(Update: See the latest on the Deployment Framework here.)

It has been several months since I initially posted on the topic of using NAnt to coordinate the deployment of BizTalk 2004 solutions, with updates here and here.  Since that time, I've been heads-down in a BizTalk project and have had a chance to refine the practice quite a bit.

So, I took the opportunity to introduce quite a few improvements into the sample I had previously released - hopefully turning it into a "template" that can be used on your projects more easily.  To that end, I've created a GotDotNet workspace where you can download the current version of the build template, and I invite any feedback/suggestions/participation.

The major changes I've introduced can be categorized as follows:

• Split binding files: On a BizTalk project with multiple developers, I believe it is beneficial to split your binding files into content related to orchestrations vs. content relating to "everything else" (send ports, receive ports, etc.)  The new build template sample does just that.
• Improved scripts: Many of the WMI-related scripts I relied on in the past were from BizTalk SDK samples.  These were modified to get better/different error handling behavior.
• More generic handling of send ports, receive ports, receive locations, etc. in the script - though it still requires some maintenance.
• Better NAnt citizen: I'm still not a NAnt expert by any stretch, but thanks to Duncan Millard I realized that I had been remiss in applying quotes to a good deal of the text names I was using (i.e. &quot).  In addition, I started using built-in NAnt tasks for at least a few items where I had overlapped functionality, though not in cases where I needed slightly different behavior.  Finally, the build file naming convention proposed by the build template is "BizTalkSample.sln.deploy.build" (rather than "BizTalkSample.sln.build") to avoid conflicting with a NAnt file that would actually be used to build the project (e.g. with Cruise Control), as opposed to deploying it.
• Use of "Deploy with NAnt" in an MSI scenario: This is the biggest (and I believe most useful) change...Let me explain further.

As you study the build script, you will notice that it now also tackles such tasks as deploying virtual directories and applying permissions to them, patching binding files to match “local conditions”, etc. The initial goal here is to ensure that in a team development scenario, the amount of “out of band” setup required for any given developer to establish their BizTalk 2004 project environment is minimal – and that the current topology for the project is communicated efficiently (i.e. through the build file, as opposed to email or word-of-mouth.)

After a development team has spent weeks or months with this build script, refining it to represent their exact situation and deployment needs, a question might arise: Why not take this well-tested script and use it for production (or just non-development-machine) deployments as well?

To do this, it is helpful to agree that a two-phase deployment of BizTalk-related projects is both acceptable and useful. The first phase is an MSI-based installation that simply installs all of the BizTalk-related assets in a specific directory, but doesn’t deploy them to the BizTalk server. The second phase occurs when the user goes to the Start-Programs-YourProjectName menu, and chooses “Deploy YourProjectName”. The user also has the option to un-deploy (leaving the assets still on the file system until the MSI is uninstalled) or redeploy (to support the case where a single file is “patched”.)  See a picture of the Start menu created by the build template's MSI file.

The reason that two phases are useful is that if the deployment to BizTalk fails, we would like a complete log of the results (which in this sample occurs in the DeployResults directory) for analysis/diagnostics, and we would rather not have the MSI simply roll back. The same holds true of un-deployment.

Dual-purposing the NAnt file you have been maintaining for “real” deployments can be summarized as follows:

• Use the build template's approach (see BizTalkSample.sln.deploy.build in the download) for the “server.deploy” and “server.undeploy” targets, which make the build file able to work with “co-located” binaries in addition to the standard developer tree.
• Create a Visual Studio Setup project that deploys your BizTalk assemblies according to the pattern shown in the BizTalkSample.Setup project. Notice that this setup project includes the NAnt build file, the “DeployTools” directory (with our scripts), as well as a subset of NAnt in a subdirectory of our installation (so that we don’t rely on NAnt being present on the target server.)  It also includes a few convenience batch files referenced by our Start menu entries for invoking NAnt with specific targets.

The BizTalkSample.Setup project in the build file can be built, installed, and deployed to try this out.  (You will want to make sure you haven't already deployed in "developer" mode, because the "Deploy BizTalk Deployment Sample" passes a flag to the NAnt script that will skip the undeploy phase in order to speed up "clean" installations.)  Note that as your deployment grows more sophisticated - multiple hosts, servers, etc. - you will find the NAnt script requires additional properties to govern these variations between server deployments and developer-desktop deployments.

