Exam AWS Certified Solutions Architect - Professional topic 1 question 14 discussion

The only answer we can be assured of working is A. B doesn't help for data corruption (question doesn't specify physical or logical so have to address both) as logical corruption will be replicated. C. We don't put backups on ephemeral storage. D. Glacier Standard retrieval is 3-5 hours, and whilst expedited can be used to retrieve within 5 minutes, this is only for archives up to 250MB - the question doesn't give the size of the database, so I think we should not choose this option.

A.. coz RDS backup or snapshot is written to S3.. If automatically backup, transaction logs are also being backup in 5 mins interval to S3.

D is clearly the most correct choice, but none of these solutions are optimal given the information. A does not backup often enough to meet the RPO of 15 minutes. It is otherwise a reasonable answer. B does not provide point-in-time restoration of data corruption. Will not work. C is hourly backups again, which does not meet RPO. Also, backups on ephemeral storage? D meets RPO, but is possibly slow because Glacier. We don't know if it will meet RTO, but we do know it will meet RPO and nothing else will, so we have to choose it even though it'd make more sense to store in S3 standard and setup a retention policy to move to glacier or discard.

B is the correct one

A. Take hourly DB backups to S3, with transaction logs stored in S3 every 5 minutes.

I would go with the Option B. the Option A & C states "Take Hourly Backup" which will exceed the RPO requirement. they just wanna lose the data for 15 minutes not more than that. The option D is also a good solution according to the scenario but saving the backup in glacier would also exceed the RTO requirements.

RPO for backup and RTO for restore So D is correct

D is using Glacier.. 15 min RPO VS 3-5 hours retrieval from Glacier.. ? Should be A

a) hourly back up does not meet RPO 0f 15 minutes d) glacier takes longer time to retreive, this satisfies RPO

A - Best Answer: Meets the requirement and is more cost effective than B B - If corrupted data is replicated is wrong this answer C - We cannot backup to an instance store volume. Instance store volumes are ephemeral (temporary) D - Won't meet the RTO, because Glacier's standard retrieval time is typically within 3 to 5 hours

A - Best Answer: Meets the requirement and is more cost effective than B B - If corrupted data is replicated is wrong this answer C - We cannot backup to an instance store volume. Instance store volumes are ephemeral (temporary) D - Won't meet the RTO, because Glacier's standard retrieval time is typically within 3 to 5 hours

A - doesn't fit RTO (hourly backups) whereas we need 15 minute backups. B - replication will not help in case of data errors, plus, it's expensive, considering RPO, RTO C - could work, but Instance store volume aren't scalable, - bad solution D - could work, however RTO must be less than 3 hours, only with Glacier Expedite retrieval. Glacier Standard would fail as it's typically 3-5 hours, which doesn't fit the requirements.

A works

If we are at 10:30am, that would mean that incident occured at 8am. Based on the backup schedule, we will be restoring from the 7am backup and have transaction logs upto 9:55am. Therefore total amount of data loss, will be 5 minutes, satisfying the RPO of 15 min and the RTO. So my answer would be A.

A. Take hourly DB backups to S3, with transaction logs stored in S3 every 5 minutes.

A Correct

With A , I will not be able to meet the RPO requirement. if we use the S3 backup 2 hours back to restore the DB, the RPO will be 30 minutes from the moment it was corrupted. Transaction logs cannot be used to restore a database . So the answer is D .