Feb 21st, 2008, 10:25 PM
I thought something like that existed already in the open source space, but I could be mistaken. The JobExecution Id should be unique to the execution, so it should be safe as a key.
Right, I have a few clients with a similar scenario, but they created their own table to record this kind of information separate from the framework. It's really scheduling meta-data, not processing meta-data. With that being said, I understand it's a common scenario, and I think long-term we'll address it, and a few other issues that make integration with various ways of launching a job easier. However, for right now scheduling meta-data must be stored separate from the framework.