The build template sample also discusses the use of NUnit as a post-deployment verification process.  See the documentation in the sample for a full discussion --  basically all we are doing is making sure our MSI includes a subset of our unit tests, NUnit, and appropriate test files.  This is then wired up to the "Verify Deployment" Start menu option.  This can be an extremely effective means of ensuring your operations team has a means of testing the system interactively (not to replace standard automated heartbeat tests you might have…)

Peter Provost proposed a great generic approach to building BizTalk NAnt files a couple of weeks ago - and I shamelessly stole his approach to reverse the "orchestration ordering" property (very slick!)  If you do not wish to pursue the MSI option just discussed, and you do not wish to integrate the "out of band" setup actions discussed above (i.e. virtual directory creation, etc.), you might prefer his approach or you might want to push the build template sample provided here in that direction (i.e. more property driven.)  However, I believe tackling the kinds of concerns that we do here will generally mean that a real-world project will no longer have a generic-looking NAnt script. 

Again, you can download the build template sample here.  Enjoy, and I hope you find it useful.

A lot has been said regarding the MSMQT adapter for BizTalk 2004 already, but below are a few recent observations that may be of help to you. 

When people ask me what MSMQT is, my short answer goes something like this: "MSMQT is the name of the BizTalk 2004 adapter that implements the MSMQ network protocol directly within BizTalk.  It allows BizTalk 2004 to send/receive MSMQ messages directly, and move messages to/from the MessageBox very quickly - without external (DTC) transaction coordination.  Only private queues are supported for receives."

Using the MSMQT adapter you can:

• Use Send Ports to send to public or private queues on remote machines - but the queues must be transactional.
• Use Send Ports to send to local queues, which will be private since that is all MSMQT supports.
• Use Receive Locations to receive from locally defined queues, which will be private since that is all MSMQT supports.  These queues can be either transactional or non-transactional.  You cannot monitor remote queues (i.e. receive locations can't reference remote queues.)

You cannot using the System.Messaging MSMQ APIs/COM MSMQ APIs/MSMQ C Libraries or the MSMQ MMC console on a BizTalk 2004 machine that is running the MSMQT adapter!  This is because these all rely on the native MSMQ Service running.

You can use the aforementioned APIs (though not the MMC console) from remote machines (that are not running the MSMQT adapter) to put messages in MSMQT-defined queues.

To test MSMQT queues on the BizTalk 2004 server where they are defined, you can:

1. Create a Receive Port and a File Receive Location that will monitor a particular directory...
2. Create a static one-way Send Port that specifies MSMQT & the queue you want to test as the destination.  In the "Filters" portion of the Send Port configuration, define BTS.ReceivePortName = ReceivePortNameFromStep1

Now, when we you drop a file on your directory, it will be routed to your queue.  The receive location could also use HTTP posts, in which case the WFETCH tool from the IIS Resource Kit (or a similar tool) will let you easily get test messages to the queue.

Note: If you have existing applications that used to send to public MSMQ queues, but will now be sending to MSMQT queues, the most common problem is that they are referencing those queues via the PathName property, rather than the FormatName property (with 'DIRECT=OS:machine\private$\qname' syntax.)  This is only a problem for users of the COM/C-Library APIs - the System.Messaging.MessageQueue class actually allows format names to be used with the Path property.

See this Microsoft Support discussion for lots of additional insight into MSMQT.

There are several QFEs for BizTalk 2004 that you may want to know about, in case you encounter the situations described below.

Situation: You attempt to create or edit Receive or Send ports within the BizTalk Explorer in Visual Studio, and CPU consumption hits 100%.  Memory consumption climbs until the IDE crashes with an OutOfMemoryException.

Fix: Ask PSS for the hotfix associated with KB870619 (also known as hotfix 1185)

Situation: You have a CDATA section in an inbound xml document within an orchestration.  The CDATA section contains flat file data, with vital trailing (or leading) whitespace.  After you execute a transform in your orchestration, the CDATA designation is stripped (although the flat data is still there) - and the leading/trailing whitespace is now lost.  Curiously, using the "Test Map" feature (by right-clicking on a map in the solution explorer) doesn't exhibit this behavior.

