CiaB 7.0.1: bootstrap fails on traefik

[garloff]

cd /opt/configuration/environments/manager
./run.sh traefik
[...]

TASK [osism.services.traefik : Create traefik external network] ************************************************************************
ok: [manager.systems.in-a-box.cloud]

TASK [osism.services.traefik : Copy docker-compose.yml file] ***************************************************************************
An exception occurred during task execution. To see the full traceback, use -vvv. The error was: ansible.errors.AnsibleUndefinedVariable: {{ hostvars[inventory_hostname]['ansible_' + internal_interface]['ipv4']['address'] }}: 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'ansible_vlan100'. 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'ansible_vlan100'. {{ hostvars[inventory_hostname]['ansible_' + internal_interface]['ipv4']['address'] }}: 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'ansible_vlan100'. 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'ansible_vlan100'
fatal: [manager.systems.in-a-box.cloud]: FAILED! => {"changed": false, "msg": "AnsibleUndefinedVariable: {{ hostvars[inventory_hostname]['ansible_' + internal_interface]['ipv4']['address'] }}: 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'ansible_vlan100'. 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'ansible_vlan100'. {{ hostvars[inventory_hostname]['ansible_' + internal_interface]['ipv4']['address'] }}: 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'ansible_vlan100'. 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'ansible_vlan100'"}

PLAY RECAP *****************************************************************************************************************************
manager.systems.in-a-box.cloud : ok=8    changed=0    unreachable=0    failed=1    skipped=0    rescued=0    ignored=0   

[garloff]

Is something wrong with the network (netplan) config?
If I add ip link add link eno1np0 name vlan100 type vlan id 100; netplan apply, this script starts to work.

[garloff]

And ip link add link eno1np0 name vlan101 type vlan id 101 is needed as well. bootstrap.sh && deploy.sh then succeeds.

[garloff]

And this somehow needs to be persisted, so the system survives a reboot.

[garloff]

I wonder if this is a netplan bug. The settings in 01-osism.yaml look correct to me and according to the docs, I would have expected netplan to bring up these vlan ifaces.

[garloff]

I could not find a way how to tell netplan to bring up the vlan links, so I injected a custom systemd unit into the boot process. This is certainly not the intended way to do things.

dragon@manager:~$ cat /etc/rc.network 
#!/bin/bash
if test "$1" = "start" -o "$1" = "restart"; then
        ip link add link eno1np0 name vlan100 type vlan id 100
        ip link add link eno1np0 name vlan101 type vlan id 101
        netplan apply
fi
dragon@manager:~$ cat /etc/systemd/system/rc-network.service 
# SPDX-License-Identifier: LGPL-2.1-or-later
# Injected to work around missing vlan links in netplan
# (c) Kurt Garloff <[email protected]>, 4/2024
[Unit]
Description=/etc/rc.network setup
Documentation=man:systemd-rc-local-generator(8)
ConditionFileIsExecutable=/etc/rc.network
After=network-pre.target

[Service]
Type=forking
ExecStart=/etc/rc.network start
TimeoutSec=0
RemainAfterExit=yes
GuessMainPID=no

[Install]
WantedBy=multi-user.target

[berendt]

Please paste /etc/netplan/01-osism.yaml.

[garloff]

Looks correct to me ...

# This file describes the network interfaces available on your system
# For more information, see netplan(5).
---
network:
  version: 2
  renderer: networkd

  bonds:
    {}


  bridges:
    {}


  ethernets:
    eno1np0:
        dhcp4: true


  tunnels:
    {}


  vlans:
    vlan100:
        addresses:
        - 192.168.16.10/24
        id: 100
        link: eno1np0
    vlan101:
        id: 101
        link: eno1np0


  vrfs:
    {}

[berendt]

This was probably fixed with osism/ansible-collection-commons#637.