ISC · Information Systems and Data Management · 7 practice questions
System Availability: Why Backups Do Not Prove Recovery
Recovery evidence must show that users can resume the committed operations within the promised time. Below: judge whether a recovery test proves availability when it stops before testing user access.
Try one first
Hint
Match each existing control to the specific availability commitment first, then identify which commitment is still not actually demonstrated.
Answer D. The contractual commitments are time-bound recovery and transaction re-creation limits, and those requirements have not been validated in the migrated environment. Existing controls (redundant AZs, 5-minute replication, CPU alerts, and maintenance notices) align with uptime, RPO, and communications, but a tabletop exercise does not demonstrate that the organization can restore service within 1 hour or meet the 15-minute re-creation tolerance. An end-to-end failover and recovery test (with documentation) is the first needed action to prove compliance with the commitments.
Why not A: Better customer communication can reduce disputes about scheduled downtime, but the organization already sends 72-hour notices and the availability commitments exclude preannounced maintenance. Extending notice does not address the critical issue that recovery time and transaction re-creation have not been demonstrated in the new environment.
Why not B: This is tempting because synchronous replication reduces RPO, but the contract allows up to 15 minutes of transaction re-creation and the current 5-minute replication already appears to meet that tolerance. More important is proving the recovery process and timing on the new architecture; switching to synchronous replication is costly and does not itself demonstrate that the 1-hour recovery objective can be met.
Why not C: Proactive monitoring is relevant to availability, so this is superficially plausible. However, the stem gives no evidence of capacity-related outages or repeated CPU saturation, and lowering the alert threshold may increase noise without addressing the primary unvalidated risk: actual failover and recovery capability after migration.
What should you do?
The situation
Harbor Co. promises to resume branch order processing within four hours of disruption, losing no more than 15 minutes of transactions. It has nightly backups, successful backup logs, and transaction logs replicated every ten minutes. A recovery test restored application and database services in three hours but excluded DNS, network failover, and branch connectivity. Availability requires branch users to access and process orders. What should management do?
Test the full recovery path through successful branch order processing and document elapsed recovery time.
Best This tests the committed boundary, including omitted dependencies, and establishes whether users can resume processing within four hours.
Report that the data-loss objective is supported but the restoration-time objective remains unverified.
Acceptable This accurately describes the evidence, but management still needs a complete test to resolve the recovery gap.
Conclude both objectives are supported because services recovered within three hours.
Wrong Application and database recovery does not establish that branch users can connect and process orders.
Use successful backup logs instead of further recovery testing.
Wrong Backup completion does not demonstrate restoration timing or whether DNS, network failover, and branch access work.
Prioritize synchronous replication to eliminate transaction loss.
Wrong The commitment permits 15 minutes of loss. Eliminating data loss would not resolve the untested user-access steps.
Key points
- Remember: the recovery clock does not stop when servers start; it stops when users can perform the promised work.
- RTO measures restoration time. RPO measures acceptable data loss.
- Test DNS, network failover, and user connectivity when they are required to resume processing.
How the exam traps you
- Treating three-hour application recovery as proof of meeting the four-hour commitment. Include branch access and transaction processing before concluding that recovery meets the deadline.
- Treating successful backup logs as proof of timely restoration. Use timed recovery testing, not backup completion, to demonstrate restoration capability.
- Requiring zero data loss when the commitment allows 15 minutes. Evaluate the stated tolerance. Ten-minute log replication supports the 15-minute recovery-point objective.
Question 2
Hint
Focus on what actually went wrong in the facts: Was the problem unauthorized access, improper disclosure, bad processing, or inability to use the system when needed?
Answer C. Availability addresses whether the system is accessible for operation and use as committed or agreed. The facts show authorized HR users cannot access the payroll system during the critical month-end window because capacity is insufficient; data remain secure and processing is accurate after recovery. Therefore, availability is the primary trust services concern.
Why not A: Candidates may associate outages with security incidents, but the stem explicitly rules out unauthorized access or a breach. The outage is due to insufficient computing capacity, which points to availability rather than a failure of security controls.
Why not B: This is tempting because payroll delays may suggest a timeliness issue, and processing integrity covers correctness and timeliness. However, the stem emphasizes lack of access (an availability issue) and states that calculations are accurate once processing resumes, so processing integrity is not the primary problem here.
Why not D: Payroll data are indeed sensitive, making confidentiality an important concern in general. Here, however, the facts state the data remain secure and do not describe any improper disclosure, so confidentiality is not the primary issue.
Question 3
Hint
Ask which control most directly helps the system keep running or return to service after an interruption.
Answer A. System availability concerns whether systems remain operational and can be recovered after disruptions. A tested failover environment together with backup restoration procedures are controls specifically designed to ensure continuity and recovery, directly supporting availability. The other choices primarily address access control or data confidentiality rather than keeping systems running or restoring service.
Why not B: This control mainly prevents unauthorized access and strengthens authentication (security/confidentiality), not the system's ability to remain available or to be recovered after downtime. It doesn't ensure service continuity or recovery mechanisms.
Why not C: Encryption protects the confidentiality of data at rest and reduces risk of disclosure, but it does not directly support system uptime or recovery following an outage.
Why not D: Limiting developer access reduces risk of data exposure and supports security and confidentiality controls (and sometimes integrity), but it does not primarily ensure availability or system recovery after interruptions.
Question 4
Hint
Ask whether the control described mainly helps restore the system after a problem, or helps keep the system up and reachable in the first place.
