Microsoft AZ-104 Practice Test 5
Question 40
A line-of-business Windows application needs a shared SMB file path in Azure that multiple servers can mount. Which service should be provisioned?
- Azure Blob index tags
- Azure Files share
- Azure Container Registry
- Azure DNS private zone alone
Explanation
Correct Answer
B:Azure Files supplies managed file shares accessible over SMB in supported deployments. Configure networking, share permissions and identity-based authentication as needed for the Windows servers.
Incorrect Answers
A:Blob index tags help categorize and query blob metadata. They do not expose a conventional SMB-mounted shared filesystem to Windows servers.
C:Container Registry stores and distributes OCI container images rather than presenting an SMB file share. It cannot satisfy the file path requirement.
D:A private DNS zone provides name resolution but does not host file content or implement SMB file operations. Name resolution complements a storage endpoint.
Question 41
An operations team needs a point-in-time view of an Azure Files share before an application upgrade, without copying the full share to another account. Which feature fits?
- Azure Files share snapshots
- Azure RBAC Owner role
- A virtual network peering connection
- Storage account access-key rotation
Explanation
Correct Answer
A:A share snapshot captures a read-only point-in-time view of the Azure Files share and can support file-level restoration. Retention and backup protection should still be planned; snapshots alone are not a complete offsite disaster recovery strategy.
Incorrect Answers
B:Owner grants management authority but does not create a recoverable point-in-time data view. It may increase access risk without preserving files.
C:Peering joins virtual network address spaces and routes traffic; it does not record file-share data before an upgrade.
D:Access-key rotation changes credentials and should not be confused with file content protection. It cannot restore prior file versions after a bad deployment.
Question 42
A user uploaded a frequently accessed blob into the archive tier and now expects immediate online reads. What is the administration implication?
- Archive blobs remain fully online at hot-tier latency
- The blob must be rehydrated to an online tier before its contents can be read normally
- Assigning Reader at the subscription instantly rehydrates it
- Changing the blob's name bypasses rehydration
Explanation
Correct Answer
B:An archived blob generally requires rehydration to a supported online access tier before ordinary content read. Administrators should plan the operation and time requirements rather than promise immediate retrieval.
Incorrect Answers
A:Archive tier stores data offline, so immediate hot-tier read latency is not a valid expectation. Availability and rehydration timing differ from online tiers.
C:Azure RBAC permissions can authorize reading but do not change the object's offline archive state. Authorization and tier rehydration are separate conditions.
D:Renaming an object cannot transform offline archived content into an online tier. The supported rehydration workflow must be followed.
Question 43
A Bicep file deploys an App Service plan and an app. The app must use the plan created in the same deployment. How should the Bicep author express the relationship?
- Hard-code a plan resource ID belonging to another subscription
- Use a symbolic reference to the declared plan resource so the dependency is inferred
- Run the app deployment before the plan to force retries
- Remove the plan from the template and leave it undocumented
Explanation
Correct Answer
B:Bicep symbolic resource references let properties use the output or identity of a declared resource and normally establish the required deployment order. Verify the plan's type and app location constraints as part of deployment.
Incorrect Answers
A:Hard-coding an unrelated resource ID makes the deployment depend on the wrong plan and undermines environment portability. It is not the correct reference to a newly declared dependency.
C:Deploying a dependent app before its plan is ready can produce validation or resource-not-found failures. Deliberate retries are less reliable than declaring dependencies correctly.
D:Removing the plan omits a required application dependency from the declarative deployment. Reproducible infrastructure needs the relationship captured explicitly.
Question 44
Before an administrator applies a Bicep change that may replace production settings, which Azure deployment operation previews the changes without applying them?
- az deployment group what-if
- az vm delete
- az storage blob upload
- az network nsg rule create
Explanation
Correct Answer
A:The deployment what-if operation compares declared infrastructure with current Azure resources and reports predicted additions, changes, or removals. Review its results and limitations before executing the actual resource-group deployment.
Incorrect Answers
B:Az vm delete performs a destructive compute action rather than previewing a Bicep plan. Using it would change the environment before impact assessment.
C:Blob upload transfers storage data and cannot compare ARM deployment intent with existing infrastructure. It addresses a different operational plane.
D:An NSG rule creation makes an immediate network-policy change. It does not produce a read-only preview of a Bicep deployment.
Question 45
One Bicep deployment must use different resource names for development and production, while retaining the same template logic. What is the best input design?
- Maintain completely unrelated copy-pasted templates for each environment
- Use parameters and environment-specific parameter files
- Encode names into a private storage account key
- Read the requested environment from an NSG priority
Explanation
Correct Answer
B:Bicep parameters separate reusable declarative structure from environment-specific values. Parameter files can supply distinct names, SKUs or regions while the deployment remains consistent and reviewable.
Incorrect Answers
A:Separate copied templates drift over time and make feature changes harder to reconcile. The requirement is environmental configuration, not different deployment logic.
C:Storage keys are credentials and must not be misused as configuration values for naming infrastructure. That would expose secrets and hinder review.
