Micro pruning is a term that is specific to using the deduplicated database (DDB). It removes specific references and data from disk storage based on retention criteria.
Pruning occurs and decrements reference counts in the DDB once all references to an archive file (the way Commvault tracks data) become zero. The data is then physically deleted.
- This feature uploads data in 4 GB chunks (could be less depending on the size of the job) and breaks the chunk into 8 MB parts (default size: nCloudPageModeSubFileSizeKB) for upload to the cloud.
- The entire part is deleted when all deduped records in it are prunable, and then the 8 MB part is pruned (micro pruning).
- The 4 GB chunk is deleted when all dedupe records for all part files are prunable.
- For cloud storage configured with deduplication, micro pruning is enabled by default. It ensures that data will be pruned from cloud storage, as soon as the backup job meets the retention rules, without sealing the corresponding deduplication database (DDB).
Note: When Object Lock is used, no micro pruning occurs. Once the retention period is reached, all data is deleted. Using a 60-day WORM policy as an example, micro pruning does not occur until day 120, and then the entire set of data is deleted.
Note: Due to the way Commvault prunes deduplicated data and ECS reclaims data, a large amount of unreclaimable garbage could remain in ECS until all dedupe records for all part files are pruned.
- Macro pruning is the method used for nondeduplicated data. Once archive files associated to jobs meet the retention period, the data is immediately removed from disk or storage. Macro pruning also occurs when the DDB is sealed, and all jobs tied to that DDB have met retention. (The aging process is then like non-deduplicated data because Commvault no longer uses the DDB process previously described.)
- The archive (chunk) maximum size is 4 GB (could be less depending on the size of the job), which is broken into 32 MB (default size: nCloudMaxSubFileSizeKB) parts for upload to the cloud.
- The entire archive is deleted when pruned.
- Monitoring the capacity utilization of ECS is highly recommended. You can monitor the capacity utilization of storage pools and nodes, as well as the entire VDC, from the ECS Portal Monitor > Capacity Utilization page.