Troubleshooting server bond problems

Bond exception troubleshooting when doing bond verification, the original bond is done on the same network card, and after deleting the configuration, in the bond

Ifenslave-d bond0 eth5

Stop the eth5 network port, how to eth0 to make up the heat.

By default, this interface 0 4 is the mac that joins the bond0 group and displays the mac address of the active physical interface.

So this error occurs in ifdown, which means your bond uses my address, so you can't give me to down.

Server bond failure.

Knowledge: when initializing bond, the mac address of bond will choose one of the network cards.

One of the situations encountered in production is. One of the network cards in bond belongs to down and is kicked out with ifenslave because bond uses its mac address

It causes the mac address of bond to change, and the server cannot communicate with ping due to the mac address admission on the switch.

The important thing, if you don't think it's easy to change it, do it in two or three minutes and open the failover of the database. After discovering that the server was down.

Take the lead in going to the remote management card to turn off failover, and then turn on mac address access.

Ifenslave-c bond0 eth2 uses a physical interface eth2 as the active interface

# ifenslave-d bond0 eth0 / / removing eth0 from bond0 is not recommended

In the future, I can switch the bond to another network card without using ifdown eth0 or ifconfig eth0 down.

You can directly use this ifenslave-c bond0 eth2 to use a physical interface eth2 as the active interface