Share reserved compute resources with Anyscale machine pools
Using machine pools in an Anyscale cluster is in developer preview.
A machine pool defines a fixed-size group of provisioned compute resources. Anyscale recommends using machine pools to leverage compute reservations in your cloud provider account. Because of the limited supply of GPUs in many cloud regions, Anyscale recommends reserving GPUs and configuring machine pools to share them across your GPU workloads.
For an example, see Tutorial: Create and use an Anyscale-managed machine pool.
All machine pools require that the head node for your cluster uses an on demand virtual machine in your Anyscale cloud. The driver and Anyscale control plane must communicate with all nodes in your machine pool.
What type of machine pools are available on Anyscale?
The following table provides an overivew of Anyscale machine pool offerings:
| Type | Description |
|---|---|
| Anyscale-managed machine pools | You configure the Anyscale global resource scheduler to manage reserved compute resources in the cloud provider account where you deployed your Anyscale cloud. See Configure an Anyscale-managed machine pool. |
| Customer-managed machine pools | You install Anyscale's machine manager on your on-prem compute instances. See Customer-managed machine pools. |
How do machine pools manage priority and share compute infrastructure?
Anyscale uses a global resource scheduler to manage resources in machine pools. See What is the global resource scheduler?.