Skip to content

Conversation

@donald
Copy link
Collaborator

This is designed to be used on a routeri to router link so use xmit hash of layer 2+3 in a LAG bond to get a better load balance.

This is designed to be used on a router-to-router link so use a xmit
hash policy of layer 2+3 in a LAG bond to get a better load balance.
@donald Donald Buczek (donald) merged commit 70370cc into master Oct 7, 2026
@pmenzel
Copy link
Contributor

Is there a way to test the improved balancing?

@donald
Copy link
Collaborator Author

Is there a way to test the improved balancing?

Yes. Here's a tool:

root@bka:~# mxrouterctl exec /home/buczek/monitor-ifload.pl

Before: net02_tx used to be fixed to zero:

net01_rx: 17894       
net02_rx: 11749       
gnz_rx  : 29643       
net01_tx: 13164       
net02_tx: 0           
gnz_tx  : 13164       

After:

net01_rx:    7466574
net02_rx:      19187
gnz_rx  :    7485761
net01_tx:    1217434
net02_tx:    1838649
gnz_tx  :    3056083

Note, that the accumulation interval has been increased from a few seconds to 5 minutes, so the numbers have other magnitudes. And you need to wait 5 minutes if you want to reproduce. Sometimes a link is dominated by a specific connection, e.g. backup traffic to the Proxmox backup server.

Our rx side (the GNZ sending side) is not balanced, but that is a problem on the other end and F.M. want to look into it.

D.

Sign in to join this conversation on GitHub.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants