[lxc-users] LXD static IP in container

Thu Feb 20 00:02:18 UTC 2020

When I factor out the bonded network configuration, your LXD profiles 
and containers look just like mine.  There is an existing non-LXD 
bridge, a profile which describes it as "bridged", and containers which 
use that profile.

On my system, the NIC gets two IP addresses:  one configured by the 
container (the desired static IP) and another configured using DHCP by 
LXD on the host.  dhclient is not installed in the container.

This seems to be as described in
https://blog.simos.info/how-to-make-your-lxd-containers-get-ip-addresses-from-your-lan-using-a-bridge/
I thought that this showed the container getting an IP from the the LAN 
using network.services using DHCP, but I think that it is LXD getting 
the IP address.

It seems that there must something configured differently on the host.

On 2/12/20 8:08 AM, Michael Eager wrote:
> On 2/11/20 4:57 PM, Joshua Schaeffer wrote:
>> Not sure this will help but I provided my configuration for LXD below. 
>> I use Ubuntu so you'd have to translate the configuration network 
>> configuration portions over to RedHat/CentOS. My containers' configure 
>> their own interfaces (static, dhcp, or whatever), LXD simply defines 
>> the interface. These are the basic steps that I do:
>>
>>  1. On the LXD host I setup bridges based on the vlan's that I want a
>>     NIC to connect to. Those vlan interfaces use a bond in LACP mode. If
>>     you don't use vlan's or bond's in your setup then just create the
>>     bridge from a physical Ethernet device.
>>  2. I then create a profile for each bridge corresponding to a vlan.
>>  3. When I create a container I can assign those profiles (one or
>>     multiple) to create the network devices.
>>  4. Inside the container I configure the network device just like any
>>     other system; physical, VM, container, or otherwise.
>>
>> I do not use LXD managed network devices. All my network devices are 
>> managed by the host operating system. Again, if you don't use vlan's 
>> or bond's then you can jump straight to creating a bridge.
>>
>> Here's the details of the steps:
>>
>> Step 1:
>> Create the network devices that the LXD containers will use.
>>
>> lxcuser at blllxc02:~$ cat 
>> /etc/network/interfaces.d/01-physical-network.device
>> # This file contains the physical NIC definitions.
>>
>> ############################
>> # PHYSICAL NETWORK DEVICES #
>> ############################
>>
>> # Primary services interface.
>> auto enp3s0
>> iface enp3s0 inet manual
>>      bond-master bond-services
>>
>> # Secondary services interface.
>> auto enp4s0
>> iface enp4s0 inet manual
>>      bond-master bond-services
>>
>> lxcuser at blllxc02:~$ cat /etc/network/interfaces.d/02-bonded.device
>> # This file is used to create network bonds.
>>
>> ##################
>> # BONDED DEVICES #
>> ##################
>>
>> # Services bond device.
>> auto bond-services
>> iface bond-services inet manual
>>      bond-mode 4
>>      bond-miimon 100
>>      bond-lacp-rate 1
>>      bond-slaves enp3s0 enp4s0
>>      bond-downdelay 400
>>      bond-updelay 800
>>
>> lxcuser at blllxc02:~$ cat /etc/network/interfaces.d/03-vlan-raw.device
>> # This file creates raw vlan devices.
>>
>> ####################
>> # RAW VLAN DEVICES #
>> ####################
>>
>> # Tagged traffic on bond-services for VLAN 28
>> auto vlan0028
>> iface vlan0028 inet manual
>>      vlan-raw-device bond-services
>>
>> # Tagged traffic on bond-services for VLAN 36
>> auto vlan0036
>> iface vlan0036 inet manual
>>      vlan-raw-device bond-services
>>
>> # Tagged traffic on bond-services for VLAN 40
>> auto vlan0040
>> iface vlan0040 inet manual
>>      vlan-raw-device bond-services
>> ...
>>
>> lxcuser at blllxc02:~$ cat /etc/network/interfaces.d/04-bridge.device
>> # This file creates network bridges.
>>
>> ##################
>> # BRIDGE DEVICES #
>> ##################
>>
>> # Bridged interface for VLAN 28.
>> auto vbridge-28
>> iface vbridge-28 inet manual
