Question
Why is there a difference in estimated vs completed data size for Linux Block backups
Answer
For all new Linux Block level backups, we use a new algorithm to determine the actual amount that needs to be backed up called hasher. When the backup starts, LVM takes a snapshot and we report this size as "estimated data." Once the backup actually starts transferring data we examine the blocks to see which blocks were actually changed and only move those over to the destination storage. We do not move the non-changed blocks as there is no sense in transferring the data twice. The transferred blocks is the actual amount that was backed up in an incremental but, as sometimes only a few blocks in a large file have changed, you can get a large difference between the estimated and actual backed up data.