VMware bridge network loop problem?

[DavidVermillion]

I have a two port Mellanox Connect X-3 in my vmware host, with one end connected to my 10gb switch. I've got my workstation connected to the other 10gb port. I was hoping to add another host at 10gb by making a sort of chain.

I used the first port, connected to my main switch to put all of my VMs on one Vswitch. This works fine, all of my VMs can access the network and the network can access my VMs.

I used the other port on the nic to directly connect via fiber to a Mellanox Connect X-2 in the 3rd host. I created another vswitch assigned only to that 2nd port. I then created a freeBSD vm with two nics, one for the primary vswitch, and another for the vswitch assigned only to the 2nd port.

I created a bridge between these two nics, which theoretically should have connected the 3rd host to the main network. What actually happened was that my entire network went down, until I killed the freeBSD vm. I did the same thing with a linux VM, and got the exact same result when I brought up the bridge.

What am I missing here? I realize this is a stupid way to go about it. I know it probably won't perform well, but I just wanted to TRY it instead of having to buy a 500+ dollar switch.

 

[whitey]

Ya gotta love how VMware professes they participate in NO STP shinanigans but I have seen in a few occasions due to OS misconfigurations loops originate from a virtual network segment. Creating a bridge using two vnics on the same vswitch/vlan is a sure-fire way to create a loop is all I know. Seen core switch sup mods go ballistic a time or two in our enterprise data centers when this happened. I realize not a direct answer but at least some background.

Other than that you don't wanna hear my other answer :-D

 

[DavidVermillion]

I know, I know. I'm like someone who gets a hold of some propane and some pipe and instead of installing gas plumbing makes a potato gun. No one ever in their right mind would do what I'm trying to do...but dammit, what else are homelabs for?