I am developing a batch application on Java EE platform where Jberet (jsr-352 implementation) is used on WildFly application server. Jberet by default store all the batch data in local file system (under standalone/data) in XML format. And this is being displayed on jberet-ui with jberet-rest components. We have a requirements where we would like to store this batch data in a NoSQL database (e.g. mongodb or DynamoDB) rather than file system. What is the best way to implement this?
WildFly currently supports in-memory (default), and jdbc batch job repository backed by a relational database.
JBeret in standalone Java SE environment supports MongoDB-based batch job repository. This type of job repository has not been integrated into WildFly to support the configuration and management of batch subsystem. If you think this is important feature to have in future WildFly, please submit a feature request in WildFly JIRA.