If you plan on using datapower to write a modified value to the AppIdentity MQMD header, ensure that your Put Message Options on your destination are correct.

You want to use “2052″ which represents “MQPMO_SET_CONTEXT_ALL” along with “MQPMO_NO_SYNCPOINT”. You need to tell MQ that it should use the message headers from the written message. In Datapower, (for some unknown reason) in the URL Builder for MQ, this is the “User Identity” on/off toggle.

‘No Syncpoint’ tells MQ that this is a single message transaction that isn’t dependent on the writing on subsequent messages. This of it like “I want all 5 messages I write to the queue to be committed all together or none at all”.

If you don’t set your Put Message Options correctly, you won’t see your modified headers written onto the message in the Queue.

If you are using DataPower to pull messages from MQ which contains persistent messages, do yourself a favor and set the ‘Units of Work’ on the QM in DP to 1.

This will enable DataPower to use SyncPoints from MQ and allow you transactionality on a per-message basis.

Without this setting, you can get very strange behavior when you want to discard a message. You could see the entire batch of messages re-appear on your input queue.

My theory is that MQ implicitly creates a syncpoint when DataPower connects and one end it disconnects. If it doesn’t get confirmation that all the messages were successfully processed, then the only recourse is for the QM to rollback the transaction, causing all the messages datapower successfully processed to return to the queue.

In WTX, I have a COBOL EBCDIC input with a bunch of filler fields. The Input Type Tree was created by the Cobol Copybook Importer.

I had an ‘invalid input’ occur when I submitted a cobol message to my WTX map. The map failed on a filler item where the input wasn’t spaces, but NULL.

This is apparently valid, but by default, the WTX COBOL Importer will create thsi field of type CHARACTER which the null’s will cause to fail.

I switched the type of the item to ‘BINARY’ (because who cares about filler data) and got an error

“failed item presentation test for TYPE”

Looking into the documentation, I had to ensure that a binary type had a constant value. By default, WTX creates the item with a 0-XX type for the max and min. I changed the value of the min to the max and everything worked ok.

I was using WTX 8.3 to build a type tree off a COBOL copybook. When I then attempted to run a map based on the tree I got the error “One or More Inputs was Invalid”.

My Input data was in EBCDIC format and RFHUtil was my saviour. It was able to display the copybook and the data side-by-side so I could see what values should be where. (Sidenote: Good EBCDIC editor is VEDIT ).

I turned on Map Tracing via the not-well-defined ‘Map Settings’ and InputContentTrace and saw that my data wasn’t matching up with the items in the input. It was trying to put data contained in byte 1000 into an object defined at byte 1050.

Then I noticed that the number of bytes that my imported data in WTX was ‘off’ was the exact same number of bytes as a group.

I opened the Generated type tree and sure enough, a group was missing. Once I added it back as a Child Component in the Root ‘Record’ group (along with a few other missing groups) my input successfully validated.

Now, I don’t know a ton about the structure of cobol copybook or what could have caused it, but if you are in WTX and don’t see your input data lining up properly it’s probably because your WTX Root Object is missing them.

Just had a problem where a WSDL file that was uploaded to the DataPower device did not show up in the local:// drop down for selection.

Upon further inspection, we found that there was whitespace before the
<? xml .. ?> tag and whitespace before the <wsdl:definitions...> tag.

Once this extra whitespace was removed, the WSDL appeared in the drop downs and worked fine.

I’ve posted a question to the Datapower external newsgroup asking about how Siteminder can be used to authenticate users. The documentation is pretty thin on the subject.

https://www.ibm.com/developerworks/forums/thread.jspa?threadID=315235

Basically, i have a username and password that I’d like Siteminder to authenticate. I’m sure that Datapower invokes a URL and posts the information to the given URI. It then inspects the response cookie and determines if Siteminder was successful.

The usual Siteminder usage pattern is for someone to hit a page with a web browser and then be issued a challenge.. So it’s possible Siteminder requires this step into order to validate.

Yes, I’m still alive. Things have been pretty slow this year and I haven’t had all that much to say here at DanZrobok.com. It looks like business is picking up in the 4th quarter and I find myself with an opportunity to work with iTKO Lisa, an SOA Testing tool.

I’ll get into more details about the product later on.

From IBM Redbooks, Integrating WebSphere Service Registry and Repository with WebSphere Process Server and WebSphere ESB This IBM® Redpapers™ publication describes how you can integrate IBM WebSphere® Service Registry and Repository with IBM WebSphere Enterprise Service Bus and IBM WebSphere Process Server to create a flexible and robust infrastructure for the execution of SOA solutions.

From Developerworks, Migrating to WebSphere Process Server V6.2 This tutorial shows you how to migrate WebSphere Process Server configuration data, application data, and databases from V6.1.2.3 to V6.2.0.1. The tutorial also describes the sub-tasks involved and shows you how to use migration log files for troubleshooting. Everyone's favourite task: Migrating between versions.

If you are planning on upgrading your WebSphere Servers to a later fixpack level, be sure to start with the Deployment Manager. If you do not, your Deployment Manager will not be able to contact your ‘future level’ node agent.

“Server cannot be started because the node agent for server Node1 on node x is not active”

Attempting to call syncNode to the deployment manager will result in:

ADMU0127E: The version of the Deployment Manager is earlier than that of this node. Deployment Manager version earlier than that of a node is an unsupported configuration.Upgrade the Deployment Manager to the same or later version as that of the node.

If you are upgrading Solaris, you will see a bunch of patch numbers fly by along with a bunch of return codes. When I first saw the return codes, I was concerned that something was not updating properly. I saw lots of Return Code 2 and Return Code 8.

When I checked out what the error codes represent, I was relieved. Code 2 is for a patch that is already applied. Code 8 is when a patch attempts to modify a package that is not installed.

Just incase you were thinking of writing some IBM Certifications, there will be a new promotion in July called Take It Again which allows you to rewrite a failed certification test free.

I have to re-write my WebSphere Integration Developer, WebSphere Process Server and DataPower certifications and am waiting for this promo to start before I book my tests.

You can import and export your WebSphere Application Server or WebSphere Process Server profiles into .CAR archive files.

Importing and exporting version 6 WebSphere Application Server profiles provides the specific wsadmin (or RAD/WID) steps to perform.

The infocenter has a great article about how to use “Industry Standard Schemas” with WebSphere Process Server.

This is required reading.

You will need to know about the Sequence object when dealing with

• xsd:any
• xsd:anyType
• xsd:anySimpleType
• xsd:anyAttribute

This used to be completely undocumented and now, the documentation is actually really good. Enjoy your undefined datatype.

I still stick with the rule that undefined types are the scourge of good integration.

The Visual Configuration Explorer

When configuration problems occur in a complex software environment, you can spend a lot of time hunting for subtle changes that might have introduced errors. The new Visual Configuration Explorer tool, which runs as a plug-in to the IBM Support Assistant, simplifies configuration troubleshooting by letting you visualize, explore, and analyze the configurations of various software stacks offline.

You can also use this plugin to compare two configurations and look for changes. Very good if your server ‘magically’ breaks one day or you are trying to track down a change from a while ago.