Improve logging in pipes-iterator-jdbc

It is possible now to forget that an id column is needed for some use cases, for example. We should add more logging to let the user know what has been configured and what has not been configured.

We should add a logged backoff to use the fetchkey column for the id column if now id column is specified. We should also allow one column to be more than one feature, e.g. one column could be fetchkey, emitkey and idkey.