Cluster build failing with API and ingress VIPs in a MachineNetwork allocation pool

level=error msg=failed to fetch Master Machines: failed to load asset "Install Config": failed to create install config: [platform.openstack.apiVIPs: Invalid value: "10.226.212.71": apiVIP can not fall in a MachineNetwork allocation pool, platform.openstack.ingressVIPs: Invalid value: "10.226.212.72": ingressVIP can not fall in a MachineNetwork allocation pool]

It was said in the support case that this validation was added intentionally to avoid IP conflicts. For the reasons below, a workaround is needed. It has also been seen that a similar install on 4.14 does not trigger this validation which suggests there may be a bug.

Reasoning for workaround or bugfix:

Securing a New Subnet: Obtaining a new subnet from OpenStack for API and Ingress VIPs poses a challenge. Additionally, dedicating a subnet solely for API and Ingress VIPs might result in underutilization.

Inter-subnet Connectivity: Establishing connectivity between the two subnets is a concern.

Upgrade Management: Managing both subnets during upgrades from Openshift and Openstack introduces complexities.

Automation Adjustments: Significant changes in automation processes are required to accommodate IP assignment modifications.


This is fairly critical for the customer as they cannot currently rebuild in a DR scenario and they cannot do rebuild testing.

    4.12

Always

    1.
    2.
    3.