Support Matrix

Primary Storage Support

HPE Ezmeral Unified Analytics Software supports the following storage platforms for primary storage:

HPE Ezmeral Unified Analytics Software	HPE Ezmeral Data Fabric
1.5.0, 1.5.2	7.6.1¹, 7.7.0, 7.8.0

¹HPE Ezmeral Data Fabric 7.6.0 is deprecated and HPE Ezmeral Data Fabric 7.6.1 is a patch release that replaces release 7.6.0, see Deprecation of Release 7.6.0.

To learn more about the primary storage support in HPE Ezmeral Unified Analytics Software, see Preparing HPE Ezmeral Data Fabric to be Primary Storage for HPE Ezmeral Unified Analytics Software.

Tools and Frameworks

HPE Ezmeral Unified Analytics Software supports the following Tools & Frameworks in the versions listed:

HPE Ezmeral Unified Analytics Software	Airflow	EzPresto	Feast	Kubeflow	Livy	MLflow	HPE MLDE	Ray	Spark Applications	Spark History Server	Spark Operator	Superset
1.5.0, 1.5.2	2.9.2	0.287	0.39.0	1.8.0	0.8.0.5	2.13.2	0.28.1	2.24.0	3.5.1	3.5.1	1.3.8.7-hpe	4.0.1
1.4.1	2.8.4	0.285	0.34.1	1.8.0	0.8.0	2.10.2	0.28.1	2.9.3	3.5.0	3.5.0	1.3.8.7-hpe	3.1.0
1.4.0	2.7.3	0.285	0.34.1	1.8.0	0.8.0	2.10.2	0.28.1	2.9.3	3.5.0	3.5.0	1.3.8.7-hpe	3.1.0
1.3.0	2.7.3	0.281	0.34.1	1.8.0	0.8.0	2.8.1	0.26.7	2.7.0	3.5.0	3.5.0	1.3.8.4-hpe	3.0.1
1.2.0	2.7.0	0.281	0.34.1	1.7.0	0.7.0	2.7.1	N/A	2.6.1	3.4.1 3.3.1	3.4.1	1.3.8.3-hpe	2.1.1
1.1.0	2.6.1	0.269	0.31.0	1.7.0	0.7.0.302	2.4.0	N/A	2.4.0	3.4.0 3.3.1	3.4.0	1.3.8.2-hpe	2.1.0
1.0.0	2.5.1	0.269	0.29.0	1.6.0	0.7.0	2.1.1	N/A	2.2.0	3.3.1	3.3.1	1.3.8.2-hpe	2.0.1

Notebook Images

The following table lists the default notebook images and their packages in HPE Ezmeral Unified Analytics Software 1.5.2:

Notebook Images	Libraries	General Packages
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-scipy:ezua-1.5.2-175c1e34`	SciPy 1.11.3 complete package list	Conda Python 23.3.1-1 JupyterLab 3.6.6 kfp 2.5.0 kfp-pipeline-spec 0.2.2 kfp-server-api 2.0.5 feast 0.33.1 mlflow 2.13.2 kserve 0.11.2 kubeflow-katib 0.16.0 kubeflow-kale 0.7.0 presto-python-client 0.8.3 ray 2.24.0 The following packages are installed in a separate Ray kernel, ray[tune] 2.24.0 ray[default] 2.24.0 ray[client] 2.24.0 ray[serve] 2.24.0
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-pytorch-full:ezua-1.5.2-175c1e34`	PyTorch (CPU) torch 2.1.0 torchvision 0.16.0 torchaudio 2.1.0 complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-pytorch-cuda-full:ezua-1.5.2-175c1e34`	PyTorch (CUDA) cuda 12.1 torch 2.1.0 torchvision 0.16.0 torchaudio 2.1.0 complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-tensorflow-full:ezua-1.5.2-175c1e34`	Tensorflow (CPU) tensorflow 2.13.0 complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-tensorflow-cuda-full:ezua-1.5.2-175c1e34`	TensorFlow (CUDA) tensorflow 2.13.0 cuda 11.8 complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-data-science:ezua-1.5.2-175c1e34`	complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/codeserver:ezua-1.5.2-175c1e34`	code-server 4.17.1 (Visual Studio Code) Python extension 2023.18.0 NOTE This is not the Python language version. complete package list

The following table lists the default notebook images and their packages in HPE Ezmeral Unified Analytics Software 1.5.0:

