Microsoft MB-500 Practice Test 8
Question 70
The following issue is under review by an engineering team during user acceptance testing: A service credential was mistakenly committed to a repository with several read-only collaborators. Which engineering response best addresses this failure?
- Retain the approved prior artifact and use the tested rollback or recovery procedure
- Revoke the exposed credential and move secret handling to a controlled vault-backed process
- Resolve the conflict against reviewed business behavior and run the merged build and tests
- Pin the pipeline to an immutable build identifier and record artifact checksums
Explanation
Correct Answer
B:Revoke the exposed credential and move secret handling to a controlled vault-backed process. Removing a value from the latest commit does not undo potential exposure in prior history. The supplied symptoms make this the decisive implementation step.
Incorrect Answers
A:Recovery needs known-good release provenance and compatible database-change handling. The proposed action does not remove the stated failure; removing a value from the latest commit does not undo potential exposure in prior history.
C:A text merge is not proof that the combined X++ semantics remain correct. Applying that approach would leave the observed path uncorrected because removing a value from the latest commit does not undo potential exposure in prior history.
D:Names alone cannot prove that two deployments use the same binary artifact. The reported symptoms instead require the behavior described here: removing a value from the latest commit does not undo potential exposure in prior history.
Question 71
A new business rule must reject duplicate external document numbers for a defined business scope. During the next test cycle, a batch operations engineer diagnosing recurring failures must select a corrective action. Which is best?
- Introduce a table extension field with suitable metadata and validate deployment synchronization
- Use a data map or supported mapping layer to align semantically equivalent fields
- Design an appropriate unique key or index and enforce the intended scope
- Express the selection as a query with explicit ranges and appropriate company scope
Explanation
Correct Answer
C:Design an appropriate unique key or index and enforce the intended scope. Data integrity requires a database-backed constraint where the key semantics are well-defined. The proposed change is tied to the observed failure, not merely to a general deployment preference.
Incorrect Answers
A:Extensions are the supported way to add custom fields without replacing the vendor's table. Applying that approach would leave the observed path uncorrected because data integrity requires a database-backed constraint where the key semantics are well-defined.
B:Mapping must express field relationships without renaming unrelated source tables. The proposed action does not remove the stated failure; data integrity requires a database-backed constraint where the key semantics are well-defined.
D:A query object should capture filtering semantics instead of scanning all records in memory. The reported symptoms instead require the behavior described here: data integrity requires a database-backed constraint where the key semantics are well-defined.
Question 72
A test engineer during a synthetic load trial is evaluating a proposed fix for this problem: An emergency fix is being copied directly into a shared release branch without a reviewer. Which result demonstrates the intended behavior?
- A clean build agent produces a working package from the documented commit and dependencies
- The target deployment points to the exact immutable build and checksum accepted in testing
- The old credential is rejected, the new secret is not in source, and the pipeline accesses it through approved controls
- The pull request shows authorized review and the resulting build links to the approved commit
Explanation
Correct Answer
D:The pull request shows authorized review and the resulting build links to the approved commit. A developer's successful manual deployment cannot substitute for review traceability. This observes the intended end-to-end behavior rather than only an implementation activity.
Incorrect Answers
A:Reusing the engineer's cached output would conceal the missing dependency. Even a successful observation could leave the original defect unresolved because a developer's successful manual deployment cannot substitute for review traceability.
B:Equal folder names do not guarantee equal deployable content. Passing that check would not establish the stated outcome; a developer's successful manual deployment cannot substitute for review traceability.
C:A repository search with no current match does not prove the leaked key is safe. The required acceptance test must go further: a developer's successful manual deployment cannot substitute for review traceability.
Question 73
In a retest by an integration tester preparing an end-to-end test, the original problem was: A batch processes ten thousand qualifying rows with identical field changes using a slow per-row loop. Which observation offers the strongest proof of resolution?
- The required validations pass and measured query or row-operation counts drop on a representative workload
- A concurrent test preserves both intended changes or returns a controlled conflict without data loss
- A multi-record test produces only each record's own payload with bounded memory behavior
- A controlled update succeeds in the correct transaction while the failure path rolls back
Explanation
Correct Answer
A:The required validations pass and measured query or row-operation counts drop on a representative workload. In the stated validation context, the decisive fact is: A batch processes ten thousand qualifying rows with identical field changes using a slow per-row loop.
