10 exam-style questions with answers and explanations, straight from our 1,030-question bank. Tap an answer to check yourself. When you're ready, take the scored version in the free practice test.
These 10 free EX210 questions are organized by exam domain, so you can see how each part of the Red Hat Certified Specialist in Cloud Infrastructure blueprint is tested. Reveal the answer and explanation under each question.
Domain 1: Manage the Red Hat OpenStack Platform control plane
Question 1
During an incident, an administrator changes a Nova API setting in a generated configuration file on an RHOSP 16.1 Controller and restarts the service. The change works until the next director deployment, which regenerates the file with its earlier value. The setting has an appropriate director configuration parameter. Where should the approved change be made so it survives subsequent deployments?
Show answer & explanation
Correct answer: A - In the director environment configuration, followed by deployment and verification.
Domain 3: Manage user security
Question 2
Mira authenticates successfully through Keystone's read-only IdM backend and receives an unscoped token. She needs to create volumes only in the forecast project. This cloud's Cinder policy requires a project-scoped token containing the custom role volume_creator. That role already exists, but neither Mira nor any of her groups has a role assignment on forecast. What is the smallest sufficient authorization change?
Show answer & explanation
Correct answer: A - Assign volume_creator to Mira on forecast, then obtain a token scoped to that project.
Domain 4: Manage application deployment resources
Question 3
A cloud image occupies 2.4 GiB as a QCOW2 file, while qemu-img info reports a 28-GiB virtual disk. Glance records min_disk=28 GB and min_ram=4096 MB. The application team will boot directly from this image, not from a Cinder volume, and must retain both image minimums. All listed flavors are compatible with the image's architecture, and eligible hosts have sufficient capacity. Which flavor can be used?
Show answer & explanation
Correct answer: C - Two vCPUs, 6144 MB of RAM, and a 40-GB root disk.
Domain 5: Manage storage in Red Hat OpenStack Platform
Question 4
One storage host fails in a three-host Ceph cluster serving Cinder volumes. The replicated pool has size=3, min_size=2, and a host-level CRUSH failure domain. Each affected placement group is active+undersized+degraded with two usable replicas. Client I/O continues, the surviving hosts are healthy, and no capacity limits have been reached. Operations must maintain service without weakening the configured replication policy. Which recovery plan meets that requirement?
Show answer & explanation
Correct answer: D - Keep I/O running while restoring the failed host and verifying recovery to active+clean.
Question 5
A new Cinder volume type named gold is intended to select the existing SSD-backed Ceph backend. Volume creation with this type fails at backend selection. The scheduler's inventory shows the intended backend is healthy, has free capacity, and advertises volume_backend_name='ceph-ssd'. The gold type instead has the extra specification volume_backend_name='gold'. Other scheduling constraints are satisfied, and no backend advertises 'gold'. Backend settings must remain unchanged. Select the corrective command.
Show answer & explanation
Correct answer: D - openstack volume type set --property volume_backend_name=ceph-ssd gold
Domain 6: Manage networking
Question 6
A newly commissioned Compute node in an RHOSP 16.1 ML2/OVN cloud cannot provide connectivity to a VLAN provider network that works on other nodes. The provider network uses physical_network='edge' and segmentation_id=240. The new node's br-ex has the intended uplink, and the switch permits VLAN 240. Its ovn-bridge-mappings value is currently 'storage:br-storage'. Which replacement value connects the provider network through br-ex while preserving the existing storage-network mapping?
Show answer & explanation
Correct answer: B - storage:br-storage,edge:br-ex
Domain 7: Manage compute node operations
Question 7
Compute-07 stops reporting heartbeats after losing both management and migration-network connectivity. Its instances use shared Ceph RBD disks. The power controller still reports the host as powered on, and storage monitoring shows client writes from that host. Replacement capacity is available, and evacuation has been requested. Before allowing replacement instances to start, the operator must:
Show answer & explanation
Correct answer: B - Fence Compute-07 and verify that it can no longer execute guests or write shared storage.
Domain 8: Monitor operations
Question 8
At 14:20, an STF 1.5 dashboard still displays 2% CPU usage for a Compute host. The displayed value comes from a sample timestamped 14:03; Prometheus has no newer sample for that host. Local collectd output is current, and AMQ Interconnect is receiving the host's metric messages. The Smart Gateway receiver disconnected at 14:03. How should the operator interpret the displayed value?
Show answer & explanation
Correct answer: C - Current CPU usage is unknown; restore the Smart Gateway ingestion path before using this graph for capacity decisions.
Domain 9: Automate cloud application deployment
Question 9
An Ansible 2.9 play has two consecutive template tasks in its tasks section: one renders /etc/httpd/conf.d/site.conf, and the other renders /etc/httpd/conf.d/security.conf. Each task notifies the same handler, named restart httpd, which runs service: {name: httpd, state: restarted}. Both tasks report changed on the first successful execution. On a second successful execution, both files already match their templates and neither task reports a change. There are no other notifications or handler flushes. How many times does the handler restart httpd on these two executions?
Show answer & explanation
Correct answer: D - First execution: 1; second execution: 0.
Domain 10: Troubleshoot operations
Question 10
A Heat deployment fails with this event chain:
billing/app_group: CREATE_FAILED - child stack failed
app_group/worker: CREATE_FAILED - Nova server entered ERROR
Nova server fault: NoValidHost
The worker's flavor requires hw:numa_nodes=2, and the Nova NUMA topology filter is enabled. Every eligible Compute host has only one NUMA node, although CPU and RAM totals are ample. Project quotas have headroom, and the image is accessible. The workload's two-NUMA-node requirement cannot be relaxed. Which correction addresses the originating failure?
Show answer & explanation
Correct answer: A - Provide an eligible Compute host that satisfies the required NUMA topology.