How to Replicate Aurora MySQL Cross-Region for DR at Lowest Cost?

SAP-C02 By islever Editorial Last updated

Design a solution to ensure business continuity. Determine a cost optimization strategy to meet solution goals and objectives.

Short answer

A Use an Aurora global database with a secondary Aurora cluster in another Region, which meets the 1-hour RPO and provides the lowest-cost valid DR replication.

Question

A company runs an ecommerce application in a single AWS Region. The application uses a five-node Amazon Aurora MySQL DB cluster to store information about customers and their recent orders. The DB cluster experiences a large number of write transactions throughout the day.
The company needs to replicate the data in the Aurora database to another Region to meet disaster recovery requirements. The company has an RPO of 1 hour.
Which solution will meet these requirements with the LOWEST cost?
  1. Modify the Aurora database to be an Aurora global database. Create a second Aurora database in another Region. Correct Answer
  2. Enable the Backtrack feature for the Aurora database. Create an AWS Lambda function that runs daily to copy the snapshots of the database to a backup Region.
  3. Use AWS Database Migration Service (AWS DMS). Create a DMS change data capture (CDC) task that replicates the ongoing changes from the Aurora database to an Amazon S3 bucket in another Region.
  4. Turn off automated Aurora backups. Configure Aurora backups with a backup frequency of 1 hour. Specify another Region as the destination Region. Select the Aurora database as the resource assignment.

Key Takeaways

The question tests cross-Region DR for Aurora and the trap is choosing DMS CDC or hourly backups because they seem cheaper while ignoring restorability and Aurora backup constraints.

Aurora Global Database is the valid cross-Region disaster recovery solution for an Aurora MySQL cluster with a one-hour RPO. The page explains why the lower-cost-sounding DMS-to-S3 and hourly backup options do not provide a directly restorable Aurora database.

Common mistake: A common wrong choice is DMS change data capture to S3 in another Region; it can replicate changes, but the resulting Parquet or CSV files are not a restorable Aurora disaster recovery copy.

Expert Analysis

Why the Answer Is Correct

Aurora Global Database is AWS's managed cross-Region replication feature for Aurora. It creates a secondary Aurora cluster in the DR Region and continuously replicates storage, giving an RPO far below the required one hour and a directly promotable Aurora database. For a five-node write-heavy MySQL cluster, this avoids building and operating a separate CDC and ETL pipeline. Although the secondary Region has instance and cross-Region transfer costs, among the options that actually deliver restorable Aurora DR it is the lowest-cost valid solution. The question's LOWEST cost requirement must be read together with the DR and RPO requirements, not as permission to store unusable change files.

Why the Other Options Are Wrong

B uses Backtrack, which only rewinds the same cluster in place, and a daily Lambda snapshot copy gives an RPO of up to 24 hours, not 1 hour. C's DMS CDC task to S3 can capture changes within the RPO window, but S3 Parquet or CSV change files are not an Aurora backup and cannot be promoted or restored directly as a database in the DR Region; using them for DR requires ETL and reload, which does not meet the stated objective. D suggests turning off Aurora automated backups, which is not possible for Aurora, and hourly backup copies would still be snapshot-based rather than continuous replication for a reliable one-hour RPO. That makes D unimplementable as written and still weaker for DR than an Aurora global database.

Community Comment Notes

TonytheTiger focused on the key requirements, including RPO under one hour, DR, and lowest cost, and pointed to the Aurora Global Database documentation, which aligns with A. Wardove argued that DMS-to-S3 is "not a Disaster Recovery" because Parquet or CSV files are not directly restorable into a database. kjcncjek noted that "You can't disable automated backups on Aurora," which undermines option D. kgpoj observed that Global Database can achieve a one-second RPO but is more expensive; that cost point is real, yet C and D do not provide a valid restorable Aurora DR target, so cost alone cannot make them correct. Some learners selected C, but their rationale stops at replication-to-S3 rather than database recovery.

Official Reference

https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/aurora-global-database.html

This link opens the official vendor documentation in a new tab.

