Paul Somers Connected Systems Blog

Could not create SSOSQL. To fix the problem, reinstall SSO or try 'regasm SSOSQL.dll' from a Visual Studio command prompt.Error Code: 0x80131700

2012-01-29T15:13:00.000-08:00

You have installed BizTalk?

You have installed SQL server?

You have installed .NET4.0 ?

Everything is fine.

Then you reboot, or you restart SQL Server.

Enterprise Single Sign-on Does not want to start…. ???

It errors with:

Could not create SSOSQL. To fix the problem, reinstall SSO or try 'regasm SSOSQL.dll' from a Visual Studio command prompt.Error Code: 0x80131700

OR

Windows could not start the Enterprise Single Sign-On Service service on Local Computer.
Error 0x80131700: 0x80131700

OR you are getting:

Creation of Adapter FILE Configuration Store Entries failed.
An Error occurred while attempting to access the SSO database.

The problem is .NET 4 does something to break SSO.

This issue occurs after installing .NET Framework 4.0. The registration of the assembly used by ENTSSO to access SQL Server (SSOSQL.dll) does not specify the correct version of the .NET Framework. When .NET Framework 4.0 is installed, the assembly will try to use the newer framework and then fail to load

DO NOT do what it says in the event log it will NOT work fully.

The fix is: http://support.microsoft.com/kb/2252691

However you need to BE SURE you do it correctly.

Is your environment 64bit? Most are these days. Did you install 64bit SSO?

Is it in:

64bit: C:\Program Files\Common Files\Enterprise Single Sign-On
32Bit: C:\Program Files (x86)\Common Files\Enterprise Single Sign-On

If you have 64bit, follow the 64BIT instructions ONLY

1. Open a command window
2. Go to C:\Windows\Microsoft.NET\Framework64\v2.0.50727
3. Type each of the following and hit ENTER

32bit: regasm “C:\Program Files\Common Files\Enterprise Single Sign-On\win32\ssosql.dll”
64bit: regasm “C:\Program Files\Common Files\Enterprise Single Sign-On\ssosql.dll”

Note On a 64-bit server, regasm will need to be run for both the 32-bit and 64-bit versions of ssosql.dll.

Do both of these steps and you day will be brighter.

Dynamic SMTP Failure: Unknown Error Description

2011-12-12T20:19:00.000-08:00

I was sending an email dynamically, from an orchestration.

I set up a nice payload message with all of these nicely distinguished fields, that contained the to/ from, subject and body of the email, so that other parts of my system could send an email, just by generating this message.

My message assignment shape had:

EmailSendMessage = EmailReceiveMessage;
EmailSendMessage(SMTP.EmailBodyText) = EmailReceiveMessage.Body;
EmailSendMessage(SMTP.CC) = EmailReceiveMessage.CopyTo;
EmailSendMessage(SMTP.From) = EmailReceiveMessage.From;
EmailSendMessage(SMTP.Subject) = EmailReceiveMessage.Subject;
EmailSendMessage(SMTP.MessagePartsAttachments) = 0;

EmailSendPort(Microsoft.XLANGs.BaseTypes.Address) = "mailto:" + EmailReceiveMessage.SendTo;

The message itself was not sent, just the properties of this message, so the message I was constructing as receive message didn’t matter, so I made it the same type of the EmailReceiveMessage. It compiled, and I deployed it.

First Try:

Event Type: Error
Event Source: BizTalk Server 2009
Event Category: (1)
Event ID: 5754
Date: 13/12/2011
Time: 1:16:31 PM
User: N/A
Computer: [computer]
Description:
A message sent to adapter "SMTP" on send port "XXX.Email.Orchestrations_1.0.0.0_XXX.Email.Orchestrations.SendEmail_EmailSendPort_43e93d0db20c465a" with URI "mailto:email@address.com" is suspended.
Error details: Unknown Error Description
MessageId: {B66F52BA-DFF0-4274-B4B2-3B1F51E862E0}
InstanceID: {C7A4CA9E-E606-4D04-9001-D34974B4D971}

For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.

I then checked and you needed to have the body text charset:

EmailSendMessage(SMTP.EmailBodyTextCharset) = "UTF-8";

Second try, same error…

Several modifications later I get the message: Unknown Error Description

Which I must say is not a great deal of use when you are trying to figure out what’s wrong.

I then decided I’d make a different message type for sending, and used a transform. I used my send message and copied and pasted the schema, I changed the target name space of course.

SAME ERROR…..

I was annoyed, and tried a bunch of things, then EUREAKA! …

The send schema which I copied and pasted has the same properties promoted as distinguished fields.

This was BAD it seems, the instant I removed these properties from being promoted, and changing nothing else… everything worked….

BAD BAD BAD... schema properties, who would have thought…

DO NOT HAVE PROMOTED PROPERTIES ON YOUR SCHEMA WHEN SENDING TO THE SMTP ADAPTER.

ABA bank payment file format (Australian Bankers Association)

2011-11-11T20:38:00.000-08:00

I'm currently working on an application for a un-named organisation. As part of this, I need to export files for processing in the Australian defacto standard for Electronic Funds Transfer (EFT) files - the ABA format.

