On Tuesday, 10/13/2009 at 10:02 EDT, "Mrohs, Ray" <ray.mr...@usdoj.gov> wrote: > Hi, > Our network firewalls have a policy of dropping all connections that are > idle for 30 minutes. That is causing our VM-to-VM link to drop and > restart throughout the day. The network guy asked if there is an RSCS > keepalive option, since changing firewall policies is like moving > mountains. TCPIP already had keepalive in effect with INTERVAL 20 , > SENDGARBAGE FALSE. I set SENDGARBAGE TRUE on one of the VMs, and now the > drops happen about every 23 minutes, i.e., I somehow made it worse. ITO > value is at the default setting (100). Does anyone have experience with > this?
As you have discovered, KEEPALIV=YES is the default for TCPNJE links. The firewall may be ignoring keepalives. Firewall Folks need to look at their logs. Note that keepalives have the undesirable side effect of bringing down a link when there is a transient error, even when there is no file transmission in progress. Alan Altmark z/VM Development IBM Endicott
