6.3 vSphere failover plans
Once all the protected virtual machines have at least one restore point stored into Veeam Cloud Connect, a tenant can use the failover capabilities he subscribed to.
In order to complete a full failover, a Failover Plan must be configured. A Failover Plan is a group of virtual machines that Veeam Backup & Replication has to manage as a single entity, following the boot order and delays configured in the plan itself. When it comes to Veeam Cloud Connect, a Cloud Failover Plan has additional options.
6.21: Create a Cloud VMware Failover Plan
After giving the new Failover Plan a unique name (unique for the tenant, as multiple tenants at the service provider can have the same name for their failover plans without any problem), the tenant selects the virtual machine he wants to add to this plan:
6.22: Select replica VM's
Only virtual machines replicated to Veeam Cloud Connect are shown in the selection screen, and only those with at least one complete restore point available can be selected. Replica VM's are added to the Cloud Failover Plan, where boot order and delay can be configured:
6.23: Replica VM's added to Failover Plan
In the next step, a tenant configures the different default gateways for each available network:
6.24: Specify default gateways for cloud networks
When a full failover is initiated, the network extension appliance at the Service Provider enables the different internal networks and set the configured default gateway's IP address as its own address. The result is that any virtual machine connected to a cloud network can reach its original gateway IP without changing its original IP address. The NEA is in fact simulating the original gateway of the tenant's production network.
The next step is the configuration of Public IP Addresses. The network extension appliance has one external interface, connected to a network that can usually reach internet. In addition to its primary IP address, a service provider can assign additional IP addresses to be used by tenants to publish services running on the failed over VM's.
First, a tenant enables the option Assign public IP addresses to use during full site failover, then by using the Add button, he creates publishing rules like in a firewall:
6.25: Add a public IP address mapping rule for a Windows VM
For a Windows VM, the internal IP address of the replica VM is automatically recognized. However, even non-windows VM's can be published by manually writing their IP address in the box (lamp is a linux VM):
6.26: Add a public IP address mapping rule for a Linux VM
Once the Failover Plan is configured, it is stored directly at the service provider, and what the tenant sees locally is read in real time from the service provider. This is done because, upon a failure, a tenant may lose his entire environment, thus also the Veeam Backup & Replication installation and its configuration like the failover plans. By storing the Failover Plan directly at the service provider, it can be initiated by the service provider or by the tenant (using the Cloud Portal) without the need for a local Veeam installation at the tenant.