Proxmox LAG setup.....

[techie_pro]

Hi
I want to created a Link aggregation group(LAG) setup in DELL EMC server (2x10G eth ports + 4x1G eth ports).


Both the PC must be able to ping 192.168.22.C, physically the node is connected to a switch (plug and play) which is further connected to LAN.
If I remove the eth cable between eth1 and switch the traffic must switch towards the second port and vice versa (still I must be able to ping 192.168.22.C)
Diagram for reference. Can someone help ?

 

[VictorSTS]

LAG is configured as layer2, so 2 nic's of your server should be added as a bond0 and then vmbr0 has to be configured to use bond0 as it's port. You can't have different IP addresses on each nic that you use for the bond.

This is a similar setup:

auto lo
iface lo inet loopback

iface eno1 inet manual

iface eno2 inet manual

auto bond0
iface bond0 inet manual
      bond-slaves enp6s18 enp6s19
      bond-miimon 100
      bond-mode active-backup

auto vmbr0
iface vmbr0 inet static
        address  10.0.10.10/24
        gateway  10.0.10.254
        bridge-ports bond0
        bridge-stp off
        bridge-fd 0

I suggest that you start using bond mode "active-backup" as it's the easier to begin with.

 

[techie_pro]

Thanks VictorSTS.
One more query, is there a way to know which ethernet port is active and which is backup on a particular instant of time ?

 

[VictorSTS]

Not from the GUI, but the one set as primary will always be the running nic as long as it is in working state. The backup link is only used if primary fails (i.e. it has no link).

This may help you finding out the current state of your nics:

https://forum.proxmox.com/threads/d...on-vmbrs-with-linux-vlans.132545/#post-583763

 

[LnxBil]

One more query, is there a way to know which ethernet port is active and which is backup on a particular instant of time ?

The current information is in /proc/net/bonding/bond0. I recommend having an extenal monitoring that does the aquisition of data over time in order to get what you want.