Mokhtar Naamani 6d9153d042 pioneer and nodejs fixes vor 3 Jahren
..
bin b2832192a1 update helios vor 3 Jahren
lib 47bed3abf6 storage-node: drop ipns and discovery vor 3 Jahren
paths 7a87e3a9cb colossus: always accept uploads vor 3 Jahren
test 77c8a67d86 Autofix linter issues vor 4 Jahren
.eslintrc.js 34358429f8 move storage node into storage-node/ folder vor 4 Jahren
README.md 8b34fc03d2 storage-node: prettier formatting vor 4 Jahren
api-base.yml 8b34fc03d2 storage-node: prettier formatting vor 4 Jahren
package.json 6d9153d042 pioneer and nodejs fixes vor 3 Jahren

README.md

Storage Nodes for Joystream

Development

Run a development server (an ipfs node and development chain should be running on the local machine)

$ yarn colossus --dev

This will expect the chain to be configured with certain development accounts. The setup can be done by running the dev-init command for the storage-cli:

yarn storage-cli dev-init

Command-Line

$ yarn colossus --help
  Colossus - Joystream Storage Node

  Usage:
    $ colossus [command] [arguments]

  Commands:
    server        Runs a production server instance. (discovery and storage services)
                  This is the default command if not specified.
    discovery     Run the discovery service only.

  Arguments (required for server. Ignored if running server with --dev option):
    --provider-id ID, -i ID     StorageProviderId assigned to you in working group.
    --key-file FILE             JSON key export file to use as the storage provider (role account).
    --public-url=URL, -u URL    API Public URL to announce.

  Arguments (optional):
    --dev                   Runs server with developer settings.
    --passphrase            Optional passphrase to use to decrypt the key-file.
    --port=PORT, -p PORT    Port number to listen on, defaults to 3000.
    --ws-provider WS_URL    Joystream-node websocket provider, defaults to ws://localhost:9944

To run a storage server in production you will need to enroll on the network first to obtain your provider-id and role account.

API Packages

Since it's not entirely clear yet how APIs will develop in future, the approach taken here is to package individual APIs up individually. That is, instead of providing an overall API version in api-base.yml, it should be part of each API package's path.

For example, for a foo API in its version v1, its definitions should live in ./paths/foo/v1.js and ./paths/foo/v1/*.js respectively.

Note: until a reasonably stable API is reached, this project uses a v0 version prefix.

Interface/implementation

For reusability across API versions, it's best to keep files in the paths subfolder very thin, and instead inject implementations via the dependencies configuration value of express-openapi.

These implementations line to the ./lib subfolder. Adjust app.js as needed to make them available to API packages.

Streaming Notes

For streaming content, it is required that stream metadata is located at the start of the stream. Most software writes metadata at the end of the stream, because it is when the stream is committed to disk that the entirety of the metadata is known.

To move metadata to the start of the stream, a CLI tool such as qtfaststart for MP4 files might be used.