4 LAN Interface Question
-
What is the best way to setup the LAN side of my firewall so that all 4 LAN interfaces (2 per box) have the same VIP.
I have tried using LAGG and unless I am doing something wrong I get no options to add interfaces to the menu.
Thanks

 -
Why do you want to use LAGG interfaces? If it is for redundancy, you already will achieve that with CARP.
-
We need no single point of failure hence each firewall running with 2 switches.
Is it possible to create a CARP across multiple interfaces on the same box?
LAN1:
192.168.1.1 - firewall 1
192.168.1.2 - firewall 2LAN2:
192.168.1.3 - firewall 1
192.168.1.4 - firewall 2CARP LAN1:
192.168.1.10CARP LAN2:
192.168.1.11GATEWAY CARP:
192.168.1.20is the above possible with carp?
-
Separate CARP instances should be setup for LAN and WAN sides.
If you want multiple interfaces in each, then you would have to use LAGG, and then CARP on top of that, but that is pointless.
A traditional pfSense cluster with 1x SYNC + 1x LAN, and 1x WAN running CARP represents no single point of failure.
If the BACKUP pfSense stops hearing updates from the MASTER on ANY CARP interface, it will fail-over and become MASTER.Adding LAGG to the setup will add redundancy to the redundancy, and will complicate the setup a lot. It and will not allow you to detect when one member of the LAGG goes down, so you could end up with degraded performance, and CARP wouldn't fail-over since it doesn't see it as a problem.
-
It and will not allow you to detect when one member of the LAGG goes down
Well you could look for traps from the switch/stack doing the LACP for LACP issues but it really seems like overkill but it depends on the application.
Everything always comes down to the endpoints. Unless you are going to LACP to two NICs in every endpoint to two different switches (You can LACP a group across stack members or sometimes with multi-chassis trunking), when the switch that the endpoints are connected to has a problem, those endpoints lose connectivity.
On all of your LANs:
X.X.X.1 CARP
X.X.X.2 Master interface
X.X.X.3 Backup interfaceAll clients pointed to .1 for routing, DNS, etc.