A datawarehouse stack (or suite) usually consists of three layers. These are usually referenced as ETL (loading), Database & Reporting (interface). In addition, there are somewhat more advanced tools for performance and expert needs. These consist of Cubes and Statistical Analysis Tools.
As far as interoperability goes, the ETL tools and the reporting tools need to support whatever database you are using. However, since there are only two big open source databases, there is usually no problem mixing different solutions.
As for specifics -
1 - ETL
Data loading can be achieved by open-source tools such as Pentaho's Data Integration or Talend (an eclipse extension). I would suggest googling "open source etl" to tailor the solution for your specific needs.
2 - DB
You'll need a relational database (RDBMS). The two most prominent open-source players are PostgreSQL (used by Stack Overflow) and MySQL. While MySQL has a larger user base, Postgres is gaining more an more popularity ever since implementing several crucial features that were missing in earlier versions.
3 - Reporting
Pentaho offer reporting platform. So is BIRT (another eclipse extension). Again, Google is your friend for specific comparisons. Note that when if you choose Pentaho for both the ETL and Reporting tools you are likely to enjoy a better integration.
You've also mentioned Mondrian, which is a tool to generate MDX queries over an RDBMS. MDX is the standard language for querying cubes.
At this point of time, assuming you are starting from scratch, I would recommend setting up the first two layers of the data warehouse - ETL & DB. You can later add any number of reporting tools above.
