Just wanted to say that we encountered the ESXi 6.0 Update 2 bug, https://kb.vmware.com/kb/2144968
, that Gostev described in his "The word from Gostev" the previous two weeks.
We have a ship with a DELL Vertex onboard, where we have 2 blades running ESXi 6.0.0 Update 2, build 3620759. Veeam is a virtual server running on one of the blades. This VM has Hardware version 11, and VMXNET3 network adapter.
The Vertex also has a 10GbE internal network. So far we have solved this temporarily by this step: "Disable hardware LRO for VMXNET3 on the ESXi host".
The Veeam Server is running Windows Server 2012 R2, and I tried to find if "The virtual machines have Large Receive Offload (LRO) for VMXNET3 NICs enabled", but all I can find on
vmxnet3 Ethernet Adapter Properties in the OS is "Large Send Offload V2", not "Large Receive Offload". Maybe I am looking in the wrong place.
There is no Veeam supportcase for this incident.
Hope VMware gets better quality assurance in the future.