# Deadline Cloud queues
A queue is a farm resource that manages and processes jobs.
To work with queues, you should already have a monitor and farm set up.
**Topics**
+ [Create a queue](#create-queue)
+ [Create a queue environment](create-queue-environment.md)
+ [Associate a queue and fleet](associate-a-queue-and-fleet.md)
## Create a queue
1. From the [Deadline Cloud console](https://console.aws.amazon.com/deadlinecloud/home) dashboard, select the farm that you want to create a queue for.
1. Alternatively, in the left side panel choose **Farms and other resources**, then select the farm you want to create a queue for.
1. In the **Queues** tab, choose **Create queue**.
1. Enter a name for your queue.
1. For **Description**, enter the queue description. A description helps you identify your queue's purpose.
1. For **Job attachments**, you can either create a new Amazon S3 bucket or choose an existing Amazon S3 bucket.
1. To create a new Amazon S3 bucket
1. Select **Create new job bucket**.
1. Enter a name for the bucket. We recommend naming the bucket `deadlinecloud-job-attachments-[MONITORNAME]`.
1. Enter a **Root prefix** to define or change your queue's root location.
1. To choose an existing Amazon S3 bucket
1. Select **Choose an existing S3 bucket** > **Browse S3**.
1. Select the S3 bucket for your queue from the list of available buckets.
1. (Optional) To associate your queue with a customer-managed fleet, select **Enable association with customer-managed fleets**.
1. If you enable association with customer-managed fleets, you must complete the following steps.
**Important**
We strongly recommend specifying users and groups for run-as functionality. If you don't, it will degrade your farm’s security posture because the jobs can then do everything the worker's agent can do. For more information about the potential security risks, see [Run jobs as users and groups](security-best-practices.md#job-run-as-user).
1. For Run as user:
To provide credentials for the queue's jobs, select **Queue-configured user**.
Or, to opt out of setting your own credentials and run jobs as the worker agent user, select **Worker agent user**.
1. (Optional) For Run as user credentials, enter a user name and group name to provide credentials for the queue's jobs.
If you are using a Windows fleet, you must create an AWS Secrets Manager secret that contains the password for the Run as user. If you don't have an existing secret with the password, choose **Create secret** to open the Secrets Manager console to create a secret. For more information, see [Manage access to Windows job user secrets](https://docs.aws.amazon.com/deadline-cloud/latest/developerguide/manage-access-windows-secrets.html) in the *Deadline Cloud Developer Guide*.
1. Requiring a budget helps manage costs for your queue. Select either **Don't require a budget** or **Require a budget**.
1. Your queue requires permission to access Amazon S3 on your behalf. You can create a new service role or use an existing service role. If you don't have an existing service role, create and use a new service role.
1. To use an existing service role, select **Choose a service role**, and then select a role from the dropdown.
1. To create a new service role, select **Create and use a new service role**, and then enter a role name and description.
1. (Optional) To add environment variables for the queue environment, choose **Add new environment variable**, and then enter a name and value for each variable you add.
1. (Optional) Choose **Add new tag** to add one or more tags to your queue.
1. To create a default conda queue environment, keep the checkbox selected. To learn more about queue environments, see [ Create a queue environment](create-queue-environment.md). If you are creating a queue for a customer-managed fleet, clear the checkbox.
1. Choose **Create queue**.
# Create a queue environment
A queue environment is a set of environment variables and commands that set up fleet workers. You can use queue environments to provide software applications, environment variables, and other resources to jobs in the queue.
When you create a queue, you have the option of creating a default conda queue environment. This environment provides service-managed fleets access to packages for partner DCC applications and renderers. The default environment For more information, see [Default conda queue environment](#conda-queue-environment).
You can add queue environments using the console, or by editing the json or YAML template directly. This procedure describes how to create an environment with the console.
1. To add a queue environment to a queue, navigate to the queue and select the **Queue environments tab**.
1. Choose **Actions**, then **Create new with form**.
1. Enter a name and description for the queue environment.
1. Choose **Add new environment variable**, and then enter a name and value for each variable you add.
1. (Optional) Enter a priority for the queue environment. The priority indicates the order that this queue environment will run on the worker. Higher priority queue environments will run first.
