How-To LAG (I think)…
dataresilience last edited by
So I am trying to add a little bit of redundancy to my setup and I need to understand a few pieces in pfSense that was not needed until now.
Here is the setup.
a) I have two NetGear XSM2772S 10GbE L2+ managed switches stacked together across two DAC cables at port 17 & 18.
b) I am running the 2.1 release of pfSense on a small Atom based flipped Supermicro (where ports face the rack front).
c) I have a dual port (supported) 10GbE Intel card in the Atom.
d) The WAN is the onboard 1GbE port and should remain so.
e) The LAN currently is the second onboard 1GbE port.
f) The setup is using VLAN's as well.
The LAN is attached to one of the stacked switches and can be reached from any port on any of the two switches due to the netgear stacking.
All well and done. However, I would now like to bond the two 10GbE ports in the pfSesne having each of the ports, DAC-attached to switch port 19 in the XSM7224S, and instead use this as the 192.168.1.1 GW LAN replacing the 1GbE onboard and having the two 10GbE ports act as one IP address (LAG?) in a loadbalanced or roundrobin or whatever fashion.
The end result should be that if I have a server (having its two ports bonded) attached to both switches I could lose/service any of the two switches and still be in business.
- Could this be done?
- Is this the way you would connect things?
- How should I configure pfSense to play with me in this game?
- What are the steps I need to take?