We applied the C3EFIX4 and ask the client application to use the original GMO_SYNCPOINT option and it seems to fix the problem. Is it because the GMO_NO_SYNCPOINT caused this delay?
Our MQ application is to simulate the old SNA communication with SRVCONN and server on TANDEM and client on UNISYS. The server MQPUT a message and MQGET the reply with the waitinterval of 10 secs.
...