I have seen that Exception, and if I recall, it may be been when I ran with a corrupted pipeline input file (from pub.flow:restorePipelineFromFile). But then when I sent another request through our driver and saved the pipeline, that seemed to have cleared up this Exception.
I do have a related question though. We are using the wM6.0.1 WmDB package against Oracle9i to handle transactional requests. The basic design looks like this:
SEQUENCE (parent of try/catch - exit on success)
SEQUENCE (try block - exit on failure)
SEQUENCE (catch block - exit on done)
What I’m seeing is that execSQL in the try block uses a $dbAlias to open a database handle to Oracle. Then if there is an error, I jump down into the catch block. Here, the rollback service no longer has acdess to the same database handle. Therefore, it can’t rollback the SQL statement which caused the Exception.
Has anyone successfully done this type of logic? Thanks.