Deduplication - Advanced
Use this dialog box to modify the advanced deduplication options.
- Temporarily disable deduplication
- Use this check box to temporarily suspend deduplication during backups for diagnostics and maintenance purposes. When you clear the check box, the signature generation and data deduplication is resumed.
- To continue client backups without deduplication during DDB recovery, use Allow backup jobs to run to deduplication storage policy copy when DDB is an unusable state option on the Media Management Configuration dialog box. For more information, see How do I continue my client backups during DDB recovery?.
- Deduplication Options
- Select options to perform deduplication operations in DASH (Deduplication Accelerated by Streaming Hash) mode. In this mode, hash signatures generated for data segments are effectively used to accelerate data transfer.
Applies To: Primary Copy.
- Enable DASH Full (Read Optimized Synthetic Full)
- DASH Full is a read optimized Synthetic Full operation. When the first full backup is complete, changed data blocks are protected during incremental or differential backups. DASH Full operation, reads the signatures from the meta data and updates the DDB and index files for existing data rather than physically copying data. DASH Full significantly reduces the time it takes to perform full backups.
- Enable Deduplication on Clients
- Select this option to enable source-side deduplication on the storage policy copy.
- When this option is selected Use Storage Policy Settings option is enabled by default on the associated subclient properties. All clients associated to this storage policy will honor the source-side deduplication.
- Enable DASH Copy (Transfer only unique data segments to target)
This option is available only on secondary copies with deduplication.
DASH ((Deduplication Accelerated by Streaming Hash mode) Copy is enabled by default and Disk Read Optimization is the default method used.
Note: If the source copy is without deduplication and Network Read Optimized copy option is not selected, or if DASH Copy is disabled, then all data blocks are transmitted and signature generation/comparison is done on the destination copy's MediaAgent.
- Disk Read Optimized Copy
- Optimizes data transfer by using existing data signatures.
For disk read optimization the source copy must be deduplication enabled. Disk Read Optimized copy uses existing deduplicated block signatures on the source copy for comparison against existing signatures on the destination copy's DDB. Only unique blocks are transmitted. During DASH Copy with Disk Read Optimization:
- The existing signatures are read from the data chunk's meta data (that contains the data block signatures) available on the source copy.
- The signature is compared against the destination copy's deduplication database (DDB).
- If the signature already exists, the destination copy DDB is updated to reflect another copy of data exists on the destination storage.
- If the signature does not exists (unique data block), the destination copy's DDB is updated with the new signature, and the data block is copied to the destination copy.
- Network Optimized Copy
- Optimizes data transfer by performing data deduplication on the source.
For network read optimization, the source storage policy copy can be with or without deduplication. Network Optimization copy reads each data block to create a signature for comparison against existing signatures on the destination copy's DDB. Only unique blocks are sent over the network. During DASH Copy with Network Read Optimization:
- Each data block is read from the source copy and a signature is generated.
- The signature is compared against the destination copy's DDB.
- On primary copy, when this option is enabled, source-side cache is configured on all clients associated to this storage policy.
- On secondary copies, use source-side cache option when both source and destination MediaAgents are in WAN network environment. If source and destination MediaAgents are in network environment other than WAN, enabling source-side cache on those MediaAgents might degrade the performance of the DASH Copy.
Source-Side cache for DASH Copy can be performed using:
Tip: For faster DASH copy performance, move the Job Results directory to a faster disk on the source MediaAgent. For instructions, see Changing the Path of the Job Results Directory.
- Limit the Max cache size to n MB
Use this option to set the maximum size of the source-side cache. The range of the size:
- For backup jobs is: 1 GB to 128 GB.
- For auxiliary copy jobs is: 8 GB to 128 GB.
- Optimize for High latency networks by avoiding remote lookups.
- Use this option to increase data protection operation performance when clients or source MediaAgent (for DASH Copy) are in delayed network or in high latency environments like WAN, and Data Mover and DDB MediaAgents are in fast network environments like LAN.
- On primary copy, enable this option to configure high latency optimization on all clients associated with this storage policy.
When high latency optimization is enabled, the client compares the signature against the local cache. The DDB is not looked up by the client or source MediaAgent (for DASH Copy).
- If the signature exists in the local cache the data block is discarded.
- If the signature does not exists in the local cache, the signature is updated in the local cache that can be referred by further backup jobs and the data with the signature is transmitted to the data mover MediaAgent.
The data mover MediaAgent with a local or remotely hosted DDB compares signatures against the DDB.
- If the signatures exists in the DDB, the Data Mover MediaAgent discards the data blocks and adds the additional entries in the DDB.
- If the signatures are not available in the DDB, the DDB will be updated with new signatures and data will be written to the disk.
Restriction: This option is not supported for a storage policy copy configured with cloud storage library or when Use DDB Priming option with Source-Side Deduplication option is enabled on Deduplication - Settings tab.