Seedbank is a librarian for your Amazon Glacier archives.
It allows you to create archives with more metadata including file lists, Markdown descriptions, and timestamps. It also integrates with Git to track changes to your archives over time. You can upload copies of your seedbank repository to Glacier automatically.
Amazon Glacier limits the length of archive descriptions to 1KB of ASCII characters. Amazon's suggestion is to store metadata in another database, but that doesn't make sense if you're a consumer user just trying to store old archives. Seedbank enables you to store the metadata for your archives both in Git and Amazon Glacier itself so that even if the repository for your archives gets lost, you can still recover all of your metadata.
Starting in an empty directory, initialize a seedbank repository.
> sb init
In addition to creating seedbank's config file (seedbank.json,) sb init also initializes the git repository and makes the initial commit.
Note: it's recommended that you add a remote (preferably a private repo) for the git repository. Seedbank will offer to push to it automatically after you have it set up.
Creating an archive is just as easy.
> sb create ~/somedir
Building archive of ~/somedir...
Using provided archive description '~/somedir/description.md'.
and then some time later...
Archive [uid] for ~/somedir created.
Notice that the above refers to the provided description for the archive. If you place a description.md (with that exact name, no capital letters) in the target folder for archive creation, it will be included with the archive's metadata. That is, after all, the whole point of seedbank. Descriptions are not required, though, and the archive's file list will still be included.
sb list [-n N] lists all of your archives in chronological order. The -n option shows only the last N recent archives. It also warns you about archives that have been created, but not uploaded.
sb upload {uid} uploads an archive to Amazon Glacier. Automatically chooses between multipart and singlepart uploads.
NOT YET IMPLEMENTED: sb download {uid} initiates a retrieval from Glacier. Ideally you could check (with sb check) to see the status of any outstanding jobs.
As with git, you do not have to provide the full uid in to reference an
individual archive, but only enough bytes for the reference to be unambiguous.
All local archives are saved in the (.gitignored) local folder in the seedbank repository.
It's tedious to have to keep your current working directory in the seedbank repo while you archive other folders, so seedbank supports setting a default seedbank repository for the system.
You can create a file called ~/.sbrc.json with the following contents:
{
"default_repo" : "~/seedbank"
}
Seedbank will attempt to load this repository when you use commands like sb list instead of defaulting to the current directory. In the future we'd like to create a way to override this with --here, but that's not currently implemented.