Issue with priority of interfaces in bonding

[kr5ch]

Hello, I have the following interfaces: 10G NIC (set up as eth0) and a 1G NIC (set up as eth1) bonded together in active-backup. The bond works fine but it always defaults to eth1 for some reason, every time I reboot the server it switches back to eth1.

 

I'm viewing the active slave using /sys/class/net/bond0/bonding/active_slave and changing it using

 

echo eth0 > /sys/class/net/bond0/bonding/active_slave

 

but I would really like the bond to always use eth0 when it's up. Any way to force it ?

 

Thanks in advance for the help.

[Herbiewalker]

I am looking at this same setup with a 10G NIC in a PCI slot, and the MotherBoard 1G NICs. Haven't been successful in getting all the traffic to prioritize the 10G connection....

 

Hopefully someone has the answer!

[kr5ch]

Bump, anybody has an idea ?

[Vr2Io]

On 3/14/2021 at 12:17 AM, kr5ch said:

echo eth0 > /sys/class/net/bond0/bonding/active_slave

You may add it in "go" file then execute once system reboot.

Edited March 24, 2021 by Vr2Io

[FirbyKirby]

I was able to solve this using the thread https://forums.unraid.net/topic/102194-priority-of-active-backup-bond/. Linking here so others may benefit.

[kr5ch]

1 hour ago, FirbyKirby said:

I was able to solve this using the thread https://forums.unraid.net/topic/102194-priority-of-active-backup-bond/. Linking here so others may benefit.

Thanks @FirbyKirby, that did the trick for me too.