Create Queue

Create a new queue to define the resources required for AI task execution and assign the responsible operators.

Prerequisites

• You have successfully created a Kubernetes cluster. For details, see [Create Cluster](CCE/Operation guide/Cluster management/Create cluster.md).
• Ensure CCE AI Job Scheduler and CCE Deep Learning Frameworks Operator components are installed successfully; otherwise, cloud-native AI functionalities will be unavailable.

Limitations

• Before assigning GPU resources to the queue, ensure that the CCE GPU Manager component is installed.

Operation steps

1. Sign in to the Baidu AI Cloud official website and enter the management console.
2. Go to Product Services - Cloud Native - Cloud Container Engine (CCE) to access the CCE management console.
3. Click Cluster Management - Cluster List in the left navigation pane.
4. Click on the target cluster name in the Cluster List page to navigate to the cluster management page.
5. On the Cluster Management page, click Cloud-Native AI - Queue Management.
6. On the Queue Management page, click Create Queue.
7. On the Create Queue page, complete the configuration of basic queue information:

• Queue Name: Customize the name for the queue. It must be between 1-63 characters, consist only of digits, lowercase letters, and “-”, and must start and end with a letter or digit.
• Resource Overcommitment: If the queue lacks sufficient resources, tasks can still proceed by being overcommitted in the cluster. High-priority tasks in the cluster can preempt resources being used by overcommitted tasks.
• Remarks: Add custom remarks for the queue, up to a maximum of 300 characters.

8. Complete the configuration of queue quota information:

• CPU: Specify the CPU resource size for the queue. The value must be greater than 0 and can include up to two decimal places.
• Memory: Indicate the memory resource size for the queue. The value must be greater than 0 and can include up to two decimal places. Units can be switched between Gi and Mi.
• Exclusive GPU: Define GPU resources allocated as entire cards for the queue. The value must be a positive integer.
• Shared GPU: Define GPU resources allocated by memory size for the queue. The value must be a positive integer.
• Exclusive NPU: Define NPU resources allocated as entire cards for the queue. The value must be a positive integer.
• Shared NPU: Specify NPU resources allocated by memory size for the queue. The input must be a positive integer.

Note: To specify GPU and NPU resources, you must first install the CCE GPU Manager and CCE NPU Manager components

9. Complete the resource object configuration.

• Object type: Specify the type of resource objects available to the queue. Currently, only "User" is supported.
• Object selection: Specify the IAM users who can utilize the queue’s resources. Multiple selections are supported.

10. Click the OK button to complete creating the queue.