Fix: Ask PSS for the hotfix associated with KB841563

Situation: You have a scope shape, and scope timeouts aren't working as expected.  Specifically, for blocking calls in an expression shape within the scope (like a DCOM call, etc.) where the timeout is being exceeded, you see 100% CPU consumption in the BizTalk service and the orchestration never terminates.

Fix: Ask PSS for the hotfix associated with KB811250.  (Note: this problem may have been addressed in the fix rollup released in April - I'm not sure what the ordering was here.)

(Update: See the latest on the Deployment Framework here.)

I realized today that in the "deploy.orchestrations" target of the NAnt build file (discussed here) there is a bug in the ordering of operations.  The original file imported the binding file prior to deploying the orchestration - which doesn't work for a "first time" deployment. 

In addition, Hermo Terblanche made a good point (in comments) regarding piplines which are used by Send/Receive ports - i.e. the ports need to be removed prior to removing/updating the corresponding pipeline assembly. 

Therefore, an additional dependency has been added to the "undeploy.piplines" target - namely, "remove.ports".  This target will remove exactly one Send and one Receive port, the names of which are derived from the binding file.  This section of the build file will have to be customized for your purposes, just like the names and ordering of orchestrations.  Note that in my sample, the orchestration uses “Specify Now” ports, and removing the Send/Receive ports is probably not necessary.  However, for “Specify Later” ports (the more typical case) that use custom pipelines, you will get the error referenced in Hermo's comment if you do not remove the ports prior to updating the pipeline assembly.

The zip file with a new build file is still located here.  Current NAnt output from the sample will look like this.  Enjoy!

(Update: See the latest on the Deployment Framework here.)

One of the more complicated aspects for a developer using BizTalk 2004 is the large number of steps required during the edit/run/debug cycle.  Since your BizTalk artifacts can't be run "in place" (that is, where they were compiled) but must instead be deployed to a local (or remote) BizTalk server, life can get quite complicated.

If you have done much BizTalk 2004 development, you know the routine quite well at this point.  If you have orchestrations, schemas, transforms, pipelines, and (say) C# components all partitioned into separate assemblies - and you have a number of orchestrations with "Call/Start Orchestration" shape dependencies that introduce specific start/stop-ordering - you can spend a lot of time doing the whole stop/unenlist/undeploy/redeploy/enlist/start/bounce BizTalk routine. 

Worse, you might not get it right - which can lead to hours spent debugging problems (over the course of your project) that don't really exist.

To alleviate this problem, it can be quite helpful to use a tool like NAnt to coordinate the "update my BizTalk server" process that must occur after each build.  (NAnt is a large topic - suffice to say it is an Xml-driven software build system.)  As long as your NAnt build file (and BizTalk bindings file) are kept up to date, the whole process can be reduced to:

1. Comple your solution (which might have multiple orchestrations, schemas, transforms, pipelines, and external components in separate assemblies)
2. Choose "Deploy to BizTalk" from your Tools menu.
3. Wait 60 seconds or so, enjoying the feeling that you have a consistent & reliable deployment mechanism.

In addition, you can of course use NAnt to kick off full unattended builds for nightly builds or continuous integration systems (like Cruise Control).  Since Visual Studio.NET is the only supported way to build BizTalk projects - despite that tempting XSharpP.exe file sitting in your BizTalk installation directory - this part of your NAnt build file must defer to calling devenv.exe in command-line fashion.  (A short "getting started" for NAnt & directions for using with a VS external tool can be found here.)

