Clovertech

WSAEWOULDBLOCK & WSAEINVAL

› Clovertech Forums › Read Only Archives › Cloverleaf › Cloverleaf › WSAEWOULDBLOCK & WSAEINVAL

  • Creator
    Topic
  • July 14, 2005 at 2:34 pm #47900
    Rick Martin
    Participant

      While testing several new interfaces in 5.3 REV1, we have to Stop/Restart threads in order to get a connection to the receiving system (TCP/IP sockets over VPN).  

      When the outbound thread is initially brought up, we get the WSAEWOULDBLOCK connect error followed by the WSAEINVAL connect error.  The thread waits for the timeout, but seems to never connect unless we bring the thread down and up; and in some cases have to bring other threads in the same process down in order for the outbound thread to connect to the receiving system.  

      Is this a known problem?  Any known workarounds?  Anyone else experiencing the same thing?

      Thanks, Rick

    • Creator
      Topic
    Viewing 4 reply threads
    • Author
      Replies
      • August 8, 2005 at 10:58 pm #57019
        Mark Perschbacher
        Participant

          We have had some issues with establishing TCP/IP connections between Cloverleaf and foreign systems.  It sounds like this is an outbound thread, what are you using for protocol under Thread properties, and under PDL?

        • August 9, 2005 at 3:46 pm #57020
          Rick Martin
          Participant

            Yep, it’s an outbound thread using PROTOCOL: pdl-tcpip and mlp_tcp.pdl.

            Have you been able to do anything about your issues?

          • August 12, 2005 at 10:21 pm #57021
            Mark Perschbacher
            Participant

              I have spent hours troubleshooting with our partners for the foreign systems, and truthfully, haven’t got very far.  The general observation is that some systems Cloverleaf seems to communicate well with, and others, not so well.  By turning on the EO logging, you can tell alot about what is happening connection-wise, but in most cases it has come down to we send an ORU, but never get the ACK back.  Sorry I don’t have any more useful suggestions.

            • August 12, 2005 at 11:03 pm #57022
              Jim Kosloskey
              Participant

                Can I hazard a guess this/these is/are Windows based receiving systems?

                I hazard this guess because these kinds of situations happen most often, in my experience, with Windows systems. It seems to me that many (maybe most) Windows programmers who program TCP/IP at the socket level are really unprepared for the challenge. This is complicated by Windows not being the best (in my opinion) environment to deploy mission critical TCP/IP based real time solutions.

                All I can suggest is in my experience the problem has been with the receiving system and rarely with Cloverleaf. I generally prove that I have everything OK on my end (the Process log usually suffices for that) and challenge the other end to provide the same proof (they rarely can, but if they make the effort sometimes they magically correct the issue).

                Jim Koloskey

                email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.

              • August 16, 2005 at 6:27 pm #57023
                Mark Perschbacher
                Participant

                  Jim, in our environment, yes Cloverleaf is loaded on a Win2003 server.  The problem we are having is in the sending of ORU’s to a UNIX server.  What is really strange about the situation is that we also have an HL7 interface directly from our lab system, also a Win2X server, to this same UNIX box, and it works fine.  It does appear to be more of an OS problem than Cloverleaf related.

              • Author
                Replies
              Viewing 4 reply threads
              • The forum ‘Cloverleaf’ is closed to new topics and replies.