Linuxbridge-agent down, VMs fail with No valid host was found

    This is a repost of information from another thread

    Linuxbridge agent was not happy - wouldn't come up to in the output of 'nova service-list'. When @nmeadows  tried to start up a simulation, the VMs would fail with the error message:

    "state changed from BUILD to ERROR with message: No valid host was found. Exceeded max scheduling attempts 3 for instance e748eec0-95c5-414c-9a7d-77a46518ac77: File "/usr/lib/python2.7/dist-packages/nova/scheduler/", line 140, in run_instance"


    Neil had a bare-metal server with only two ports. For installation, he needed to use the instructions at and the section 'Prepare for an Interface-Constrained Installation'.

    If you subsequently make configuration changes in /etc/virl.ini, enabling dummy interfaces for example, the scripts that then apply those changes do not take into account the configuration that is currently active. When the script runs, it will reconfigure /etc/network/interfaces but those changes are not applied until the system has been rebooted.

    For a period in time, you will have both the original and the new interface settings active, until system reboot. On reboot, original settings are discarded.

    If you make changes to the network elements, you need to reset Openstack so that it is up to date with the network changes - this is vital.

    The following command sequence was used to bring the system back to operational state:

    edit /etc/virl.ini - ensure all of the settings are correct

    vinstall vinstall

    vinstall salt

    sudo salt-call state.sls <<< this will build the /etc/network/interfaces file

    sudo reboot

    On reboot

    sudo salt-call state.sls openstack

    sudo salt-call state.sls openstack.reset

    sudo salt-call state.sls openstack-restart

    Then check that all openstack services are up and running:

    nova service-list neutron net-list