On Tue, 26 Sep 2006 18:36:33 -0700, Tom Cluster <[EMAIL PROTECTED]> wrote:
>I'm trying to configure my TCP/IP virtual machines so that when >they're restored to our DR environment they will work correctly >without any changes. While not directly related to your original HOST question I'll throw out what we do during our DR exercises in case it helps anyone else out. Like you, our VM TCPIP configuration is simple but we do have Linux guest s to contend with. At our base site, VM TCPIP and all the Linux guests go through VSWITCHes to the OSA, and on to outboard gateway IP addresses tha t are different depending on the VLAN(s) the Linux guests belongs to. At DR, our VM runs as a 2nd level guest. The outboard DR gateway has a different IP address than at our base site. The goal was to not have to update all the Linux TCPIP stacks with new DR gateway IP addresses. We accomplish that by *NOT* defining the OSA addresses used by the VSWITCHes to our DR VM userid. This makes the VSWITCHes disconnected. The DR VM userid uses OSA addresses (which leads to the DR gateway) that do not exist at our base site and gives them to a TCPIPDR stack that only runs a t DR. The TCPIPDR stack has several NICDEFs to each VSWITCH, with the LINK S set to the required VLAN numbers, and the HOME statement gives those link s the IP addresses of the gateways from our base site needed by the guests connected to the VLANs on that VSWITCH. This allows the TCPIPDR stack to masquerade as the various base site gateway IP addresses that the normal VM TCPIP and Linux guests expect. The TCPIPDR stack has PRIROUTER on one of the TCPIPDR interfaces to each VSWITCH and on the DR OSA interface. The TCPIPDR GATEWAY statments route everything to the appropriate VSWITCH LINK and to the DR gateway. At DR all we have to do after the system comes up is XAUTOLOG TCPIPDR. (DO NOT bring up TCPIPDR at the home site as the IP addresses and PRIROUTERs involved will cause havoc.) Brian Nielsen (just back from a DR exercise last week)
