Note: this question has changed while investigating problem to stay up to date.
Note 2: This question is invalid, the described problem is gone. Askubuntu prevents me from deleting it.
Here is my /etc/netplan/01.yaml
config on my new 18.04 headless server. The file describes two of my hardware interfaces i may use:
network:
version: 2
renderer: networkd
ethernets:
enp2s0:
optional: true
dhcp4: yes
wifis:
wlx0013eff10948:
access-points:
"the-network":
password: "the-password"
dhcp4: yes
The problem is that Wifi connection sometimes can die. Or there is no connectivity at boot time. The same conditions are for Ethernet interface.
How can i tweak my system to adapt routing table and and etc to these changes? For now, it seems that any hardware change breaks system's "default gateway".
As a result, i can have one or both my interfaces with an assigned IP addresses without working ping
in my system. And there is not problem with access point or router behind cable - my setup just cannot survive "changes" and switches..
ifconfig summary
docker0
(br)
enp2s0
(no inet)
lo
tun0
(openvpn - there is client daemon that needs to be connected to my VPN)
wlx0013eff10948
(inet 192.168.1.x) = shows as connected, but internet is not working
Less relevant information
wlx0013eff10948 is rtl8192cu
USB dongle. I only installed linux-generic-hwe-18.04
for it to appear and work (tested with wicd-curses
). As it turns out, there is no problem with WIFI - just the way system handles the lost wifi (or cable) connection.
Sometimes i can see this output appearing in machine's tty:
[ 232.2332] rtlwifi: AP off, try to reconnect now
Exactly at the state when i don't have wifi cable and wifi connection has died, but should be re-established. Which is not happening