With all the there is at least one blog that did not get the attention that it deserves: VMware has released a. That is an important one and highly recommended for all the folks that use nested virtualization on top of ESXi. If you run virtual ESXi (vESXi) hosts on top of physical ESXi hosts then you probably know that the network portgroup that the vESXi hosts connect to must be configured for promiscuous mode. This ensures that the vESXi hosts do not only receive the network packets for their own management interfaces, but also for the nested VMs that they run themselves.

But it has a drawback: Because of the promiscuous mode all network packets will be copied to all vESXi hosts all the time, and that leads to a lot of unnecessary network traffic and a notable compute overhead. With the MAC learning filter the physical ESXi host will learn and remember what MAC addresses are really connected to a specific virtual network port and will forward only the packets that are destined for these MAC addresses.

Esxi Nic Teaming

All other packets will be dropped regardless of the promiscuous mode being enabled. William has very described the problem and the implemented solution in his blog post, so. It should become clear from William's pictures and description, but it is not explicitly mentioned in his post: Unlike the (another must-have for nested virtualization setups!) the MAC learning dvFilter VIB package must be installed on the physical hosts that run virtual hosts, but is not needed on the virtual hosts! However, both packages have one annoying thing in common: They are both provided as VIB files only and have not officially been made available as Offline bundles or in an Online Depot. And that means that there is no official way to use the package with PowerCLI ImageBuilder or to create an ImageProfile for vSphere Auto Deploy that contains the package.

Sure, you can create an Offline Bundle yourself using VIB2ZIP from my, and I already showed the necessary steps in the. However, this is not really necessary, because I also made the package available through my, so you can just pull it from there. Time doctor for mac os.