I'm using BizTalk of course, and it can handle this weird format, however all I had was the sample file, which for a fixed width flat file is not great.

I found the format documented here: http://ddkonline.blogspot.com/2009/01/aba-bank-payment-file-format-australian.html

Apparently the banks have all agreed on this format, which is firstly Ancient in origin and format, and design Highlighting that it’s a flat file, of fixed field lengths..

If this was re-factored into a XML format, it would be much easier to generate, highly flexible, and they could expose an interface, via a simple web service to accept this format. All authentication could be done via a secure https web service, with encryption on the web service.

Let’s understand that this is used to effect payments from a company’s bank account to individuals, this is highly sensitive and needs to be secured.

This format of this file is NOT encrypted in anyway; it is open, readable, and modifiable. There are no check digits, no certificate of authentication or any of the modern features you would expect in such a file.

By Exposing a WCF endpoint, which had authentication via certificate, using an https/transport encryption/security would handle some of this requirement; the rest is in the detail of the message itself.

Currently what happens is we output this file to the file system, and then someone picks it up... and processes it...

If a WCF or web service endpoint was open by the bank, we could securely communicate with this, and send the payment file, someone could log onto their secure interface and approve the transfers still, however there would be zero chance of someone modifying this file before it got there....

Before we got involved, this was just popped onto the file system somewhere, generated via a different method…

I am not one to mess around when it comes to security, and this smells to me. The banks need to provide an interface. I’ll happily build it, securely and flexible enough for all platforms to communicate with it.

Handy Hint: BtsCompile

2011-10-02T13:25:00.000-07:00

Are you looking for the build action on an orchestration, you know BtsCompile
For the uninitiated, it is great for having a BizTalk Artefact that’s broken, or not
complete and having it not form part of the compile. It works for maps, schemas,
you name it.

I generally use it to set it for some unfinished orchestrations not to include in the
build/compile, so it does not break things....

Well, as it happens, in BizTalk 2009, it can disappear, for some reason, however I have
discovered the way of making it re-appear.

Simple copy and paste ANY orchestration in your solution, and paste it back, now the
BuildAction property appears, not just for that orchestration, of ALL of them.

It’s as if it knows, if you are copying and pasting an existing orchestration you
are probably not going to want to have it build right away.

Fixed Field Settings on Map

2011-09-06T11:28:00.000-07:00

On a map, you can set the value of a destination node to a fixed value, from within the mapper, simply by setting the value property when clicking the destination node on the right hand side schema. I’ve see people use the string functiods to pre-set a value, however you do not need to do this.

There is a catch I found in BizTalk 2009. If you set the fixed value on a map page other than the first page, it does not set, worse yet, if that field on the destination schema happens to be a promoted property, it does not promote the value, and will fail when you try and use the value inside your orchestration as desired.

You will not know why, the shape will just fail, with some weird exception.

Always use the first page.

DefaultPipelines.XMLReceive, Attempted to read or write protected memory

2011-08-31T19:05:00.000-07:00

I got the error of death.... I mean we were this close to formatting the machine.

Event Type: Error
Event Source: BizTalk Server 2009
Event Category: BizTalk Server 2009
Event ID: 5719
User: N/A
Description:
There was a failure executing the receive pipeline: "Microsoft.BizTalk.DefaultPipelines.XMLReceive, Microsoft.BizTalk.DefaultPipelines, Version=3.0.1.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" Source: "XML disassembler" Receive Port: "GetReferenceDelta" URI: "mssql://zz11aazz//My?" Reason: Attempted to read or write protected memory. This is often an indication that other memory is corrupt.

I thought that the xmlpipeline was broken, so I tried the passthrough...

There was a failure executing the receive pipeline: "Microsoft.BizTalk.DefaultPipelines.PassThruReceive, Microsoft.BizTalk.DefaultPipelines, Version=3.0.1.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" Source: "Unknown "
Receive Port: "GetMyDelta" URI: "mssql://server/uri" Reason: Attempted to read or write protected memory. This is often an indication that other memory is corrupt.

I can no longer receive ANY messages into BizTalk from SQL...

I tried the file receive, I tried xml and pass-through pipelines, nothing works, and it all fails....

I can no longer receive ANYTHING at ALL into BizTalk.... Error of death...

I looked, I tried updates to the Adapter packs... nothing worked....

I was about to trash my environment and format and start again....

THEN!!! I thought, let me remove all updates I had done to BizTalk, I removed one, and tested, removed one and tested.

Finally finding that..... The offending update was cumulative update 1:

http://support.microsoft.com/kb/2429050

I removed it, and my BizTalk came back to life..... PHEW!

Encoded Multi Part Messages to SOAP adapter

2011-07-30T03:32:00.000-07:00

I had one of those old/non conforming to the standards web service of kinds I needed to call.
The ones you really hate, and wish they’d just upgrade them to WCF... but hey thats 90% of integration, old rotten systems... so I persisted.

It did accept xml messages, however was very fussy as it seemed to employ the good old, no xml translator, the one that is more of a text base, string manipulation xml decoder, where if you send quite valid xml, it will fail because they didn’t support it.
This kind of set up is... sadly... quite common....
It would not handle xml tags that had no content, but had a closing tag.

