The Veeam backup repository is a storage location where Veeam Backup & Replication stores backup files and virtual machine metadata. Veeam tends to be flexible with backup repository setup and enables a variety of storage types. Even so, there are some best practices that organizations should adhere to when implementing a Veeam backup repository.
Several best practices for Veeam backup repositories are tied directly to the use of Veeam backup proxies. A backup proxy reads the source data that an organization backs up and transfers that data to the backup repository.
Do the math
Repositories need to be sized to match the proxies that are in use. Veeam recommends dividing the number of proxy cores by three and then rounding up. For example, an organization with 32 proxy cores would result in 11 cores, rounded up.
The Veeam repository requires RAM. Veeam recommends that users multiply the number of repository cores by 4 GB to reach the required amount of RAM. The 11 cores from the previous example, multiplied by 4 GB, would be 44 GB. According to Veeam, no repository should have less than 2 cores and 8 GB of RAM.
Know your hardware options
The backup repository setup process enables users to limit the maximum concurrent tasks and data transfer rates. These limitations are not necessary if the repository is running on appropriately sized hardware, but users may need to throttle undersized repositories to prevent overruns and timeouts from occurring.
Veeam backup repositories can reside on physical or virtual hardware. However, Veeam recommends physical hardware whenever possible. This not only improves performance, but it also helps to avoid accidentally storing the backup on the same system that is being backed up.
If an organization does choose to store a backup repository on a virtual machine, then Veeam recommends storing the repository on a physical disk rather than in a VMDK file. Otherwise, the repository could be lost if the VMDK file is corrupted.
To scale out or not to scale out?
Organizations should adhere to the 3-2-1 rule when designing their repository architecture. The 3-2-1 rule loosely states that an organization needs to have three copies of its data -- the original data plus two backups, stored on two different types of media with at least one copy off site.
When an organization creates a Veeam backup repository, consider if a scale-out repository is justified. Scale-out backup repositories are essentially repositories that use multiple storage tiers. Scale-out repositories are best suited to enterprise environments.
To create a scale-out repository, users define a performance tier, a placement policy and an optional capacity tier. The performance tier is a collection of existing repositories that are brought together within the scale-out repository and treated as performance extents.
The placement policy enables users to control whether backups will store data locally or optimize the backups for performance. It is worth noting that if organizations choose the performance option and then later lose an extent that contains a full backup, restoring incremental backups will become impossible, even if the increments are stored on other repositories.
A capacity tier is optional, but a good idea for large organizations that have a lot of data. A capacity tier enables admins to copy backups to object storage, which increases the storage space that is available to the repository.