Backing up to two rotating removable drives and online storage

[mphelpsmd]

I’m looking at using duplicacy to back up my CentOS 6.9 system to 2 rotating removable drives and an online backup (Backblaze B2).

My sense is that I would do something like:

$ cd /path/to/server/data
$ duplicacy init server_data /path/to/mounted/removable/drive_1
$ duplicacy backup
$ duplicacy add removable_drive_2  server_data  /path/to/mounted/removable/drive_2
$ duplicacy copy -to removable_drive_2
$ duplicacy add b2_backups server_data b2://my_b2_bucket
$ duplicacy copy -to b2_backups

I would be swapping out the two removable hard drives, with one of them being in the system backing up the data while the other one sits in my office at work (ie, offsite). And the B2-stored data would be the other layer of backups.

How would I manage swapping the removable drives? I’m thinking I would have them mounted at different mount points. So if I removed drive 1 and connected drive 2, then I would run:

$ cd /path/to/server/data
$ duplicacy backup -storage removable_drive_2

Am I on the right track here?

Thanks so much in advance.

[gilbertchen]

Basically right, but you should use different repository ids for the removable drives, otherwise the revision numbers will be messed up when copied to the b2 storage. And don't forget the -copy option for the add command.

$ cd /path/to/server/data
$ duplicacy init drive1 /path/to/mounted/removable/drive_1
$ duplicacy backup
$ duplicacy add -copy default removable_drive_2  drive2  /path/to/mounted/removable/drive_2
$ duplicacy copy -to removable_drive_2
$ duplicacy add -copy b2_backups b2 b2://my_b2_bucket
$ duplicacy copy -to b2_backups

[TowerBR]

I use a similar strategy, two rotating removable drives and a remote storage (Dropbox), but instead of using copy, I use backup for all three, to prevent an error (lack of chunk, corrupted chunk, etc.) propagates to other backups.

[mphelpsmd]

Thanks so much for the information.

I still don't feel I have a good grasp on a couple of the principles of Duplicacy, however.

For instance, when you say "you should use different repository ids for the removable drives", what exactly is a repository id? My understanding is that the term "repository" in duplicacy means the directory tree to be backed up.

Second, how do I have duplicacy periodically keep backing up to my various backup destinations? Do I use the backup or copy command? And do I put this in a cron job or something similar once the first backup has been done?

And third, especially for the remote backups (Backblaze B2 in this case), how do I encrypt the data? I know that there is the -e option. Does that just encrypt with a password or can you specify a key? And does duplicacy by default use ssh for B2, or is there another step that needs to happen in order to encrypt the data transmission for cloud-based backups? (Most of the examples I find here use an sftp server, rather than B2.)

Thanks again for the support. I'm going to try experimenting with small test set of files to ensure that I have everything right before attempting to back up the several terabytes.

[gilbertchen]

For instance, when you say "you should use different repository ids for the removable drives", what exactly is a repository id? My understanding is that the term "repository" in duplicacy means the directory tree to be backed up.

The repository id is what identifies a repository on the storage. Multiple repositories backing up to the same storage should have different repository ids. Repository ids are actually uses as the names of the subdirectories under snapshots on the storage to save the snapshot files for each repository.

Second, how do I have duplicacy periodically keep backing up to my various backup destinations? Do I use the backup or copy command? And do I put this in a cron job or something similar once the first backup has been done?

Currently this is not supported by the CLI version. A cron job or a scheduler task on Window is the best option.

And third, especially for the remote backups (Backblaze B2 in this case), how do I encrypt the data? I know that there is the -e option. Does that just encrypt with a password or can you specify a key? And does duplicacy by default use ssh for B2, or is there another step that needs to happen in order to encrypt the data transmission for cloud-based backups? (Most of the examples I find here use an sftp server, rather than B2.)

When you specify the -e option, Duplicacy will ask you for a password to encrypt the storage. And if you're using the B2 backend, then Duplicacy will also ask you for the B2 credentials for accessing your B2 bucket.

[TowerBR]

The repository id is what identifies a repository on the storage. Multiple repositories backing up to the same storage should have different repository ids. Repository ids are actually uses as the names of the subdirectories under snapshots on the storage to save the snapshot files for each repository.

So just to confirm that I understood:

For example, in this preferences file:

        "name": "NOTE2_arquivo_1M_fix_copy-dropbox",
        "id": "NOTE2_arquivo",
        "storage": "dropbox://NOTE2_arquivo_1M_fix_copy-dropbox",
        ...

name = storage name id = repository id (usually the name of the local folder or something related) storage = storage URL

And - obviously - there is no "name" for snapshots.

@gilbertchen, the names above make sense to you, thinking of logic: repository - storage - storage URL?