EG: <salary currency="en-AU"></salary>
It needed to have them encoded differently:
<salary currency="en-AU"/>
Now both of these examples are valid xml, however good old string parser went the easy route and didn’t cater for an end tag.
The problem is that you can’t exactly tell biztalk to not put the end tag, it’s going to put xml, valid xml, and thats what it does....
Further more... it also complained about white space in my xml message.... ??? you say???
<salary currency=""><salary> <standout logoid="”1111"><standout>
The white space it didn’t like was between the tags, <salary>[--- white space ---]<standout>

Now how on earth do you cater for this?

A custom pipeline component to the rescue, just before it sends the message I would “FixUP” the xml by removing the end tags, and white space between tags, using my old favourite, c#.net .
So I go to the trouble of making a custom pipeline component, hook it up to a custom send pipeline, and then attach the pipeline to the send port.

BANG, nothing happens... it stil fails for the same reason.

At this point I’m a little annoyed, and more p’d off than i’d care to mention.
I spend the next hour trying to figure out why......

Firstly the web service is a SOAP web service, I have to use the SOAP adapter because it’s a multipart message, I check with Yossi. http://blog.sabratech.co.uk/2009/08/biztalk-wcf-adapter-and-multipart.html to discover that even if I went to WCF and fudged a wcf-http port to work it would not make much difference.

I then try and pump the message out to the file system, using a send port in the orchestration. I get a message, with only the first part. It throws me for a second, then I realise that the FILE adapter does not support multi part messages. Weird, considering all messages are multi part messages, most with one part, but I prevail.

How do I see what is going to the adapter? I really need to see the message, I mean I have tested by component, and it is good, it works and does the trick, its .net, so I of course have a .TEST project to test it...

How... Ah..ha... You can stop the send port, not unenlist it, but stop it, the message will still go to the port however sit there waiting for you.

I can then see the message from the admin console, ALL parts of the message, a much better way to debug messages BTW, I notice the message part has been encoded, the multi part message takes two strings, the second of which is my xml message. So the message is wrapped in a <string> tag, with the contents.... no longer in XML format, they are html encoded. My lovely <salary> tag is now a horrid, & lt;Salary& gt; the xml is hardly readable.... I cringe, it’s destroyed my lovely xml...

Now I can see what’s going on, before it even gets to my custom pipeline, my xml is encoded, therefor my pipeline which is looking for xml, does not find any, and does not work.... UGH!

Interestingly when I assign the value of the xml in the outbound message, I do it from a message assignment shape, I have the xml in my grasp...
I do a cheeky thing in BizTalk, turn the message into an xml document, which I can turn into text. Luckily I wrote the pipeline by using a separate class for the tagging and white space cleaning. So I can call the methods from the message assignment shape, and I process the xml, clean and massage it.

I then set the xml document back up by loading the NEW xml into to, and then set the value of the parameter of the outbound message to the xmldocument.

The message is now encoded before it goes to the port. I send the message to the port and wolla, response message, in xml... that tells me it worked.

This is a bit of a hack and a workaround, it does work, however I would have preferred the pipeline to work.

I do note with interest that only the SOAP adapter is able to process this multipart message, because it’s an OLD soap web service, I did hear of talk of scrapping the SOAP adapter, however multipart messages are VERY common on soap web services, and this would be foolish to scrap it. I’d like to see support for multi part messages in the WCF set of adapters, with more backwards compatible support, so I don’t have to use an out dated SOAP adapter.

Exception: System.EnterpriseServices.TransactionProxyException

2011-06-16T15:11:00.000-07:00

I get an exception when configuring BizTalk 2010.

BizTalk application server.
Separate SQL server for the databases.

Exception: System.EnterpriseServices.TransactionProxyException

Looking at http://support.microsoft.com/kb/293799

When a System.EnterpriseServices.TransactionProxyException exception is triggered during a transaction completion, it cannot be caught from other application domains. Instead, you receive a System.Runtime.Serialization.SerializationException exception that resembles the following:

Unhandled Exception: System.Runtime.Serialization.SerializationException: Type 'System.EnterpriseSer vices.TransactionProxyException' in Assembly 'System.EnterpriseServices, Version=2.0.0.0, Culture=ne utral, PublicKeyToken=b03f5f7f11d50a3a' is not marked as serializable.

Gave me a hint to the problem, however was not the solution…… or any where near to it be careful following this I did not.

I saw that the databases were created during the configuration, and then deleted; only leaving the management database in a weird state. So the communication to the database server was fine.

I did some more investigation and found the dtctester tool… great for testing if my dtc was set up correctly, as this is one thing that BizTalk leverages extensively.

