ORA-03135 Connection Lost Contact Errors

Since our upgrade from 10.2.0.4 to 11.2.0.3, we are experiencing occaisional overnight job failures throwing the ORA-03135 Connection lost contact errors. It doesn't happen often, but every once in a while. Even so, it is bugging me because I can't find anything wrong!

We have a dual server setup, with one IA64 HPUX 11.31 server hosting the database, and a second IA64 HP-UX 11.31 running all of the application code. I'm seeing errors in the Alert log that is pointing me to a specific cjq0 trace file, and the contents of that trace file are pretty much as follows:

*** 2013-08-15 02:05:45.541
Waited for process J001 to initialize for 107 seconds
*** 2013-08-15 02:05:45.544
Process diagnostic dump for J001, OS id=19302
-------------------------------------------------------------------------------
os thread scheduling delay history: (sampling every 1.000000 secs)
0.000000 secs at [ 02:05:45 ]
NOTE: scheduling delay has not been sampled for 0.312200 secs 0.000000 secs from [ 02:05:41 - 02:05:46 ], 5 sec avg
0.000000 secs from [ 02:04:46 - 02:05:46 ], 1 min avg
0.000031 secs from [ 02:00:45 - 02:05:46 ], 5 min avg
loadavg : 0.31 0.17 0.09
[code]....

There were jobs that completed successfully just prior to this failure, and subsequent jobs ran successfully as well. I'm stumped - not sure what's going on here. I have reviewed a lot of posts relative to the ORA-03135, and followed as many of the recommendations as I can, but I'm still seeing this errors periodically. Their frequency has not changed since I started applying these changes, so I know I haven't stumbled on the cause as yet.