Question 1

A company IT department has strict requirements around data protection and wants you to explain
how to use the Max Target and Mm Target options for an HPE Ezmeral Data Fabric volume.
What should you explain?
The Max Target sets the desired RAID level for data In the primary volume. The Mm Target sets the
desired RAID level for data In associated mirror and snapshot volumes.
A. The Max Target sets the desired number of replicas for data; the Min Target sets a replica number
at which the fabric still tries to reach the Max Target, but less aggressively.
B. The Max Target and Min Target work together to set the erasure coding level for data in the
volume. C. The Max Target is the total number of blocks and the Min Target is the total minus
redundant blocks.
D. The Max Target sets the desired number of replicas for "hot" (frequently used) data; the Min
Target sets the desired number of replicas for "warm" (less frequently used) data.

Question 2

You want to set up high availability (HA) for the HPE Ezmeral Container Platform control plane. This
includes HA for gateways and HA for the platform. You are not planning to use the embedded
Ezmeral Data Fabric.
Which hosts will you require at minimum?

• A. Two hosts with 500 GB for ephemeral storage and 500 GB for OS storage Two hosts with 500 GB for OS storage
• B. Three hosts with 500 GB for ephemeral storage and 500 GB for OS storage Two hosts with 500 GB for OS storage
• C. Three hosts with 500 GB for OS storage Two hosts with 500 GB for OS storage and 500 GB for ephemeral storage
• D. Four hosts with 500 GB for OS storage and 500 GB storage for ephemeral storage

Question 3

What is a setting that you are allowed to change after you deploy a Kubernetes cluster?

• A. The number of hosts assigned to a role
• B. The type (data fabric or non-data fabric)
• C. The pod network range
• D. The authentication settings

Question 4

You need to enable the Al/ML project option for a Kubernetes tenant on HPE Ezmeral Container
Platform. What Is a requirement?

• A. That the Kubernetes cluster for the tenant is running Kubernetes version 1.18.6 or above
• B. That the Ezmeral ML Ops Installation package was run on the platform and cluster before the tenant was added
• C. That the platform has enough available Ezmeral ML Ops licenses to match the core quota on the tenant
• D. That the Kubernetes cluster for the tenant has both Spark operator and Istio enabled on it

Question 5

What is one way that HPE Ezmeral ML Ops can help companies overcome their AI/ML challenges?

• A. It enforces isolation between notebook environments and training environments.
• B. It provides acceleration for machine learning algorithms to eliminate the need for CPUs.
• C. It provides notebook and training environments with many AI/ML libraries and frameworks.
• D. It provides a custom HPE Ezmeral library for deep learning and neural network algorithms.

Question 6

What ate the network requirements for the HPE Ezmeral Container Platform controller and gateway?

• A. The controller and gateway must be on the same routable network as the Kubernetes or EPIC hosts.
• B. The controller and gateway must be on routable networks, but do not have to be on the same network as each other or other hosts.
• C. The controller and gateway must be on the same routable network as each other, but not necessarily other hosts.
• D. The controller and gateway must be on the same non-routable private network as each other and all other hosts.

Question 7

What Is one way that HPE Ezmeral Container Platform helps customers to reduce costs?

• A. By making it easy to install bare metal Kubernetes hosts and clusters, it eliminates the costs of many virtualization licenses.
• B. It embeds all control plane functions in the same host systems that run container workers, reducing hardware costs.
• C. It includes its own container orchestration software so that customers do not have to purchase expensive Kubernetes licenses.
• D. It provides a streamlined solution with fewer components than typical DIY Kubernetes, reducing hardware and operational costs.

Question 8

What is one requirement tor setting up Platform High Availability (HA)?

• A. Gateway HA must be configured before you enable Platform HA.
• B. The controller and the shadow controller must be on different subnets.
• C. The controller and the gateway must be on the same subnet.
• D. Any virtual clusters that have been created on EPIC workers must be deleted.

Question 9

What is a use case for Apache Spark?

• A. Processing and analyzing data quickly
• B. Establishing secure communications
• C. Storing data and ensuring data protection
• D. Making Kubernetes easier to manage

Question 10

You have an Ezmejal Container Platform solution that manages an internally created Kubernetes
cluster. The cluster uses AD for Its authentication settings, and users can successfully log into the
platform with their domain credentials.
You have created a Tenant on the cluster which has the "AI/ML project" setting enabled on it. You
have also created a Notebook cluster within that project. However, users receive an error when they
try to log into the Notebook endpoint (Jupyter Hub) with their domain credentials.
What should you check?

• A. That the users have a manually created external account that assigns them to the Cluster Admin role in the cluster
• B. That the Notebook cluster has a "cluster" connection to the AD server hostname
• C. That the users have a manually created external account that assigns them to the Admin status in the AI/ML Project
• D. That the Notebook cluster has a "secret" connection with this value: "hpecp-ext-auth-secret"

Question 11

Refer to the exhibit.

You are logged in as a Tenant Admin named "admin 1." You have uploaded a file into the location
shown In the exhibit. Now you want to apply that tile from the Web Terminal. How do you obtain
access to the file in the Web Terminal?

• A. Copy the file to the Web Terminal using the scp@admin 1 :/TenantShare/apps/ command within the terminal.
• B. Set up a Source Control that references both the Web Terminal and the TenantShare.
• C. Set up a Data Tap between the Web Terminal and the TenantShare/apps directory.
• D. Access the file in the Web Terminal's "/bd-fs-mnt/TenantShare/apps" directory.

Question 12

You want to create a custom dashboard with monitoring information collected by HPE Ezmerat
Container Platform. What step should you complete first?

• A. Exec Into the epic-monitoring pod on the controller and enable Kibana.
• B. Exec into the epic-nagios pod on the controller and start the Nagios service.
• C. Access each host managed by the platform and export logs in the "var/log/bluedata" directory.
• D. Create a user account and assign that account the "monitoring" role.

Question 13

How does HPE Ezmeral Container Platform regulate the use of GPUs on EPIC hosts?

• A. It lets Tenant Admins create flavors, which specify the number of GPUs required by a virtual node using that flavor.
• B. It creates virtual GPUs and allocates these virtual GPUs to the virtual nodes that request them.
• C. It allows Site Admins to reserve a certain number of GPUs for the tenant within an EPIC tenant's settings.

Question 14

A Tenant Member plans to use KubeDirettor to deploy a Spark 2.4.5 cluster. You need to give cluster
members access to data stored in a volume In Ezmeral Data Fabric. What are two required steps?

• A. A Tenant Admin must add another role to the Spark 2.4.5 application for the dtap sidecar.
• B. The Tenant Member must specify the "hpecp.hpe.com/dtap: hadoop2" pod label under each role when running the cluster.
• C. The Tenant Member must create an FS mount to the volume's mount path.
• D. A Tenant Admin must create a DataTap to the volume's mount path.
• E. The Tenant Member must reference the Data Fabric namespace when running the cluster.

Question 15

For which use case should you enable Istlo on a Kubernetes cluster?

• A. Connecting a KubeDirector cluster to external Kubernetes components such as secrets and config maps
• B. Converting stateful applications such as Spark into a form that can run on Kubernetes
• C. Training ML algorithms to perform complex tasks such as clustering and image recognition
• D. Creating more sophisticated routing rules for services running on the cluster