I ran the 'dtctester' tool, and it came up with the following:

Executed: dtctester.exe
DSN: PMS
User Name: *******
Password: ********
tablename= #dtc16032
Creating Temp Table for Testing: #dtc16032
Warning: No Columns in Result Set From Executing: 'create table #dtc16032 (ival
int)'
Initializing DTC
Beginning DTC Transaction
Enlisting Connection in Transaction
Error:
SQLSTATE=25S12,Native error=-2147168242,msg='[Microsoft][ODBC SQL Server Driver]
Distributed transaction error'
Error:
SQLSTATE=24000,Native error=0,msg=[Microsoft][ODBC SQL Server Driver]Invalid cur
sor state
Typical Errors in DTC Output When
a. Firewall Has Ports Closed
-OR-
b. Bad WINS/DNS entries
-OR-
c. Misconfigured network
-OR-
d. Misconfigured SQL Server machine that has multiple netcards.
Aborting DTC Transaction
Releasing DTC Interface Pointers
Successfully Released pTransaction Pointer.

Ok, so my problem was DTC….

I found a great response to a similar problem that told me to look at a couple of DTC issues. (http://dbaspot.com/sqlserver-server/215054-msdtc-doesnt-seem-work.html)

http://support.microsoft.com/kb/817064
http://support.microsoft.com/kb/301600

If you are running Win2K3 SP1, you will need to reset the DTC security parameters:

http://support.microsoft.com/kb/899191

Since you are beginning the transaction on a remote host, that host will also need to properly configure its DTC services.

Finally, if the communication must transit any firewalls, then you will need to restrict RPC ports on both Client and Server for Internet Ports, and then authorize these Ports in the firewall ACLs.

http://support.microsoft.com/kb/300083
http://support.microsoft.com/kb/250367/EN-US

I checked the firewall, no problem, I checked the other things, and made a change to the DTC security not via the method mentioned but via the component services.

Everyone should know by now to change the security settings under component services for the dtc on the sql server and the BizTalk server.

Go to the Component services/My Computer/Distributed Transaction Controller/LocalDTC

Properties/security

Ensure that everything in here is checked, (Remote admin is optional)

I usually select no Authentication required, and ensure that the network service account is running the service.

I did this, thinking I’ve found the problem…..

Nope same problem re occurs…

I find out that the SQL server is clustered, so I go to the cluster manager, and look at the dtc settings, and open the localdtc from there, the same thing, all the settings are correct because I had already changed them.

Then I notice, CLUSTER DTC, and expand that and mange that…. The cluster has its OWN instance of DTC, of course, it does not use the local DTC at all.

I look at the security settings of the cluster DTC, and change those to match, and wolla… it works…. Nice to know if the SQL server is clustered…

Always look at DTC, if your BizTalk does not wish to install to a remote SQL server.

2011-04-21T13:16:00.000-07:00

Your WCF-SQL adapter is not comming up after you have installed it in BizTalk?

You installed the WCF LOB Adapter SDK? From the ASDK folder on the installation cd.

You Installed the adapter Pack? From the AdapterPack folder on the installation cd.

You did this: http://soa-thoughts.blogspot.com/2010/08/wcf-sql-adapter-table-operations.html

Which version did you install?x86, x64, were they both x86, or x64 ?

Still not comming up... What version of BizTalk did you install? x86 or x64?

If you installed x86, then you need to install x86 version of both of these, and then.... Go to the BizTalk Administration Console, Platform Settings, Under Adapters, right click and say new adapter...

See this again: http://soa-thoughts.blogspot.com/2010/08/wcf-sql-adapter-table-operations.html

Do you NOW see WCF-SQL adapter there ??

If in doubt install x86.... Handy Hint....

Exam 70-595 BizTalk 2010 Exam Released

2011-03-30T14:01:00.000-07:00

After much pushing, and some prodding of the people who look after exams at Microsoft. Several members of the community got involved with Microsoft to create a BizTalk 2010 Exam.

It has now been released... have a look here

The more interesting thing to note, this exam will be quite different from exams on this topic from previous exams as the community created it, the community they knows in depth about the product.

Just look at the main areas:

Configuring a Messaging Architecture (20 percent)
Developing BizTalk Artifacts (20 percent)
Integrating Web Services and Windows Communication Foundation (WCF) Services (14 percent)
Implementing Extended Capabilities (13 percent)
Deploying, Tracking, and Supporting a BizTalk Solution (16 percent)

With a large focus on Architecture and actual development, without forgetting about deployment and support for BizTalk Solutions, an area that was sadly lacking any real content however significantly needed.

I have passed previous exams on BizTalk, and was really pushing for an exam that was not like other ones and set the mould for new exams.

Go on, get out there and have a go... here

To the Cloud......

2011-03-20T21:35:00.000-07:00

I have just returned from the MVP summit held in Redmond each year, there were some highlights and some low lights, as with each summit. The highlights are always all the cool new stuff, most of which I’m not allowed to talk about publicly, which is great (NOT), it does mean I can’t say much on the blog. I can only say what has already been announced publicly at PDC, Microsoft is doing the cloud thing, the next off the conveyer belt is composite apps, what’s in the box, wait and see, I will say it’s interesting, and then more interesting when you add in the comments of the MVP’s present when they told us how interesting it was.

We are talking about the cloud here, and of course I want to run apps on it, I want to run workflow in the cloud, I’ve wanted this since they had it a few years back, and then took it away because it was so limited. It makes sense, in the right scenario.

