High availability (HA) deployment mode includes active and standby nodes to ensure that services are not interrupted.
An HA deployment runs on three ESXi hosts and includes at least nine virtual machines. These include two load balancers, three database nodes, two controllers, the OpenStack Management Server, and at least one compute driver. An additional compute driver is created for each compute cluster that you add to your deployment.
To deploy in HA mode, at least 11 contiguous IP addresses must be available on the management network. Each of the preceding virtual machines requires one IP address, and two additional IP addresses are required for private OpenStack endpoints. At least two contiguous IP addresses must be available on the API access network, both of which are used as public OpenStack endpoints.
- In VDS networking mode, an additional two virtual machines are required for DHCP.
- If you want to deploy Ceilometer, an additional five virtual machines are required. These virtual machines each require one IP address on the management network.
To ensure that you can upgrade VMware Integrated OpenStack successfully, plan management and API access networks that include twice the number of IP addresses required for your deployment. The extra IP addresses will be used during the upgrade procedure.
