Version: $ openshift-install version ./openshift-install 4.5.14 built from commit 9893a482f310ee72089872f1a4caea3dbec34f28 release image quay.io/openshift-release-dev/ocp-release@sha256:95cfe9273aecb9a0070176210477491c347f8e69e41759063642edf8bb8aceb6 Platform: Openstack What happened? Our customer's Openstack doesn't support qcow2 image formats. Instead we convert the rchos image to raw format and follow https://github.com/openshift/installer/blob/master/docs/user/openstack/customization.md#image-overrides documentation to manually specify the image location. However the installer still tries to do the qcow2 check and gets the following error: time="2020-10-15T13:04:07+02:00" level=fatal msg="failed to fetch Terraform Variables: failed to generate asset \"Terraform Variables\": failed to get openstack Terraform variables: Bad request with: [POST https://image.cloud.muni.cz/v2/images], error message: {\"message\": \"Provided object does not match schema 'image': 'qcow2' is not one of [None, 'raw']\\n\\nFailed validating 'enum' in schema['properties']['disk_format']:\\n {'description': 'Format of the disk',\\n 'enum': [None, 'raw'],\\n 'type': ['null', 'string']}\\n\\nOn instance['disk_format']:\\n 'qcow2'<br /><br />\\n\\n\\n\", \"code\": \"400 Bad Request\", \"title\": \"Bad Request\"}" our config: apiVersion: v1 baseDomain: vega.io compute: - architecture: amd64 hyperthreading: Enabled name: worker platform: {} replicas: 3 controlPlane: architecture: amd64 hyperthreading: Enabled name: master platform: {} replicas: 3 metadata: creationTimestamp: null name: vega-cluster networking: clusterNetwork: - cidr: 10.128.0.0/14 hostPrefix: 23 machineNetwork: - cidr: 10.0.0.0/16 networkType: OpenShiftSDN serviceNetwork: - 172.30.0.0/16 platform: openstack: clusterOSImage: file:///var/tmp/img.raw apiVIP: 10.0.0.5 cloud: openstack computeFlavor: standard.large externalDNS: null externalNetwork: public-cesnet-78-128-250-PERSONAL ingressVIP: 10.0.0.7 lbFloatingIP: 78.128.250.207 octaviaSupport: "1" region: "" trunkSupport: "1" publish: External pullSecret: '{"auths":{"...... sshKey: | ssh-rsa ......
This is a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=1882649. In your case it's failing because the cloud restrict the image upload to raw only. Mike is working on a patch to set the glance disk format based on the file extension. In the meantime, there is a workaround. You need to upload the image to glance yourself and reference it from `clusterOSImage` in your `install-config.yaml`. *** This bug has been marked as a duplicate of bug 1882649 ***