Integration Server 6.x Log/Audit Store/... Rolling Best Prac

Our Sun Solaris installations of the 6.1/6.5 Integration Servers and Brokers log and write audit store information to local file systems. It seems that we keep getting bitten by one or more of these (more often Audit Store) running out of available space (either server or hardware enforced). I would like to know what “Best Practices” are recommended for managing log/audit store/… “rolling”.

Besides the logs and audit store, are there other sets of files that we should be managing?

Is there a minimum set of files that should not be shrunk/archived to tape/etc. while the server is running? How do we know which are actively in use (or potentially to be) and which are historical?

Is there a similar set that should be dealt with during a server down time or restart?

Should these logs be maintained within the Integration Server (scheduled flows) or by local server scripts?

Are there utilities already developed for handling this maintenance task?

Are any special (other than the obvious) considerations in a clustered environment?

Anything I’ve forgotten to ask?

Thanks in advance,