How Should I use Granular Backup?

You don’t have to use Granular Backup to back up everything; it’s there for added flexibility and value.
Most users create two or more tasks, one for the full backup, whether compressed / deduped or not. The other one runs more frequently and backs up only specific folders from inside the VM, via Granular Backup. This is faster and more economic than backing up the whole VM, especially when it’s a big one. Because it uses a real CPU and not a virtualized one, it’s also more efficient.

If you are concerned about disk space, you could use higher levels of compression and more deduplication workers ie. CPU cores (Enterprise Edition allows more than 2) to cut storage usage.

The beauty of using several backup tasks with BackupChain is:

• you can use different schedules
• you can use different processing settings: compression or not, dedupe or not
• you can use different retention settings.
• all of the above in all kinds of useful combinations

For example, a good customer of ours described how he set up his backups plans across almost 50 servers:

He runs a full backup as-is, no processing, every week to one device, and keeps only few backups per file: this allows for immediate access to the full VHDX, so you can boot it instantly
The second task backs up to another device, hourly, and creates backups as zipped data files (a SQL folder, file folders, etc). It keeps the last 24 hours allows for immediate access of crucial files.
Then there is a backup task to our cloud, everything nightly or weekly with default settings (dedupe and compression on, 10 revisions). In case all else fails.
Yet another task writes to another device everything nightly using default settings, compressed with dedupe, up to 10 revisions. This option saves space and keeps a longer history, but requires a restore operation.

In terms of license value and efficiency it’s recommended to leave everything on the host, unless you have pass-through disks, which we would almost never recommend using.