I am very confused at this. We are having intermittent connection issues between 2 VMs that reside on teh same subnet. to give you a quick run down:
ESXi 4.1, Broadcom NICs, Nexus 4000 switches
Cisco ACE 20
vswitch1 Nic1 goes to physical_switch1
vswitch1 Nic2 goes to physical_switch2
web1 - 192.168.1.1 VLAN1
server1 - 192.168.1.2 VLAN2
When WEB1 & Server1 are on the same ESX host, same vswitch (different portgroups due to different vlans), same physical_switch, and under heavy web traffic load.... there will be packet loss. All criterial must be met.
So far I have tested many senerios, and the above config gives me problems.
The following Configs work fine:
If the web server is not under load like off hours
If you force the portgroups to use seperate Physical switches (I.e. web1 on sw1, server1 on sw2)
If they are on seperate ESX hosts
Please let me know what you think. FYI, this happens on all webservers that are connected to the ACE.