On Thursday, 10/02/2008 at 08:20 EDT, "O'Brien, Dennis L" <Dennis.L.O'[EMAIL PROTECTED]> wrote:
> A couple of months ago, someone said that IBM has an internal version of > IPGATE with some bugs fixed. Is there any chance of getting that on the > VM Download page? Unlikely in the extreme. First, make sure you have OverridePrecedence in AssortedParms of PROFILE TCPIP. *Everyone* should have that. There are two RFCs that conflict with each other and the later of the two is more common. OverridePrecedence induces VM TCP/IP to respect the later RFC. If that's not the problem, then I suspect that you have routers or traffic monitors that are terminating idle connections. You'd need to see the socket trace of both ends to get evidence and a network sniffer trace to confirm it. Or just ask your network folks what the policy is for idle connections between Point A and Point B. You could also write a Q&D REXX client/server program or PIPE to establish an idle TCP connection (not using IPGATE) to see if it is disrupted. Have the program log the disruption and reestablish the connection. See if there is a pattern. If so, you have empirical evidence of an Outside Force. If it appears at random, network traces are needed (after trying with OverridePrecedence). Alan Altmark z/VM Development IBM Endicott