Incorrect Answers
B:A single-user test cannot expose the race condition. Passing that check would not establish the stated outcome; an unmeasured speed claim does not establish safe set-based equivalence.
C:A multi-record test produces only each record's own payload with bounded memory behavior. This would not resolve the reported condition: A batch processes ten thousand qualifying rows with identical field changes using a slow per-row loop. The relevant validation for this case is: The required validations pass and measured query or row-operation counts drop on a representative workload.
D:A successful read outside the transaction does not verify update integrity. The required acceptance test must go further: an unmeasured speed claim does not establish safe set-based equivalence.
Question 74
A batch job stores growing per-record lists in a long-lived class member and memory steadily rises. Before accepting the remediation, which result should an integration tester preparing an end-to-end test verify?
- Representative volumes complete within the resource budget and yield correct company-scoped rows
- Memory use stabilizes over a representative large batch with correct record isolation
- Load tests show bounded retries and no lost updates under concurrent posting
- The plan shows reduced examined rows and the same authorized result set
Explanation
Correct Answer
B:Memory use stabilizes over a representative large batch with correct record isolation. A ten-record smoke test cannot expose unbounded retained state. This observes the intended end-to-end behavior rather than only an implementation activity.
Incorrect Answers
A:A tiny dataset cannot reveal temp-storage scaling problems. Even a successful observation could leave the original defect unresolved because a ten-record smoke test cannot expose unbounded retained state.
C:A single-threaded benchmark cannot establish concurrency safety. Passing that check would not establish the stated outcome; a ten-record smoke test cannot expose unbounded retained state.
D:A smaller UI page size can hide a large backend scan. The required acceptance test must go further: a ten-record smoke test cannot expose unbounded retained state.
Question 75
A solution consultant preparing a reproducible defect is evaluating a proposed fix for this problem: Two feature branches both modify the same extension method and the merge reports an overlap. Which result demonstrates the intended behavior?
- The pull request shows authorized review and the resulting build links to the approved commit
- The target deployment points to the exact immutable build and checksum accepted in testing
- A reviewed merge commit builds and both feature-specific tests pass on the result
- The old credential is rejected, the new secret is not in source, and the pipeline accesses it through approved controls
Explanation
Correct Answer
C:A reviewed merge commit builds and both feature-specific tests pass on the result. Choosing the newest file mechanically can silently discard a feature. This observes the intended end-to-end behavior rather than only an implementation activity.
Incorrect Answers
A:A developer's successful manual deployment cannot substitute for review traceability. The required acceptance test must go further: choosing the newest file mechanically can silently discard a feature.
B:Equal folder names do not guarantee equal deployable content. Passing that check would not establish the stated outcome; choosing the newest file mechanically can silently discard a feature.
D:A repository search with no current match does not prove the leaked key is safe. Even a successful observation could leave the original defect unresolved because choosing the newest file mechanically can silently discard a feature.
Question 76
After a candidate repair, a migration engineer reviewing a failed mapping needs acceptance evidence for this issue: A warehouse team wants to reproduce a screen-driven process across builds. Which outcome demonstrates the required behavior?
- The test passes in isolation and when run in a different order with predictable fixtures
- The test observes the completed batch state and verifies committed records and error logs
- The run report lists the expected methods, outcomes, and build identifier
- The replay reaches the expected final document state in a clean test environment
Explanation
Correct Answer
D:The replay reaches the expected final document state in a clean test environment. A recording opening the right menu alone does not prove the process completed. This observes the intended end-to-end behavior rather than only an implementation activity.
Incorrect Answers
A:Repeatedly running the same order masks shared-state coupling. Even a successful observation could leave the original defect unresolved because a recording opening the right menu alone does not prove the process completed.
B:Checking that a job ID was returned proves only scheduling. Passing that check would not establish the stated outcome; a recording opening the right menu alone does not prove the process completed.
C:Merely compiling the test project is not an execution result. The required acceptance test must go further: a recording opening the right menu alone does not prove the process completed.
Question 77
An independent developer checking extension behavior is evaluating a proposed fix for this problem: Users should filter a new list by a relationship already represented in the table model. Which result demonstrates the intended behavior?
