[omniORB] COMM Exception

[Richard Hardgrave]

Hello,
	We are running omniORB 3.0.3 on Solaris 2.5.1 on SPARC
Ultra 1's and 2's.  We have been experiencing exceptions
(CORBA::SystemException) under heavy loads and frequent
servant invocations.
	The trace output, below, may be a result of another
problem happening earlier, but I would like to know just
what it is trying to tell me.  I'm particularly curious about
the "close socket no. -1" statement.  Is this something that
can occur under normal circumstances, or has the system
already gone to hell in a handbasket, at that point?
	The process invoking this servant is multithreaded.
All calls to the methods in the servant are going to execute
sequentially, right?  I'm just starting to debug this, so
only bother responding if something obvious is jumping out
at you.
	I suppose it wouldn't hurt to upgrade to 3.0.4 ?

Thanks,

Richard

omniORB: tcpSocketMTfactory Rendezvouser: unblock from accept()
omniORB: tcpSocketMTfactory Rendezvouser: accept new strand.
omniORB: tcpSocketMTfactory Rendezvouser: block on accept()
omniORB: tcpSocketMTfactory Worker: start.
connect from 131.101.50.135
ll_recv: 84 bytes
4749 4f50 0100 0000 0000 0048 0000 0000 GIOP.......H....
0000 0001 0100 0000 0000 000e fe3d 19cf .............=..
fda9 3e00 0000 0000 0000 0000 0000 000b ..>.............
7265 7175 6573 745f 6573 0000 0000 0007 request_es......
6e6f 626f 6479 0000 0000 0008 6465 6661 nobody......defa
756c 7400                               ult.
omniORB: throw omniConnectionBroken (minor 146) from tcpSocketMTfactory.cc:1101
omniORB: tcpSocketStrand::~Strand() close socket no. -1
omniORB: throw COMM_FAILURE from remoteIdentity.cc:178
ll_send: 72 bytes
4749 4f50 0100 0001 0000 003c 0000 0000 GIOP.......<....
0000 0001 0000 0002 0000 0023 4944 4c3a ...........#IDL:
6f6d 672e 6f72 672f 434f 5242 412f 434f omg.org/CORBA/CO
4d4d 5f46 4149 4c55 5245 3a31 2e30 0000 MM_FAILURE:1.0..
0000 0092 0000 0001                     ........
ll_recv: 84 bytes