Before installing VMware Data Recovery, ensure the system and storage requirements are available in your environment.

Data Recovery requires vCenter Server and the vSphere Client. Data Recovery does not work with similar VMware products such as VirtualCenter Server. You can download the vSphere Client from your vCenter Server.

Virtual machines to be backed up and the backup appliance must both be running on ESX/ESXi 4 or later. To use all features, the ESX/ESXi host that runs the backup appliance must be managed by vCenter Server. Data Recovery can be used without vCenter Server, but features such as automated backups may not be supported.

When using Data Recovery with vCenter Servers running in linked mode, login to the vCenter Server with which the Data Recovery appliance is associated.

You can store backups on any virtual disk supported by ESX/ESXi. You can use technologies such as storage area networks (SANs) and network attached storage (NAS) devices. Data Recovery also supports Common Internet File System (CIFS) based storage such as SAMBA.

When adding hard disks to the backup appliance, consider how many disks most virtual machines to be backed up have. Each backup appliance can back up 100 virtual machines, but a maximum of 8 virtual machines can be backed up simultaneously. Each disk on each virtual machine may be hot-added for the backup to occur. In the default configuration, the backup appliance has a SCSI adapter #0 and a SCSI disk #0 attached to the SCSI adapter. Since the first SCSI adapter has a system disk at SCSI 0:0, only 14 SCSI disks can be hot-added. As the backup of a virtual machine completes, that virtual machines disks are removed and subsequent backups can begin. In the default configuration, if the total number of disks for the virtual machines being backed up reaches 15, the disks are backed up over the network instead of through hot-add. If you are working with virtual machines with a greater number of disks, consider adding additional disks to the appliance. For example, if each virtual machine in your environment has 3 disks, some of the virtual machine disks are backed up over the network, and performance may be negatively affected. By adding a dummy disk of 1 MB to another SCSI bus adapter, the total available SCSI bus locations for hot-adding increases to 30, so all 8 virtual machines in the example given here can be backed up simultaneously using hot-add. Additional disks should be added in the sequence SCSI 1:0, SCSI 2:0, SCSI 3:0, and so on for as many or as few instances as are required. A virtual machine, such as the backup appliance may have up to 4 SCSI adapters, enabling a maximum of 60 available SCSI bus locations for hot-adding disks, which is sufficient for most environments.

See the most recent vSphere documentation for information about setting up a vSphere 4.0 or later environment including ESX, ESXi, vCenter Server, and the vSphere client.

The amount of storage required varies, depending on how much deduplication can save disk space as a result of running similar virtual machines. Even with space savings, Data Recovery requires an absolute minimum of 10 GB of free space. This space is used for indexing and restore point processing, so even if the virtual machines to be backed up are very small, they may fail to complete if less that 10 GB of disk space is available. While a minimum of 10 GB is acceptable, having at least 50 GB is highly recommended for typical usage. The more diverse the set of virtual machines to be protected, the more space is required for each virtual machine. The amount of space required is also affected by the frequency of backup, the length of time the backups are kept, and the number of virtual machines to be backed up.

For initial setup, provide storage space equal to the amount of used disk space on all virtual machines being protected. For example, if you are protecting 10 virtual machines, each with one 20 GB virtual disk, and those virtual disks are on average 50% full, then you should provide at least 100 GB of storage available for the deduplication store. Over time, the amount of space the deduplication store consumes typically reaches an equilibrium as data being updated is roughly equal to aging restore points being removed by the retention policy.

Deduplication stores can be stored on thin-provisioned or thick-provisioned virtual disks. Using thin-provisioning may result in decreased performance because space as allocated as it is required. Therefore, it may be best to use larger thick-provisioned disks sized to avoid the potential performance impact from growing a thin-provisioned disk. If the space available on a thick provisioned disk becomes unavailable, you can extend the disk using the vSphere Client.

Deduplication stores can be stored in all HCL supported storage and CIFS based network shares, and they are compatible with storage that is capable of deduplication. While any supported format may be used, virtual disks (VMDKs) or RDMs are recommended for deduplication stores because they provide the most well-understood and consistent performance. CIFS shares are also supported, but the performance of such shares varies across providers, and as such, is not an ideal solution. Furthermore, in many cases, virtual disks and RDMs perform better than network-based deduplication stores. Deduplication stores can be stored in RDM with either virtual or physical compatibility.

While CIFS can be used, do not use CIFS shares that are:

On a server that has another role. For example, do not use CIFS shares hosted on a vCenter Server.