So, how about a sample that shows using NAnt to coordinate a reasonably complex deployment of inter-related BizTalk projects?  Available for download is a zip of a BizTalk solution that contains the following projects:

• BizTalkSample.Components (a C# project, with a class called from orchestration)
• BizTalkSample.Orchestrations (that contains two orchestrations, one which calls the other via a Call Orchestration shape)
• BizTalkSample.Pipelines (a send and receive pipeline, which are not needed but included to illustrate the deployment aspects)
• BizTalkSample.Schemas (which includes two schemas)
• BizTalkSample.Transforms (which contains a single transform used by one of the orchestrations)

The item I hope you will find most interesting, however, is a file called BizTalkSample.sln.build.  This is a NAnt build file that has a deployment target which captures the following dependency tree (only partially blown out here, for brevity - but this no doubt seems familiar to you if you've been using Biztalk 2004 for awhile…)

Deploy
   Deploy Schemas
      Undeploy Schemas
         Undeploy Orchestrations
            Unenlist Orchestrations
               Stop Orchestration
         Undeploy Transforms
            Undeploy Orchestrations
   Deploy Components
   Deploy Pipelines
      Undeploy Pipelines
         Undeploy Orchestrations
   Deploy Transforms
      Undeploy Transforms
         Undeploy Orchestrations
   Deploy Orchestrations
   Bounce BizTalk

Perhaps more illustrative is the output of the deployment itself - a sample of which can be seen here.  Note that the NAnt script relies heavily on the EnlistOrch.vbs and StopOrch.vbs script files that ship as samples with BizTalk 2004, as well as the BTSDeploy command line utility.

The BizTalkSample.sln.build script included in the download should represent a BizTalk project organization of sufficient complexity to act as a template for your own projects.  You will want to maintain the list (and correct ordering!) of the names of your orchestrations within the build targets for "deploy.orchestrations" and "unenlist.orchestrations" (this isn't deriveable automatically…)

Note that the build script relies on a set of naming conventions that will be evident once you spend some time with it - namely, that directory names correspond to assembly names.  In addition, the standard "Development" and "Deployment" configurations in BizTalk projects have been replaced with "Debug" and "Release", in order to not create inconsistencies with standard .NET projects (and allow for one "Configuration" property within the build script.)  This replacement was accomplished with a file-level search/replace.

There is probably room for more sophistication in this build file.  It takes a somewhat pessimistic view of what must happen for each edit/run/debug/deploy cycle, but I've found that despite the 60 seconds spent executing the script, your net productivity gain will be quite high given the time you won't waste trying to figure out what aspect of your deployment isn't correct.  Leave comments with suggestions for improvements, if you like.

In the latest round of documentation that was released for BizTalk 2004 there is an "orchestration operator" defined that was not previously documented: 'succeeded()'

The documentation states that this operator can be used to determine the outcome of a transactional scope or orchestration.  When might this operator be needed?

Well, it turns out that the orchestration compiler has some interesting rules about what you can do in an exception handler that might not be entirely intuitive at first (though as you reflect on analogies to C# or other exception-enabled languages, it begins to make sense.)

Suppose that you have defined a Request/Response port as the means of interacting with an orchestration, and you want to ensure that some response is generated regardless of the failure conditions you encounter.  Your first attempt might look like this (I know mine did…)  Stretch this JPG out to full size to see it clearly (IE will shrink it.)

This will generate a compiler error:
error X2162: must receive before sending a fault message on an implemented port

What is going on here?  It sure seems as if we have received a message already - we did it in the Rcv_SomeDoc shape.  However, we have the Snd_ResponseDoc shape inside of Scope_WorkThatMightFault, and the orchestration compiler is assuming that we might have already executed that shape prior to the catch block executing (i.e. prior to an exception being raised.)  A Request/Response port must only have one response for any given request…and our Snd_FaultDoc shape has the potential to violate this rule.  It sure would be nice if X2162 could be more explanatory in this regard…

How do we overcome this?  It isn't terribly obvious…We must wrap the Snd_ResponseDoc shape in an (additional) transactional scope, and check in our catch block to ensure that the associated transaction did not succeed before performing Snd_FaultDoc.  See this diagram (Acrobat required). 

What we are doing here is structuring the flow such that exactly one response will be sent for the original Rcv_SomeDoc shape.  The way we do this is to use a Decide shape, with an expression such as "!succeeded(Transaction_SndResponse)" in the rule branch.  The Snd_FaultDoc will be in the 'true' side of the branch (i.e. we did not successfully perform Snd_ResponseDoc), while the 'false' side will likely be empty. 

