Skip to main content
Version: Next


Running VAST means spawning a process of the vast executable. A VAST process can operate in two modes:

  1. Server: runs continuously and listens on a network socket accepting connections.
  2. Client: connects to the server to (1) submit a request and receive a response, (2) publish data, or (3) subscribe to data.

A server contains a special component called the node that acts as container for pluggable components implemented as actors. In the future, VAST will be able to connect multiple nodes together to create a distributed system.

A standard deployment consists of a server close to the data sources and multiple clients that publish events and submit queries:

Start a server

The start command spins up a VAST server that blocks until told to stop:

vast start

By default, a VAST server listens on localhost and TCP port 5158.

Usually you would invoke vast start only for testing purposes in a terminal. In production you would typically use a service manager, e.g., systemd on Linux.

Stop a server

There exist two ways stop a server:

  1. Hit CTRL+C in the same TTY where you started VAST.
  2. Send the process a SIGINT or SIGTERM signal, e.g., via pkill -2 vast. Sending VAST a SIGTERM is the same as (1).

Option (3) comes in handy when you are working with a remote VAST server.

Spawn a client

Every command except for start is a client command that interacts with a server. Run vast help for a list of available commands.

To select a specific VAST server to connect to, configure the endpoint, e.g., by providing --endpoint=host:port on the command line, exporting the environment variable VAST_ENDPOINT=host:port, or setting the configuration option vast.endpoint: host:port in your vast.yaml.

Client connection failure

In the event of a connection failure, the clients will try to reconnect. This process can be tuned by the two options in the configuration file:

# The timeout for connecting to a VAST server. Set to 0 seconds to wait
# indefinitely.
connection-timeout: 5m

# The delay between two connection attempts. Set to 0 to try connecting
# without retries.
connection-retry-delay: 3s