Network completely 'dies' upon adding second vSwitch

[breakaway9000]

Hi All,

I'm running VMWare ESXi 4.1 on an IBM x3500 MT7977. The server has dual gigabit. Initially, only one network port was used (vmnic0) - it was used for both management and also Virtual Machines (I know this is bad, but it was a test build so hey).

Then I decided to try out a routing applicance. To do this, I needed to

- Create a second vSwitch (vswitch1) and connect it to vmnic1; This vSwitch is created in addition to my exsiting vSwitch (vswitch0) connected to vmnic0.

- Plug in the second network adapter (i.e. vmnic1) into the switch with a network cable

- Put both the switches on promiscuous mode

So that the routing appliance would work. Then, I used this guide. I've read multiple guides on running routing applicances virtualized, the settings seem pretty similar so I set it up exactly as described. So far so good right?

Wrong. About 2 minutes after I completed the aforementioned changes, lost all network functionality in my entire network. Not only did the vSphere client lose connection to the ESXi host on my workstation, but other machines on the network also lost communication between each other / the internet - pings etc would not work.

However the ESXi 4.1 host was still functioning perfectly - I could ping the internet from it.

I disconnected the most recently connected network cable (i.e. vmnic1 / vSwitch1) and about 2 minutes later all network functionality was restored.

I'm doing something really silly here aren't I? I feel like I'm missing something unbelievably simple. Can someone please point me in the correct direction?

NOTE: The second NIC (vmnic1) doesn't use VLAN tagging, also it isn't on a spearate netowrk, it's on the same network as vnic0 and vSwitch 0. The reason for this is that I want to only use the routing applicance for a couple of the virtual machines housed on the host, the rest need direct access to the network without a router inbetween.

Thanks in advance!

[taylorb]

If it killed connectivity outside the VM environment, then you likely had something horrible running on your routing appliance that was messing with routing, DNS, etc.  Perhaps you accidentally set the IP of the device to the IP of the default gateway or something heinous like that?

[breakaway9000]

Hi taylorb, thanks for the reply.

However I don't think that's it - the first time I tried this, I thought the same thing, so I actually set up the vSwitch and walked away. Never powered on the routing VM. When I returned about 10 minutes later, the network was completely down, including other physical workstations / servers on the subnet. Pulling the network cable fixed this problem though.

However when I went to the ESXi host and logged into the command prompt, I was able to ping the internet (I think I might have been able to ping other systems on my LAN also, but I can't 100% remember).

What else could be causing this? Does the second network adapter, upon connection to the network need an IP address of its own? I suppose it would, wouldn't it? How would I go about setting this IP address (I've looked everywhere and can't seem to find it) - or is that not how additional vSwitches / vmnics work?

TIA!