Connection
module: streamtasks.connection
Used to create connections external to the process.
It implements servers and clients for multiple connection methods.
connecting to a server
The easiest way to connect to a server is to call connect with a url. The connect function returns a link connected to the server. It best can be managed with the AutoReconnector.
If no url is specified a connection will be made to the unix socket corresponding to the current node name. See environment variables for more information about the node name.
The following url schemes are supported:
unix:///[some path]- specify a path to a unix sockettcp://[hostname]:[port]- specify a tcp connection
When the url string is not a valid url, it will be interpreted as a node node. The node name will then be transformed into a unix socket path with get_node_socket_path. A connection will be attempted to this unix socket path.
If a username or password is specified in the url, it will be sent as handshake data. Any query parameter will also be included in the handshake data.
If a query parameter cost, with an integer value is specified, it will be used to set the cost of the connection.
The handshake data can be used to authenticate and configure a connection.
AutoReconnector
The AutoReconnector can be used to reestablish connections when closed. it takes a link connecting to the rest of the system and a connect function returning a link.
It can be run by calling run.
starting a server
Before starting a server a server of a specific type must be created. This can be done by calling get_server with a link connecting the server to the rest of the system and a url. The url works the same way it does for the connection. get_server returns a ServerBase. The server can be started by calling run.
In case of TCP servers the hostname and port is used for binding the socket.
In case of Unix socket servers the path is used to create the unix socket.