I also want to access my applications inside the organisation (on premise), provide a rich integration layer to them, to enable my cloud apps to communicate with my on premise systems.

This kind of application is called a hybrid model, and it is/will become very common. I would like to use the same technology I use in the cloud to do the whole integration and workflow as I do to access my on premise applications. Currently I use BizTalk for my on Premise applications and then have to write something different to enable my cloud applications to do this, hence the use of the term “Hybrid” it’s using a bit of both. This is currently possible by various means of the service bus in windows Azure and the new bits to enable BizTalk to expose a port or Orchestration on the service bus to accept connections. I can then establish a communications pattern into my organisation’s “legacy” on premise applications.

My problem with this approach is detailed in my recent webcast at http://www.cloudcasts.net/Default.aspx?category=BizTalk+Light+and+Easy my on premise middleware still needs to exist, and it needs to scale in line with my cloud system, it’s not 1:1 more like 2-3(cloud instances):1(on Premise) but it needs to scale, hence I need to still invest in on premise hardware, however I don’t want to have to scale it I want to leverage the cloud to scale on demand, and scale back when I don’t need it. It’s one of the key selling factors for using the cloud.

Whilst I cannot put everything in the cloud, it’s never going to happen, I want to have the option of scaling to the cloud and then scaling back to on premise when I have low load levels, hence justifying my on premise costs.

I do have this for websites in the cloud; this is a little more difficult for an integration platform that needs to access legacy systems and is written in a non-cloud friendly way.

I would love to provide this on the cloud, but this is one ask that is some time away which ever provider you look at. My view is whoever cracks this will dominate the cloud market.

The rest of the detail will come…. I don’t know when and I can’t say how but it’ll come wait and see, with more announcements coming…. It’s how you leverage the cloud to work for you that will make the real difference in adopting a cloud/hybrid model or not.

Integration is a hard sell enough, to add cloud to the mix makes it even harder, I’m not the only one out there trying this on, customers are not buying yet, and the amount of convincing, assurances and explaining needed is staggering.

BizSpark Event Results

2011-02-06T14:22:00.000-08:00

Following the BizSpark event, I needed to recall the results and more the criteria for the event.

One of them was number of Windows Azure investments used. This was perhaps one of the major ones, there were perhaps 3 of the top major criteria.

Our solution was 100% hosted and built for Windows Azure, it used:

Web Roles x 4, (Two Fully working Websites)
Web Service(s) x 2, including hosting of Windows Workflow in the cloud. A https wcf endpoint for doing credit checks.

We used and had working by the end of the 2 days:

Azure Blob storage
Azure Table Storage
Azure Queue Storage

We even had a windows phone 7 application that worked with our application.

Our solution worked end to end, had quite a complex architecture and did what we said it would do.

So we were on the ball with using the technology required, and it worked.

I'm not being bias here, but from the look at the other presentations our solution used the most Azure features, was actually built in the two days and was actually running on Windows Azure.

The winning entries had all been developed prior, one of which was not even running on windows Azure. It did look nice, was much more complete than any of the other solutions, and had several months of development from two people who formally worked for Microsoft developing the platform of which the application lived on.

Given several of the judges were from Microsoft this did not sit well.

Don't get me wrong, I liked the two entries that came first and second, their solutions and concepts were great.

I think for future competitions it needs to be fair, given the majority of people had not developed a windows azure application before and their presentations were not fully working, as expected, they should be judged on this factor alone. If you want to bring along something you have spent several months developing and put it forward then you should sit in a different category.

Fair is Fair, this event was not fair and those people who worked on the solutions may not have such a great first experience with Windows Azure.

WIndows Azure BizSpark Event

2011-02-04T17:22:00.000-08:00

We are here at the BizSpark event, working on a solution that we are only given two days to develop. We have to get it done and presented in less than this time.

We are working hard at it, and have two services running, with https endpoints, as the data is personal.

Twitter is working hard with #WABizCamp. To make some noise....

Let’s see how we go, we have a great plan and a lot of work to go.....

The Future of Middleware?

2011-01-30T20:57:00.001-08:00

Today I read a rather interesting and profound statement in the book recently released "BizTalk 2010 Recipes" by Mark Beckner

It was so interesting that I'd like to share it with you, it speaks of the future of which I’d tend to agree with, "in the decade ahead, middleware will be more important and relevant than ever before."

"Why does middleware like this have such staying power? You’d think that newer advances in technology like web services, SOA, and software as a service (SaaS) would render applications much more inherently interoperable and that the pain and complexity of systems integration would be a thing of the past.

The truth is that enterprises of all sizes still experience tremendous cost and complexity when extending and customizing their applications. Given the recent constraints of the economy, IT departments must increasingly find new ways to do more with less, which means finding less expensive ways to develop new capabilities that meet the needs of the business. At the same time, the demands of business users are ever increasing; environments of great predictability and stability have given way to business conditions that are continually changing, with shorter windows of opportunity and greater impacts of globalization and regulation. These factors all put tremendous stress on IT departments to find new ways to bridge the demanding needs of the users and businesses with the reality of their packaged applications.

