ClusterJPA with non-jta Data Source backup?

[Format Dynamics]

I'm currently using openjpa with mysql connector/j to my mysql cluster (using dbcp2 and jdbc:mysql:loadbalance).
I'm doing this in my persistence file using:
<property name="javax.persistence.nonJtaDataSource" value="java:/comp/env/jdbc/mydatasource" />

I want to start using clusterjpa with the NDB brokerfactory, but so far all the examples I have seen have the backup JDBC data source specified directly in the persistence file (url,user,pass).
Is it possible to use my non-jta data source with pooling instead?

On a side note, is there any updated docs on compiling/running with clusterjpa? The last official stuff seems from 7.1.2. Or has nothing really changed since then?

Thanks!

-Tony