Farm Queue¶
Omniverse Farm Queue is responsible for receiving and collecting tasks from Users, as well as offering processing Agents with the information they need in order to execute them.
Installation¶
Queue can be installed from the Omniverse Launcher, by navigating to the Exchange header menu, then selecting Queue from the Apps section of the left-hand sidebar.
After first installing Queue, the Install button will then allow you to launch it and send you on your journey to automation.
Interface Overview¶
# |
Button Name |
Action |
|---|---|---|
1 |
Copy Queue URL |
Copies the selected Queue URL in order to paste it in the Agent configuration UI. |
2 |
Quit |
Shuts down the Queue background process. Once closed, no new task can be submitted, and Agents currently executing tasks will no longer be able to report their progress, or pick new tasks. |
Usage¶
Once Queue is launched, tasks can be submitted to it and their progress can be monitored via a web interface by clicking the Open Dashboard button.
The very first time Queue is installed, the web interface will not display any task history or report the progress of any Agent since no content has been submitted for automation yet. Let’s fix this by submitting our first task!
Omniverse Create comes bundled with the ability to submit render jobs to Queue, making it easy to set up a cluster for distributed rendering. In the next steps, we will be using it to submit a list of frames to render, a task that an Agent will pick up and process in the next steps.
After installing and starting Omniverse Create from the Launcher, open a scene a from the menu bar at the top of the screen and select Rendering > Movie Capture. Then, from the Movie Capture extension:
Make sure the scene you wish to render and the assets it contains are stored in a location that will be accessible to the environment where Agents will run.
This can be a local drive if you intend to run the Agents on your own machine, or a shared network drive or other federated storage if Agents will run on a separate environment.
This can also be a Nucleus server, for collaborative workflows and document sharing purposes.
Configure the render settings to your liking
In the settings section, make sure the the localhost Queue (or any other desired Queue available on the network) is selected.
Optionally, add a description to the task so you can more easily identify it later. This can be useful once your tasks start scaling, or if you share resources with others. Information you may want to capture include:
Which renderer you used to submit lower-scale render samples
The resolution of your test renders, for rapid iterations over different lighting options
etc.
Once you are satisfied with your settings, submit a task to Queue by clicking the Submit to Queue button at the bottom of the Movie Capture window.
A few seconds after the task is submitted, the Queue dashboard should automatically update to show that the task has been created, and that its status is set to Submitted while waiting for an Agent to pick it up.
In the next section, we will configure an Agent to execute this task, and produce stills from our scene.
In this guide we will be using Agents to produce images from a scene for illustrative purposes, although it is worth noting that Agents are not limited to rendering tasks, and can execute any type of work. This may include typical batch processing workflows such as converting assets from one format to another, generating animation caches, running simulations or training machine learning models.