This leads back to the reason why middleware—certainly not sexy as technologies go—continues to deliver tremendous value to both businesses and IT departments. As the technology’s name suggests, it sits in the middle between the applications you use and the underlying infrastructure; this enables IT departments to continue to innovate at the infrastructure level with shifts like many-core processing, virtualization, and cloud computing. Instead of having to continue to continually rewrite your LOB applications to tap into infrastructure advances, you can depend on middleware to provide a higher level of abstraction, so you can focus your efforts on writing the business logic, not plumbing code. Using middleware also helps future-proof your applications, so that even as you move ahead to the nextgeneration development tools and platforms (including the current trends toward composite applications and platforms as a service), you can still leverage the existing investments you’ve made over the years.

So, in the decade ahead, middleware will be more important and relevant than ever before. "

Burley Kawasaki

Director of Product Management, Microsoft Corporation

AZURE: Error: (400) Bad Request

2010-10-26T15:50:00.000-07:00

Have you been playing with Azure storage, trying to create Blobs, Tables or Queue’s ??

Have you been getting any of the following errors:

[Azure Blob Storage]
Microsoft.WindowsAzure.StorageClient.CloudBlobClient ccb = account.CreateCloudBlobClient();
CloudBlob cb = ccb.GetBlobReference("myblob2871618");
cb.DeleteIfExists();

FAILS with: StorageClientExecption
"One of the request inputs is out of range."
"The remote server returned an error: (400) Bad Request."
Error code: BadRequest

[Azure Table Storage]
Microsoft.WindowsAzure.StorageClient.CloudTableClient cct = account.CreateCloudTableClient();
cct.DoesTableExist("MyTable");
"Exception of type 'Microsoft.WindowsAzure.StorageClient.StorageClientException' was thrown."
StatusCode: System.Net.HttpStatusCode.BadRequest

[Azure Queue Storage]
Microsoft.WindowsAzure.StorageClient.CloudQueueClient ccq = account.CreateCloudQueueClient();
CloudQueue cq = ccq.GetQueueReference("myqueue");
cq.Exists();

"One of the request inputs is out of range."
"The remote server returned an error: (400) Bad Request."
System.Net.HttpStatusCode.BadRequest

Well, the solution is as I found a little simpler than imagined.

When I created my Storage account online I created it with my standard Camel Case name. PaulStorage. On my connection string I had the name of the storage and my key.

So it looked like:

DefaultEndpointsProtocol=https;AccountName=PaulStorage;AccountKey=XYZ…..

When you try and open the connection for my Cloud Storage Account, and pass this in, it will work fine, when you try and do anything like the statements above you get the errors listed.

I knew that for other things like Queue’s in Azure you need to specify them in lowercase or they don’t work. I asked around and found that Azure was very fussy when it came to case. Even though I had specified the Storage account as Camel case I needed to pass it in as lower case to have it work:

DefaultEndpointsProtocol=https;AccountName=paulstorage;AccountKey=XYZ…..

Once I made this remarkable at the time, discovery.. Everything started working.

One thing to note is that development storage is a lot more forgiving and you don’t have to worry about any of this.

Development storages connection string is as simple as it gets:

UseDevelopmentStorage=true

Azure Queue Explorer

2010-10-25T15:00:00.000-07:00

Having played a great deal with Queues in the past, the opportunity to expand the abilities of a real world solution to use the concept of Queues in the cloud as a great way of doing what Queues do best, hold onto something for someone later on to pick it up, in a managed and ordered fashion.

It seems Azure Queues do just that, and I was so happy that I made a tool that enabled me to manage my Queues:

View and Create Queues
Send and Receive Messages
Peek what messages I Have
Delete Messages from the Queue

I have made this publicly available for all to use and play with, along with the all important source code, it's no trade secret on how to use them, it's just nice to have something ready to go that does it for me.

Check out the link here: Azure Queue Explorer

Source code

BizTalk 2010 On MSDN

2010-10-01T17:44:00.000-07:00

You can now download from MSDN the full suite of BizTalk 2010. It came up today.

BizTalk Server 2010 Branch Edition (x86 and x64) - DVD (English) 660 (MB)

BizTalk Server 2010 Developer Edition (x86 and x64) - (English) 552 (MB)

BizTalk Server 2010 Enterprise Edition (x86 and x64) - DVD (English) 965 (MB)

BizTalk Server 2010 RFID Enterprise (x86 and x64) - DVD (English) 81 (MB)

BizTalk Server 2010 Standard Edition (x86 and x64) - DVD (English) 965 (MB)

Host Integration Server 2010 Developer Edition (x86 and x64) - (English) 152 (MB)

Host Integration Server 2010 Enterprise Edition (x86 and x64) - CD (English) 321 (MB)

Interestingly the developer edition which is a smaller download, actually contains the same contents as the Enterprise Edition.

What you have to do in SQL Azure that you take for granted in SQL Server

2010-10-01T15:02:00.000-07:00

I have been using the cloud quite extensively and have had to come up with strategies for many things to adapt them to the restrictions on the cloud. I'll be doing a few posts on this as I go.

The first of these was the database setup in SQL Azure.

You usually need to store your data somewhere, I’m not going to go into the merits of Table Storage or SQL Azure, and I’ll cover that separately as the choice comes with penalties each side and needs a whole post about that.