- The form filters related rows correctly as the parent selection changes
- The form shows the approved translated label in each user's language setting
- Selecting two different orders invokes validation against the respective selected record IDs
- The existing invoice form renders the field and correctly persists values for the intended record
Explanation
Correct Answer
A:The form filters related rows correctly as the parent selection changes. A one-time hard-coded filter could pass a demo but fail for another parent. This observes the intended end-to-end behavior rather than only an implementation activity.
Incorrect Answers
B:A translation in a test document does not confirm the UI uses a label reference. Even a successful observation could leave the original defect unresolved because a one-time hard-coded filter could pass a demo but fail for another parent.
C:A test using only one open order cannot reveal stale-row behavior. Passing that check would not establish the stated outcome; a one-time hard-coded filter could pass a demo but fail for another parent.
D:A standalone custom screen would not validate the required form extension. The required acceptance test must go further: a one-time hard-coded filter could pass a demo but fail for another parent.
Question 78
After a candidate repair, a source-control reviewer validating the candidate build needs acceptance evidence for this issue: A nightly export sends the full customer master even though downstream only needs records changed since the prior run. Which outcome demonstrates the required behavior?
- The imported account values match the source mapping and rejected rows carry actionable diagnostics
- After controlled insert and update tests, only expected changed keys appear in the next export
- An invoice with multiple lines imports with valid header and line relationships
- The entity appears in the intended target environment and its staging job can run
Explanation
Correct Answer
B:After controlled insert and update tests, only expected changed keys appear in the next export. Comparing total file size alone does not prove incremental correctness. This observes the intended end-to-end behavior rather than only an implementation activity.
Incorrect Answers
A:A staging row count does not prove correct target-field mapping. The required acceptance test must go further: comparing total file size alone does not prove incremental correctness.
C:Successful header-only import cannot prove dependent child sequencing. Passing that check would not establish the stated outcome; comparing total file size alone does not prove incremental correctness.
D:Seeing it only in development does not establish target availability. Even a successful observation could leave the original defect unresolved because comparing total file size alone does not prove incremental correctness.
Question 79
A financial controls reviewer signing off acceptance tests is evaluating a proposed fix for this problem: A custom service endpoint allows a guest user to invoke an inventory adjustment. Which result demonstrates the intended behavior?
- A controlled call works with rotated tokens and no secret is present in source or logs
- Requests validate against the partner's contract and responses map into expected fields
- Authorized calls complete while a guest request is rejected without changing inventory
- The user receives a correlation identifier and the eventual job outcome can be reconciled
Explanation
Correct Answer
C:Authorized calls complete while a guest request is rejected without changing inventory. Hiding the endpoint URL does not provide a security boundary. This observes the intended end-to-end behavior rather than only an implementation activity.
Incorrect Answers
A:A single successful request with a hard-coded token is not an acceptable control. Even a successful observation could leave the original defect unresolved because hiding the endpoint URL does not provide a security boundary.
B:A generic JSON test does not exercise the SOAP message contract. Passing that check would not establish the stated outcome; hiding the endpoint URL does not provide a security boundary.
D:Merely accepting a queue message is not proof that processing finished. The required acceptance test must go further: hiding the endpoint URL does not provide a security boundary.
Question 80
The following issue is under review by a test engineer during a synthetic load trial: A dashboard queries joined invoice and supplier fields but should not store a duplicate transactional copy. Which engineering response best addresses this failure?
- Design an appropriate unique key or index and enforce the intended scope
- Express the selection as a query with explicit ranges and appropriate company scope
- Design the data entity and stable key mapping needed for import and matching
- Model the read projection with an appropriate view and relationships
Explanation
Correct Answer
D:Model the read projection with an appropriate view and relationships. A view can expose a relational projection without creating a second authoritative transaction table. This directly addresses the configuration or execution path described.
Incorrect Answers
A:Data integrity requires a database-backed constraint where the key semantics are well-defined. Applying that approach would leave the observed path uncorrected because a view can expose a relational projection without creating a second authoritative transaction table.
B:A query object should capture filtering semantics instead of scanning all records in memory. The reported symptoms instead require the behavior described here: a view can expose a relational projection without creating a second authoritative transaction table.
C:Data entities require explicit identity and mapping to avoid accidental duplicate inserts. The proposed action does not remove the stated failure; a view can expose a relational projection without creating a second authoritative transaction table.