Using a LXD macvlan Network : "Error: Failed to start device "eth0": Parent device 'enp2s0' doesn't exist"i

LXD
,
1

Hello ,
i just trying to make some containers to access from network and finded this tutorial :

but facing error when launch container
Error: Failed to start device “eth0”: Parent device ‘enp2s0’ doesn’t exist

this is fresh installation ubuntu-22.04.3-live-server-amd64

$ lxd init --minimal
$ lxc profile copy default macvlan-profile
$ lxc profile device remove macvlan-profile eth0
Device eth0 removed from macvlan-profile
$ lxc profile device add macvlan-profile eth0 nic nictype=macvlan parent=enp2s0
Device eth0 added to macvlan-profile
$ lxc launch images:ubuntu/22.04 cPanel --profile macvlan-profile
Creating cPanel
Starting cPanel                           
Error: Failed to start device "eth0": Parent device 'enp2s0' doesn't exist
Try `lxc info --show-log local:cPanel` for more info
$ lxc info --show-log local:cPanel
Name: cPanel
Status: STOPPED
Type: container
Architecture: x86_64
Created: 2023/10/09 19:48 UTC

Log:




$ ifconfig 
enp0s25: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.1.1  netmask 255.255.255.0  broadcast 192.168.1.255
        inet6 fe80::223:24ff:fe12:d147  prefixlen 64  scopeid 0x20<link>
        ether 00:23:24:12:d1:47  txqueuelen 1000  (Ethernet)
        RX packets 93643  bytes 136594271 (136.5 MB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 46551  bytes 3349163 (3.3 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
        device interrupt 19  memory 0xf0500000-f0520000  

enp0s25.1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::223:24ff:fe12:d147  prefixlen 64  scopeid 0x20<link>
        ether 00:23:24:12:d1:47  txqueuelen 1000  (Ethernet)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 75  bytes 20894 (20.8 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 1000  (Local Loopback)
        RX packets 174  bytes 15421 (15.4 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 174  bytes 15421 (15.4 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lxdbr0: flags=4099<UP,BROADCAST,MULTICAST>  mtu 1500
        inet 10.134.3.1  netmask 255.255.255.0  broadcast 0.0.0.0
        inet6 fd42:2f46:32ae:ba6a::1  prefixlen 64  scopeid 0x0<global>
        ether 00:16:3e:b2:76:4b  txqueuelen 1000  (Ethernet)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 0  bytes 0 (0.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

$ lxc profile show default
config: {}
description: Default LXD profile
devices:
  eth0:
    name: eth0
    network: lxdbr0
    type: nic
  root:
    path: /
    pool: default
    type: disk
name: default
used_by: []



$ lxc network show lxdbr0
config:
  ipv4.address: 10.134.3.1/24
  ipv4.nat: "true"
  ipv6.address: fd42:2f46:32ae:ba6a::1/64
  ipv6.nat: "true"
description: ""
name: lxdbr0
type: bridge
used_by:
- /1.0/profiles/default
managed: true
status: Created
locations:
- none

that’s all

2

So you’ve got it configured to use your host enp2s0 but don’t have such a device, error looks accurate. Looks like you want enp0s25 instead.

1 Like
3

Thanks so much.

now how i can set static IP for each container

4

By using whatever the network configuration mechanism is for the distro you’re using.
With macvlan, LXD has no control over IP allocation, so it must be done inside the instance.

1 Like
5

now container and lxd server cant ping each other :smiling_face_with_tear:
but can ping other containers

6

That’s normal behavior for macvlan.

7

Thanks friends for the discussion. I used to use the default FanNetworking and it worked extremely fine without my worrying about it. When I came to Incus, the situation seems different. Now I use the following steps to create or copy an instance with the desired profile, and position it on the the correct cluster node (with the --target switch or move it later).

  1. Create a profile for each node of the cluster
    1.1 Copy the default profile to create a new profile
incus profile copy default macvlan-profile-hpz2mini

1.2 Modify the eth0 of this new profile as indicated by @mohamed.daif (Note: eno1 is the nic of cluster node hpz2mini.)

incus profile device remove macvlan-profile-hpz2mini eth0
incus profile device add macvlan-profile-hpz2mini eth0 nic nictype=macvlan parent=eno1
  1. Create or Copy to make a new instance based on the desired profile
incus copy u2204-template jupyterhub01 --profile macvlan-profile-hpz2mini
  1. If the newly create instance is placed on the wrong cluster node and cannot be stared, move it to the node corresponding to the profile of this instance:
incus move instance_name --target cluster_node_name

This solution might not be the best. However, it works and solved the problem raised by @mohamed.daif, which I encountered after migrating from LXD to Incus.