What I do want to cover is the differences in the SQL Script you need to have to setup your databases in the first place in SQL Azure. The first thing you notice is that it’s a lot of extra work, and you notice this pretty early on. Many of the statements you are used to in SQL Server don’t work, or are partially supported, or not supported at all.

Take for example a simple user, you get by default an super user with full rights to the database. This is great to set it all up, but I want my application to have only read access to my database, so I need a new user, NEVER run your application with the privileges of the super user.

I then need to create the user, so I do the click on the database/security/users and say create user, a nice script comes up:

-- =================================================
-- Create User as DBO template for SQL Azure Database
-- =================================================
-- For login , create a user in the database
CREATE USER
FOR LOGIN
WITH DEFAULT_SCHEMA =
GO

-- Add user to the database owner role
EXEC sp_addrolemember N'db_owner', N''
GO

The first thing about SQL azure is there are NO popup wizards to help you, it’s all script, however you should ideally not be using the admin console, I always use script. In SQL Server my script for this would look like so:
SET QUOTED_IDENTIFIER ON
SET ANSI_NULLS ON
SET NOCOUNT ON
SET TRANSACTION ISOLATION LEVEL READ UNCOMMITTED
SET ANSI_WARNINGS OFF
SET XACT_ABORT ON

/*
*********************************************************************************************************
-- TAB SIZE 4
IDENT SIZE 4
---------------------------------------------------------------------------------------------------------
Purpose: Create a new SQL Server Login, assign a default database
and associates the Login with a database read permission

Notes: none.

Revision History:
----------------------------------------------------------------------------------------------------------
| Date | Changed By | Reference | Comments
----------------------------------------------------------------------------------------------------------
| 20100804 | Paul Somers | Create a new login with DB access | Initial release
***********************************************************************************************************
*/
USE master
GO

IF NOT EXISTS (SELECT * FROM syslogins WHERE [NAME] = N'NewReader')
BEGIN
CREATE LOGIN NewReader WITH PASSWORD=N'Password@2',
DEFAULT_DATABASE=GEO1,
DEFAULT_LANGUAGE=us_english,
CHECK_EXPIRATION=ON,
CHECK_POLICY=ON
END
USE GEO
GO
IF NOT EXISTS (SELECT * FROM sysusers WHERE [name] = N'NewReader')
BEGIN
CREATE USER NewReader FOR LOGIN NewReader
END
USE GEO
GO
EXEC sp_addrolemember N'db_datareader', N'NewReader'
GO

SET QUOTED_IDENTIFIER OFF
GO
SET XACT_ABORT OFF
GO
SET ANSI_NULLS ON
GO

The very next thing you notice about SQL Azure is that if you run this it will not work, whilst this works fine in SQL Server. You get all sorts of errors.

Take the simple, USE statement. USE master

Msg 40508, Level 16, State 1, Line 1
USE statement is not supported to switch between databases. Use a new connection to connect to a different Database.

You can’t use it… ? Between databases

Ok, so I need to connect to master run that bit and then run the master part script? Ok so we try that, and then discover that there are MANY parameters that are not allowed, well all of them to be precise.

IF NOT EXISTS (SELECT * FROM syslogins WHERE [NAME] = N'NewReader')
BEGIN
CREATE LOGIN NewReader WITH PASSWORD=N'Password@2'
--DEFAULT_DATABASE=GEO1,
--DEFAULT_LANGUAGE=us_english,
--CHECK_EXPIRATION=ON,
--CHECK_POLICY=ON
END

My traditional way of creating a user will NOT work, or will be extra cumbersome.
It even complains about this statement:

Msg 40530, Level 16, State 1, Line 3
The CREATE LOGIN statement must be the only statement in the batch.

It actually wants you to do a single line statement in the master database:

CREATE LOGIN NewReader WITH PASSWORD=N'Password@2'

That done, I have to switch over to the database I want to give the user permissions to so I run this:

CREATE USER [DataReader]
FOR LOGIN [DataReader]
WITH DEFAULT_SCHEMA = dbo
GO

EXEC sp_addrolemember 'db_datareader', 'DataReader'

To finish it off and make it more robust for repeated deployments I can use:
IF EXISTS (SELECT * FROM sys.database_principals WHERE name = N'DataReader')
DROP USER [DataReader]
GO

So my final steps are:

Connection to master:
CREATE LOGIN NewReader WITH PASSWORD=N'Password@2'

Connection to my specified database where I want to give the user permsission:

IF EXISTS (SELECT * FROM sys.database_principals WHERE name = N'DataReader')
DROP USER [DataReader]
GO
CREATE USER [DataReader]
FOR LOGIN [DataReader]
WITH DEFAULT_SCHEMA = dbo
GO
EXEC sp_addrolemember 'db_datareader', 'DataReader'

It’s a bit of an effort, however if you use a database project in VS you need to keep this in mind.

BizTalk 2010 Released

2010-09-27T07:34:00.000-07:00

I noticed with not too much fan fare that BizTalk 2010 has finally been released, a few days earlier than expected, we were expecting end of this month, however we got it in on the 27th.

http://www.microsoft.com/biztalk/en/us/default.aspx

You can now download the Developer edition for free, there is no license required or cost involved with using the developer edition.

