Run operations once after deployment - just like you do it with migrations!
Take your CI/CD to the next Level with One-Time Operations for Laravel! 🚀
Create specific classes for a one-time usage, that can be executed automatically after each deployment. Same as migrations they get processed once and then never again. Perfect for seeding or updating some data instantly after some database changes or feature updates.
This package is for you if...
- you regularly need to update specific data after you deployed new code
- you often execute jobs just only one single time after a deployment
- you sometimes forget to execute that one specific job and stuff gets crazy
- your code gets cluttered with jobs, that are not being used anymore
- your co-workers always need to be reminded to execute that one job after some database changes
- you often seed or process data in a migration file (which is a big no-no!)
Require this package with composer:
composer require timokoerber/laravel-one-time-operations
Create the required table in your database:
php artisan migrate
Now you're all set!
php artisan operations:make <operation_name> // create new operation file
php artisan operations:make <operation_name> -e|--essential // create file without any attributes
php artisan operations:process // process all new operation files
php artisan operations:process --sync // force synchronous execution
php artisan operations:process --async // force asynchronous execution
php artisan operations:process --test // dont flag operations as processed
php artisan operations:process --isolated // run command isolated
php artisan operations:process --queue=<name> // force queue, that the job will be dispatched to
php artisan operations:process --tag=<tagname> // only process operations, that have the given tag
php artisan operations:process <operation_name> // re-run one specific operation
php artisan operations:show // show all operations
php artisan operations:show pending // show pending operations
php artisan operations:show processed // show processed operations
php artisan operations:show disposed // show disposed operations
php artisan operations:show pending processed disposed // use multiple filters
The One-Time Operations work exactly like Laravel Migrations. Just process the operations after your code was deployed and the migrations were migrated. You can make it part of your deployment script like this:
...
- php artisan migrate
- php artisan operations:process
...
By default, the following elements will be created in your project:
- the table
operations
in your database - the directory
operations
in your project root directory
If you want to use a different settings just publish and edit the config file:
php artisan vendor:publish --provider="TimoKoerber\LaravelOneTimeOperations\Providers\OneTimeOperationsServiceProvider"
This will create the file config/one-time-operations.php
with the following content.
// config/one-time-operation.php
return [
'directory' => 'operations',
'table' => 'operations',
];
Make changes as you like.
To create a new operation file execute the following command:
php artisan operations:make AwesomeOperation
This will create a file like operations/XXXX_XX_XX_XXXXXX_awesome_operation.php
with the following content.
<?php
// operations/XXXX_XX_XX_XXXXXX_awesome_operation.php
use TimoKoerber\LaravelOneTimeOperations\OneTimeOperation;
return new class extends OneTimeOperation
{
/**
* Determine if the operation is being processed asynchronously.
*/
protected bool $async = true;
/**
* The queue that the job will be dispatched to.
*/
protected string $queue = 'default';
/**
* A tag name, that this operation can be filtered by.
*/
protected ?string $tag = null;
/**
* Process the operation.
*/
public function process(): void
{
//
}
};
Provide your code in the process()
method, for example:
// operations/XXXX_XX_XX_XXXXXX_awesome_operation.php
public function process(): void
{
User::where('active', 1)->update(['status' => 'awesome']) // make active users awesome
}
By default, the operation is being processed asynchronously (based on your configuration) by dispatching the job OneTimeOperationProcessJob
.
By default, the operation is being dispatched to the default
queue of your project. Change the $queue
as you wish.
You can also execute the code synchronously by setting the $async
flag to false
.
(this is only recommended for small operations, since the processing of these operations should be part of the deployment process)
Hint: If you use synchronous processing, the $queue
attribute will be ignored (duh!).
If you don't need all the available attributes for your operation, you can create a cleaner operation file with the --essential
or -e
option:
php artisan operations:make AwesomeOperation --essential
php artisan operations:make AwesomeOperation -e
You can provide a custom class layout in /stubs/one-time-operation.stub
, which will be used to create a new operation file.
Use the following call to process all new operation files.
php artisan operations:process
Your code will be executed, and you will find all the processed operations in the operations
table:
id | name | dispatched | processed_at |
---|---|---|---|
1 | XXXX_XX_XX_XXXXXX_awesome_operation | async | 2015-10-21 07:28:00 |
After that, this operation will not be processed anymore.
For each operation a OneTimeOperationProcessJob
is being dispatched,
either with dispatch()
oder dispatchSync()
based on the $async
attribute in the operation file.
By providing the --sync
or --async
option with the operations:process
command, you can force a synchronously/asynchronously execution and ignore the attribute:
php artisan operations:process --async // force dispatch()
php artisan operations:process --sync // force dispatchSync()
Hint! If operation:process
is part of your deployment process, it is not recommended to process the operations synchronously,
since an error in your operation could make your whole deployment fail.
You can provide the --queue
option in the artisan call. The given queue will be used for all operations, ignoring the $queue
attribute in the class.
php artisan operations:process --queue=redis // force redis queue
If you work with a Multi-Server Architecture you can use --isolated
option to make sure to only run one instance of the command (Laravel Isolatable Commands).
php artisan operations:process --isolated
You can provide the $tag
attribute in your operation file:
<?php
// operations/XXXX_XX_XX_XXXXXX_awesome_operation.php
protected ?string $tag = "awesome";
};
That way you can filter operations with this specific tag when processing the operations:
php artisan operations:process --tag=awesome // run only operations with "awesome" tag
This is quite usefull if, for example, you want to process some of your operations before and some after the migrations:
- php artisan operations:process --tag=before-migrations
- php artisan migrate
- php artisan operations:process
You can also provide multiple tags:
php artisan operations:process --tag=awesome --tag=foobar // run only operations with "awesome" or "foobar" tag
Hint! operations:process
(without tags) still processes all operations, even if they have a tag.
If something went wrong (or if you just feel like it), you can process an operation again by providing the name of the operation as parameter in operations:process
.
php artisan operations:process XXXX_XX_XX_XXXXXX_awesome_operation
You might want to test your code a couple of times before flagging the operation as "processed". Provide the --test
flag to run the command again and again.
php artisan operations:process --test
So you don't have to check the database or the directory for the existing operations,
you can show a list with operations:show
.
Filter the list with the available filters pending
, processed
and disposed
.
pending
- Operations, that have not been processed yetprocessed
- Operations, that have been processeddisposed
- Operations, that have been processed and the files were already deleted
php artisan operations:show pending // show only pending operations
php artisan operations:show pending disposed // show only pending and disposed operations
The whole idea of this package is, that you can dispose the operations once they were executed, so your project won't be cluttered with files and code, you won't be using anymore.
So you just need to delete the files from your repository
The deleted operations will be shown as DISPOSED
when you call operations:show
, so you still have a history on all the processed operations.
composer test
Copyright © Timo Körber | www.timokoerber.com
"One-Time Operations for Laravel" is open-sourced software licensed under the MIT license.