1. Choose **Create queue environment**.
## Default conda queue environment
When you create a queue associated with a service-managed fleet, you have the option of adding a default queue environment that supports [https://docs.conda.io/en/latest/](https://docs.conda.io/en/latest/) to download and install packages in a virtual environment for your jobs.
If you add a default queue environment with the Deadline Cloud [console](https://console.aws.amazon.com/deadlinecloud/home), the environment is created for you. If you add a queue another way, such as the AWS CLI or with CloudFormation, you'll need to create the queue environment yourself. To ensure you have the correct contents for the environment, you can refer to queue environment template YAML files on GitHub. For the contents of the default queue environment, see the [default queue environment YAML file](https://github.com/aws-deadline/deadline-cloud-samples/blob/mainline/queue_environments/conda_queue_env_from_console.yaml) on GitHub.
There are other [queue environment templates](https://github.com/aws-deadline/deadline-cloud-samples/tree/mainline/queue_environments#the-sample-queue-environments) available on GitHub that you can use as a starting point for your own needs.
Conda provides packages from *channels*. A channel is a location where packages are stored. Deadline Cloud provides a channel, `deadline-cloud`, that hosts conda packages that support partner DCC applications and renderers. Select each tab below to view the available packages for Linux or Windows.
------
#### [ Linux ]
+ Autodesk Arnold for Cinema 4D
+ `cinema4d-c4dtoa=2025`
+ Autodesk Arnold for Maya
+ `maya-mtoa=2024.5.3`
+ `maya-mtoa=2025.5.4`
+ `maya-mtoa=2026.5.5`
+ Autodesk Maya
+ `maya=2024`
+ `maya=2025`
+ `maya=2026`
+ `maya-openjd`
+ Autodesk VRED
+ `vredcore=2025`
+ `vredcore=2026`
+ Blender
+ `blender=3.6`
+ `blender=4.2`
+ `blender=4.5`
+ `blender=5.0`
+ `blender-openjd`
+ Chaos V-Ray for Maya
+ `maya-vray=2025.7`
+ `maya-vray=2026.7`
+ Foundry Nuke
+ `nuke=15`
+ `nuke=16`
+ `nuke=17`
+ `nuke-openjd`
+ Maxon Cinema 4D
+ `cinema4d=2025`
+ `cinema4d=2026`
+ `cinema4d-openjd`
+ Maxon Redshift for Maya
+ `maya-redshift=2025.4`
+ `maya-redshift=2026.2`
+ SideFX Houdini
+ `houdini=19.5`
+ `houdini=20.0`
+ `houdini=20.5`
+ `houdini=21.0`
+ `houdini-openjd`
------
#### [ Windows ]
+ Adobe After Effects
+ `aftereffects=24.6`
+ `aftereffects=25.1`
+ `aftereffects=25.2`
+ `aftereffects=25.6`
+ `aftereffects=26.0`
+ Autodesk Arnold for Cinema 4D
+ `cinema4d-c4dtoa=2025`
+ `cinema4d-c4dtoa=2026`
+ Maxon Cinema 4D
+ `cinema4d=2024`
+ `cinema4d=2025`
+ `cinema4d=2026`
+ `cinema4d-openjd`
+ Unreal Engine
+ `unrealengine=5.4`
+ `unrealengine=5.5`
+ `unrealengine=5.6`
+ `unrealengine=5.7`
+ `unrealengine-openjd`
------
**Note**
For **Cinema 4D**, the Linux conda package does not support substance 3D materials. Jobs with this material fail with one of the following errors:
```
Commandline: ./modules/io_substance/source/substance_framework/src/details/detailsengine.cpp:794: SubstanceAir::Details::Engine::Context::Context(SubstanceAir::Details::Engine&, SubstanceAir::RenderCallbacks*): Assertion `res==0' failed.
```
```
/home/job-user/.conda/envs//Lib/deadline/cinema4d_adaptor/Cinema4DAdaptor/adaptor.sh: line 44: 10832 Segmentation fault (core dumped) $C4DEXE ${ARGS[*]}
```
We recommend that you submit jobs with substance materials to Windows instead.