This is a pretty subtle bit of enforcement that the orchestration compiler is performing.  It is somewhat analogous to a typical language compiler ensuring that all code paths have a return value for non-void functions or methods.  And, of course, even though it is not enforced, it is certainly the case that 'catch' and 'finally' blocks in standard languages often have to be aware of what has or hasn't taken place in the associated 'try' block.  The orchestration compiler (apparently) just has some well-defined & strict rules it wants orchestrations to adhere to (such as "exactly one response for each request emanating from a Request/Response port".)

There is a somewhat similar case that is described briefly in the BizTalk documentation.  Imagine we wish to make reference to a message or variable in our 'catch' block that was initialized within the associated scope.  In this case, the orchestration compiler will assume that we might not have gotten around to initializing that variable/message prior to the exception being thrown - and a compiler error will be generated as a result: error X2109: use of unassigned local variable 'Variable_Blah'

In this case, we can wrap the portion of the scope's work that is responsible for initializing the variable/message of interest in an (additional) transactional scope (i.e. "Scope_InitWork"), and we can use a Decide shape with an expression such as "succeeded(Transaction_InitWork)" in the rule branch.  This will allow the orchestration to compile…

Wow! One might start to agree with Charles' boss...

(See here for the current version of the naming conventions!)

One of the primary benefits of the BizTalk Orchestration model is the great transparency you can get when a software implementation is pictorial.

Regardless of how well a developer comments code, there will always be a need to maintain a separate set of artifacts (UML diagrams, Visio diagrams with random shapes of your choosing, prose documents, whiteboard discussions, etc.) that are used to convey what the code is actually doing especially when working with a business audience.  This is true if for no other reason than that a discussion of interesting functionality will often take place at a different level of granularity than raw code can support

Round-trip engineering tools - that attempt to keep code in sync with diagrams often seem to suffer from a lack of fidelity that renders them ineffective.

With BizTalk Orchestration, the diagram is the implementation (at least at a particular level) of a piece of functionality.  Yes, you can disappear into components and lose sight of what might happen.  Yes, there is a code representation (xlang/s) underneath the orchestration but it seems to be completely isomorphic with the diagram.

So the opportunity exists to use an orchestration diagram in several interesting ways within a project lifecycle:

1. As a way to capture an initial high-level design, using all the orchestration shapes as intended but not yet bothering with real maps and schemas.  Stubbing out schemas (so you can declare messages and variables to be of proper types) and maps will allow you to flesh out the orchestration diagram(s) quite a bit, using the compiler as just a way to check for consistency.  All of the external system interactions, communication patterns, decision points, parallel vs. joined flows, etc. can be represented at this point in a shell orchestration.
2. As a way to gain consensus with the development team & business sponsor about whether the right functionality is indeed going to be built.  The high level design just described is a great tool for this discussion.  Put your orchestration(s) up on a wall with a projector and do a walk-through with as many of the project stakeholders as makes sense.  Or use a tool like CutePDF to print the orchestration as a PDF to send around via email.  (Of course, once Microsoft ships the Visio add-on for BizTalk 2004 orchestrations, this will represent another option for non-VS.NET users.  This has the added benefit of allowing you to exclude what you might consider to be lower-level detail by setting the Report to Analyst switch on various orchestration shapes to False.)
3. As a way to estimate work.  The various shapes in your initial orchestration can often represent reasonable granularity for time estimates.
4. And finally, as a way to guide project work...Rather than starting with the entire orchestration that you created to support steps 1-3, you might find it easier to create a new orchestration that represents the path(s) you are tackling at a particular point.  You can cut/paste portions of that original orchestration or simply use it as a reference for what comes next it serves as your outline.

To help realize some of these benefits, naming conventions within an orchestration are quite important

While the naming conventions are good practice for variables, Messages, Multi-Part types, etc. they are even more import for the workflow shapes.  The goal is to ensure that the intent of each shape is clear, and that the text associated with the shape conveys as much as possible given the space constraints.  In this way, a non-technical audience will be able to use the orchestration as documentation.

(See here for the current version of the naming conventions.)

Respond with comments & the document will remain updated per your feedback!