Hi,
We’re using the process engine in an embedded scenario. Currently we’re using JTA and a separate datasource for Camunda engine, but would like to not have to do this, since it has a negative performance impact and requires a database environment supporting prepared transactions (not always possible). Ideally we could just use the same data source, because the camunda data is located within the same database as the rest of our application. Is there any way to do this without JTA or spring transactions? Basically what I want is to just reuse the underlying JDBC connection that we’ve already acquired for our transaction. We also don’t really want to use spring because we are already using Guice for dependency injection.
I realize that this would require some way to call into our code from Camunda timers to begin a transaction when a timer is firing. Is there any way I can hack this up myself even? Can I write my own transaction interceptor, like the JTATransactionInterceptor that essentially calls into our code to do the transaction management, and expose the JDBC connection acquired somehow? We actually have already created an interceptor for when timers need to do work that would require calling back into our code, so I guess all I’m looking for is a way to get our JDBC connection into your code.
Thanks