[omniORB] UNIX Transport Question

Brian Neal bgneal at gmail.com
Tue May 16 10:30:23 BST 2006


We are using omniORB 4.0.6 under linux.

We are trying to use the UNIX domain socket transport. We have several
applications on the same processor, and some others are on other
processors. We are starting our applications that run together on the
same processor with the following command line arguments:

-ORBendPoint giop:unix:
-ORBendPoint giop:tcp::
-ORBunixTransportDirectory /tmp

We are then relying on the default client transport rule of "*
unix,tcp,ssl", hoping that the applications on the same processor will
use the UNIX transport to communicate with each other, and the
applications on the other processors will use TCP. Is this the proper
recipe for that? :-)

We aren't sure this is working. Is there a trace level we can use to
see what transport got picked?

We are also dumping out stringified IOR's on some of our objects. We
then use catior (and catior -o) on these strings, but we don't see any
obvious UNIX transport related stuff in the output. Should there be?

Thanks.



More information about the omniORB-list mailing list