Inbound Processing

The following list describes the Inbound processing file poll protocols and their capabilities:

• Polling for file(s) in a specified directory.

• Instead of renaming the original files, create a work directory relative to the input directory. Name of the work directory is configurable via the workArea setting in file:address, with the default being "filebc-in-processing".

• Move each input file to work directory, and tag the original file names with a UUID.

• Inbound worker thread reads files from the work directory and processes contents as follows:

  • If multiple records per file setting is false (default), the content of the file will be treated as the payload and set in the normalized message content.

  • If multiple records per file is true, create separate normalized messages one for each record. How a record is read from the file depends on maxBytesPerRecord and recordDelimiter settings.

• All successfully processed input files will be archived, if the "archive" setting is true.

  • A "processed_files" directory will be created relative to the input directory

  • Original files will be renamed with “_processed” suffix and put into the "processed_files" directory

Error Handling

In case of errors while processing an input file:

• An "error" folder will be included in the "new" folder created by the Binding Component in the input directory.

• Original UUID tagged file (from the work directory) will be moved to the errors directory and a additional file with "_error" suffix will be generated. "_error" file will contain details of the failure.

Recovery

Recovery is a mechanism used to prevent message loss in case of system crash.

On startup, Inbound Processor will first check the work directory and put any unprocessed files in the queue for the worker thread. This ensures that any files left unprocessed in the work directory will be processed when the systems restarts.