Partial File Updating not working

[tdcooper]

Hi,

I cannot seem to get Partial File Updating to actually partially update a large (47GB) file! The large file is part of a "reverse incremental" backup set of virtual machines. I keep 7 days of backups, so the folder structure contains the main full backup (approx 47 GB) + 6 incrementals. Each evening my backup software creates a small incremental file (Approx 2GB), then inserts the changes into the full backup to create a synthetic full and then deletes the oldest of the small incremental files.

I am currently syncing from a server to a remote PC with an eSATA drive hanging off it. I am running the Extremesync service on the remote PC, and have selected the appropriate option for this in the profile. I have also selected the "Partial File Updating" option, which automatically selects the supporting options for this. The only other option I have configured is to limit the ZIP file sizes to 1GB each.

So... I have scheduled the sync to run after my nightly backup, and what I was hoping was the most that would happen would be the new incremental would be produced + only the parts of the full backup file that have been modified by the synthetic backup would be copied.... but instead, every time the job runs it rezips and copies the entire fullbackup file! My main purpose for doing this is to send this eSATA drive to Amazon S3 to "pre-load" the files for me.... and then I can use partial file updating across the slow link to keep the backups in sync. But, at the moment it does not appear that this will work. I've attached my latest log file... perhaps this might give you some clue as to why?

[superflexible]

I doesn't work because the name of the big backup file changes every day. It has to remain the same file with the same filename. Different files cannot be recognized as being related.

[tdcooper]

I doesn't work because the name of the big backup file changes every day. It has to remain the same file with the same filename. Different files cannot be recognized as being related.

Turns out this was an issue with my backup software.... Veeam Backup and Replication, it renames the synthetic full each run. I've fixed this now, so it keeps the same file name.... but it's still zipping and copying the entire file, even though only a GB or 2 is changing. Do you think I need to create an entirely new profile and try again? Alternatively, is there a way to reset the database so it starts again as if it's the first run?

Attached last couple of logs.

[superflexible]

Apparently the backup files are not eligible for block-level processing, because every block in them is different every day (they are a stream of bytes and not an array of blocks).

What you need to do is back up the original database files directly with our software. Those are block-based database files. The backup files are not block-based.