Home / ubuntu / Questions / 1033847
Accepted
n8tlarsen
Asked: 2018-05-09 20:19:33 +0800 CST

Configure bonded 802.3ad network using netplan on Ubuntu 18.04

  • 772

I am running into some issues configuring netplan on Ubuntu 18.04 server to bond my four hardware ethernet ports named eno1, eno2, eno3, eno4 using the 802.3ad protocol. I've consulted the netplan man page and put together the following config file /etc/netplan/50-cloud-init.yaml:

network:
  version: 2
  renderer: networkd
  ethernets:
    eports:
      match: 
        name: eno*
  bonds:
    bond0:
      interfaces: [eports]
      addresses: [192.168.1.101/24]
      gateway4: 192.168.1.1
      nameservers:
        addresses: [8.8.8.8, 8.8.4.4]          
      parameters:
        mode: 802.3ad            
        lacp-rate: fast            
        mii-monitor-interval: 100

Upon running the command sudo netplan --debug apply I receive the following information:

** (generate:6473): DEBUG: 00:39:14.911: Processing input file //etc/netplan/50-cloud-init.yaml..
** (generate:6473): DEBUG: 00:39:14.911: starting new processing pass
** (generate:6473): DEBUG: 00:39:14.911: eports: setting default backend to 1
** (generate:6473): DEBUG: 00:39:14.911: bond0: setting default backend to 1
** (generate:6473): DEBUG: 00:39:14.912: Generating output files..
** (generate:6473): DEBUG: 00:39:14.912: NetworkManager: definition eports is not for us (backend 1)
** (generate:6473): DEBUG: 00:39:14.912: NetworkManager: definition bond0 is not for us (backend 1)
DEBUG:netplan generated networkd configuration exists, restarting networkd
DEBUG:no netplan generated NM configuration exists
DEBUG:device eno2 operstate is up, not replugging
DEBUG:netplan triggering .link rules for eno2
DEBUG:device lo operstate is unknown, not replugging
DEBUG:netplan triggering .link rules for lo
DEBUG:replug eno3: unbinding 0000:03:00.0 from /sys/bus/pci/drivers/igb
DEBUG:replug eno3: rebinding 0000:03:00.0 to /sys/bus/pci/drivers/igb
DEBUG:replug eno1: unbinding 0000:01:00.0 from /sys/bus/pci/drivers/igb
DEBUG:replug eno1: rebinding 0000:01:00.0 to /sys/bus/pci/drivers/igb
DEBUG:Cannot replug bond0: cannot read link /sys/class/net/bond0/device: [Errno 2] No such file or directory: '/sys/class/net/bond0/device'
DEBUG:netplan triggering .link rules for bond0
DEBUG:replug eno4: unbinding 0000:04:00.0 from /sys/bus/pci/drivers/igb
DEBUG:replug eno4: rebinding 0000:04:00.0 to /sys/bus/pci/drivers/igb

I'm not sure what to make of the statement

Cannot replug bond0: cannot read link /sys/class/net/bond0/device: [Errno 2] No such file or directory: '/sys/class/net/bond0/device'

since the directory /sys/class/net/bond0 was generated by the netplan apply command.

I checked my ifconfig output and my network devices seem to be configured correctly with the exception that no address is set for bond0:

bond0: flags=5123<UP,BROADCAST,MASTER,MULTICAST>  mtu 1500
    ether XX:XX:XX:XX:XX:XX  txqueuelen 1000  (Ethernet)
    RX packets 112768  bytes 7785014 (7.7 MB)
    RX errors 0  dropped 54  overruns 0  frame 0
    TX packets 18854  bytes 2337896 (2.3 MB)
    TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

eno1: flags=6211<UP,BROADCAST,RUNNING,SLAVE,MULTICAST>  mtu 1500
    ether XX:XX:XX:XX:XX:XX  txqueuelen 1000  (Ethernet)
    RX packets 290  bytes 19322 (19.3 KB)
    RX errors 0  dropped 0  overruns 0  frame 0
    TX packets 55  bytes 6820 (6.8 KB)
    TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
    device memory 0xdf400000-df47ffff  

eno2: flags=6211<UP,BROADCAST,RUNNING,SLAVE,MULTICAST>  mtu 1500
    ether XX:XX:XX:XX:XX:XX  txqueuelen 1000  (Ethernet)
    RX packets 73991  bytes 29824155 (29.8 MB)
    RX errors 0  dropped 0  overruns 0  frame 0
    TX packets 20848  bytes 2110417 (2.1 MB)
    TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
    device memory 0xdf300000-df37ffff  

eno3: flags=6147<UP,BROADCAST,SLAVE,MULTICAST>  mtu 1500
    ether XX:XX:XX:XX:XX:XX  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
    device memory 0xdf200000-df27ffff  

eno4: flags=6147<UP,BROADCAST,SLAVE,MULTICAST>  mtu 1500
    ether XX:XX:XX:XX:XX:XX  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
    device memory 0xdf100000-df17ffff  

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 2923  bytes 184477 (184.4 KB)
    RX errors 0  dropped 0  overruns 0  frame 0
    TX packets 2923  bytes 184477 (184.4 KB)
    TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

The ether XX:XX:XX:XX:XX:XX statements are in place of each interfaces's mac address. In the original output, all addresses are the same.

What am I missing to successfully configure my system?

networking 18.04 netplan network-bonding

2 Answers

  1. Best Answer

    After some digging, I discovered that Ubuntu 18.04 uses a utility called cloud-init to handle network configuration and initialization during the boot sequence. The file /etc/cloud/cloud.cfg.d/50-curtin-networking.cfg and other .cfg files are used to reconfigure cloud-init settings. My config file settings are as follows:

    network:
  version: 2
  ethernets:
    eports:
      match:
        name: eno*
      optional: true
  bonds:
    bond0:
      interfaces: [eports]
      addresses: [192.168.1.101/24]
      gateway4: 192.168.1.1
      nameservers:
        addresses: [8.8.8.8, 8.8.4.4]
      parameters:
        mode: 802.3ad
        lacp-rate: fast
        mii-monitor-interval: 100

    The optional: true parameter prevents the system from waiting for a valid network connection at boot time which will save you the hassle of waiting 2 minutes for your machine to boot. After updating the config file run the following command to update your configuration.

    cloud-init clean -reboot

    Alternatively running the following allows for some debug information without rebooting your machine; however, a reboot will be required to commit the changes during early boot stages.

    cloud-init clean
cloud-init init
cloud-init status
    • 5

  2. The netplan.io site seems to indicate that cloud-init is one of several utility configured by netplan. The problem with netplan maybe that you need to specify dhcp4 = false when you specify the physical interfaces to later be bonded. If you do not addresses may be attached to the individual interfaces by DHCP before they are bonded. Thus at bonding time there are conflicts with existing IP logic.

    I suspect your cloud-init (earlier in boot sequence?) workaround may work by bonding interfaces before the DHCP process occurs. Or with the clean/init sequence resets the DHCP address then bonds before the DHCP process can be reapplied. However, that can mean that the workaround may stop working if the maintainers make even minor changes to the boot logic/sequence.

    • 0