nanopack / hoarder

A simple, api-driven storage system for storing code builds and cached libraries for cloud-based deployment services.

nanopack

Hub for nanopack resources and content

hoarder's Issues

CLI

Needs a basic CLI

Pluggable backends

Design in a way to allow for pluggable backends (S3, Postgres, ect).

For MVD the first backend will just be the Local FS

Hoarder responds with '200' to any HEAD request

Hoarder stores entirety of data before writing

Prior to storing to a file, hoarder performs an ioutil.ReadAll which reads into memory the entire blob. If the data is large enough, this causes system/performance issues.

Unconfirmed: When writing a file, hoarder server doesn't respond to further requests.

Can we support versioning for each blob

Hi,

I find this project is what I need. But only one little defect, which is version is not supported for the blob. So do we have the plan to have this feature?

Thanks in advance.

Keep oldest builds by default

Turn off garbage collection (GC- removing oldest builds) by default.

Allow for GC to be enabled/configured via a config file providing an interval between runs, and amount to GC.

Make generic all the things

Remove libs, and call builds blobs.

Turn this into a key value store, where the key is the file name and the value is just the binary data. Then it can be used for w/e anyone wants to use it for.

Refactor

Refactor the application structure, breaking things into packages where appropriate.

Running server with --insecure=false requires two token headers

Because hoarder is using nanoauth's DefaultAuth object, "X-NANOBOX-TOKEN" is the expected header to authenticate, however, hoarder is authenticating again by checking 'x-auth-token', requiring a user to specify both (-H "x-auth-token: 123" -H "X-NANOBOX-TOKEN: 123") in a request to avoid a 401 response.