The VMware Knowledge Base provides support solutions, error messages and troubleshooting guides
Search the VMware Knowledge Base (KB)
View by Article ID
Reference architecture sample for VMware Data Recovery (2017324)
Purpose
Although the instructions for setting up and administering VMware Data Recovery (VDR) are explained in the documentation which is available on VMware.com, this document is intended to present an example of a reference architecture for a VDR installation.
It can be used as a starting point for a new installation, or as a guide when re-architecting an existing environment to reduce time, and possibly address other issues such as frequent failed Integrity Checks / Catalog Rebuilds.
The objective is to design an architecture that is robust, as efficient and as fast as possible, and which is subject to as few destination datastore issues as possible.
You can tailor this example to suit the number, variety, and size of the virtual machines in your environment, according to your available storage and other constraints.
Notes:
The reference architecture described in this article is intended as an example only. There are many combinations of appliances and datastores which will work well depending on the situation and environment.
If you are experiencing frequent issues with Integrity Checks, or other issues such as Destination Index or Catalog issues, you may wish to try re-architecting your environment using the principles described in this article.
Remember to consider a backup strategy for the destination datastores themselves.
Also note that, "VMware Data Recovery natively supports backups to disk and currently there is no officially supported method to migrate these backups to other media", as stated in the VMware Data Recovery FAQ.
It is always a best practice to thoroughly and regularly test all aspects of your backup and recovery strategy to ensure that it meets your requirements.
Resolution
Example environment
This example uses version 2.x of the VDR appliance.
Managed by vCenter. (Requires at least vCenter 4.1 Update 1 or higher, and ESX/ESXi 4.1 Update 1 or higher.)
6 x Windows 2008 and 2008 R2: Total disk size of 300 GB, of which 280 GB is used
4 x Windows 2003: Total disk size of 160 GB, of which 100 GB is used
2 x Linux virtual machines: Total disk size of 400 GB, of which 300 GB is used
2 x Windows XP virtual machines: Total disk size of 30 GB, of which 30 GB is used
Sample architecture
In order to define our goals for the reference architecture, we can begin with these statements based on the VDR 2.x documentation:
Maximum supported destination datastore sizes vary depending on whether the datastore is a .vmdk file or not
Deduplication datastores should be sized based on the used space of the protected virtual machines
Therefore our goals are to optimize deduplication and reduce the complexity of the integrity checking / catalog management functions within the appliance.
We will deploy two appliances in this example, as this will allow us to use more than two deduplication datastores in the environment. The two destination datastores per appliance are thick-provisioned .vmdk files:
Appliance A, deduplication store 1: 300 GB, intended for W2008 / W2008R2 machine backups
Appliance A, deduplication store 2: 120 GB, intended for W2003 machine backups
Appliance B, deduplication store 1: 320 GB, intended for Linux machine backups
Appliance B, deduplication store 2: 50 GB, intended for XP machine backups
Getting started in your environment
The best approach to get started is to prepare a list of the virtual machines that you wish to protect.
If you are using a spreadsheet, some suggested column titles are:
VM Name
OS Name and Version
Used Storage
After preparing this list, use it to organize your appliance and destination datastore deployment accordingly.
Notes and considerations
VDR maximizes destination datastore space usage, because of the grouping of similar OS types.
The sizing is slightly more than needed to back up the used space of the virtual machines. For more information on sizing, see the Deduplication Store Sizing and Deduplication Store Formats sections in the VMware Data Recovery Administration Guide.
Datastores can be expanded as needed, and additional appliances can be deployed as needed.
You can have up to 10 VDR appliances per vCenter.
It is possible to connect NFS storage to the ESX/ESXi host for the appliance's use, and then create a .vmdk destination datastore on that NFS datastore. However, the NFS storage should be provided by a device that is in the VMware approved compatibility guide for NAS storage.
