I am using ACE TAO as a CORBA implementation. I would like to know if anyone has any options for setting the Maximum message size and Maximum number of connections.
omniORB has two options for them: giopMaxMsgSize and maxGIOPConnectionPerServer .
Even if TAO has ORBMaxMessageSize and ORBLingerTimeout , this will crash my server. I don’t even know if they are the right options.
This is what the TAO debug output says.
TAO (30232|3086943952) ORB_Core: Unable to initialize Codeset Manager TAO (30232|3086943952) - Completed initializing the process-wide service context TAO (30232|3086943952) - Default ORB services initialization begins TAO (30232|3086943952) - Default ORB services initialization completed TAO (30232|3086943952) - We are the default ORB ... TAO (30232|3086943952) - Initializing the orb-specific services TAO (30232|3086943952) - Setting primary connection timeout hook TAO (30232|3086943952) - Default_Resource_Factory - unable to find codeset manager factory. TAO (30232|3086943952) - ORB_Core: Codeset Manager not available TAO (30232|3086943952) - Loaded default protocol <IIOP_Factory> TAO (30232|3086943952) - Loaded default protocol <UIOP_Factory> TAO (30232|3086943952) - Loaded default protocol <SHMIOP_Factory> TAO (30232|3086943952) - Loaded default protocol <DIOP_Factory> TAO (30232|3086943952) - Created new ORB <> TAO (30232|3086943952) - Transport_Cache_Manager_T::purge, Cache size after purging is [0] TAO (30232|3086943952) - IIOP_Connector::begin_connection, to <wnpcls.econz.co.nz:4000> which should block TAO (30232|3086943952) - Transport_Connector::wait_for_connection_completion, transport [171625212], Connection not complete. TAO (30232|3086943952) - Transport_Cache_Manager_T::bind_i, Transport[171625212] @ hash:index{-1408233282:0} TAO (30232|3086943952) - Transport_Cache_Manager_T::bind_i: Success Transport[171625212] @ hash:index{-1408233282:0}. Cache size is [1] TAO (30232|3086943952) - Transport_Connector::wait_for_connection_completion, going to wait for connection completion on transport[171625212] TAO (30232|3086943952) - Leader_Follower[171625212]::wait_for_event, (leader) enter reactor event loop TAO (30232|3086943952) - IIOP_Connection_Handler::open, The local addr is <172.16.1.30:46404> TAO (30232|3086943952) - IIOP_Connection_Handler::open, IIOP connection to peer <172.16.1.30:4000> on 6 TAO (30232|3086943952) - Leader_Follower[171625212]::wait_for_event, (leader) exit reactor event loop TAO (30232|3086943952) - Transport_Connector::wait_for_connection_completion, transport [6], wait done result = 1 TAO (30232|3086943952) - IIOP_Connector::make_connection, new connected connection to <wnpcls.econz.co.nz:4000> on Transport[6] TAO (30232|3086943952) - Transport[6]::register_handler TAO (30232|3086943952) - Transport_Connector::connect, opening Transport[6] in TAO_CLIENT_ROLE TAO (30232|3086943952) - Transport_Connector::connect, got an existing connected Transport[6] in role TAO_CLIENT_ROLE TAO (30232|3086943952) - Muxed_TMS[6]::request_id, <1> (30232|3086943952) Error in writing request header TAO (30232|3086943952) - Transport[6]::generate_request_header, error while marshalling the Request header TAO (30232|3086943952) - Transport[6]::make_idle TAO (30232|3086943952) - IIOP_Acceptor::open, address==:4011, options=(null) TAO (30232|3086943952) - Transport_Connector::connect, got an existing connected Transport[6] in role TAO_CLIENT_ROLE TAO (30232|3086943952) - Muxed_TMS[6]::request_id, <2> (30232|3086943952) Error in writing request header TAO (30232|3086943952) - Transport[6]::generate_request_header, error while marshalling the Request header TAO (30232|3086943952) - Transport[6]::make_idle TAO (30232|3086943952) - Connection_Handler[6]::close_connection_eh, purging entry from cache TAO (30232|3086943952) - Transport[6]::cleanup_queue_i, cleaning up complete queue TAO (30232|3086943952) - Transport[6]::cleanup_queue_i, discarded 0 messages, 0 bytes. TAO (30232|3086943952) - Connection_Handler[6]::close_connection_eh, removing from the reactor TAO (30232|3086943952) - Connection_Handler[6]::close_connection_eh, cancel all timers TAO (30232|3086943952) - Transport[6]::cleanup_queue_i, cleaning up complete queue TAO (30232|3086943952) - Transport[6]::cleanup_queue_i, discarded 0 messages, 0 bytes. TAO (30232|3086943952) - Connection_Handler[6]::close_connection_eh TAO (30232|3086943952) - Transport[6]::cleanup_queue_i, cleaning up complete queue TAO (30232|3086943952) - Transport[6]::cleanup_queue_i, discarded 0 messages, 0 bytes. TAO (30232|3086943952) - Destroying ORB <>
The code I use is as follows:
#define SC_MAX_SIZE (1024*1024*4) char *corbaARGV[12] = { argv[0] }; int args = 1; char msize[16]; snprintf(msize, sizeof(msize), "%d", SC_MAX_SIZE); corbaARGV[args + 0] = "-ORBMaxMessageSize"; corbaARGV[args + 1] = strdup(msize); args += 2; int timeout = config().GetInteger("Corba_TimeOuts"); mainlog << notice << "CORBA timeout = " << timeout << " seconds." << endl; char stimeout[16]; snprintf(stimeout, sizeof(stimeout), "%d", timeout); corbaARGV[args + 0] = "-ORBLingerTimeout"; corbaARGV[args + 1] = strdup(stimeout); args += 2;
None of the two works.
Now I also stumbled upon a test sample in TAO / tests / Oneway_Timeouts, which uses policies. Maybe someone shed light on RELATIVE_RT_TIMEOUT_POLICY_TYPE, CONNECTION_TIMEOUT_POLICY_TYPE, BUFFERING_CONSTRAINT_POLICY_TYPE.
I can not find documentation about this.