Connected to a virtual machine.

Shared to multiple services or servers. If multiple appliances use a single CIFS share, miscalculations in space requirements may result, which may cause the appliance to run out of disk space.

Note

Striping results in a loss of space efficiency across deduplication stores. Protecting virtual machines in separate deduplication stores typically provides better results than using striping to combine disks to create one large deduplication store.

There are special considerations when using thin provisioned virtual disks as the data recovery destination disk. vSphere automatically freezes any virtual machine whose thin provisioned disk usage exceeds its hosting VMFS datastore's capacity. Therefore, VMware recommends using one of two strategies to avoid running out of space for the Data Recovery destination disk.

Use alarms to identify when space is limited on a thin-provisioned disk and add more space as required.

Use smaller thick provisioned virtual disks and extend the disk as required.

Different components of Data Recovery communicate among each other over TCP. As a result, ensure the appropriate ports are open in your environment for normal operation.

The backup appliance connects to vCenter Server web services. By default, this connection is established using ports 80 and 443.

The Data Recovery client plug-in and File Level Restore (FLR) client connect to the backup appliance using port 22024.

The backup appliance connects to VMware ESX or VMware ESXi using port 902.

ESX/ESXi servers that were added to vCenter using a DNS name must have a name that is resolvable. In some cases, using DNS names creates problems. If problems arise with resolving DNS names, consider adding ESX/ESXi servers using IP addresses instead.

Data Recovery completes operations using permissions that are granted through a role that is assigned to a user. The specific permissions that must be assigned to roles vary based on the task to be completed.

If your backup infrastructure uses Network Block Device (NBD) technology, the following minimum permissions must be assigned to the role that the appliance runs as:

VirtualMachine->Configuration->Disk change tracking

VirtualMachine->Provisioning->Allow read-only disk access

VirtualMachine->Provisioning->Allow virtual machine download

VirtualMachine->State->Create snapshot

VirtualMachine->State->Remove snapshot

Global->DisableMethods

Global->EnableMethods

Global->License

If your backup infrastructure uses SCSI hot-adding, the role that the appliance runs as must have all permissions required by NBD, as well as the following additional minimum permissions:

Datastore->Allocate space

VirtualMachine->Configuration->Add existing disk

VirtualMachine->Configuration->Add new disk

VirtualMachine->Configuration->Add or remove device

VirtualMachine->Configuration->Change resource

VirtualMachine->Configuration->Remove disk

VirtualMachine->Configuration->Settings

There are special considerations to be aware of when establishing Data Recovery in your environment. Data Recovery is supported for use with:

Ten Data Recovery backup appliances for each vCenter Server instance.

Each backup appliance protecting up to 100 virtual machines.

VMDK or RDM based deduplication stores of up to 1 TB or CIFS based deduplication stores of up to 500 GB.

CIFS shares with passwords limited to 64 characters or less. CIFS share passwords must conform to the Latin 1(ISO 8859-1) standard. Double-byte characters are not supported.

If a third-party solution is being used to backup the deduplication store, those backups must not run while the Data Recovery service is running. Do not back up the deduplication store without first powering off the Data Recovery Backup Appliance or stopping the datarecovery service using the command service datarecovery stop.

Up to two deduplication stores per backup appliance.

vCenter Servers running in linked mode. For this configuration to perform as expected, log in to the vCenter Server with which the Data Recovery appliance is associated.

Data Recovery does not support:

IPv6 addresses. IPv4 addresses are required for the Data Recovery appliance.

Hot adding disks with versions of vSphere that are not licensed for hot plug.

Restoring VMware View linked clones. Data Recovery can back up VMware View linked clones, but they are restored as unlinked clones.

Backing up virtual machines that are protected by VMware Fault Tolerance.

Backing up virtual machines that use VMware Workstation disk format.

Backing up virtual machines with 3rd party multi-pathing enabled where shared SCSI buses are in use.

Raw device mapped (RDM) disks in physical compatibility mode in virtual machines to be backed up.

Using older versions of the vSphere Client plug-in or older versions of FLR with the current version of Data Recovery.

Multiple backup appliances on a single host.

Using Data Recovery to backup Data Recovery backup appliances. While this is not supported, this should not be an issue. The backup appliance is a stateless device, so there is not the same need to back it up as exists for other types of virtual machines.

Backup of virtual machine disks that are marked as Independent.

Backup of Storage Virtual Appliances (SVAs).