Synthetic Backups of Very Large Files

[QSS-Support]

Hi,

Apologies if this has been asked before. I'm trying to understand how the Synthetic Backups work as I need to backup very large files (200+GBs each). In testing myself, it seems to create complete copies of the files at the destination each time with different timestamps in the filenames, so for example, I ended up with :-

file1-mondays timestamp.ext (inital full copy = 200+GBs)
file1-tuesdays timestamp.ext (second version = 200+GBs)
file1-wednesdays timestamp.ext (third version = 200+GBs)
etc.

with each version being almost the same size (200+GBs each). And of course, I very quickly ran out of disk space.

Is this how it's supposed to work or have I selected some wrong options to change the behavior ? I was hoping that it would wold work like this :-

file1-mondays timestamp.ext (first run full copy = 200+GB)
file1-tuesdays timestamp.ext (only contains changed blocks or incrementals, say 10+GBs of changes)
file1-wednesdays timestamp.ext (only contains changed blocks or incrementals, say 10+GBs again)
etc.

Can Synthethic backup work like my second example or only like my first example ?

Thanks.

Regards,

CK

[superflexible]

Hello,
if the source file has a block-oriented internal structure, and only some blocks change, then Synthetic Backup will copy only the changed blocks every time.

However some files such as SQL backup files are not block-based. If they are a stream of backup data, then the complete file will be different every time it is generated. In that case, Synthetic Backup does not work.

If you think it should work, but still have problems, please post log files here or send them to support@syncovery.com. Also make sure you test the latest version 7.61a.

[QSS-Support]

Thanks for the reply.

To help me understand how it works, when Synthetic backup runs and only sends across the changed blocks, how are these changes stored ?

For example, let say the original file is 200GBs in size.
After the initial run, the target folder now contains a copy of the 200GB file.

During the next run, only 500MBs of blocks were changed.
Assuming Synthetic backup recognises that only these 500MBs was changed and transfers these 500MBs of deltas to the target, what happens at the target end ? Does it create a new file to store only the 500MBs but leaves the original alone ? Or does it swap out the changed blocks from the target file ie. takes the blocks that were changed from the original file and write that to a file then integrate the new deltas into the 200GB file (I think this is called Reverse Incremental ??) ? Or does it do something completely different ?

Thanks.

CK

[superflexible]

It creates a new zip file with the changed blocks. When you look inside the zip, it will become very clear. Each modified block region is put into the zip file as a separate item, named in such a way that it is human readable.