I am using VMWare ESXi 5.
The ESXi 5 host has a NIC with the MAC XXXX:XXXX:XXXX
I have created a console port for that card.
I have added a new guest named COMP and manually assigned the MAC address of the physical NIC (XXXX:XXXX:XXXX - same as the NIC of the ESXi server).
Now COMP reports it has a NIC with IP Y.Y.Y.Y, same IP in VMWare port definition.
COMP will not respond to pings.
When I logged in to the ESXi host via SSH and tried to ping other computers connected to it, COMP responded to pings for the rest of the day. After a pingless night (or no communication to COMP) - COMP returned to the point of not responding to pings again.
- What's happening there?
- IS there a way to resolve this issue such that COMP will always respond to pings?
Here are the full details: I am trying to test driver setup of a NIC I personally develop.
Up until now I had to install a new machine for the test. Lately I encountered VMware, and I thought it can be set up to virtualize a machine such that it's NIC behaves just like the physical NIC of the virtual machine.
This question led me to believe this setup was possible: Can't ping host from vmware guest using bridged networking.
A MAC address conflict is what's happening. Since you have two objects with the same MAC address, packets bound for either device will end up at one or the other, and where you "intend" for them to go doesn't matter a whole lot.
Don't assign the same MAC address to your "COMP" as you have for your VMWare server. MAC addresses should be unique, when you have duplicates on your network things like this happen.
To address your most recent content update, if you want to "pass" the physical NIC through to the guest. This article describes it for ESXi 4 and this article walks you through it for ESXi 5. You use DirectPath I/O to present the physical NIC directly to the guest. When you do this, the host will not use that NIC at all.
The way you're trying to do it now is just basically causing you to arp poison yourself.