>>      bridge_ports vlan0028
>>      bridge_stp off
>>      bridge_fd 0
>>      bridge_maxwait 0
>>
>> # Bridged interface for VLAN 36.
>> auto vbridge-36
>> iface vbridge-36 inet manual
>>      bridge_ports vlan0036
>>      bridge_stp off
>>      bridge_fd 0
>>      bridge_maxwait 0
>>
>> # Bridged interface for VLAN 40.
>> auto vbridge-40
>> iface vbridge-40 inet manual
>>      bridge_ports vlan0040
>>      bridge_stp off
>>      bridge_fd 0
>>      bridge_maxwait 0
>>
>> Step 2:
>> Create profiles for the network devices. Technically not required but 
>> helps to setup new containers much more quickly.
>>
>> lxcuser at blllxc02:~$ lxc profile list
>> +----------------------+---------+
>> |         NAME         | USED BY |
>> +----------------------+---------+
>> | 1500_vlan_dns_dhcp   | 5       |
>> +----------------------+---------+
>> | 28_vlan_virt_mgmt    | 15      |
>> +----------------------+---------+
>> | 40_vlan_ext_core_svc | 0       |
>> +----------------------+---------+
>> | 44_vlan_ext_svc      | 4       |
>> +----------------------+---------+
>> | 48_vlan_ext_cloud    | 0       |
>> +----------------------+---------+
>> | 80_vlan_int_core_svc | 2       |
>> +----------------------+---------+
>> | 84_vlan_int_svc      | 4       |
>> +----------------------+---------+
>> | 88_vlan_int_cloud    | 0       |
>> +----------------------+---------+
>> | 92_vlan_storage      | 0       |
>> +----------------------+---------+
>> | default              | 15      |
>> +----------------------+---------+
>>
>> lxcuser at blllxc02:~$ lxc profile show 28_vlan_virt_mgmt
>> config: {}
>> description: ""
>> devices:
>>    mgmt_net:
>>      name: veth-mgmt
>>      nictype: bridged
>>      parent: vbridge-28
>>      type: nic
>> name: 28_vlan_virt_mgmt
>>
>> Step 3:
>> Create the container with the correct profile(s) to add the network 
>> device(s) to the container.
>>
>> lxcuser at blllxc02:~$ lxc init -p default -p 28_vlan_virt_mgmt -p 
>> 44_vlan_ext_svc ubuntu:18.04 bllmail02
>>
>> Step 4:
>> Connect to the container and setup the interface the same way you 
>> setup any other system. The example below is set to manual but just 
>> change to however you want to setup your device.
>>
>> lxcuser at blllxc02:~$ lxc exec bllmail02 -- cat 
>> /etc/network/interfaces.d/51-container-network.device
>> auto veth-mgmt
>> iface veth-mgmt inet manual
>> ...
>>
>> auto veth-ext-svc
>> iface veth-ext-svc inet manual
>> ...
>>
>> lxcuser at blllxc02:~$ lxc exec bllmail02 -- ip link show veth-mgmt
>> 316: veth-mgmt at if317: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc 
>> noqueue state UP mode DEFAULT group default qlen 1000
>>      link/ether 00:16:3e:f6:e5:ec brd ff:ff:ff:ff:ff:ff link-netnsid 0
>> lxcuser at blllxc02:~$ lxc exec bllmail02 -- ip -4 addr show veth-mgmt
>> 316: veth-mgmt at if317: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc 
>> noqueue state UP group default qlen 1000 link-netnsid 0
>>      inet 10.2.28.129/22 brd 10.2.31.255 scope global veth-mgmt
>>         valid_lft forever preferred_lft forever
>>
>> lxcuser at blllxc02:~$ lxc exec bllmail02 -- ip link show veth-ext-svc
>> 314: veth-ext-svc at if315: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 
>> qdisc noqueue state UP mode DEFAULT group default qlen 1000
>>      link/ether 00:16:3e:21:ac:dc brd ff:ff:ff:ff:ff:ff link-netnsid 0
>> lxcuser at blllxc02:~$ lxc exec bllmail02 -- ip -4 addr show veth-ext-svc
>> 314: veth-ext-svc at if315: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 
>> qdisc noqueue state UP group default qlen 1000 link-netnsid 0
>>      inet 192.41.41.85/26 brd 192.41.41.127 scope global veth-ext-svc
>>         valid_lft forever preferred_lft forever
>>
>> -- 
>> Thanks,
>> Joshua Schaeffer
> 
> Thanks.
> 
> That's a lot to unpack and translate from Ubuntu to CentOS.
> 
> -- Mike Eager
> 
> 
> _______________________________________________
> lxc-users mailing list
> lxc-users at lists.linuxcontainers.org
> http://lists.linuxcontainers.org/listinfo/lxc-users