Important

You are browsing documentation for version 6.0 of OroCommerce, supported until 2028. Read the documentation for the latest LTS version to get up-to-date information.

See our Release Process documentation for more information on the currently supported and upcoming releases.

Session Storage

By default, the Oro application is configured to store sessions in files. If multiple servers serve your application, you must use a database to make sessions work across different servers. The recommended database for best performance is Redis. See Configure Redis Servers for more details.

Session Locking Impact on Application Availability

The Oro application requires shared session storage in any distributed environment (more than one web node). By default, session data is locked to prevent race conditions and ensure data consistency.

This works well for consecutive requests (classic web browsing), but causes problems when multiple parallel requests run within the same session. A common B2B case is real-time price and inventory checks against a back-end ERP, which may respond slowly. To keep these checks from blocking the interface, they typically run in parallel over AJAX.

In production, this can critically affect availability: each parallel request hits the session lock and queues behind the others. With many concurrent users generating dozens of such requests, ERP performance directly limits Oro availability — and a slow ERP can overflow the request queue.

There are a few options to overcome availability issues for this kind of scenario:

• Use stateless endpoint without session initialization. Such an approach has a significant downside as it will allow accessing data without authentication stored in the session.

• Close the session before accessing any 3rd party system (recommended approach):

  public function myAction(Request $request)
  {
      $session = $request->getSession();
      if (null !== $session && $session->isStarted()) {
          $session->save();
      }
  
      // do controller work here
  }