README

Coordinate execution of one-time commands

This library can be used to enqueue arbitrary commands that need to be executed on all stages, this includes:

• database migrations
• data fixtures / transformations
• one-time bug fixes
• ...

Usage

Install via composer:

composer require wwwision/command-jobs

and run pending command jobs via

vendor/bin/command-jobs run

Configuration

Upon first run, the CLI will create a config.yaml file in the root folder with the following contents:

commandResults:
  storage: filesystem
  options:
    path: /some/path/to/commandResults.yaml

Note

All configuration options allow to refer to environment variables in the form %env(SOME_VARIABLE)%

Store command results in a database

To store command results in a MySQL/MariaDB database, the following configuration can be used:

commandResults:
  storage: database
  options:
    dsn: '%env(DATABASE_URL)%'
    tableName: command_jobs_results

Note

If the database storage is configured, the doctrine/dbal package has to be installed

Options

root

By default, the run command will create a command-jobs folder in the current directory, that contains YAML files for Command Definitions, Command Jobs, Command Results and a .gitignore file that excludes Command Results from being added to VCS. The --root option allows to specify a different root directory for those three files:

vendor/bin/command-jobs run --root=/some/other/folder

stop-on-error

Commands that fail or return an exit code other than 0 lead to an error (and the run command to exit with an error code as well). By default, remaining pending Command Jobs will still be executed. It might make sense to stop processing of those jobs entirely in this case:

vendor/bin/command-jobs run --stop-on-error

default options

The run command supports default options that control verbosity and styling of the output:

Usage:
  vendor/bin/command-jobs run [options]

Options:
  -r, --root=ROOT       Root directory
      --stop-on-error   Intercept execution upon errors
  -h, --help            Display help for the given command. When no command is given display help for the list command
      --silent          Do not output any message
  -q, --quiet           Only errors are displayed. All other output is suppressed
  -V, --version         Display this application version
      --ansi|--no-ansi  Force (or disable --no-ansi) ANSI output
  -n, --no-interaction  Do not ask any interactive question
  -v|vv|vvv, --verbose  Increase the verbosity of messages: 1 for normal output, 2 for more verbose output and 3 for debug

Concepts

Command Definition

In order for commands to be enqueued/executed, they have to be explicitly defined. This prevents invalid or potentially dangerous commands to be executed automatically. Furthermore, it allows commands to have a description and to be reused. A Command Definition consists of:

• id – a unique string, that identifies it. It might make sense to use dots for namespaces to group definitions (e.g. "db.migrate.products-initial")
• description – a string, that explains the intention of a command (e.g. "Creates required product tables – this command is idempotent")
• cmd – the actual command to execute. This is represented as an array that consists of an element for the command itself and one for each argument (e.g.["some-command", "arg1", "\"arg2 with spaces\""])

The add-definition CLI command can be used to add Command Definitions:

Usage:
  vendor/bin/command-jobs add-definition [options] [--] [<id> [<description> [<cmd>...]]]

Arguments:
  id                    ID of the command definition
  description           Description of the command definition
  cmd                   The command to execute (command and arguments, separated by spaces)

Options:
  -r, --root=ROOT       Root directory
      --timeout=TIMEOUT  Timeout (in seconds) for the command to run
  -h, --help            Display help for the given command. When no command is given display help for the list command
      --silent          Do not output any message
  -q, --quiet           Only errors are displayed. All other output is suppressed
  -V, --version         Display this application version
      --ansi|--no-ansi  Force (or disable --no-ansi) ANSI output
  -n, --no-interaction  Do not ask any interactive question
  -v|vv|vvv, --verbose  Increase the verbosity of messages: 1 for normal output, 2 for more verbose output and 3 for debug

Command Job

To enqueue commands, a corresponding Command Job has to be added. Each Command Job is executed once (unless if fails) when vendor/bin/command-jobs run is invoked. A Command Job consists of:

• id – a unique string, that identifies it. This is the timestamp at the time of creation in the format YmdHis (e.g. 20250925151814)
• commandDefinitionId – the id of an existing Command Definition

The add-job CLI command can be used to add Command Jobs:

Usage:
  vendor/bin/command-jobs add-job [options] [--] [<definition>]

Arguments:
  definition            ID of the command definition

Options:
  -r, --root=ROOT       Root directory
  -h, --help            Display help for the given command. When no command is given display help for the list command
      --silent          Do not output any message
  -q, --quiet           Only errors are displayed. All other output is suppressed
  -V, --version         Display this application version
      --ansi|--no-ansi  Force (or disable --no-ansi) ANSI output
  -n, --no-interaction  Do not ask any interactive question
  -v|vv|vvv, --verbose  Increase the verbosity of messages: 1 for normal output, 2 for more verbose output and 3 for debug

Command Result

The result of any execution of a Command Jobs is stored as a Command Result. By default, the results are stored in a commandResults.yaml file.

Important

The commandResults.yaml file is specific to a single installation, it must not be added to the VCS It is therefore excluded via .gitignore file by default

A Command Result consists of:

• commandJobId – the id of the Command Jobs that was executed
• commandDefinitionId – the id of the corresponding Command Definition
• executionTime – date and time of the execution
• executionDurationInMilliseconds – execution runtime in milliseconds
• success – a boolean flag that is true if the exit code of the command was 0 and otherwise false
• output - the string output of the command – or the exception message if it failed

Contribution

Contributions in the form of issues or pull requests are highly appreciated

License

See LICENSE