Struct datacake_node::DatacakeNodeBuilder

pub struct DatacakeNodeBuilder<S = DCAwareSelector> { /* private fields */ }

Build a Datacake node using provided settings.

Implementations

impl<S> DatacakeNodeBuilder<S>where
    S: NodeSelector + Send + 'static,

pub fn new(
    node_id: NodeId,
    connection_cfg: ConnectionConfig
) -> DatacakeNodeBuilder<DCAwareSelector>

Create a new node builder.

pub fn with_node_selector<S2>(self, selector: S2) -> DatacakeNodeBuilder<S2>

Set a node selector.

This is used by systems to select a specific set of nodes from the live membership set with a given consistency level.

pub fn with_cluster_id(self, cluster_id: impl Display) -> Self

Set the cluster id for the given node.

pub fn with_data_center(self, dc: impl Display) -> Self

Set the data center the node belongs to.

pub async fn connect(self) -> Result<DatacakeNode, NodeError>

Starts the Datacake cluster, connecting to the targeted seed nodes.

When connecting to the cluster, the node_id must be unique otherwise the cluster will incorrectly propagate state and not become consistent.

Typically you will only have one cluster and therefore only have one cluster_id which should be the same for each node in the cluster. Currently the cluster_id is not handled by anything other than chitchat

No seed nodes need to be live at the time of connecting for the cluster to start correctly, but they are required in order for nodes to discover one-another and share their basic state.