Best practice for backup second disk?


I would like to use duplicacy to back up my second hard drive on a regular basis. This has the following structure/directories:


I want to backup everything except “d:\archive”.

Do I create a backup for the complete disk “d:” and exclude the directory “d:\archive” or do I create a separate backup for each directory “d:\data”, “d:\video”, “d:\images” and “d:\projects”?

What are the advantages/disadvantages of the two options?

If I create a separate backup for each directory, what do I have to use as “Backup ID”? Always the same ID or does each backup/directory get its own ID.

I don’t really understand the topic “Backup ID” in general. What exactly is the effect of this?

Thanks, Alex

Yes, this would be the most straightforward

You can do that too, if you for example want separate cadence of backups for those folders. You can go the third way, and create another folder, put symbolic links to the stuff you want to back up into that folder, and create only one backup task for that folder: duplicacy follows first level links.

Assuming you don’t need to have separate schedules or retention policies for these folders, the difference will be what happens with new folders in the root of your D: drive you might create in the future: The “exclude archive” option will have new folders backed up, while if you create separate backup schedules for the existing folders—new folders will be unprotected. But then maybe that’s your intention, and in this case, had you gone the “all but archive option” you would need to add new folders to exclusion list to avoid them being backed up.

It’s up to you.

Separate names.

Different IDs.

Think of it as a name of a dataset; duplicacy will keep the version history of the dataset. You may have a dataset “Music”, that backs up stuff from your entire drive daily by file types, and never prunes, and another dataset name “Documents”, that collects document files from specific folders and backs up hourly, and prunes after two weeks, etc. So later, you can ask—hey, give me the version history of dataset “Music”, or Restore my file “Taxes.pages” from Documents dataset to its state as of Friday, 13, 2013.

Many thanks. I understand all except the "Backup ID” thing.

Why does duplicacy not use the automaticly the “Directory” name for ""Backup ID” name?
In which case do I have different backups, with the same ""Backup ID”?

And what happens, when i create a backup for each directory, and later, i decide to make a new structure of the directories. For example i cretae “d:\multimedia” and move “d:\music” and “d:\images” to the new “d:\multimedia” directory as subdirs.

And what is there to consider, when i work with Windows and Linux (Dual-Boot on Disk1) and want to make duplicacy backups sometimes from Windows and sometimes from Linux? (Both times the same directories of the second hard drive.)

You may have same directory names that contain different data, e.g. on different machines, or on the same machine, e.g. c:\users\alex\documents and c:\users\james\documents.

But I do agree, duplicacy shall offer the default unique (across the backup store) name that is guaranteed to work, and not ask user to do unnecessary work coming up with the unqiue name.

Personally, I name them as <machine-name>-<description>, such as “Alex’s MacBook Pro - All Users”, or “TrueNAS - Archive”. etc.

Never. This shall never be allowed to happen. The backup ID must be uniquely identifying the source dataset. This is why I think duplicacy is in the best position to come up with the unique one, since it has access to all the other backup IDs that have been configured on that specific storage (those are subfolders under the snapshots folder on the destination)

Assuming the multimedia folder is included in the backup set, Duplicacy will proceed backing up your files. The backup history before day X will have data in separate folders, and after—in one. Not sure if I understand your question though. What do you expect to happen?

Would that be possible?

Ah, I see.

Then yes, you can have same snapshot ID for the same subset of data. But make sure that no other data is included in that dataset and that the filters if any are configured identically.

Ohhh…but this is not very safe and errors are preprogrammed.

What would happen in such a case if i do not take exactly the same data/directories or have different filters?

But when i take a different/new id, i think it would also be ok. The data will still be deduplicated. I have just then no common history, Windows has his and Linux has his. Right?

The files that are included in one but not the other will appear in revisions that are done from one and not the other.

Yes, and yes. That seems to be a preferable way to handle that, avoiding potential confusion in the future if something changes in either of configurations.

I’m still confused about the backup-id. When i create a new backup job in the gui, i can see the already existing id’s on the right side via the symbol and i could also select one of them. Why is this function there if the id’s should only be used one time?

For example, you may have got a new computer, moved all data from the old one, erased and sold the old one and want to continue the backup history on the new one. You would then create a backup of the same files and folders to the same snapshot ID and continue as if nothing happened.

Ok, now i understand it.