It works quite nicely with Visual Studio 2010, and comes complete with all the adapters required.

You can download the evaluation version here: Evaluation Version

Or the Developer Version here: Developer Edition

You may not see it on MSDN as yet for

BizTalk 2010 Beta Public Availability

2010-06-13T17:50:00.000-07:00

You can now download the beta of BizTalk 2010 from the BizTalk site at Microsoft: www.microsoft.com/biztalk

Please have a look, I've been running it for a while now and it looks good thus far.

If you encounter any issues please contact me: paul@somers.com

BizTalk Best Practices

2010-06-12T19:29:00.000-07:00

I conduced some training recently on BizTalk, and we were covering topics around BizTalk Best practices.

I looked at what was avilable on the internet, and edited this, added my own flavour on this and put together a best practices for BizTalk, with naming standards.

This would apply to pretty much all versions of BizTalk right up to 2010.

Please download it here: [DOWNLOAD]

What’s New in BizTalk 2009 R2…

2010-03-10T19:54:00.000-08:00

We have had BizTalk 2006 R2, and now we will soon have BizTalk 2009 R2, so don’t get confused… and BizTalk is far from being replaced it only gets better.

I’ve been looking at 2009 R2 recently, and I can tell you a few things about what’s coming, within reason, as most of this is still under wraps I can’t spill all of the beans.

As an obvious start, there is support for Windows 7 and windows 2008 r2, also support for Visual Studio 2010 as expected, from the previous release a platform shift is required...

The BizTalk Administration console has gone for a few enhancements, making it easier to use particularly in live production environments, where lots of things might be happening at once, adding to the cool things in 2009, when HAT went away and tracking was available from the administration console.

The BizTalk mapper has had a few enhancements, things you would have thought should have been there long ago, and there are still many improvements here that never made it out of the product group that would make life much nicer.

FTP you say, we have been asking for SFTP for a long time, I hear whispers that it might be in this release… along with a whole bunch of new FTP features, keep your mouth open for this one, it looks good to me.

There are some new features introduced, and some enhancements to existing features, not as many as I would have liked… These lie round EDI, there is a focus on making this better in each release as we saw in 2009. EDI is not dead, just as Cobol is not dead.

There are plans to deprecate some features, why you would want to do this? However there are plans, not major or drastic, and you will have to wait and see all of them.

The SOAP Adapter is on the cards for removal, replaced by the WCF-BasicHttp, there is some contention amongst the industry as this adapter still provides some features not found in WCF.

SQL Adapter, this has been coming for some time and nothing thats not expected with most adapters, the old SQL Adapter may go and be replaced by a WCF SQL Adapter.

What does this mean if you want to upgrade? You need to be aware of what’s going, and if you are using any of this, you need to adjust your solution to use the alternatives.

All in all it shows the continued investment in the product going forward.. It’s good news.

BizTalk vNext features or WF/WCF vnext features?

2010-02-25T14:56:00.000-08:00

Look at the next version of BizTalk 2009 R2, coming soon, and then think what else would I like to see...

Then understand that the team that wrote BizTalk is the same team that made WF and WCF.

I'd turn it around to say, look at WF and WCF where the bulk of the effort is. What would I like to see here, as post Dublin we finally have a product that works, with nice hosting and manageability. This is really the way to go. .

So ask what you would like to see more of in this stack. As it's still not 100% usable, and very clunky in areas here BizTalk is seamless.

I'd like to see a mapper that works for WCF endpoints. Defined for contracts, to map incoming formats into that of the contract.

This would involve identifying the incoming format first, like matching it to a different contract/schema, for the map, and then applying the transformation. Sounds a lot like a BizTalk port.

Then mapping inside of WF, to construct an outgoing message from a different incoming format.

WF is far tooo clunky for this...

You will see more ws-* wcf adapters, however the other features, like debug orch in vs, never going to happen. It's already there in wf so why put it in BizTalk....

I've been waiting for that new mapper I saw 3 years ago to appear in BizTalk, still no sign of it....

Where does Dublin, WF and WCF fit with BizTak going forward?

2010-01-20T07:59:00.000-08:00

I previously spoke about where BizTalk and Dublin (Insert Name), WF and WCF all fit…. Well here is my view.

As far as where does Dublin fit here I can only touch on this, Dublin could host the workflow, much like BizTalk hosts the orchestration and the communication to end points. Dublin workflows could call BizTalk to kick off the back end communication and orchestration process, and get a result when they are done. In this way the workflow/human workflow can interact with back end systems, in a correctly architected manner, you can of course cut the corners here and call wcf services hosted, not a good idea, you could call back end oriented workflows that would be hosed in Dublin. WF can’t talk to SharePoint, and it perhaps can’t talk to many back end systems, whose functionality live in BizTalk. For example WF can’t send a fax.

There are a few fax adapters that can. WF can’t map a document from one format to the format that the end system is expecting; it has no concept of this. WCF can’t do this, and WF can’t do this. BizTalk will be here for few more years still.