Authentication and connectivity
Aether CLI needs two things before it can do useful work:
- A reachable service endpoint.
- Whatever authentication context your deployment expects.
Service endpoint
The primary endpoint setting is:
- Environment variable:
AFS_AETHER_SERVER_ENDPOINT - Config key:
[bridge].server_endpoint
Example with environment variables:
export AFS_AETHER_SERVER_ENDPOINT='https://grpc.aetherfs.io'
Example with config:
[bridge]
server_endpoint = "https://grpc.aetherfs.io"
If this is not set correctly, mount and runtime commands will fail to reach the remote service.
Worker identity
Some deployments use a worker or client identity label.
Relevant setting:
- Environment variable:
AFS_AETHER_WORKER_ID - Config key:
[bridge].worker_id
Example:
export AFS_AETHER_WORKER_ID='alice-laptop'
Auth model
The exact credential source depends on how your environment is set up.
Common options include:
- Environment variables.
- Shell login context.
- A local credential helper.
- Service-issued tokens or identity federation.
The rule for Aether CLI users is simple: make sure the local environment you launch aether from has the credentials your deployment expects.
If your deployment uses the built-in CLI token path, the supported forms are:
AFS_AUTH_TOKEN='Bearer <tenant-uuid>:<principal>'aether --auth-token-file ./token.txt ...aether --auth-token-stdin ...
Connectivity guidance
When connectivity fails, work through these questions in order:
- Is
AFS_AETHER_SERVER_ENDPOINTor[bridge].server_endpointset correctly? - Can your machine reach that hostname and port?
- Are your local credentials present in the shell or environment that launches
aether? - Are you trying to inspect a running local runtime with
aether status, or are you expecting it to validate connectivity before anything is running?
That last distinction matters: aether status is for the running local runtime, not a generic preflight reachability probe.
First commands to use
aether doctor
aether mount --help
aether config --help
Use doctor first for local preflight checks such as FUSE availability, kernel support, and mount-directory sanity.
If the problem is endpoint reachability, remote auth, or session access, doctor is not enough. Move quickly to a real mount attempt with explicit endpoint/auth settings and readable logs.