Hi:
We are relatively new to webMethods and are using Enterprise
Server 5.0, sp 8 (Sun Solaris 8) ADK 4.2i(4).

We had a developer report that a previously mapped Operation
data field in the Enterprise Integrator suddenly became
unmapped, while other fields retained their mapped targets.
Can someone describe to us the ways in which a previously
mapped integration data field can become “unmapped” other
than by manually removing the mapping at the integration
folder level from within the Enterprise Integrator?

Thanks in advance for your help,
Randy

Randy,

I don’t have an further information for you other than to let you know that we have experienced similar behavior with the mappings sometimes becoming umapped. We too are on the same versions of ES and ADK.

I think it may have to do with the generation/re-generation of code as components and operations are built, so once the design is stabilized it is a good idea to go back through and validate that all the mappings are there. It is probably a bug, but one that is hard to reproduce for webMethods support.

Maybe others out there have seen deterministic behavior that causes this to occur, but I for one have not been able to pinpoint the exact reason this happens.

Steve

Hi, Randy and Steve.

Areyou referring to the behavior in which the entire GUI “flow” for a component will disappear? If so, I have found a reliable work-around.

Ignore your first instinct to delete the component!. In the left-pane, highlight the component and select the “Script” tab. If all is right, you will see Java code. That code is the low(er)-level representation of your entire script. Use CTRL-C to copy that code to the clipboard.

Next, create a new component and immediately go to the “Script” tab. Paste the clipboard contents using CTRL-V. Save the component and switch back to the GUI view. If all looks right (as it should), make a note of the hosed component’s properties such as triggers and other traits. Delete the original, return the duplicate component, and re-set the properties.

That should do the job. I haven’t tracked down the exact cause of this behavior, but its frequency is inversely proportional to the amount of available memory on the local machine.

Hope this helps.