Creating a queue in AWS PCS - AWS PCS
PrerequisitesTo create a queue in AWS PCS

Creating a queue in AWS PCS

This topic provides an overview of available options and describes what to consider when you create a queue in AWS PCS.

Note

You can configure custom Slurm settings on queues to implement partition-specific scheduling policies and resource management. For more information, see Configuring custom Slurm settings in AWS PCS.

Prerequisites

  • An AWS PCS cluster - queues can only be created in association with a specific AWS PCS cluster.

  • One or more AWS PCS compute node groups - a queue must be associated with at least one AWS PCS compute node group.

To create a queue in AWS PCS

You can create a queue using the AWS Management Console or the AWS CLI.

AWS Management Console
To create a queue using the console

  1. Open the AWS PCS console.

  2. Select the cluster for the queue. Navigate to Queues and choose Create queue.

  3. In the Queue configuration section, provide the following values:

    1. Queue name – A name for your queue. The name can contain only alphanumeric characters (case-sensitive) and hyphens. It must start with an alphabetic character and can't be longer than 25 characters. The name must be unique within the cluster.

    2. Compute node groups – Select 1 or more compute node groups to service this queue. A compute node group can be associated with more than 1 queue.

  4. (Optional) In the Additional scheduler settings section, you can add parameter name and value pairs to configure additional Slurm settings. For a complete list of supported parameters, see Custom Slurm settings for AWS PCS queues.

  5. (Optional) Under Tags, add any tags to your AWS PCS queue

  6. Choose Create queue. The Status field will show Creating while AWS PCS creates the queue. Queue creation can take several minutes.

Recommended next step

  • Submit a job to your new queue.

AWS CLI
To create a queue using AWS CLI

Use the following command to create your queue. Make the following replacements:

  1. Replace region-code with the AWS Region of the cluster. For example, us-east-1.

  2. Replace my-queue with the name for your queue. The name can contain only alphanumeric characters (case-sensitive) and hyphens. It must start with an alphabetic character and can't be longer than 25 characters. The name must be unique within the cluster.

  3. Replace my-cluster with the name or ID of your cluster.

  4. Replace compute-node-group-id with the ID of the compute node group to service the queue. For example, pcs_abcdef12345.

    Note

    When you create a queue, you must provide the ID of the compute node group and not its name.

aws pcs create-queue --region region-code \
    --queue-name my-queue \
    --cluster-identifier my-cluster \
    --compute-node-group-configurations \
    computeNodeGroupId=compute-node-group-id
Example – Creating a queue with custom Slurm settings
aws pcs create-queue --region region-code \
    --queue-name my-queue \
    --cluster-identifier my-cluster \
    --compute-node-group-configurations \
    computeNodeGroupId=compute-node-group-id \
    --slurm-configuration \
    'slurmCustomSettings=[{parameterName=Default,parameterValue=YES}]'

For more information, see Custom Slurm settings for AWS PCS queues.

It can take several minutes to create the queue. You can query the status of your queue with the following command. You won’t be able to submit jobs to the queue until its status reaches ACTIVE.

aws pcs get-queue --region region-code \
    --cluster-identifier my-cluster \
    --queue-identifier my-queue
Recommended next step

  • Submit a job to your new queue