Inefficient partial downloads of larger files

[parthy]

Issue description

We are hosting files in the 10-15 GiB range and have a use case (hybrid Torrent/HTTP downloads) where clients frequently request partial content from the entire file range. Downloading a chunk from the end of the file can take up to 10 seconds and more.

Proposed mitigation

A more efficient implementation of the partial content feature using some form of seek instead of skipping through the entire file could be beneficial.

Deployment

We are currently running Sonatype Nexus RepositoryOSS 3.68.1-02 on a Linux (NixOS) server.

[nblair]

Thanks for opening an issue @parthy. I've marked this as an enhancement request, we'll keep it open to gauge demand for this type of feature.