Is filename encryption repeatable?

[rossmcm]

i.e. If I have file and folder encryption turned on, will the resulting encrypted names be the same for the same file and path backed up from 2 different PC's?

Here's the reason this is important:

I have a regular sync to S3. The files are encrypted and compressed to a local folder (using another tool, not syncovery). This folder has a structure that mirrors the folder the file came from, i.e. file

Code: Select all

    C:\Projects\Project1\Project1.pas


ends up as

Code: Select all

    C:\__Backup\Projects\Project1\Project1.pas.7z


If I use syncovery to sync this folder to S3, this file becomes:

Code: Select all

    C:\__Backup\asdfasdfa\ewrtwertwert\hjlkhjlkkl


So, what happens if I run the same profile from another PC with the same folder structure and using the same encryption keys? Will I end up with an identical image on S3 (including the file and folder names)?

Changing the contents of a file will not change the encrypted filename, right? In other words, if I make changes to Project1.pas, Project1.pas.7z will be different, but the encrypted filename must not change so that the sync operation overwrites (and updates) the existing file.

[superflexible]

For safety purposes, the results of any encryption must not be predictable or deterministic.

Additional random values are used to ensure that the encrypted file name is always different.

An exception is made for folder names, these are always the same encrypted name, in order to prevent duplicate folders from causing problems.

But I don't see any problem there. Syncovery recognizes the names.

[rossmcm]

OK, I'm not so sure - I think the salting means I can't update the file.

Taking my example again, we have:

Code: Select all

C:\Projects\Project1\Project1.pas                  // my source file to be backed up
C:\__Backup\Projects\Project1\Project1.pas.7z      // an encypted 7Zip of the source file
S3:\__Backup\asdfasdfa\ewrtwertwert\hjlkhjlkkl     // S3 object (from Syncovery profile)

if I modify Project1.pas, generate a new Project1.pas.7z, and then run the Syncovery profile again, will file hjlkhjlkkl be overwritten? I.e will I end up with:

Code: Select all

S3:\__Backup\asdfasdfa\ewrtwertwert\hjlkhjlkkl  // i.e. hjlkhjlkkl is updated

or:

Code: Select all

S3:\__Backup\asdfasdfa\ewrtwertwert\ccvbnvbnnh  // i.e. hjlkhjlkkl replaced by ccvbnvbnnh

or:

Code: Select all

S3:\__Backup\asdfasdfa\ewrtwertwert\hjlkhjlkkl  // i.e. hjlkhjlkkl is ignored
S3:\__Backup\asdfasdfa\ewrtwertwert\ccvbnvbnnh  // and ccvbnvbnnh created

And (this is where it gets a bit tricky): What if the same filename and path is backed up to S3 from another computer? Will this backup store the file in the same path with the encrypted folder names (I presume and hope so). Will the name of the file be the same (I don't think so, unfortunately)?

(NOTE I'm not being lazy, I intend to do some more experiments, but this must be a question other users have and I'm interested to hear David's reply).

[superflexible]

Whether or not multiple versions of the same file remain depends on the Versioning settings.

If versioning is not used, Syncovery will upload the modified file with a different encrypted name and delete the old one.

If you back up from multiple computers to the same folder, just make sure you do not use "Cache Destination File List". Once that is avoided, Syncovery will always look at the destination first and then do the appropriate actions.

Encrypted folder names are the same from all machines.

The encryption password must be identical on all machines too.

In general I do recommend backing up each machine to its own destination folder, but I guess if the dataset is the same, you can also use a mutual backup destination. Just make sure that one machine doesn't delete another machine's files.

[rossmcm]

Thanks David, for those tips (especially the cache setting - I would have missed that). I'm going to do some experiments with a small dataset and report back.