ClusterTieredStorageConfig
Defines the configuration for managed tier checkpointing in a HyperPod cluster. Managed tier checkpointing uses multiple storage tiers, including cluster CPU memory, to provide faster checkpoint operations and improved fault tolerance for large-scale model training. The system automatically saves checkpoints at high frequency to memory and periodically persists them to durable storage, like Amazon S3.
Types
Properties
The percentage (int) of cluster memory to allocate for checkpointing.
Specifies whether managed tier checkpointing is enabled or disabled for the HyperPod cluster. When set to Enable, the system installs a memory management daemon that provides disaggregated memory as a service for checkpoint storage. When set to Disable, the feature is turned off and the memory management daemon is removed from the cluster.