Exam Strategy

For SAP-C02 DR questions, first verify that the option creates a restorable workload in the target Region, then compare cost and RPO. Eliminate choices that only copy change files or snapshots when the requirement is database recovery, and watch for unsupported actions such as disabling Aurora automated backups.

Frequently Asked Questions

Why is DMS CDC to S3 not a valid Aurora DR solution here?

DMS writes change data as CSV or Parquet in S3, not a restorable Aurora cluster. Restoring would require ETL and reload, so it does not meet the database DR requirement.

Can Aurora automated backups be disabled to use hourly cross-Region backups?

Aurora automated backups cannot be disabled; retention is managed at the DB cluster level. That makes option D not implementable as written.

More SAP-C02 FAQ →

Discussion 70 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.

From public community threads 30 of 70

  1. YodaMaster 👍 35 Answered C
    Good luck for the exams. I know I'm gonna fail coz it takes me 3 hours just to read the questions. >:(
  2. SkyZeroZx 👍 21
    if you got far it means you are persistent, good luck on your exam
  3. yog927 👍 12
    The trick is not to read the question first. Here is what I do: 1. Read all options. 2. Eliminate the incorrect ones, and settle on 1 or 2 options. 3. Scroll through the question last.
  4. teo2157 👍 6 Answered A
    Github copilot While AWS Database Migration Service (DMS) can be used to replicate ongoing changes from the Aurora database to an Amazon S3 bucket in another region using a change data capture (CDC) task, it's important to note that DMS does not create a standard SQL dump or backup file that can be directly restored to an Aurora database. The data migrated to S3 by DMS is in Apache Parquet format, a columnar storage file format optimized for speed and for a small footprint. This format is not directly restorable to an Aurora database. If you need to restore the data to an Aurora database, you would need to use a service like AWS Glue or Amazon Athena to read the data from S3 and then insert it into Aurora. This process could be complex and time-consuming, and might not meet your RPO of 1 hour.
  5. TonytheTiger 👍 5 Answered A
    My head hurts after the reading the last 2 questions and 45 mins later, still confuse. I am looking at the key requirements, RPO <1h, meet DR requirements, and LOWEST Cost. After reading the link below and all of the comments, I think Option A fulfil all the requirements in the question. https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/aurora-global-database.html
Show 25 more comments
  1. kjcncjek 👍 5
    its can't be D You can't disable automated backups on Aurora. The backup retention period for Aurora is managed by the DB cluster. so answer is C
  2. i_am_robot 👍 4 Answered A
    Amazon Aurora Global Database is designed for globally distributed applications, allowing a single Amazon Aurora database to span multiple AWS regions. It replicates your data with no impact on database performance, enables fast local reads with low latency in each region, and provides disaster recovery from region-wide outages. This setup would allow your application to meet the RPO of 1 hour because data written to the primary AWS region is replicated to the secondary region typically in less than a second. Other options like using AWS Database Migration Service (DMS) or configuring backups to another region or enabling backtrack would not be as efficient or cost-effective. DMS and Backtrack may add complexity and additional cost. Turning off automated Aurora backups and configuring them manually could potentially risk data loss if not managed properly.
  3. ggrodskiy 👍 4
    Correct A. We have an RPO of 1 hour, which means you can tolerate losing up to 1 hour of data in case of a disaster. However, you also need to consider the cost and the recovery time objectives (RTO) of your solution. Using AWS DMS will cost more than using Aurora global database, and it will take longer to recover your data from S3 to a new database. Aurora global database will replicate your data to another Region with low latency, and it will allow you to fail over to the secondary DB cluster in minutes if the primary Region is unavailable. Therefore, Aurora global database is a better solution for your requirements.
  4. dkx 👍 4
