Running nodes locally

There are several ways to run a Corda node locally for testing purposes.

Starting a Corda node using DemoBench

See the instructions in Demobench.

Starting a Corda node from the command prompt

You can run a node by opening a terminal / command prompt window in the node’s directory and running the following command:

java -jar corda.jar

By default, the node will look for a configuration file called node.conf and a CorDapps folder called cordapps in the current working directory. You can override the configuration file and workspace paths on the command line (e.g. ./corda.jar --config-file=test.conf --base-directory=/opt/corda/nodes/test).

Setting JVM arguments

There are several ways to set JVM arguments for the node process (particularly the garbage collector and the memory settings). They are listed here in order of increasing priority - if the same flag is set in a way mentioned later in the list below, it will override anything set earlier.

  • Default arguments in capsule: The capsuled Corda node has default flags set to -Xmx512m -XX:+UseG1GC - this gives the node a relatively low 512 MB of heap space, and turns on the G1 garbage collector, ensuring low pause times for garbage collection.

When devMode is explicitly set to false, the default node memory size will be enlarged to 4G: -Xmx4G -XX:+UseG1GC.

  • Node configuration: The node configuration file can specify custom default JVM arguments by adding a section like the one below:
custom = {
   jvmArgs: [ "-Xmx1G", "-XX:+UseG1GC" ]
  • Capsule specific system property: You can use a special system property that Capsule understands, to set JVM arguments only for the Corda process, not the launcher that actually starts it:
java -Dcapsule.jvm.args="-Xmx1G" -jar corda.jar

Setting a property like this will override any value for this property, but not interfere with any other JVM arguments that are configured in any way mentioned above. In this example, it resets the maximum heap memory to -Xmx1G but it does not touch the garbage collector settings. This is particularly useful for either setting large memory allowances that you don’t want to give to the launcher, or for setting values that can only be set on one process at a time - for example, a debug port.

  • Command line flag: You can set JVM arguments in the command prompt that apply to the launcher process and the node process as in the example above. This overrides any value for the same flag set any other way, but leaves any other JVM arguments alone.

  • OutOfMemoryError handling: In addition to the JVM arguments listed above, the capsuled Corda node has two flags that cause the node to stop on out-of-memory error and generate the corresponding diagnostic files:

-XX:+HeapDumpOnOutOfMemoryError -XX:+CrashOnOutOfMemoryError

With CrashOnOutOfMemoryError the node which is running out of memory is expected to stop immediately (fail-fast) to preserve ledger consistency and avoid flaws in operations.

Unlike for arguments related to memory and GC, to completely replace the default out-of-memory error arguments, you must explicitly add at least one out-of-memory error related argument into the custom.jvmArgs section. For example, the following configuration turns off HeapDumpOnOutOfMemoryError and does not invoke the CrashOnOutOfMemoryError option:

custom = {
   jvmArgs: [ "-Xmx1G", "-XX:+UseG1GC", "-XX:-HeapDumpOnOutOfMemoryError" ]

Command-line options

You can optionally start a node using the following command-line options:

  • --base-directory, -b: The node working directory where all the files are kept (default: .).
  • --config-file, -f: The path to the configuration file. Defaults to node.conf.
  • --dev-mode, -d: Runs the node in development mode. Unsafe in production. Defaults to true on MacOS and desktop versions of Windows, otherwise defaults to false.
  • --no-local-shell, -n: Do not start the embedded shell locally.
  • --on-unknown-config-keys <[FAIL,INFO]>: How to behave on unknown node configuration. Defaults to FAIL.
  • --sshd: Enables SSH server for node administration.
  • --sshd-port: Sets the port for the SSH server. If not supplied and SSH server is enabled, the port defaults to 2222.
  • --verbose, --log-to-console, -v: If set, prints logging to the console as well as to a file.
  • --logging-level=<loggingLevel>: Enable logging at this level and higher. Possible values: ERROR, WARN, INFO (default), DEBUG, TRACE.
  • --help, -h: Show this help message and exit.
  • --version, -V: Print version information and exit.


clear-network-cache: Clears the local copy of the network map - it will be restored from the server or the file system on node start-up.

initial-registration: Starts an initial node registration with the compatibility zone to obtain a certificate from the Identity Manager Service (formerly Doorman).


