In a client server application between a Win10 LV2015 PC (client) and a CRio RT Linux server, i experience some RTO's (Retransmission TimeOut) with Modbus protocol.
On table, the config is as simple : PC - switch - CRio and the problem occurs (one every several minutes) when two or more CRio are used. Immediatly after TCP retransmission, response is given by CRio but after several minutes, problem is back.
Timing is 80 ms acquisition cycle for CRio and 40 ms Modbus Query cycle for PC. RTO's are 300 ms.
May TCP Nagle algorythm be involved on the CRio side ?
How can i disable Nagle in CRio for a try ? When i try to use TCP_NoDelay_Linux.vi, i can't find "libc.so.6" dll.