    A. No, because this will create more database resources and the question is asking about 'replication of data', not 'replication of databases' B. No, because a daily copy does not meet the RPO of 1 hour requirement C. Yes, because using DMS CDC to replicate ongoing changes addresses the large number of write transactions throughout the day and the RPO of 1 hour requirement D. No, because this just writes the active DB cluster's logs to a different region on an hourly basis. Note, this question and answer choices are poorly worded.
  5. Wardove 👍 4 Answered D
    not C because.. DMS is not cheap and moreover, DMS S3 Target support either .csv or parquet format.. good luck with restoring this data into a database from s3 this is not a Disaster Recovery this is purely "playing around with data in a Disaster situation"
  6. BrijMohan08 👍 4 Answered D
    option D provides the most cost-effective solution by leveraging Aurora backups with a 1-hour frequency and cross-Region replication to meet the disaster recovery requirements with the desired RPO.
  7. kgpoj 👍 4 Answered C
    Ask for 1 hour RPO, cheapest Aurora Global can do 1 second RPO, but expensive DMS with CDC can do 1-hour RPO and cheaper
  8. KCjoe 👍 4
    I thought it has 304 questions, how come there is no more next page?
  9. longngo0924 👍 3
    Before considering the cost, please consider the ability of solution. B. Backtrack feature is mainly use for solved incorrect data or configuration but don't clone to new DB, just roll-back to a PITR. C. How can create a S3 as a target for DMS in other regions? It must be the same region with DMS. https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Target.S3.html#CHAP_Target.S3.Prerequisites D. Cannot turn off automatic backup of Aurora, the automatic backup range is 1 to 35 days. Disable require 0 day but don't have any option which is 0 day. So, the answer A is reasonable.
  10. SkyZeroZx 👍 3 Answered A
    Amazon Aurora Global Database is designed for globally distributed applications, allowing a single Amazon Aurora database to span multiple AWS regions. B: Backtrack is a feature that lets you quickly move your Aurora database back in time without needing to restore data from a backup. C: AWS DMS (Database Migration Service) is generally used for database migrations, not for continuous replication for disaster recovery purposes.
  11. javitech83 👍 3 Answered C
    Correct is C. Is cheaper and we do not need a RTO, just RPO
  12. yorkicurke 👍 3
    any news about your exam?
  13. Jonalb 👍 3 Answered A
    AAAAAAAAAAAAAAA
  14. 4845c28 👍 2 Answered A
    C is wrong. CDC task type only captures changes from source databases after we start the task. This is usefull when target already contains up to date data. In option C there is nothing mentioned about backing up database first. And on top of that using DMS for backing up data is like using fork to eat the soop, technically you can do it but come on A is wrong becase you cannot disable default backups. Global database may not be the cheapeast option but is the only correct one
  15. aviathor 👍 2 Answered C
    Although I also lean towards C, the problem is that I think the solution is not complete with only the CDC. We would also need a backup from which to recover the databases before applying the changes.
  16. khksoma 👍 2
    I dont think it can be C. Check the pre-req for S3 to be the target in this link. https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Target.S3.html
  17. MegalodonBolado 👍 2
    C looks more like a workaround than an architected solution. Also, I don't know how to confirm RPO < 1 hour if data amount wasn't provided. Vote: A
  18. Maria2023 👍 2 Answered C
    I vote for C. A appears to be a costly option for DR with a RPO of 1 hour
  19. trungtd 👍 2 Answered C
    Option A will replicate "the whole database", we only need "data"
  20. 07c2d2a 👍 2
    did you actually see any of the same questions on the test?
  21. career360guru 👍 2 Answered C
    As there is no RTO C is best and most cost-effective.
  22. SmileyCloud 👍 2 Answered C
    C - lowest cost. Otherwise A.
  23. dankositzke 👍 2 Answered C
    C is the least worst answer
  24. Jackhemo 👍 2
    C is the answer.
  25. yorkicurke 👍 2
    i wonder myself

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Cost-Effective Image Processing with SQS and S3 Lifecycle Policies

S-grade

Modernize centralized firewall appliances with a Gateway Load Balancer and Auto Scaling

S-grade

Use Lambda over a CloudWatch Logs stream to count unique users per client

S-grade

Serve global game assets with Route 53 latency records and health checks

S-grade

Practice All SAP-C02 Questions

Access 85 questions with complete answers and detailed explanations.

View Full SAP-C02 Practice Test →

← Back to SAP-C02 Study Guide