Thread: OmegaT, anyone? Having trouble...

If you refer to terms, you need, yes, a glossary. But OmegaT has no ability to create/modify glossaries, it just is capable of reading a glossary (or glossaries) that must be in the /glossary subfolder of your project and must be a text file in your system's default encoding (with .tab file name extension) or a text file in UTF-8 (with .uft8 extension). Its structure is pretty simple:
term[tab]translation

Please, also re-read the documentation on this issue.

As for fuzzy matches, well, it is hard to do guesswork base on such a meager piece of information. I could only suggest that either your source text does not contain any fuzzy matches within (not a big surprise for just 3 pages), or maybe, well, you do something wrong. Try to be more informative.