Community - I have an issue with a fairly wide table and it's corresponding subjob where the subjob method is exceeding Java's method limitation of 65535 bytes. The first screenshot shows the basic flow design I use to process CDC records. Having not used joblets before, my first attempt to circumvent the limitation was to refactor some of the logic into seperate joblets (2nd screenshot). Unfortunately, this does not create new process methods for the joblets, but just adds the joblet code to the original subjobs process method, so essentially I've gained nothing. Before I break this job up and stage the data for further subjob processing (would love to avoid this), does anyone know a way to deal with this where I can keep the flow simple and avoid adding complexity to an otherwise straight-forward logical flow? Why does a joblet not offer the ability to be contained in it's own method or even internal class? This would seem like the perfect use case for it and seems logical to me. All the structs, error and exception handling are already broken out, so it would seem logical to be able to call a joblet's start, main, end and finally methods directly within the main subjob process method. Unfortunately, this is something that would need changes to the core code generation in Talend itself and not something I could "code around". Thanks, Scott Sidenote: Also, to share the connection with the joblet, I created a shared connection and used a prejob in both joblets. Talend is complaining about using multiple tPrejobs in the job. How else do you share a connection with a joblet? You can see all the warning indicators in the second screenshot.