We have an elusive LabVIEW Linux problem (LabVIEW 2011). A client application that opens many TCP/IP sockets often fails to open some of the connections. The apparent failure is timeout (error code 56). The timeout period can be set short or very long, this makes no difference. We are opening in the region of 30-100 socket connections.
We have used network monitoring tools and do not see heavy CPU/network loading or network congestion when the problems occur. Its clearly the client end that is suffering from this problem - we can use telnet or other non-LabVIEW software to successfully make connections to the server.
The error is intermittent and its not obvious what causes it to happen. Sometimes the problem can be persistent but it is usually cleared by closing down and restarting the LabVIEW session from scratch. The next time it will open all the required connections. The error occurs only when using the LabVIEW development system - it never occurs if we run it as an executable LabVIEW application.
It looks like some sort of TCP/IP resource problem - but the Linux TCP/IP OS tuning parameters appear very adequate. The fact that restarting the LabVIEW development session clears the problem points to the resource issue being inside LabVIEW rather than the OS.
Has anyone else seen anything like this?
