There is confusion from the tooling side with automatic addition of the "java:" prefix to the data source name created. Currently in the Admin API
there are 3 methods correspond to data sources

1) createDataSource
2) deleteDataSource
3) getDataSourceNames

Ideally all the methods should be using the consistent names. However, if the names retrieved from the "getDataSourceNames" used as the JNDI names that will not work in Teiid runtime as they do have "java:". If the method returns the names with "java:", the the "delete" will not work and inconsistent with "create" call. The problem is management object used in the Admin API and data source object have different discovery names.

Here we could solve by providing another method called "getDataSourceJNDINames" that has the "java:" prefix on its names, but it will further add to the confusion.

Even though not ideal, it may be OK and less confusing to resolve the data source name inside the Teiid engine by automatically adding the "java:" prefix, if the name does not already contain this prefix. If the data source is not found with prefix, then Teiid can retry lookup using the original name. Thus it will work both ways.