![]() Placing info here as I go along for the benefit of anyone else who come across this problem: They will always be in the form: 00:50:56:5X:YY:ZZ where X is 0x50 | (physicalMac & 0x0f), YY and ZZ are the last 2 bytes of the physical card’s MAC address: example: Physical NIC MAC : 00:15:17:3a:ca:05 Produces virtual MAC address: 00:50:56:5a:ca:05." We have also changed the distributed port groups load balancing policy to: "Route based on IP HASH" as recommended with no improvement.Ī Generated MAC address is then assigned to each Physical NIC for use in beaconing and all traffic that comes from the virtual switch itself. VCenter is 6.5 U1, hosts are 6.5 U1 and distributed virtual switch (DVS) is on the latest version being 6.5 networking. ![]() We found it does not matter if the port groups are on LAG or normal Uplinks. ![]() Turn off the host or change to a vswitch and the problem goes away. Turn on a host on a DVS with no VMs present and the we get flooded with MACs in the range: 00:50:56:5X:XX:XX across all port groups vlans under that DVS. ![]() On further investigation, this seems to perhaps be related to the VirtualMAC on each host and only floods the MAC across all vlans when going through a DVS but is fine on a vswitch.
0 Comments
Leave a Reply. |