Amazon SageMaker HyperPod

aws Big data clusteraws-sagemaker-hyperpod

Resilient accelerator cluster for training and serving foundation models.

Provider: aws
Kind: Big data cluster
Category: Compute
Official docs: https://docs.aws.amazon.com/sagemaker/latest/dg/sagemaker-hyperpod.html

Properties · 2 set by this component

property	type	values	required	kind default	this component
autoResume · component property	boolean	—	no	—	inherited
autoscaling	boolean	—	no	false	inherited
engine	enum	spark \| hadoop \| presto \| flink \| hpc \| ray \| dask \| kubernetes-batch	no	—	inherited
managed	boolean	—	no	false	▸ true
orchestrator · component property	enum	slurm \| eks	no	—	inherited
runtimeVersion	string	—	no	—	inherited
sharedFileSystem · component property	enum	fsx-lustre \| none	no	—	inherited
spotInstances	boolean	—	no	false	▸ false
topologyAwareScheduling · component property	boolean	—	no	—	inherited
workerNodes	integer	—	no	—	inherited
zoneRedundant	boolean	—	no	—	inherited

Rules that apply · 1

Every rule whose selector covers this component's kind or its category. Whether a rule fires depends on the design: its properties, its edges and its groups.

Stateful node without zone redundancyerror

stateful-node-without-zone-redundancy

A stateful compute node pinned to exactly one availability zone disappears with that zone, and its state goes with it. Nodes inside a cell are exempt, because there the cell is deliberately single-zone and redundancy comes from the other cells.

— all of these hold:
— the node's `stateless` is false
— the node sits inside exactly 1 availability-zone group
— this does not hold:
— the node sits inside at least one cell group

Use it in a design

nodes:
  - { id: "sagemaker-hyperpod", component: "aws-sagemaker-hyperpod" }

Other Big data cluster products · 25

Elastic High Performance Computingalibaba-e-hpcalibabacloudbig-data-cluster E-MapReducealibaba-e-mapreducealibabacloudbig-data-cluster AWS Clean Roomsaws-clean-roomsawsbig-data-cluster Amazon EMRaws-emrawsbig-data-cluster Amazon EMR Serverlessaws-emr-serverlessawsbig-data-cluster AWS Parallel Computing Serviceaws-pcsawsbig-data-cluster Azure CycleCloudazure-cyclecloudazurebig-data-cluster Azure Databricksazure-databricksazurebig-data-cluster Azure HDInsightazure-hdinsightazurebig-data-cluster Armadaarmadabig-data-cluster Generic big data clusterbig-data-clusterbig-data-cluster Daskdaskbig-data-cluster Databricksdatabricksbig-data-cluster Apache Hadoop YARNhadoop-yarnbig-data-cluster Rayraybig-data-cluster Slurmslurmbig-data-cluster Apache Spark clusterspark-clusterbig-data-cluster Volcanovolcanobig-data-cluster Cluster Directorgoogle-cluster-directorgcpbig-data-cluster Dataprocgoogle-dataprocgcpbig-data-cluster Managed Service for Apache Spark (serverless)google-dataproc-serverlessgcpbig-data-cluster IBM Analytics Engineibm-analytics-engineibmbig-data-cluster OCI AI Data Platformoci-ai-data-platformocibig-data-cluster OCI Big Data Serviceoci-big-data-serviceocibig-data-cluster OCI Data Flowoci-data-flowocibig-data-cluster

Source

1# yaml-language-server: $schema=../../schema/component.schema.json
2# Amazon SageMaker HyperPod cluster: a long-running, resilient accelerator cluster for training and
3# serving foundation models. "Amazon SageMaker HyperPod" says the cluster is orchestrated either by
4# Slurm or by an Amazon EKS control plane (one EKS cluster to one HyperPod cluster inside a VPC), and
5# that HyperPod monitors the instances and replaces faulty hardware so that running jobs resume.
6schemaVersion: "1"
7type: "component"
8id: "aws-sagemaker-hyperpod"
9kind: "big-data-cluster"
10provider: "aws"
11docs: "https://docs.aws.amazon.com/sagemaker/latest/dg/sagemaker-hyperpod.html"
12defaults:
13  managed: true
14  spotInstances: false
15properties:
16  orchestrator:
17    type: "enum"
18    values: ["slurm", "eks"]
19  autoResume:
20    type: "boolean"
21  sharedFileSystem:
22    type: "enum"
23    values: ["fsx-lustre", "none"]
24  topologyAwareScheduling:
25    type: "boolean"

The same page as markdown, for an agent: /catalogue/components/aws-sagemaker-hyperpod.md