There was a problem recovering the Timesten database. The most common cause for this is that the database checkpoint files or (more commonly) transaction log files have been damaged or corrupted. For example sometimes people manually delete or move transaction log files if they are running low on disk space (which is an extremely bad idea of course).
You need to look in the daemon log files (the field ttmesg.log and tterrors.log located in <tt_install_dir>/info) and see the exact messages produced when you try and startup this TimesTen database (which is happening implicitly when you start the cache agent). Those messages will give you more details on the cause. Depending on what they say it is possible that this database is now unrecoverable.