D:NSG priority controls firewall-rule evaluation order, not infrastructure environment selection. It provides no reliable template parameter mechanism.
Question 46
An administrator exports a resource group's existing infrastructure as an ARM template to begin infrastructure-as-code. What limitation should they anticipate?
- The exported template is automatically production-ready without review
- The export is a starting point that may require parameterization, cleanup and handling unsupported details
- Exporting immediately deletes the resources
- Export requires every resource to have an Azure Files share
Explanation
Correct Answer
B:ARM export provides a starting representation of an existing deployment. Review dependencies, secrets, resource properties, API versions and parameters, then validate the result before using it as repeatable IaC.
Incorrect Answers
A:Generated templates can contain environment-specific values and imperfect or unsupported resource representations. Treating them as ready for every production environment risks unintended configuration drift.
C:Export is an inspection and template-generation operation rather than a deletion operation. It must not be confused with resource-group cleanup.
D:Template export is unrelated to Azure Files. Resource groups can contain compute, network, storage and other supported resource types.
Question 47
A parameter file for an ARM deployment contains a production administrator password in plaintext in source control. Which correction most directly protects the secret?
- Place the password in an ordinary nonsecure output
- Use a secure parameter with protected secret retrieval, such as a Key Vault-backed deployment approach
- Convert the password into a resource tag
- Disable deployment history records
Explanation
Correct Answer
B:Use secure parameter types and a supported secret management integration rather than committing plaintext values. Protect the deployment identity's access to Key Vault and avoid passing the secret through exposed outputs or logs.
Incorrect Answers
A:Ordinary outputs are not designed for secret handling and may expose the password in deployment results. Merely changing the location inside the template does not protect it.
C:Tags are widely readable metadata intended for resource organization, not a secret vault. Moving a password into a tag would worsen exposure.
D:Deployment history and logging controls do not make a plaintext secret safe in a repository. Preserve useful auditability while protecting inputs appropriately.
Question 48
A production application must remain available if one Azure availability zone goes down. Which THREE architectural choices work together toward this requirement? Choose THREE.
- A single VM with one unmanaged disk
- Run redundant application VM instances in distinct availability zones
- Only a CanNotDelete lock on the VM
- One VM replicated only inside an availability set with no zonal architecture
- Distribute client traffic using health-aware load balancing or routing
- Replicate required application state and dependencies across zone boundaries
Explanation
Correct Answers
B:Placing independent instances in separate available zones reduces exposure to loss of any one zone. A lone VM cannot satisfy this requirement even if its disk is backed up.
E:Redundant machines need a front-end mechanism that stops directing requests to unhealthy instances. Proper health checks and supported traffic distribution reduce user impact when one zone is lost.
F:Stateless front ends are not sufficient when one zone still holds the only current database or session state. Plan state replication and dependency placement so surviving instances can serve valid requests.
Incorrect Answers
A:A single VM remains a single point of failure during a zone or instance outage. Unmanaged disks do not automatically add zonal compute redundancy.
C:A resource lock prevents certain administration actions, not physical-zone outages. It does not create another running instance.
D:An availability set distributes VMs across update/fault domains within a datacenter-level architecture, not necessarily separate availability zones. It is not equivalent to the stated zonal design.
Question 49
Several identical web-server VMs must scale out when CPU load rises and scale back when demand falls. Which Azure compute feature is intended for this?
- A virtual machine scale set with autoscale rules
- Manually change a static VM name
- An Azure Storage account SAS
- A resource-group ReadOnly lock
Explanation
Correct Answer
A:VM Scale Sets manage groups of VMs with consistent configuration and support autoscaling based on monitored demand. Configure thresholds, instance bounds, cooldown periods and a load-distribution design appropriate to the application.
Incorrect Answers
B:Renaming a VM does not add or remove compute capacity. It cannot automate a response to sustained CPU demand.
C:A storage SAS permits delegated storage access; it does not control VM instance counts or compute autoscaling.
D:ReadOnly prevents many management changes and could obstruct scaling. It provides no CPU-driven capacity policy.
Question 50
A running Azure VM is underpowered, but the desired larger size may not be available on its existing host. What should the administrator expect during resizing?
- Changing the resource group's name automatically enlarges its CPU
- A resize may require stopping and deallocating the VM, then selecting a supported size
- Azure VMs can never change size after creation
- The blob's storage access tier determines the VM's new vCPU count
Explanation
Correct Answer
B:Azure VM resizing must consider regional capacity, size-family support and host allocation. Some changes require a stop/deallocate operation and can incur downtime; plan availability and disk persistence accordingly.
Incorrect Answers
A:Resource groups organize resources and do not allocate virtual CPU or memory. A rename would not meet a compute-capacity change.
C:Azure supports resizing VMs within applicable constraints. The claim that all VM sizes are immutable is inaccurate.
D:Blob access tiers affect storage cost and retrieval behavior, not the VM SKU's vCPU and RAM allocation. Compute resizing is a separate operation.