Answer C. System availability is broader than backup and restore capability. Backups primarily support recovery after a disruption or data loss, while availability also involves keeping the system operational and accessible as needed. In this fact pattern, the single server, single site, lack of failover connectivity, and no uptime monitoring leave clear availability gaps.
Why not A: This is tempting because candidates often associate availability with disaster recovery. However, restoring data after an outage is not the same as keeping a system available for use when needed. Availability also requires measures to minimize outages and detect service interruptions promptly.
Why not B: Testing restores is a good control, so this option sounds reasonable. But even well-tested backups do not address the system's single points of failure or the absence of monitoring and failover capability. The issue is not just whether backups work; it is whether the system can remain accessible with minimal downtime.
Why not D: This distractor is attractive because it identifies a possible weakness in backup frequency. Still, it is too narrow and incorrectly treats backup frequency as the only problem. The larger availability concern is the lack of redundancy, failover, and monitoring, not just whether backups are daily or continuous.
Question 5
Hint
Focus on the control that best keeps the system usable when demand increases, not the controls that protect data or detect errors.
Answer B. Availability concerns whether systems are accessible and usable when needed. The outages occur during predictable demand spikes, so capacity monitoring combined with automatic scaling addresses the root cause by increasing processing resources before demand exceeds thresholds. This preventive control helps keep the ordering system operational during peak periods.
Why not A: This is tempting because stronger authentication improves security and access control, but it primarily protects against unauthorized access rather than preventing system outages from high traffic. MFA would not directly resolve capacity-related downtime during peak demand.
Why not C: Encryption is a valuable confidentiality control and may be required for data protection, yet it does not improve system uptime. Encrypting data protects data at rest but does not prevent outages caused by insufficient processing capacity.
Why not D: Reconciliation helps detect processing integrity issues after transactions occur, making it a detective control for accuracy. It does not prevent the system from becoming unavailable during traffic spikes and therefore does not directly address the availability objective.
Question 6
Hint
Think about whether the control only creates a recovery resource or also demonstrates that the resource can actually be used when operations are disrupted.
Answer B. Availability concerns whether the system can be used as committed or agreed, including the ability to recover from disruptions. Nightly offsite backups are an important control, but without testing that backups can be restored within the needed time frame, management cannot be confident operations can be resumed. Restoration testing is therefore a key element of supporting system availability.
Why not A: This is tempting because creating offsite backups is an important availability-related control. However, backups alone are only a partial solution: if restoration is never tested, the organization cannot reasonably conclude it can recover operations when needed.
Why not C: This choice is tempting if a student focuses only on confidentiality (encryption, access controls). But backups also serve availability by enabling recovery after an outage. Saying availability is not relevant misclassifies the purpose of backups.
Why not D: Restricting access to backups is an important security control, mainly for confidentiality and integrity. It does not demonstrate that the backups can be restored accurately and quickly enough to resume operations; availability requires tested recovery capability, not just protected storage.
Question 7
Hint
Which tier became unhealthy, and what process ran simultaneously on each application host?
Answer C. The post-incident review shows each application host runs a local maintenance script that applies patches and immediately reboots, causing all instances to fail health checks at nearly the same time. Centralized orchestration with rolling updates enforces sequential restarts so some instances remain healthy during patching, preventing simultaneous downtime while keeping automated patching in place. The database remaining responsive points to the application tier as the root cause.
Why not A: This is tempting because a single primary database can be a single point of failure and improving DB HA supports SLAs. However the database and network stayed responsive during the incidents; the observed outages are caused by the application instances rebooting simultaneously, so database clustering would not prevent these application-tier failures.
Why not B: This distractor tempts by offering capacity and tuning as a mitigation for brief outages. But adding instances or increasing health-check timeouts doesn't stop identical local maintenance scripts from rebooting every instance at once; it may mask symptoms in some cases but does not correct the lack of coordinated, sequential restarts.
Why not D: Rescheduling or disabling automatic patching might stop outages at that specific time and can be a short-term mitigation, which makes it tempting. However it doesn't address the underlying problem that each host reboots independently and can increase security risk if automation is disabled; the safer long-term fix is coordinated rolling updates so patching continues without causing simultaneous downtime.
Common questions
Do successful backup logs prove system availability?
No. They show that backup jobs completed, not that Harbor can restore branch access and processing within four hours.
Does restoring the application within three hours meet a four-hour RTO?
Not by itself. Harbor's commitment requires branch users to access and process transactions, so the test must include the connections needed for that work.
Can an RPO be supported while the RTO remains unproven?
Yes. Harbor's ten-minute log replication supports its 15-minute data-loss limit, but the incomplete recovery test does not establish timely branch-user recovery.
Watch it solved
A full CPA ISC task-based simulation on System availability, worked step by step.
System Availability & RTO/RPO | CPA ISC Simulation (TBS) on YouTube
Practice ISC like the real exam
The free ChatCPA simulator: real exam layout, timed testlets, starting with a question on this topic. No account needed to start.
Open the free simulator →More on System availability
- When to Trigger Failover: RTO and RPO
- Choose Disaster Recovery Architecture by RTO and RPO
- System Availability: Redundancy and Automatic Failover
- How to Calculate SLA Uptime at 99.95% Availability
- RTO vs RPO: Synchronous vs Asynchronous Replication
- Availability vs. Security, Confidentiality, and Integrity
- Availability Calculation: MTBF, MTTR, and Redundancy
Questions from the ChatCPA bank of 17,658 CPA exam questions, each with a written reason for every wrong answer. ChatCPA is built by Nicholas Miller, CPA (Oregon #14907). How these pages are made. Spot an error? Tell us.