AWS DevOps Engineer Professional DOP-C01 – Question351

Your system uses a multi-master, multi-region DynamoDB configuration spanning two regions to achieve high availablity. For the first time since launching your system, one of the AWS Regions in which you operate over went down for 3 hours, and the failover worked correctly. However, after recovery, your users are experiencing strange bugs, in which users on different sides of the globe see different data.
What is a likely design issue that was not accounted for when launching?

A.
The system does not have Lambda Functor Repair Automatons, to perform table scans and chack for corrupted partition blocks inside the Table in the recovered Region.
B. The system did not implement DynamoDB Table Defragmentation for restoring partition performance in the Region that experienced an outage, so data is served stale.
C. The system did not include repair logic and request replay buffering logic for post-failure, to resynchronize data to the Region that was unavailable for a number of hours.
D. The system did not use DynamoDB Consistent Read requests, so the requests in different areas are not utilizing consensus across Regions at runtime.

Correct Answer: C

Explanation:

Explanation:
When using multi-region DynamoDB systems, it is of paramount importance to make sure that all requests made to one Region are replicated to the other. Under normal operation, the system in question would correctly perform write replays into the other Region. If a whole Region went down, the system would be unable to perform these writes for the period of downtime. Without buffering write requests somehow, there would be no way for the system to replay dropped crossregion writes, and the requests would be serviced differently depending on the Region from which they were served after recovery.
Reference: http://docs.aws.amazon.com/amazondynamodb/latest/developerguide/Str…