Dedupe appliance considerations
To optimize the deduplication storage performance, select at least one GFS restore point (weekly backup is mandatory for deduplication storages). Enable the “Read the entire restore point from source backup instead of synthesizing it from increments” option to avoid creating synthetic restore points from the deduplication storage (data rehydration).
- HPE StoreOnce Catalyst / Dell EMC DataDomain DDboost + copy over WAN + backup copy job:
The previously described option provides fast processing in situations that do not involve transport over WAN links. Also, optimized integration with Catalyst and Dell EMC DDBoost allows using synthetic processing. For these environments, uncheck the “Read the entire restore point from source backup instead of synthesizing it from increments,” but leave the weekly GFS backup enabled.
- Special scenarios based on vendor guidelines + backup job with active full or synthetic full backup:
The above vendor best practices and design guides mention situations where a deduplication storage can be used as a primary backup target. Depending on the guidelines, select the incremental backup mode with active or synthetic full backup.
When ExaGrid is used as a primary backup target, it is recommended to enable GFS processing with it. Please see best practices for ExaGrid.
Inline deduplication
Disable inline deduplication setting when writing into deduplication storages.
However, for deduplication devices that have a non deduplicated landing zone (Like ExaGrid) and back up only VMs, you can save some space. To do this, enable inline deduplication.
Compression
By default, the Veeam source DataMover (Veeam proxy or agent) compression level is set to “optimal” to reduce data that needs to be transported over the network. The repository or the gateway server will decompress the data before storing it to the deduplication device. For backup copy jobs, set the compression level to “Auto” to leave the data in the way it was stored originally on the primary backup target.
Dedupe settings may check the Decompress Backup Files Before Storing Setting, this setting allows the proxy/agent to compress the data before sending over the LAN to the repo, then decompress before storing to allow better dedupe ratio on the appliance.
Block size defaults
Only two of the block size settings are recommended for deduplication storage usage at VM backups:
- Storage optimization: Local target (large blocks)
Use this option for any general deduplication storage / HPE StoreOnce Catalyst / Dell EMC DataDomain DDboost / Quantum DXi.
- Storage optimization: Local target
Use this option for ExaGrid.
ExaGrid has built-in DataMover, so firewall ports to copy out to another repo (like Cloud Connect) need to be open from the ExaGrid IP address.
Block alignment
By default, this advanced repository setting is disabled. Enable it for deduplication storages that work with fixed block length deduplication (e.g. NetApp ONTAP deduplication) or in situations where you can use ReFS/XFS Fast Cloning.
Encryption
Encryption will create random data at the backup targets; as a result, the deduplication storages will not work effectively. It is recommended to not use encryption with deduplication storages.
Health check
Health check reads all the data from the last restore point and across the backup chain. When used with deduplication storages, this data is rehydrated from the storage which may be a slow process. It is recommended to enable health check on the primary backup job before data is transported by backup copy job to a deduplication storage.
Veeam defragmentation and compact settings
If you followed the above described guidelines, you do not need these settings. It is recommended that you disable them.