Current example memory specs are insufficient to handle Tika extraction

I get an out-of-memory when extracting large documents using Tika. The current amount is 256M; I'm going to double that to 512M.