Once you’ve installed the CockroachDB binary, it’s simple to start a multi-node cluster locally with each node listening on a different port.

Running multiple nodes on a single host is useful for testing out CockroachDB, but it's not recommended for production deployments. To run a physically distributed cluster in production, see Manual Deployment, Cloud Deployment, or Orchestration.

Watch a Demo

Feel free to watch this process in action before going through the steps yourself. Note that you can copy commands directly from the video, and you can use < and > to go back and forward.

Step 1. Start your first node

$ cockroach start --background
CockroachDB node starting at 2017-02-24 22:41:23.732936605 -0500 EST
build:      beta-20170223 @ 2017/02/23 16:34:25 (go1.7.4)
admin:      http://localhost:8080
sql:        postgresql:[email protected]:26257?sslmode=disable
logs:       cockroach-data/logs
store[0]:   path=cockroach-data
status:     initialized new cluster
clusterID:  {dab8130a-d20b-4753-85ba-14d8956a294c}
nodeID:     1

This command starts a node, accepting all cockroach start defaults.

  • Communication is insecure, with the server listening only on localhost on port 26257 for internal and client communication and on port 8080 for HTTP requests from the Admin UI.
    • To bind to different ports, set --port=<port> and --http-port=<port>.
    • To bind the Admin UI to a private IP address or host, set --http-host=<private-addr>.
    • To listen on an external hostname or IP address, set --insecure and --host=<external address>. For a demonstration, see Manual Deployment.
  • Node data is stored in the cockroach-data directory. To store data in a different location, set --store=<filepath>. To use multiple stores, set this flag separately for each.

  • The --background flag runs the node in the background so you can continue the next steps in the same shell.

  • The standard output gives you a helpful summary: the CockroachDB version; the URL for the admin UI; the SQL URL for your client code; the storage locations for node and debug log data; whether the node is the first in the cluster, joined an existing cluster for the first time, or rejoined an existing cluster; the cluster ID; and the node ID.
By default, each node's cache is limited to 25% of available memory. This default is reasonable when running one node per host. When running multiple nodes on a single host, however, it may lead to out of memory errors, especially when testing against the cluster in a serious way. To avoid such errors, you can manually limit each node's cache size by setting the --cache flag in the start command.

Step 2. Add nodes to the cluster

# Start your second node:
$ cockroach start --background \
--store=node2 \
--port=26258 \
--http-port=8081 \

# Start your third node:
$ cockroach start --background \
--store=node3 \
--port=26259 \
--http-port=8082 \

These commands add two nodes to the cluster, but you can add as many as you like. For each node:

  • Set the --store flag to a storage location not in use by other nodes. To use multiple stores, set this flag separately for each.

  • Set the --port and --http-port flags to ports not in use by other nodes.

  • The --join flag connects the new node to the cluster. Set this flag to localhost and the port of the first node.

  • The --background flag runs the node in the background so you can continue the next steps in the same shell.

If you don't plan to use more than one node, you can avoid unnecessary log messages about replication by editing the default replication zone to specify one node instead of three. See here for more details.

Step 3. Use the built-in SQL client

Start the built-in SQL shell on node 1:

$ cockroach sql
# Welcome to the cockroach SQL interface.
# All statements must be terminated by a semicolon.
# To exit: CTRL + D.

Then run some CockroachDB SQL statements:

> CREATE TABLE bank.accounts (id INT PRIMARY KEY, balance DECIMAL);
> INSERT INTO bank.accounts VALUES (1, 1000.50);
> SELECT * FROM bank.accounts;
| id | balance |
|  1 |  1000.5 |
(1 row)

When you’re done, use CTRL + D, CTRL + C, or \q to exit the SQL shell.

Step 4. Open the Admin UI

The CockroachDB Admin UI lets you monitor cluster-wide, node-level, and database-level metrics and events. To start up the Admin UI, point your browser to http://localhost:8080. You can also find the address in the admin field in the standard output of any node on startup.

You can also use Prometheus and other third-party, open source tools to monitor and visualize cluster metrics and send notifications based on specified rules. For more details, see Monitor CockroachDB with Prometheus.

CockroachDB Admin UI

Step 5. Stop the cluster

You can stop the nodes (and therefore the cluster) as follows:

# Stop node 1:
$ cockroach quit

# Stop node 2: 
$ cockroach quit --port=26258

# Stop node 3:
$ cockroach quit --port=26259

For more details about the cockroach quit command, see Stop a Node.

What’s Next?

Secure your cluster with authentication and encryption. You might also be interested in:

Yes No