Closed jjakob closed 4 years ago
The service also starts correctly if 02-eth0-1 is renamed and modified to the correct name eth0.1@eth0 instead of disabling it. eth0.1@eth0 stays down in any case (with blue or black connected) so I don't understand why it's there/should have a dhcp client on it.
I wouldn't expect it to bring up eth0.1. That interface only exists in old 4.4 kernels with the Ralink driver for the ethernet switch. For newer kernels with the mainline switch driver, it is ethblack that should be configured. Without logs I cannot suggest why ethblack didn't get configured (or maybe it did, but you looked in the wrong place...??).
This all works for me, so I'm going to close this bug. If you want it pursued and have logs, feel free to re-open. Thanks.
On gbpc2-5.4, the networking.service fails to start after a fresh installation via
config
from the initramfs. The first boot after installation didn't even get an address on the connected interface (sorry, I don't have the logs). It can't bring up eth0.1. I think this is because only eth0.1@eth0 exists which is not the same name.Then I upgraded to Debian Buster and rebooted, this is the relevant log cut out (ethblack got an address this time):
Configuring eth1, disabling eth0.1@eth1 (moving /etc/interfaces.d/02-eth0-1 to 02-eth0-1.disabled) and restarting networking.service fixes the service:
Questions: