[plug] Bridging Interfaces in Ubuntu [SOLVED... at least I thought so]

Leon Wright techman83 at gmail.com
Mon Mar 26 12:34:58 WST 2012


Hi All,

This is a verbose copy of what I sent to the SAGE-AU list for any
members that are on both. I'm running out of hair to pull!

DEV cluster Setup:

2x ESXi 5 nodes (fully licensed in a cluster)
Distributed Switch with a crossover cable physical uplinks for common
networking for VMs on different nodes
2x port groups with private VLANs for simulation of WAN sites
vSwitch on one node to a physical port in the lab
Ubuntu 10.04 box for general routing

Now we're attempting to make a port available to one of the Private
VLANs and there didn't seem to be a way to add it to the distributed
switch, so we figured bridging in linux would be much simpler.
Interface on a vSwitch attached to a Physical port, Interface inside
the the Distributed port group.

What works:
Both sides can ping the Bridge IP
Broadcast traffic goes from one side of the network to the other (arp/dhcp/etc)

What doesn't work:
Everything else!

Things we've tried:
vSwitch/dVswitch security settings, Allow Promiscuous mode/mac
spoofing  - this was actually required to allow both sides to ping the
bridge IP
echo 0 > /proc/sys/net/bridge/*
altering vmx file to allow network card promiscuous mode
E1000/VMXNET3 card types
Duplicated in the physical world  "Just Works" (disclaimer, not
exactly the same setup, but wanted to make sure bridging is as easy as
I recall it).

We've run out of ideas. Anyone have something else I can try? Have a
better idea of a native vmware way to get a physical interface into a
distributed port group?

We could tie all the WANSITE1 VMs onto one node and use vSwitches, but
we really like the flexibility that the dVswithes give us. Also we
have a lot of VMs to replicate for that site, so we may run out of
resources doing that.

Thanks,

Leon
--
DRM 'manages access' in the same way that jail 'manages freedom.'

# cat /dev/mem | strings | grep -i cats
Damn, my RAM is full of cats... MEOW!!


More information about the plug mailing list