Risk with that approach is that your complimentary products could become irrelevant or redundant as the open source project evolves e.g. training material could become outdated as functionality changes, a product that relies on some API could get broken by API changes etc. By owning both the open source project & the complimentary product, you can avoid nasty surprises (and be ahead of any competitors in the "complimentary product" space when you make changes in the open source project)

I can't say I'm an expert on RedHad's business model, but I think this is what they've done with Linux. Mysql did something similar as well, but in their case they built mysql from scratch themselves rather than attaching to an existing product.