Single Cluster Architecture

When independent BlobCity nodes are connected together to form a single cluster, you
have essentially made a single cluster architecture. All clustering starts with first making a
single cluster architecture. Nodes can be added and removed from a running cluster.


Setting up a cluster, requires at-least 2 nodes to be up and running. Individually setup two
nodes of BlobCity, and apply license on each node. Do not create any datastore or
collections yet, as non-empty nodes cannot be added to clusters.

Make sure your nodes can communicate over the network. Keep the LAN ip address and
node-id of each node handy.

For ease of readability of this document, let’s assume your two nodes are:

Node IDLAN IP address

Adding the first cluster node

blobcity>add-node <node-id> <ip>

CLI into any of your two nodes and run the above command with node-id and ip of the
other node to connect the other node to your current node.

It is recommended that you use the LAN ip address of the node to connect and not the
WAN ip address. The faster the network i/o speed available or the nodes to communicate
with each other, the better it is. A minimum 1Gbps inter node communication speed is
recommended, but having higher speed will only help gain more clustering performance.

If the command returns an additional confirmation, then it means that the new node can
now used by your collections for sharing the data you store in collections. New
transactions can be initiated on node2. Data on node2 will be accessible to queries that are
initiated on node1.

Adding more nodes

The same process as used to add node2 can be used to add more nodes into the cluster.
The add-node command can now be run from either of node1 or node2. Any further
nodes added to the cluster will be connected to both node1 and node2 and all other nodes
in the cluster.