Even if you dedicate 2-3 TB HDD server for file server, there will be moment, when disk space will be exhausted. Uploader already has scaling mechanism.

Provide UPLOAD_DIGEST_PREFIX parameter to container setup.

docker run -ti \
...
-e UPLOAD_DIGEST_PREFIX=abcde \
...

This means that after file uploaded every digest will be prefixed with fixed string “abcde”. So you know that digests beginning with this “abcde” belong to this particular server. Although UPLOAD_DIGEST_PREFIX is optional, it is very recommended to set it to any uploader instance.

Then you can configure rules on you loadbalancer (nginx, HaProxy, ingress) to route “abcde” to proper server:

• Route “https://upload.example.com/file/abcde*” urls to old uploader instance.
• Route all other “https://upload.example.com/file” urls to new uploader instance.