Command Executor

The engine behind the OT-Node application.


The Command Executor is a component of the ot-node implementation which uses an approach similar to the event sourcing pattern. Essentially it allows developers to organize functionalities (code) in "commands" which can be executed in sequence to implement the protocol features, as well as enable system recovery in case of the node stopping or restarting for some reason.


The Command Executor splits business logic into commands. A command is a general abstraction with many features that can be enabled. The Command interface is described in the table below.
async execute()
Executes command and produces zero or more events
async recover(command, err)
Recover system from failure
async expired(command)
Execute strategy when the command is too late
async retryFinished(command)
This method is executed when retry command counter reaches 0
Packs data for the database
Unpacks data from the database
continueSequence(data, sequence, opts)
Makes command from the sequence and continues the execution
async handleError(operationId, errorMessage, errorName, markFailed)
Error handler for command. If an error pops up during the execution of a command, operation status is set to FAILED with the appropriate operation error message. List of operation errors can be found in the */src/constants/constants.js file.
Table 1.1 Command interface
Creating a command is done by extending this abstract class called simply Command, meaning it inherits the default behaviour of all the methods and can override them with specific behaviour.
The core command method is the execute method. This method executes the code of the command and returns one of the three results:
  • this.continueSequence(data,sequence,opts) - A list of commands taken from the execution context that will be executed after the current command is finished successfully.
    • Returns Command.empty() if the current command is the last one in the sequence.
  • Command.repeat() - Command object with the repeat flag set to true. That means that the command will be executed once again.
  • Command.retry() - Command object with the retry flag set to true. That means that the command will be executed again and the retried counter will be decreased.
Command data describes everything that is related to the specific command. This is described in the table below.
Command id (uuid)
Command name (example: helloCommand)
Command data needed for execution
Time in milliseconds when the command is ready for execution
Initial delay in milliseconds for command execution
Time in milliseconds when the command has started
Future time in milliseconds until the command needs to be executed
If the command is repeatable (repeat=true), this is the interval time in milliseconds
Command status:
  • Failed
  • Expired
  • Started
  • Pending
  • Completed
  • Repeating
Proprietary message for the command. This is useful if the command has failed
Command can have its parent. This is the parent command ID
Command can be transactional or not
If the command fails this is the number of times the command can retry to execute again
Command can carry information about the future commands that can be executed after successful completion (chain of commands)
Table 1.2 Command data parameters

Command Executor and dependency injection

Command executor is initialized on ot-node start. Commands are stored in the */src/commands directory. Commands will be injected into Awilix automatically. The naming convention of the command is in camel case and the name of the file where the command is described by using slashes(kebab case). In the following chapter we will create a simple command called PublishStartedCommand.


Let’s create a simple PublishStartedCommand and call it inhandleHttpApiPublishRequest controller method that handles the asset publishing request.
import Command from "./command.js";
class PublishStartedCommand extends Command {
async execute(command){
const {name} = command.data;
this.logger.info(`Hello from ${name}`);
return this.continueSequence(
{ ...command.data, retry: undefined, period: undefined},
default(map) {
const command = {
name: 'publishStartedCommand',
transactional: false
return command;
export default PublishStartedCommand;
PublishStartedCommand will be called before the passed assertion is validated and propagated to the network.
async handleHttpApiPublishRequest(req, res) {
Code responsible for creating operation record and notifying the client
that the operation started is intentionaly left out for simplicity reasons.
You can see the whole implementaton here:
const commandData = {
name: 'publishController',
//adding the publishStartedCommand to the command sequence
const commandSequence = [
await this.commandExecutor.add({
name: commandSequence[0],
sequence: commandSequence.slice(1),
delay: 0,
period: 5000,
retries: 3,
data: commandData,
transactional: false,
} catch (error) {
... error handling
This is the simplest command that logs the name given in the data parameter. After this command is finished, command executor continues with the next command in the sequence, in this case it continues with the execution of validateAssertionCommand and networkPublishCommand.
If we want to return some new command or list of commands, the return statement will look like in this
return {
commands: [
name: 'someCommand',
data: {
param: 'value'
transactional: false
In order to make this new command repetitive and add a delay for example, we would add these parameters to the command. The code snippet will look like in this
return {
commands: [
name: 'someCommand',
data: {
param: 'value'
delay: 10000,
period: 5000,
transactional: false

Command Executor API

Command Executor API is simple and it looks like this:
* Initialize executor
* @returns {Promise<void>}
async init() {...}
* Starts the command executor
* @return {Promise<void>}
async start() {...}
* Adds single command to queue
* @param addCommand
* @param addDelay
* @param insert
async add(addCommand, addDelay = 0, insert = true) {...}
* Replays pending commands from the database
* @returns {Promise<void>}
async replay() {...}
One of the key methods of the API is add() which is responsible for adding new commands to the array of commands for command-executor to carry out. Simple call would look like this code snippet:
await this.commandExecutor.add({
name: 'nameOfTheCommand',
transactional: false
The start() command starts the executor and all the repetitive commands listed in the constants.PERMANENT_COMMANDS. Those commands will be executed permanently throughout the node execution.

Need help with command executor?

Jump into our Discord and someone from the OriginTrail Community of developers will gladly help!