I am working on webMethods 6.1.5 Modeler. I am using 2 logical servers, but my main modeler process is in server1. Earlier the modeler was working, but I imported modified modeler process (modified in some other server) and now it is not triggering at all. As I am using the same name, before importing, I had disabled the existing process through WmMonitor and deleted the modeler process from Modeler. After importing the modified one, I generated it, updated for monitoring and then enabled the new modeler process through WmMonitor. The old one is still there but in disabled mode. So the new process should be triggered. But I am getting the following message in server log.
“BizDocAttribute ReceiverID is required in a process, but no definition is available on this server”
(Whenever I submit an EDI document, the modeler process has to be triggered. I checked EDI document and it has receiver ID and thus it is getting submitted properly.)
And the error log I am seeing a different error (not sure whether it is related to this business process or not):
Hi
Just to understand your question,
You are submitting an EDI document to TN, and a model should be invoked with this EDI document as the subscription document.
Ok.
As far as i know, for a model to get invoked from TN doc, there shud be a conversationId be extracted in the TN doc type, and that shud be unique everytime, when ever u submit the doc to TN , to get the model invoked.
InstanceAlreadyexists Exception might tell us that, the repository has that EDI doc already.
Can u try changing the Interchange control number and submit, just to make it unique.
Thanks for your prompt response. I have taken care of generating the uniconversation ID everytime by defining the instance ID query in WmEDI. Earlier it was working fine. After importing the modified model only, it is not triggering.
Hi
Before generating the model, did u delete the earlier package.
try even deleting the model related info from webMethods tables…wmprocessdefinition…and import and generate.
In the generated package , disable and enable the trigger.
Are there any other triggers which are using this instance.
Thanks for your suggestions. I have followed those. Let me give all the steps that I did.
Deleted the BP packages both from source and target servers
Deleted the modeler process from the modeler
Deleted both old and new modeler process entries from WMPROCESSDEFINITON table
Imported the modeler process freshly through modeler, generated and updated for monitoring
Enabled the newly created modeler process through WmMonitor.
But still I am getting the same error in server log:
“BizDocAttribute ReceiverID is required in a process, but no definition is available on this server”.
But this time I am not seeing the error in error log (omi.wm.devkit.UnexpectedException: javax.management.InstanceAlreadyExistsException).
I deleted the rows only from WMPROCESSDEFINITON table. I am not sure about WMPROCESSIMAGE and WMSTEPDEFINITON tables, but WMSTEPTRANSITIONDEFINITON table seems to be a dynamic table. And I didn’t see any entries in WMCUSTOMFILEDEFNITION. In Modeler guide (pg. 246), I could see only these table names but not any description.
After enabling the newly created modeler process, I reloaded the packages WmModeler and the BP packages (which were created after generating the modeler process). Could you please check the the process I followed (including the d/b tables etc.).
You really shouldn’t be editing the DB tables directly. Doing so introduces the risk of corrupting the tables and their relational integrity. The WmMonitor pages provide everything you need to maintain things.
The “omi.wm.bp.bp:init omi.wm.devkit.UnexpectedException: javax.management.InstanceAlreadyExistsException” error was because of the two models existing with the same name. You’ve corrected that so I believe from a Modeler/PRT perspective, you’re all set.
For the Bizdoc attribute error: Since TN is the component that extracts this value and places it in the pipeline, I suspect some sort of TN related issue. You mentioned that two logical servers are defined. Are there two physical servers? If so, which server is complaining? Are WmTN, WmEDI and WmEDIforTN installed on both? Basic stuff, I know, but just trying to zero in on where the trouble might be.
Actually I went ahead for deleting the entries in the d/b only because of not getting things done through WmMonitor. I clicked so many times to disable the old business process. Everytime I get the message that a request was submitted to PRT process for disabling. And there is no option to delete the already ran business processes.
There are two physical logical servers. The TN server is the source and so business process has to be triggered at it whenver a document is submitted to TN. I checked the packages WmEDI and WmEDIforTN; they are installed.
Still this error “BizDocAttribute ReceiverID is required in a process, but no definition is available on this server” has been coming in server log.
Could you/anybody have any thoughts why this eorror is coming, even though Bizdoc receiver ID is available ??
Is the second server configured to use TN? Is it using the same TN database as the first server? If not, and there are steps configured to run on this second server, then I believe that may be contributing to the problem though I’m just guessing.
The second (target) server is not connected to TN, but it subscribes the published document from the source server. Anyway, it was working fine. The only issue now is the modeler process is not getting triggered. I am not getting that error “BizDocAttribute ReceiverID is required in a process, but no definition is available on this server” also now. I tested the code without using business process (with explicit triggers) and it is working fine.
Any solution / guess what could be causing the problem for modeler process ?
I exported the modeler process and imported in my local server (which I am using both for source and target environments, ie, single server which acts as both source server and target server). Here the modeler is working without any problem. But there (in the main servers I’ve been mentining), it is not triggering at all. Could not understand when and where modeler works, though the modeler process is not changed !!
I have tested the the same modeler process on my local servers (used the same architecture: source TN server and target server) and it’s working fine. But on the main project servers, it is not at all triggering and no error is also shown on server log.
Any solution/guesses/patches ? No modeler fix was applied on this server.
I have deleted the the existing entries from WMPROCESSDEFINITON table and freshly imported it once again, now the MODLER PROCESS IS TRIGGERING AND END-TO-END FLOW IS HAPPENING FINE.
But now my concern is: there are 3 actions (through 3 services), 2 at the source TN server and 1 at the target server. The modeler process has to show all of them completed in one instance ID result. But it is giving 2 instances - one for the first 2 source server services and the other for the target service action. And that 2nd instance is coming under ‘Recently Created Processes’ with ‘Started’ status. But inside (if we click and see), it is showing as ‘completed’ only and the the interface is working end-to-end. The only thing is I am not seeing all the 3 actions together. Hope I could produce my issue in understandable format. Please ask if my point is not clear.