Answers are inline
1. Can someone explain the technical difference between Horizontal Scaling, and manually creating separate DDBs?
- The purpose of the Horizontal Scaling is Commvault's way of managing different IDA in a single Global Deduped Policy. The system will automatically detect the IDA and create associated DDB partition for the specific IDA. There are various parameters that can be tuned to manage when new DDB partition are created (i.e. 200 million Blocks before a new DDB partition is created, but existing subclients will still write to the original partition, only new clients will write to the new DDB partition)
2. The Horizontal Scaling documentation says it creates separate DDBs for "virtual machines, databases, and file system agents" ... Are NDMP or NetApp IntelliSnap backups included in one of the types mentioned?
- NDMP and NetApp Intellisnap will be classified under the FileSystem agent
3. We have 500+ front-end terabytes of Oracle flatfiles being backed up by the Linux FileSystem Agent, with its own dedicated DDB. Dedupe savings is 87%. Dedupe savings on the 68 front-end TB of general 'Unix' filesystem agent stuff is 75%. We're looking to consolidate our 14 DDBs to as few as possible, and it sounds like Horizontal Scaling would lump this in with the File System Agent stuff. Is that for the best, or does this type of data merit its own DDB?
- I believe with your current configuration leveraging the Horizontal Scaled DDB would be the best way forward. With a consolidated single Global Deduped, the system will automatically manage the specific IDA type.
Feel free to reach out if you have any other specific question