Recovery Plans are the guts of the Disaster Recovery Scenario. The recovery plan determines what VMs get started, servers that need to be powered down, scripts to be run, startup order and the overall automated section of your failover.
Multiple recovery plans can be created in order to handle failures of a single application like Exchange (Edge Transport, Mailbox, Hub Transport and CAS Servers). It can also handle an entire site failover in the event of a natural disaster or “smoking hole” event.
Chose the “Create Recovery Plan” from the SRM Module –> Recovery Plan tab.
Select the protection groups you want to plan the recovery of.
When the recovery plan networks are displayed, two networks should be mapped. The first network called the “Recovery Network” is where the VMs will start up in the event that you need to run recovery. The second network is called the “Test NetworK” and this is where the VMs will connect to in the event you just want to test the failovers.
The test network is very useful to be sure you don’t disrupt the production servers by having two identical machines on the same networks. The test network can be a “bubble” network where any VMs in the network can see each other, but nothing else. If left to “auto” a bubble network will be created during the test so that basic testing can be done during the test recovery.
You may choose to have your test network a predefined network of your choosing.
Name your recovery plan. This is a good opportunity to call it something like “Full Disaster Failure” or “Database Recovery”.
Review your selections and the click finish.
At this point the recovery plan is ready to go. In most cases additional configuration would be necessary to customize how the VMs should be recovered.
To do this, open the Recovery plan that was just created and click the “Configure Recovery” link for the VMs.
Inside the Configure Recovery Menu a list of customizable settings will be available. A popular setting change is to modify the IP settings of the virtual machine. Many times moving a VM from one site to another site requires the IP address to change for it to communicate on the network. (There are ways around this if you’re wondering)
Set the Priority Group. Here you can have a group of VMs in the Recovery plan, but you can then set priorities for VMs to be started. Maybe there are some highly mission critical servers that should be started in Priority Group 1 and some lab servers that can wait until everything else is up and running that can go in Group 5.
A dependency can be set so that other VMs must be recovered first before the VM will start. This may be useful if a web server needs to wait for a database server to be available first.
If the customization needed, isn’t available in the menu, there is always the option to run a script at any point in the process.
Once the VMs are configured, take a look at the recover steps tab in SRM. This will show all of the steps that should take place during a recovery. Additional scripts can be added from this menu at the points needed.
