Creating Protection Groups
Once you are satisfied with your array manager configuration you're ready to carry on with the next major step: configuring Protection Groups. Protection Groups are used whenever you run a test of your Recovery Plan, or when DR is invoked for real. Protection Groups are pointers to the replicated vSphere datastores that contain collections of virtual machines that will be failed over from the Protected Site to the Recovery Site. The Protection Groups' relationships to ESX datastores can be one-to-one. That is to say, one Protection Group can contain or point to one ESX datastore. Alternatively, it is possible for one Protection Group to contain many datastores—this can happen when a virtual machine's files are spread across many datastores for disk performance optimization reasons or when a virtual machine has a mix of virtual disks and RDM mappings. In a loose way, the SRM Protection Group could be compared to the storage groups or consistency groups you may create in your storage array. However, what actually dictates the membership of a Protection Group is the way the virtual machines utilize the datastores.
To configure Protection Groups follow these steps.
- Log on with the vSphere client to the Protected Site's vCenter (New York).
- Click the Site Recovery icon.
Select the Protection Groups pane and click the Create Protection Group button, as shown in Figure 9.42.
Figure 9.42 You can now create both Protection Groups and Protection Group folders.
New to this release is the ability to create folders in the Protection Groups, to allow you to more easily lay out your Protection Groups if you have a significant number of them.
In the Create Protection Group dialog box (whether you are using VR or array-based replication), if you have more than one array manager select the one associated with this Protection Group, as shown in Figure 9.43. Then select the pairing of arrays contained within the array manager configuration.
Figure 9.43 The EMC Celerra array manager configuration. You may have many array pairs, each hosting many datastores protected by replication.
- Click Next. This should enumerate all the volumes discovered on the arrays in question. If you select the volume names, you should see the VMs contained within those ESX datastores (see Figure 9.44).
Figure 9.44 Dell EqualLogic datastore containing a number of virtual machines
- In the Create Protection Group Name and Description dialog box, enter a friendly name and description for your Protection Group. In my case, I'm creating a Protection Group called "Virtual Machines Protection Group." Click Finish.
At this point, a number of events will take place. First, as the Protection Group is being created the icon of the Protection Group changes, and its status is marked as "Configuring Protection," as shown in Figure 9.45. Second, at the Recovery Site vCenter you will see the task bar indicate that the system is busy "protecting" all virtual machines that reside in the datastore included in the Protection Group (see Figure 9.46).
Figure 9.45 When Protection Groups are first created their status is modified to "Configuring Protection."
Figure 9.46 During the creation of Protection Groups each affected VM has a task associated with it.
Meanwhile, the Recovery Site's vCenter will begin registering the placeholder VMX files in the correct location in the inventory, as shown in Figure 9.47. As you can see, each Protect VM event has a "Create virtual machine" event. SRM isn't so much creating a new VM as it is registering placeholder VMs in the Recovery Site.
Figure 9.47 The Recovery Site's vCenter begins registering the placeholder VMX files in the correct location in the inventory.
You will also have noticed these "new" VMs are being placed in the correct resource pool and folder. If you select one of the placeholder files you can see it only takes up a fraction of the storage of the original VM. You should also see that these placeholders have been given their own unique icon in the vCenter inventory at the Recovery Site. This is new to SRM. Previously, the placeholder VMs just had the standard "boxes in boxes" icon, and that made them difficult to identify. Even with the new-style icon, as shown in Figure 9.48, I still recommend a separate resource pool and/or folder structure so that you can keep these ancillary placeholders separate and distinct from the rest of your infrastructure.
Figure 9.48 Creation of placeholder VMs with the new lightning bolt icon, which should make them easier to distinguish in the vCenter inventory
If you browse the storage location for these placeholders you can see they are just "dummy" VMX files (see Figure 9.49). As I mentioned before, occasionally VMware SRM refers to these placeholder VMs as "shadow" VMs. In the Virtual Machines and Template view, at the Recovery Site's vCenter the VMs have been allocated to the correct folder. SRM knows which network, folder, and resource pool to put the recovery VMs into, because of the default inventory mapping settings we specified earlier.
Figure 9.49 Placeholder VMs are created in the datastore specified in the Placeholder tab on the properties of each site.
You should know that if you create a template and store it on a replicated datastore it will become protected as well. This means templates can be recovered and be part of Recovery Plans (covered in Chapter 10) just like ordinary VMs. Templates are not powered on when you run a Recovery Plan, because they can't be powered on without being converted back to being a virtual machine. As you can see, these placeholder VMs are very different from the VMs you normally see registered to vCenter. If you try to edit them like any VM you will be given a warning (shown in Figure 9.50) that this is not a recommended action.
Figure 9.50 The warning dialog box that appears if you try to edit the placeholder VMs listed in the Recovery Site