  • --network-root-truststore, -t required: Network root trust store obtained from the network operator.
  • --network-root-truststore-password, -p: Network root trust store password obtained from the network operator.

generate-node-info: Performs the node start-up tasks necessary to generate the file, saves it to disk, then exits.

generate-rpc-ssl-settings: Generates the SSL keystore and truststore for a secure RPC connection.

install-shell-extensions: Installs a corda alias and auto completion for bash and zsh. For more information, see Shell extensions for CLI Applications.

validate-configuration: Validates the actual configuration without starting the node.

Enabling remote debugging

To enable remote debugging of the node, run the node with the following JVM arguments:

java -Dcapsule.jvm.args="-agentlib:jdwp=transport=dt_socket,server=y,suspend=y,address=5005" -jar corda.jar

This will allow you to attach a debugger to your node on port 5005.

Starting a node with JMX monitoring enabled

To enable export of JMX metrics over HTTP via Jolokia, run the following command in the command prompt:

java -Dcapsule.jvm.args="-javaagent:drivers/jolokia-jvm-1.3.7-agent.jar=port=7005" -jar corda.jar

This command will start the node with JMX metrics accessible via HTTP on port 7005.

See Monitoring via Jolokia for further details.

Starting all nodes at once on a local machine from the command prompt


If you created your nodes using deployNodes, a runnodes shell script (or batch file on Windows) will have been generated to allow you to quickly start up all nodes and their webservers. You should only use runnodes for testing purposes.

Start the nodes with runnodes by running the following command from the root of the project:

  • Linux/macOS: build/nodes/runnodes
  • Windows: call build\nodes\runnodes.bat

If you receive an OutOfMemoryError exception when interacting with the nodes, you need to increase the amount of Java heap memory available to them, which you can do when running them individually. See Starting a Corda node from the command line.


If you created your nodes using Dockerform, the docker-compose.yml file has been created and configured appropriately. Navigate to build/nodes directory and run the docker-compose up command. This will start up nodes inside a new, internal network. After the nodes are started, you can use the docker ps command to see how the ports are mapped.

Starting all nodes at once on a remote machine from the command line

By default, a Cordform task will run all the generated nodes on the same host machine. In order to run the nodes remotely, you can deploy them locally and then copy them to a remote server. If after copying the nodes to the remote machine you encounter errors related to a localhost resolution, you should follow the additional steps below.

To create nodes locally and run on a remote machine, perform the following steps:

  • Configure a Cordform task and deploy the nodes locally as described in Creating nodes locally.
  • Copy the generated directory structure to a remote machine, for example using Secure Copy.
  • Optionally, add database configuration settings if they could not be configured in the first step and the local machine does not have access to the remote database. In each top-level [NODE NAME]_node.conf configuration file, add the database settings and copy the JDBC driver .jar file (if required). Edit the top-level [NODE NAME]_node.conf files only and not the files inside the node sub-directories (for example, node.conf).
  • Optionally, bootstrap the network on the remote machine. This is an optional step when a remote machine does not accept localhost addresses, or if the generated nodes are configured to run on another host’s IP address. If needed, change the host addresses in the top-level configuration files [NODE NAME]_node.conf for entries p2pAddress, rpcSettings.address, and rpcSettings.adminAddress. Run the network bootstrapper tool to regenerate the nodes network map: java -jar corda-tools-network-bootstrapper-Master.jar --dir <nodes-root-dir>. For more information, see Network bootstrapper.
  • Run nodes on the remote machine using runnodes command.

The steps described above enable you to create the same test deployment as a deployNodes Gradle task would create on a local machine.

Database migrations

Depending on the versions of Corda and of the CorDapps used, database migration scripts might need to run before a node is able to start. For more information, see Database management.

Stability of the Corda Node

There are a number of critical resources that a Corda node needs to operate in order to ensure the transactional consistency of the ledger. These critical resources include:

  • Connection to a database.
  • Connection to Artemis Broker for P2P communication.
  • Connection to Artemis Broker for RPC communication.

Should any of those critical resources becomes unavailable, the Corda node will get into an unstable state and, as a safety precaution, it will shut itself down, reporting the cause as an error message to its log file.

Once all the critical resources the node relies on are available again, it is safe for the node operator to restart the node for normal operation.