[HECnet] Same MAC address on different nodes

Mark Pizzolato - Info Comm Mark at infocomm.com
Wed Jun 4 18:45:03 PDT 2014


On Saturday, May 24, 2014 at 3:12 AM, Jean-Yves Bernier wrote:
Node 10.1

Line = QNA-0

      Controller = Normal
      Counter timer = Off
      Protocol = ETHERNET
      Hardware address = 08-00-2B-AA-BB-CC
      Controller CSR = 174440, Vector = 120
      Priority = 5


Node 10.2

Line = QNA-0

      Controller = Normal
      Counter timer = Off
      Protocol = ETHERNET
      Hardware address = 08-00-2B-AA-BB-CC
      Controller CSR = 174440, Vector = 120
      Priority = 5

Shouldn't be AA-00-04-00-01-28 and AA-00-04-00-02-28 ?

Both nodes under simh.

The network is working, however.

Is it required to SET XQ MAC to match node address?

Without regard to how you happen to use a particular simulated Ethernet device, you should choose a unique hardware MAC address for that card.   This would reflect how hardware manufacturers assign unique, but persistent, MAC addresses to each network device that they manufacture.   Simh can't know if you've done this, and since you haven't explicitly changed the MAC address both of these simulated devices have the same hardware MAC address.

As it turns out, since you're using these Ethernet devices for DECnet hosts, due to how Phase IV DECnet was designed, the operating system changes the running MAC address to reflect the current DECnet address (or SCSSYSTEMID) that the system is running with.   Since you've got more than one simulator running on your LAN, you should set the MAC addresses uniquely in the simulator configuration files before you attempt to attach the Ethernet device to a network.   If you don't, there are timing situations which will cause the second simulators attach attempt to fail since it might detect that there is a MAC address conflict.   This could happen if both simulators are connected to the LAN but haven't yet gotten far enough along in the VMS boot for the operating system to uniquely set the MAC addresses.

- Mark

- Mark



More information about the Hecnet-list mailing list