vm only with 33gb used disk uses 100GB with backup
- Thread starter telvenes
- Start date
Do you mean the size shown in the GUI of PBS when you are in the Datastore -> {store} -> Content panel?
That shows the size of the Backup if you would restore it. Not accounting for any possible sparse allocation.
How much space the actual backup takes is hard to calculate because all backups are stored in chunks which are deduplicated. Meaning if two or more backups share exactly the same chunk, they will not store it separately but will just reference it. Once no backup references a chunk anymore, it can be removed in a garbage collection run.
The whole deduplication and usage accounting has been discussed here in the forum quite a bit already. The TL;DR is that it is a bit more complicated and more unpredictable than storing a full backup and following incremental backups because if no data changes, the only size change is in the backup's metadata which references the chunks for the backup.
The "Deduplication Factor" in the "Datastore -> {store} -> Summary panel gives you an overview of how much chunks are being used in more than one backup.
That shows the size of the Backup if you would restore it. Not accounting for any possible sparse allocation.
How much space the actual backup takes is hard to calculate because all backups are stored in chunks which are deduplicated. Meaning if two or more backups share exactly the same chunk, they will not store it separately but will just reference it. Once no backup references a chunk anymore, it can be removed in a garbage collection run.
The whole deduplication and usage accounting has been discussed here in the forum quite a bit already. The TL;DR is that it is a bit more complicated and more unpredictable than storing a full backup and following incremental backups because if no data changes, the only size change is in the backup's metadata which references the chunks for the backup.
The "Deduplication Factor" in the "Datastore -> {store} -> Summary panel gives you an overview of how much chunks are being used in more than one backup.
One reason could be that the file system inside the VM is only referencing 33GB but there might be old data that is not in use any more but still present on the disk. If the storage and the VM config support discard / trim, you can run
fstrim -a for Linux guests or Optimize-Volume -DriveLetter <YourDriveLetter> -ReTrim -Verbose for Windows.This will tell the underlying storage which areas are not in use anymore and can be nulled. Thus, the backup should see fewer data on the VM disk.
For this to work, you have to have "Discard" enabled in the VM disks setting and the bus/type of the disk needs to be SCSIO (with VirtIO SCSI Controller) or VirtIO Block if I remember correctly.
