Good morning. I have a strange situation on a pve where only a Windows Server 2016 Vm with a 500GB SSD drive runs. The space occupied is always the same or about 340GB. About 95GB of new data is copied every day when backing up to an incremental PBS even though the disk size hasn't changed at all. I am attaching some screens. I wanted to ask if it is possible somehow to understand what kind of data it is in order to intervene. Thank you very much for helping
Problem on incremental backup on PBS
- Thread starter AntInf
- Start date
PVE uses KVMs dirty-bitmapping to monitor what blocks got changed since the last backup. Not sure if its the same like the chunk size (which is always 4MB before compression) or smaller. Here it is enough if a single 512 byte block changed, then the whole 4MB chunk needs to be uploaded. So 95GB of new chunks doesn't mean the guest OS really got 95GB of changed data and alot of small changes can amplifiy a lot.
Last edited:
I continued to analyze the backups and unfortunately I always find these 95gb of data that must be copied, do you think I can not correct in any way? Do you think it may be due to the fact that the guest hosts a postgres db that clearly writes all the time? (certainly not 95GB, but probably a few megs)