Skip to content

Latest commit

 

History

History
69 lines (45 loc) · 3.99 KB

configuration.md

File metadata and controls

69 lines (45 loc) · 3.99 KB
Raw

Configuration

Configuration files

The following configuration files exist in the folder deploy/configs/gaia/:

admin-config.json
hub-config.json
reader-config.json

GAIA Admin Service

You can also use the GAIA Admin Service to remotely administer it with an API Key. It will require you to install npm with a apt install npm. Once npm is installed you can continue with the steps in the GAIA Admin Service.

Driver Selection

The Gaia hub currently supports the following drivers:

'aws' == Amazon S3
'azure' == Azure Blob Storage
'disk' == Local disk
'google-cloud' === Google Cloud Storage

Set the driver you wish to use in your config.json file with the driver parameter. Many drivers additionally accept the bucket parameter, which controls the bucket name that files should be written to.

These driver may require you to provide additional credentials for performing writes to the backends. See config.sample.json for fields for those credentials. In some cases, the driver can use a system configured credential for the backend (e.g., if you are logged into your AWS CLI account, and run the hub from that environment, it won't need to read credentials from your config.json).

:::caution The disk driver requires a *nix like filesystem interface, and will not work correctly when trying to run in a Windows environment. :::

Note on SSL

We strongly recommend that you deploy your Gaia hub with SSL enabled. Otherwise, the tokens used to authenticate with the Gaia hub may be stolen by attackers, which could allow them to execute writes on your behalf.
Configuration options are available to run the hub with an https Node.js server.
Otherwise, a reverse proxy web server such as nginx or Apache can be used.

Require Correct Hub URL

If you turn on the requireCorrectHubUrl option in your config.json file, your Gaia hub will require that authentication requests correctly include the hubURL they are trying to connect with. This is used to prevent a malicious gaia hub from using an authentication token for itself on other Gaia hubs.

By default, the Gaia hub will validate that the supplied URL matches https://${config.serverName}, but if there are multiple valid URLs for clients to reach the hub at, you can include a list in your config.json:

{
  ....
  serverName: "normalserver.com"
  validHubUrls: [ "https://specialserver.com/",
                  "https://legacyurl.info" ]
  ....
}

The readURL parameter

By default, the gaia hub drivers will return read URLs which point directly at the written content. For example, an S3 driver would return the URL directly to the S3 file. However, if you configure a CDN or domain to point at that same bucket, you can use the readURL parameter to tell the gaia hub that files can be read from the given URL. For example, the hub.blockstack.org Gaia Hub is configured to return a read URL that looks like https://gaia.blockstack.org/hub/.

Unset this configuration parameter if you do not intend to deploy any caching.

Minimum Proofs Requirement

The gaia hub can also be configured to require a minimum number of social proofs in a user's profile to accept writes from that user. This can be used as a kind of spam-control mechanism. However, we recommend for the smoothest operation of your gaia hub, to set the proofsConfig.proofsRequired configuration key to 0.

CDN & Replicated Hubs

  • https://docs.microsoft.com/en-us/azure/storage/blobs/storage-https-custom-domain-cdn
  • The hub implementation is designed to be ran from a single Node.js instance. If the hub instance is sharded (e.g. replicated hubs via load balancing), then any given bucket (identified by URI segment) must be served by the same instance, At least a couple elements of the Gaia Hub depend on this: token invalidation in-memory caching, and resource endpoint 409 contention behavior.