We're migrating to vDS. All hosts have 4 physical adapters. Our networking config has always been two vSwitches. vSwitch0 for Management (vmnic0 active, vmnic1 standby) and vMotion (vmnic1 active, vmnic0 standby), and vSwitch1 for virtual machines (vmnic2 and vmnic3 active). This has served us well for many years.
The vDS migration seems fairly straightforward. We've created a single DSwitch for ease of management (4 uplinks), and created distributed port groups. Each distributed port group uses a VLAN, with the exception of the Management port group. Our Management network is the default untagged network on each vmnic. We've gone with the default settings everywhere, with the exception that each port group has been modified with the correct teaming and failover order as listed above.
We have 3 clusters in our environment, with anywhere from 7 to 14 hosts each. Not all virtual machine VLANs are visible on every single vmnic. For example, cluster 1 sees VLAN 2, 3, and 9, but cluster 2 only sees VLAN 5, and 9. However, the appropriate host vmnic has / will be assigned to the appropriate uplink for each distributed port group / VLAN that it can see.
We started our migration last week with only Management and vMotion. This has gone well, with 12 hosts migrated so far. Once this phase is complete, we'll perform our virtual machine migration over a single night.
Finally, on to our problem! Everything seems to be functioning normally with Management and vMotion. All hosts are visible, respond to a ping, are joined and participating in the cluster, with functional vMotion. However, we see inconsistent Ethernet icon colors in the DVUplink Ethernet symbols. Some are green, some are blank (white). Some hosts have 1 white and 1 green. Some have 2 green. Some have 2 white. Yes, the 2 white hosts are functioning normally! The information link for every single port shows connected. What is the deal?
