undercloud install doesn't validate the local interface which leads to delayed and misleading errors

[Alex Schultz]

If a user does not have an undercloud.conf, the local_interface defaults to eth1. If this interface does not actually exist on the machine (ie it's baremetal and so it would be em1), the undercloud install can fail during the installation process and requires that the user understand what is going on to debug it. We should validate the local_interface to ensure that it exists and also throw a warning if the undercloud.conf is not available.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/491601

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/491601
Committed: https://git.openstack.org/cgit/openstack/instack-undercloud/commit/?id=062346e319f55ba423fd4dd239d1c4a82b61c747
Submitter: Jenkins
Branch: master

commit 062346e319f55ba423fd4dd239d1c4a82b61c747
Author: Alex Schultz <email address hidden>
Date: Mon Aug 7 14:39:03 2017 -0600

    Validate local_interface exists

    In order to improve the user experiance, we should validate that the
    provided local_interface actually exists prior to trying to configure
    the undercloud with it. If net_config_override is not used and the
    local_interface does not exist, an error will be returned to the
    user early on in the installation process. Additionally if the user
    does not have an undercloud.conf, a warning is printed to show that
    the defaults will be used.

    Change-Id: Ia2a7e7fc916f6d2b62e212cb42641494b88d6d17
    Closes-Bug: #1709177

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/instack-undercloud 7.3.0 release.