-
Notifications
You must be signed in to change notification settings - Fork 76
Description
I currently use aptly to manage a repository in S3. I have the entire .aptly folder (database) in the bucket as well so that I can manage the repos on any computer (using s3fs). This works well except for one issue, all the packages are replicated (once in the aptly database and once in the published folder), which means for large repos it uses more storage than it really needs to. Aptly allow to publish to a filesystem with three different linkmethods (copy, hardlink, or symlink). Copy is what I am currently using (which doubles the storage), hardlinks do not work with s3fs so I have tried symlinks and on the aptly side this all works great, the binaries are only stored once in the database and the published directory contains the symlinks. However apt-transport-s3 doesn't know how to follow these links so pulling packages doesn't work. My request is for apt-transport-s3 to follow symlinks when accessing the S3 bucket. Is this possible? Thanks.