Optimizing tuning and table design for large analytics DB

Can anyone point me towards good articles or books that would help a PostgreSQL novice (i.e. me) learn the optimal
approachesto setting up a DB for analytics? 

In this particular case, I need to efficiently analyze approximately 300 million system log events (i.e. time series
data).It's log data, so it's only appended to the table, not inserted and is never modified. Only 90 days worth of data
willbe retained, so old records need to be deleted periodically. Query performance will only be important for small
subsetsof the data (e.g. when analyzing a week or day's worth of data), the rest of the reports will be run in batch
mode.There will likely only be one user at a time doing ad-hoc queries. 

This is a a follow-up to the earlier suggestions that PostgreSQL will handle the volumes of data I plan to work with,
soI figured I'd give it a shot.  

Rob