Microsoft AZ-104 Exam Dumps, Practice Test Questions

CLI commands support Blob operations but still require a command-driven workflow for each ad hoc operation. Storage Explorer is the purpose-built graphical browser and transfer client for this desktop task.

Data Factory can orchestrate data movement, but building scheduled or orchestrated pipelines for occasional interactive container browsing and single-file transfers adds unnecessary automation. Storage Explorer directly supports the requested workflow.

AzCopy is useful for scripted or bulk object transfer, but it is a command-line copy tool rather than a graphical client for browsing several containers and interactively picking files. Storage Explorer better satisfies the stated workflow.

A new copy can redo already completed transfer work, may alter destination data, and ignores the job state explicitly available for targeted recovery from an interruption.

Sync can reconcile source and destination and may delete destinations depending on flags, but the case asks first to inspect and resume the interrupted original job without risking unnecessary changes.

The AzCopy jobs-clean operation removes stored job plan and log state rather than inspecting the interrupted transfer. Deleting the saved plan can prevent the intended `jobs resume` recovery, so first use job list/show and the existing job identifier instead of cleaning state.

An AzCopy include-pattern restricts which already-enumerated paths qualify for transfer; it does not tell AzCopy to descend into nested directories. Files in omitted subfolders remain outside the candidate set. For a folder tree, the copy command must enumerate recursively before file-name filtering can select matching objects below the root.

A trailing slash changes how the destination path or prefix is represented; it cannot cause the local source traversal to discover files below the top-level folder. The top-level upload succeeding shows the credentials and destination are usable. The missing operation is recursive enumeration of the local directory tree.

The ifSourceNewer overwrite policy decides when an eligible destination file should be replaced; it does not traverse nested local subfolders. With recursion disabled, files below the source directory are not enumerated for upload. The corrective setting is recursive traversal of the directory tree rather than a change to overwrite behavior.

Local staging makes bandwidth and disk space on a single client a bottleneck. It contradicts the requirement to avoid moving the complete dataset through the workstation.

A service-to-service copy still needs the appropriate access at the target endpoint. A SAS authorizing reads on the source does not automatically authorize writing to a different storage account, so a bulk transfer can fail despite being server-side.

Geo replication creates a managed secondary copy within the same storage account's redundancy system, not an independent application destination. It does not meet an account-to-account copy requirement.

The source is already readable. Extending its token does not create destination write permission and merely increases the source credential's exposure period.

Data Reader supplies download capabilities, not creating or overwriting blobs. The destination requires a suitable data write role or signed permission.

Storage encryption does not substitute for access control. The destination rejects the caller because authorization is missing, not because encryption is enabled.

AZCOPY_CONCURRENCY_VALUE controls the number of concurrent requests AzCopy makes, which can affect throughput and resource consumption, but the number 60 is not interpreted as 60 megabits per second. Resulting bandwidth depends on file sizes, available network capacity, storage latency, and retries. It is a tuning control, not the explicit data-rate limit required by the metered WAN.

The jobs resume command is useful after a transfer interruption when the job plan still exists and files remain unfinished. Here there is no interrupted job to resume, and resuming without a cap does not impose the required throughput limit. The unrelated job state and transfer rate are separate AzCopy concepts: use --cap-mbps on the actual upload operation.

Cool tier changes Blob Storage's pricing and access characteristics for data that is accessed less frequently. It does not establish a network egress or upload bandwidth ceiling for AzCopy and may incur different read or retention charges. The client's transfer rate still depends on AzCopy and network conditions; apply an explicit Mbps cap rather than manipulating storage access tier.

Deleting the destination discards all prior data and makes every transfer a full reload. It cannot preserve a target that may be newer than the source.

Archive tier affects access latency and may block normal writes, not the comparison logic for source-versus-destination modification times.

Unconditional overwrite can replace newer destination contents with stale source objects. That violates the requirement to preserve more recent target data.

Anonymous container access would expose data beyond the analyst and defeat the stated no-shared-secret policy. The existing Entra identity can provide scoped access.

Reader can expose account metadata but not necessarily blob contents under Entra data authentication. The appropriate blob data role is still needed.

An account-key SAS relies on the very shared credential the organization disallows, with an unnecessarily large bearer-token risk. Interactive Entra sign-in better satisfies the design.

Share soft delete can help restore an accidentally deleted share but does not capture earlier individual file contents before a code deployment changes them. A snapshot preserves the desired version.

A SAS delegates access to files but does not preserve file contents at a particular instant. Later corruption affects the live data, leaving no captured pre-update version.

A management lock can block certain management-plane writes, potentially disrupting operations, but does not provide file-level read-only recovery copies of the SMB share at the chosen time.

Share snapshots preserve point-in-time file contents for supported recovery scenarios but do not themselves replace undeleting a share object that is retained by Azure Files share soft delete. The question explicitly requires recovering the deleted share and its contents within the retention window, so perform the share undelete first.

Blob versioning tracks blob object versions and does not itself undelete an Azure Files share. It covers a different storage data model.

Creating an empty replacement is not the same as undeleting the protected share. It may also conflict with recovery naming while the deleted share remains retained.

A normal share snapshot does not automatically become a separate recovery-vault copy. Confusing snapshots with independent backups leaves a common deletion boundary.

More frequent snapshots improve point-in-time granularity while the share and snapshots remain available; they do not remove the shared deletion dependency.

SMB Reader grants authorized data reading, not restoration of content already permanently removed from the storage service. RBAC cannot recreate missing recovery copies.