In Cinema 4D 2025.3.3 on Linux, globalized asset paths can cause segmentation faults. Therefore, the Linux conda package contains Cinema 4D 2025.3.1 with Redshift 2025.6.0 instead. If you need features or bug fixes from Cinema 4D 2025.3.3, we recommend two options: upgrade to Cinema 4D 2026 or submit those jobs to Windows instead.
For **Cinema 4D OpenJD,** to prevent any timeout issues, we recommend you set task run timeouts to double their expected render time, instead of using the default 2 day timeout.
When you submit a job to a queue with the default conda environment, the environment adds two parameters to the job. These parameters specify the conda packages and channels to use to configure the job's environment before tasks are processed. The parameters are:
+ `CondaPackages` – a space-separated list of [package match specifications](https://docs.conda.io/projects/conda-build/en/stable/resources/package-spec.html#package-match-specifications), such as `blender=3.6` or `numpy>1.22`. The default is empty to skip creating a virtual environment.
+ `CondaChannels` – a space separated list of [conda channels](https://docs.conda.io/projects/conda/en/latest/user-guide/concepts/channels.html) such as `deadline-cloud`, `conda-forge`, or `s3://amzn-s3-demo-bucket/conda/channel`. The default is `deadline-cloud`, a channel available to service-managed fleets that provides partner DCC applications and renderers.
When you use an integrated submitter to send a job to Deadline Cloud from your DCC, the submitter populates the value of the `CondaPackages` parameter based on the DCC application and submitter. For example, if you are using Blender the `CondaPackage` parameter is set to `blender=3.6.* blender-openjd=0.4.*`.
We recommend you pin any submissions to only the versions listed in the table above, for example blender=3.6. Pinning to the major.minor version is recommended because patch releases affect the available packages. For example, when we release Blender 3.6.17, we will no longer distribute Blender 3.6.16. Any submissions pinned to blender=3.6.16 will fail. If you pin to blender=3.6, then you will get the latest distributed patch version and jobs will not be impacted. By default, the DCC submitters pin to the current versions listed in the table above, excluding the patch number, such as blender=3.6.
# Associate a queue and fleet
To process jobs, you must associate a queue with a fleet. You can associate a single fleet with multiple queues and a single queue with multiple fleets. When you associate a fleet with multiple queues, it divides its workers evenly among them. Similarly, when you associate a queue with multiple fleets, it distributes jobs evenly across those fleets.
**Note**
To use wait and save, we recommend you associate your queue only with a fleet that uses wait and save instance types. If you associate your queue with more than one fleet, and any of those fleets use spot or on-demand instance types, your fleet might not process your jobs with wait and save instances.
To associate an existing queue with an existing fleet, complete the following steps:
1. From your Deadline Cloud farm, select the **Queue** you want to associate with a fleet. The queue displays.
1. To select a fleet to associate with your queue, choose **Associate fleets**.
1. Choose the **Select fleets** dropdown. A list of available fleets displays.
1. From the list of available fleets, select the **checkbox** next to the fleet or fleets you want to associate with your queue.
1. Choose **Associate**. The fleet association status should now be **Active**.
## Stop a queue fleet association
To stop a queue fleet association, complete the following steps:
1. From your queue, select the **Associated fleets** tab.
1. Select the checkbox for the fleet you want to stop associating with the queue.
1. From the Actions dropdown, select **Eventual stop** or **Immediate stop**.
To finish processing jobs before the association stops, select Eventual stop. To immediately stop processing jobs, select Immediate stop.
1. In the confirmation window, enter **confirm** and then choose **Stop**.
1. (Optional) To disassociate the fleet from the queue, complete the following steps:
1. Wait for the association status to change to **Stopped**.
1. After the association has stopped, if you haven't already, select the checkbox for the fleet.
1. From the Actions dropdown, select **Disassociate fleet**.
1. In the confirmation window, choose **Disassociate**.
## Reactivate a queue fleet association
To reactivate a queue fleet association, complete the following steps:
1. From your queue, select the **Associated fleets** tab.
1. Select the checkbox for the fleet you want to reactivate the queue fleet association.
1. From the Actions dropdown, choose **Start**. The association status changes to Active.