I've got an FX-40 that had been running quite nicely since January (no goofy appliance, just a stripped down-to-normal JACE running 3.1.30), but very recently is suffering some lockups: I can intermittently make a platform connection to reboot, and sometimes a station connection to look around at what works and what doesn't, but rarely will both work.
A station connection always throws the little yellow box that the station is running low on disk space. I can't seem to confirm this, though - even with 118 histories, there always appears to be plenty of disk space available.
The biggest problem I find is the memory is all used up - the station's resource manager shows 128Mb of mem.total, with 128Mb of mem.used. The only clue I've been able to get from the system was a look at the TimeSync service, which showed all three time server connections faulted due to "java.net.SocketException: Too many open files"...curious.
Interestingly, this would seem to jive with the station director's output (when I can make a platform connection) that throws several errors per second pertaining to some socket exception (but I neglected to write it down the last time I did a reboot, and can't seem to make a platform connection now that I really need that info...)
Probably due to the memory lockup, none of the spy diagnostics show up (all blank screens, regardless of what I'm looking at).
The only other clue I have to offer is that this started only when I loaded the e-mail service and setup e-mail alarming two weeks ago. The e-mail alarming works fine, until the JACE locks up...
So where should I look for clues as to why the memory is all used up?