Cluster and data store backup scenarios

Backup VMs according to cluster or data store.

Data Protection for VMware domain support includes backing up VMs according to the following criteria: These criteria are implemented with the Tivoli® Storage Manager backup-archive client domain.vmfull option with the following parameters:
vmhostcluster=hostclustername,hostclustername,...,
This parameter backs up all VMs that are defined to the vCenter server that is specified by the vmchost option. The VMs must also be running on the ESX host cluster that is specified by hostclustername. This parameter not available when connecting directly to the ESXi/ESX host.
vmdatastore=datastorename,datastorename,...,
This parameter backs up all VMs that are defined to the vCenter or to the ESX server as specified by the vmchost option. The VMs must also have their configured default data store location (location of VM configuration files (.vmx) ) match the data store name that is specified by datastorename.
-vm=vmname,*,
The vmname specifies that the next set of values are VM host names to exclude from the domain. Specifying a wildcard character (*) assists with identifying VMs by their naming convention.

Cluster scenario

In this cluster scenario, the VMware environment consists of more than 3,000 VMs. Host clusters are used to manage the hardware resources. Although most of these clusters contain from 3 to 4 ESXi hosts each, some clusters contain as many as 10 ESXi hosts. However, some ESXi hosts are running from 1 to 3 VMs for larger, mission critical VMs. In order to manage the load, additional VMware hosts are frequently added or removed as needed. Each ESXi host in the cluster manages 10 - 30 VMs. Since the total number of VMs in each cluster ranges from 100 to 200, each host cluster is backed up to a dedicated vStorage Backup Server. Each server uses a dedicated data mover node to run the backup operation.

You can select host clusters to back up with the Data Protection for VMware vSphere GUI Schedule a Backup wizard. In the Data Protection for VMware vSphere GUI Getting Started window, click Define a backup task to open the Schedule a Backup wizard.

In order to back up all VMs in host cluster TivDev01, the backup schedule contains this statement:
domain.vmfull vmhostcluster=TivDev01
In order to back up all VMs in host clusters TivDev02 and TivDev03, and exclude VMs from the domain that contain VMTest* in their name, the backup schedule contains this statement:
domain.vmfull vmhostcluster=TivDev02,TivDev03;-vm=VMTest*
In order to back up all VMs in host clusters TivSVT01, TivSVT02, TivSVT03, and TivSVT04, the backup schedule contains this statement:
domain.vmfull vmhostcluster=TivSVT01,TivSVT02,TivSVT03,TivSVT04

These domain.vmfull vmhostcluster= statements can be repeated for each host cluster.

Host scenario

The VMware environment is similar to the environment described in the Cluster Scenario. However, this environment contains individual host clusters that contain too many VMs for a single data mover node backup window. A data mover can be assigned for each host or added to an existing data mover that is already processing a host cluster. This type of implementation can also be applied to individual VMs.

These host scenarios plan for a vStorage backup server (data mover) for each host on production systems that contain larger VMs or an increased data change rate.

You can select hosts to back up with the Data Protection for VMware vSphere GUI Schedule a Backup wizard. In the Data Protection for VMware vSphere GUI Getting Started window, click Define a backup task to open the Schedule a Backup wizard.

In order to back up all VMs in hosts TivProd01 and mycompany.com, the backup schedule contains this statement:
domain.vmfull vmhost=TivProd01,mycompany.com
In order to back up all VMs in hosts TivProd02.mycompany.com and TivProd03.mycompany.com, and exclude VMs from the domain that contain MailDept* in their name, the backup schedule contains this statement:
domain.vmfull vmhost=TivProd02.mycompany.com,TivProd03.mycompany.com;-vm=MailDept*
The largest and most active VMs contain MailDept at the beginning of their name. In this scenario, a separate data mover (Production06) is dedicated to these VMs. The backup schedule contains this statement:
domain.vmfull vm=MailDept*
This domain.vmfull vm= statement can be repeated for each host.

Wildcard scenario

The vmname specified by the VM keyword (vm) and VM exclusion keyword (-vm) supports wildcard characters. Specify an asterisk (*) to match any character. Specify a question mark (?) to match a single character. For example:
vm=prod1*,*testvm,*dept*,dept4?prod
This command (with wildcards) produces a list that includes only those VMs that match the domain.vmfull filter string with no duplicates. The order of the domain keywords does not affect the results. Exclusion statements are always processed at the end. As a result, the end statements override the inclusion statements.

In these wildcard scenarios, a separate data mover is defined to back up the VMs based on their naming convention. New VMs can be added and included the scheduled backup operation.

Note: The Data Protection for VMware vSphere GUI Schedule a Backup wizard does not support wildcard selection. As a result, you must set the domain.vmfull option in the backup-archive client dsm.opt (Windows) or dsm.sys (Linux) file for the data mover node.
This example backs up production VMs that contain MailDept at the beginning of their name:
domain.vmfull vm=MailDept*
This example backs up training VMs that contain TrainingVM at the beginning of their name:
domain.vmfull vm=TrainingVM*
This example backs up all VMs that are not used in production or training:
domain.vmfull all-vm;-vm=MailDept*,TrainingVM*