RMI server fails since 9.0.1.8

I'm not a SysAdmin but thought I'd post this here to see if anyone has any clues.  After we upgraded to 9.0.1.8 this summer frequently (every few days to a week) our ProcessFlow interfaces that use the IBM WebSphere Transformation Extender fail in the TXAdapter node because the RMI server is down.  My SysAdmin added a nightly CRON job that stops and restarts it and now the PF failures are less often, about every two weeks, but they still occur.  Does anyone have any ideas why this happens?