Oh no, not yet. Are you fully cognizant of the implications of the various partitioning algorithms? What if you need to join, aggregate, remove duplicates or other key-based activities downstream? You can't use Round Robin then.
How about including or generating a line number when reading the file, then partitioning two way - or splitting into two streams within the job, based on Mod(RowNumber,2)? Then use different Transformer processes to parse the different line structures, and re-join them (Join stage) based upon the li...
Welcome aboard. :D Does this occur if you use auto-generated DBoptions, supplying user ID and password through explicit properties? (Yes, I know the array size will default in this case, but let's get the password issue solved first.) If this works, what is in the DBoptions property? In particular, ...