Interfaces Members of LAGG all appear with same MAC address



  • Hi
    New fresh install of 2.2.4. Setting up a 3 link LAGG as failover. When I select the interfaces em4, em5, em6 I can see that all the mac address different. After I create the LAGG and I got back into to it, they still comes up as em4-5-6 but they all have the same mac address as em4.
    Never seen this before.
    Installation is on VMWare Esxi 6.0


  • Rebel Alliance Global Moderator

    on esxi?  Why would you be creating a lagg in pfsense?  Your failover would on your physical nics connected to your vswitch in esxi..


  • Rebel Alliance Developer Netgate

    @ianstrainor:

    Hi
    New fresh install of 2.2.4. Setting up a 3 link LAGG as failover. When I select the interfaces em4, em5, em6 I can see that all the mac address different. After I create the LAGG and I got back into to it, they still comes up as em4-5-6 but they all have the same mac address as em4.
    Never seen this before.
    Installation is on VMWare Esxi 6.0

    That is normal with LAGG/LACP. It's trying to make the switch believe it's a single aggregated port all with the same MAC.

    @johnpoz:

    on esxi?  Why would you be creating a lagg in pfsense?  Your failover would on your physical nics connected to your vswitch in esxi..

    That's also true – if you want multiple physical ports, you'd do that at the hardware level in ESX, not in the guest.



  • So LAGs can be set up in two different modes;
    Automatically over the LACP = dynamic LAG
    Manual set up = static LAG ech site entry must be matching exactly to the entries on the other side!

    New fresh install of 2.2.4. Setting up a 3 link LAGG as failover.

    Why? This is then an active/passive scenario!!! If you set it up as an failover, each link will not be used at the same time! Only one is used and if this link is down the next one is "jumping" in and over takes the link and data.

    So if you set up an LAG (LACP) and all links are in usage, and then one is failing the other were running
    on and so you will be have the benefit that all links are aggregated together to one fat pipe and the failover
    will be also integrated and there! Please think about this.

    When I select the interfaces em4, em5, em6 I can see that all the mac address different. After I create the LAGG and I got back into to it, they still comes up as em4-5-6 but they all have the same mac address as em4.

    Normally a LAG will be created as the following procedure:

    • you chose one LAN port or adapter and this will be the LAG primary port and then the MAC address
      from this port will be used for the whole LAG!
    • and the other ports will be all the secondary ports that must be added to the first one the primary port!

    Never seen this before.

    Every network switch and LAG capable device is doing it like so, because the LAG is a both ended
    thing, that must be configured on both ends identically or it will fail, either automatically over the
    LACP or manual.


  • Rebel Alliance Global Moderator

    dude he is on esxi – there should be no reason why he should be trying to setup a lagg in pfsense. NONE!!!  unless he had interfaces in multiple vswitches???  For what Fing reason?

    If he wants failover for nics on his esxi host then he would set that up in esxi, not pfsense VM..