned...' + convert(char(8),@@cursor_rows)

-- Check for an error
IF @@CURSOR_ROWS = 0

BEGIN

GOTO Done_Loop_LIST

END

-- Grab the first record
FETCH FIRST FROM _LIST

INTO @cName

SELECT @lCnt = 0

Loop_LIST:

SELECT @lCnt = @lCnt + 1
SELECT @lColumnList = @lColumnList + '[' + @cName + '],'

FETCH NEXT FROM _LIST

INTO @cName

-- Any more ?

IF @@FETCH_STATUS != -1

BEGIN

GOTO Loop_LIST

END

Done_Loop_LIST:

DEALLOCATE _LIST

-- Remove the trailing ','
print '@lColumnList: ' + convert(varchar(MAX), @lColumnList)
SELECT @lColumnList = substring(@lColumnList,1,datalength(@lColumnList)-1)

IF @lPrintIt = 1 PRINT @lColumnList

DECLARE @stm varchar(MAX)

SET @stm = 'INSERT INTO ' + @iSourceTable + ' (AssetId, ' + @lColumnList + ') SELECT ' + Convert(varchar(10), @NewAssetID) + ', ' + @lColumnList + ' from ' + @iSourceTable +
' where assetid = ' + Convert(varchar(10), @CurrentAssetID)

PRINT @stm

IF(@executeIt = 1)
BEGIN
EXEC(@STM)
END

SELECT @CurrentsubleaseID = @@identity

END