Before you configure your VMware region, it is important that you understand how OpenStack interacts with the VMware vCenter.
A VMware cluster can be mapped either to an availability zone in OpenStack, or to a host aggregate. The recommended choice is to have a cluster mapped to an availability zone, because this choice reflects the idea of a cluster as an entity with an optimized placement logic and resource grouping. If your VMware cluster has high-availability (HA) and Distributed Resource Scheduler (DRS) policies, they continue to be exploited by VMware itself, because they are transparent to OpenStack. When you define an availability zone or a host aggregate, you can decide the datastore, the datastore clusters, and the resource pool that must be used to store the disk of the images. You can use also regular expressions to easily point to a set of resources.
For example if you want to allow placement in different datastores for the same cluster to exploit different hardware characteristics of the datastores, you can create a single availability zone with multiple host aggregates where each host aggregate points to a different datastore. For more information about how to achieve this configuration, see Connecting to different datastores in the same cluster.
If you want to leverage SDRS, configure your environment by following the procedure described in Enabling Storage DRS. This can be done per availability zone or host aggregate.
Templates and virtual machines are automatically discovered and published in glance and nova after you installed and configured the OpenStack Controller.For more information, see Configuring vmware-discovery. In this way you can immediately manage these templates and virtual machines from the OpenStack Dashboard. You can also view the virtual machines from the Self-service user interface in the RESOURCES panel. Even if these instances were not created by IBM® Cloud Orchestrator, you can start, stop, resize them, or run custom actions by using the Self-service user interface. To use the discovered templates as images for deployments in IBM Cloud Orchestrator, you must modify them to meet the prerequisites documented in Managing virtual images.
If the template was created in thin provisioning mode, all the instances generated from it are thin provisioned. If you want to speed up the cloning operation of instances spawn from the same template, you can turn on the OpenStack linked clone feature. This feature can be set per availability zone or host aggregate and relies on caching the VMDK in the datastore. For more information about this feature, see http://docs.openstack.org/kilo/config-reference/content/vmware.html. Moreover, you can add disks to the image at deployment time or after the deployment occurred. Volumes can be thin provisioned or thick provisioned. For more information, see Configuring OpenStack to support thin provisioning.