Netplan not creating .network file under certain conditions

[James Denton]

OS: Ubuntu 17.10

Here's the scenario:

network:
  version: 2
  ethernets:
    ens160:
      addresses: [10.50.0.9/24]
      gateway4: 10.50.0.1
      nameservers:
        addresses: [8.8.8.8,8.8.4.4]
  vlans:
    ens160.100:
      id: 100
      link: ens160
    ens160.200:
      id: 200
      link: ens160
    ens160.300:
      id: 300
      link: ens160
  bridges:
    br-mgmt:
      interfaces: [ens160.100]
      addresses: [192.168.100.1/24]
    br-overlay:
      interfaces: [ens160.200]
      addresses: [192.168.200.1/24]
    br-storage:
      interfaces: [ens160.300]

In this configuration, a 'netplan apply' applies the following configuration:

root@ubuntu:~# brctl show
bridge name bridge id STP enabled interfaces
br-mgmt 8000.7282c779e298 no ens160.100
br-overlay 8000.ce8884059e7b no ens160.200
br-storage 8000.b6fc84976f4f no ens160.300

root@ubuntu:~# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether 00:50:56:a8:92:85 brd ff:ff:ff:ff:ff:ff
    inet 10.50.0.9/24 brd 10.50.0.255 scope global ens160
       valid_lft forever preferred_lft forever
    inet6 fe80::250:56ff:fea8:9285/64 scope link
       valid_lft forever preferred_lft forever
4: br-overlay: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether ce:88:84:05:9e:7b brd ff:ff:ff:ff:ff:ff
    inet 192.168.200.1/24 brd 192.168.200.255 scope global br-overlay
       valid_lft forever preferred_lft forever
    inet6 fe80::cc88:84ff:fe05:9e7b/64 scope link
       valid_lft forever preferred_lft forever
5: br-mgmt: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 72:82:c7:79:e2:98 brd ff:ff:ff:ff:ff:ff
    inet 192.168.100.1/24 brd 192.168.100.255 scope global br-mgmt
       valid_lft forever preferred_lft forever
    inet6 fe80::7082:c7ff:fe79:e298/64 scope link
       valid_lft forever preferred_lft forever
6: ens160.200@ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-overlay state UP group default qlen 1000
    link/ether 00:50:56:a8:92:85 brd ff:ff:ff:ff:ff:ff
7: ens160.100@ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP group default qlen 1000
    link/ether 00:50:56:a8:92:85 brd ff:ff:ff:ff:ff:ff
8: ens160.300@ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-storage state UP group default qlen 1000
    link/ether 00:50:56:a8:92:85 brd ff:ff:ff:ff:ff:ff
9: br-storage: <BROADCAST,MULTICAST> mtu 1500 qdisc noqueue state DOWN group default qlen 1000
    link/ether b6:fc:84:97:6f:4f brd ff:ff:ff:ff:ff:ff

root@ubuntu:~# ls -l /run/systemd/network
total 48
-rw-r--r-- 1 root root 34 Oct 23 16:09 10-netplan-br-mgmt.netdev
-rw-r--r-- 1 root root 57 Oct 23 16:09 10-netplan-br-mgmt.network
-rw-r--r-- 1 root root 37 Oct 23 16:09 10-netplan-br-overlay.netdev
-rw-r--r-- 1 root root 60 Oct 23 16:09 10-netplan-br-overlay.network
-rw-r--r-- 1 root root 37 Oct 23 16:09 10-netplan-br-storage.netdev
-rw-r--r-- 1 root root 50 Oct 23 16:09 10-netplan-ens160.100.netdev
-rw-r--r-- 1 root root 89 Oct 23 16:09 10-netplan-ens160.100.network
-rw-r--r-- 1 root root 50 Oct 23 16:09 10-netplan-ens160.200.netdev
-rw-r--r-- 1 root root 92 Oct 23 16:09 10-netplan-ens160.200.network
-rw-r--r-- 1 root root 50 Oct 23 16:09 10-netplan-ens160.300.netdev
-rw-r--r-- 1 root root 92 Oct 23 16:09 10-netplan-ens160.300.network
-rw-r--r-- 1 root root 142 Oct 23 16:09 10-netplan-ens160.network

Here's the problem: In our environment, there's no need to have an IP address on a bridge, and this has been reflected in the yaml for the br-storage interface. Without an IP address assigned via dhcp: true or addresses: [x.x.x.x/x], the br-storage interface is configured in a DOWN state and won't pass traffic. I can set the interface UP with iproute2 and it will pass traffic as expected.

With ifupdown, we were able to bring up the br-storage interface at boot with the following config:

auto br-storage
iface br-storage inet manual

During troubleshooting, I found that without specifying an address in the yaml, netplan is not creating the respective .network file for the interface in /run/systemd/network. I created a file named '10-netplan-br-storage.network' with the following contents:

[Match]
Name=br-storage

[Network]

Restarted systemd-networkd with the following:

systemctl restart systemd-networkd

As a result, the br-storage interface came up and bridged traffic as expected:

root@ubuntu:/run/systemd/network# ip link show br-storage
9: br-storage: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP mode DEFAULT group default qlen 1000
    link/ether b6:fc:84:97:6f:4f brd ff:ff:ff:ff:ff:ff

To resolve this, perhaps .network files can still get created when either 'dhcp: false' or 'addresses: []' or something along those lines, but the [network] block can be empty as shown above.

[Daniel Axtens (daxtens)]

Hi,

(I'm trying to tidy up some old bugs so apologies if this seems a bit out of the blue!)

This is a known issue and I'm pretty sure what you describe is an exact duplicate of LP: #1736975. (perhaps looking at the dates it would be fairer to say that that issue is a duplicate of this, as this was filed first! oh well.)

I'm going to mark it as a duplicate for now but if that's wrong please let me know.

Regards,
Daniel