Notebook Images	Libraries	General Packages
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-scipy:ezua-1.5.0-d2d18b79`	SciPy 1.11.3 complete package list	Conda Python 23.3.1-1 JupyterLab 3.6.6 kfp 2.5.0 kfp-pipeline-spec 0.2.2 kfp-server-api 2.0.5 feast 0.33.1 mlflow 2.13.2 kserve 0.11.2 kubeflow-katib 0.16.0 kubeflow-kale 0.7.0 presto-python-client 0.8.3 ray 2.24.0 The following packages are installed in a separate Ray kernel, ray[tune] 2.24.0 ray[default] 2.24.0 ray[client] 2.24.0 ray[serve] 2.24.0
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-pytorch-full:ezua-1.5.0-d2d18b79`	PyTorch (CPU) torch 2.1.0 torchvision 0.16.0 torchaudio 2.1.0 complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-pytorch-cuda-full:ezua-1.5.0-d2d18b79`	PyTorch (CUDA) cuda 12.1 torch 2.1.0 torchvision 0.16.0 torchaudio 2.1.0 complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-tensorflow-full:ezua-1.5.0-d2d18b79`	Tensorflow (CPU) tensorflow 2.13.0 complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-tensorflow-cuda-full:ezua-1.5.0-d2d18b79`	TensorFlow (CUDA) tensorflow 2.13.0 cuda 11.8 complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/jupyter-data-science:ezua-1.5.0-d2d18b79`	complete package list
`gcr.io/mapr-252711/kubeflow/notebooks/codeserver:ezua-1.5.0-d2d18b79`	code-server 4.17.1 (Visual Studio Code) Python extension 2023.18.0 NOTE This is not the Python language version. complete package list

To learn more about descriptions and uses of Notebook images, see Notebook Images Overview.

HPE Ezmeral Data Fabric

The following table lists the versions of HPE Ezmeral Data Fabric that you can connect HPE Ezmeral Unified Analytics Software to externally:

HPE Ezmeral Unified Analytics Software	HPE Ezmeral Data Fabric
1.5.0, 1.5.2	7.6.1¹, 7.7.0, 7.8.0
1.4.0, 1.4.1	7.4.0, 7.5.0, 7.6.1¹, 7.7.0
1.3.0	7.4.0, 7.5.0, 7.6.1¹
1.2.0	6.2.0, 7.0.0, 7.1.0, 7.2.0, 7.3.0, 7.4.0
1.1.0	6.2.0, 7.0.0, 7.2.0
1.0.0	7.0.0, 7.2.0

¹HPE Ezmeral Data Fabric 7.6.0 is deprecated and HPE Ezmeral Data Fabric 7.6.1 is a patch release that replaces release 7.6.0, see Deprecation of Release 7.6.0.

Operating System

HPE Ezmeral Unified Analytics Software supports the following operating systems in the versions listed:

HPE Ezmeral Unified Analytics Software	RHEL Version	Rocky Version
1.5.0, 1.5.2	8.9¹ 8.8¹	8.7²
1.4.0, 1.4.1	8.9¹ 8.8¹	8.7²
1.3.0	8.8¹	8.7²
1.2.0	8.x¹	8.x²
1.1.0	8.x¹	8.x²

¹Only RHEL 8.x is supported on GPU hosts.

²There is no GPU support for Rocky 8.x, as NVIDIA does not support the GPU operator running on Rocky 8.x.

GPU Models

HPE Ezmeral Unified Analytics Software supports the following GPU models:

HPE Ezmeral Unified Analytics Software	GPU Model
1.5.0, 1.5.2	NVIDIA A100 NVIDIA A30 NVIDIA H100-NVL, H100-PCI NVIDIA A10G¹ NVIDIA L40S¹ NVIDIA V100¹
1.4.0, 1.4.1	NVIDIA A100 NVIDIA A30 NVIDIA H100-NVL, H100-PCI NVIDIA A10G¹ NVIDIA L40S¹ NVIDIA V100¹
1.3.0	NVIDIA A100 NVIDIA A30
1.2.0	NVIDIA A100
1.1.0	NVIDIA A100

¹The A10G, L40S, and V100 GPU models do not support MIG. During the installation of HPE Ezmeral Unified Analytics Software, you must specify GPU partition size as Whole and request the number of GPU instances required for the workload. These GPU models are not split into any partitions and you will get the entire physical GPU. In this configuration, applications can use only one virtual GPU at a time.

To learn more about the GPU support in HPE Ezmeral Unified Analytics Software, see GPU Support.

Upgrade Bundle Images

The following table lists the upgrade bundle images necessary for upgrading HPE Ezmeral Unified Analytics Software to the following versions:

HPE Ezmeral Unified Analytics Software	Images
1.5.2	`ezua/ezua-system-bundle:v1.5.2-0b8e1c03e-86d69a`
1.4.1	`ezua/ezua-system-bundle:v1.4.1-bffae884e`
1.4.0	`ezua/ezua-system-bundle:v1.4.0-e8061214f`

HPE Ezmeral Unified Analytics Software 1.5 Documentation
Abstract	HPE Ezmeral Unified Analytics Software is a usage-based Software-as-a-Service (SaaS) model that operationalizes hybrid and multi-cloud modern analytical workloads through a simple user interface, easily installed and deployed in minutes. HPE Ezmeral Unified Analytics Software separates compute and storage for flexible, cost-efficient scalability to securely access data stored in multiple data platforms, enabling you to run traditional and advanced analytics workloads with open-source tools.
Published	November 2024
Edition	1.5.0