-1

Which of the following statement(s) is/are correct regarding Replication and Sharding?

Select one or more:

a. Master-slave replication reduces the chance of read-write conflicts.

b. Master-slave replication and sharding can not be combined.

c. By using sharding, each site acts as the single source for a subset of data.

d. Peer-to-peer replication and sharding can not be combined.

e. Peer-to-peer replication is a single point of failure.

3
  • What's your take on the answer(s)?
    – mustaccio
    Jan 10 at 18:34
  • This really looks like you're just posting your homework and asking us to do it for you. It's fine to ask for help, but you should show some effort on your part first. See also dba.meta.stackexchange.com/questions/490/… Jan 10 at 22:05
  • 2
    I’m voting to close this question because lack of effort
    – mustaccio
    Jan 10 at 23:54
0

a. Master-slave replication reduces the chance of read-write conflicts.

  • TRUE for the Primary -- because you would move most reads to the replicas.
  • Mostly FALSE for a single Replica -- because every replica must repeat all the writes performed by the Primary; however, since they are repeated in a simplified way (assuming binlog_format=ROW), it is not completely "false".
  • TRUE for multiple Replicas -- assuming you are spreading the reads to the various Replicas.
  • TRUE/FALSE For dual-Master and clustering -- It gets messier; I'll not elaborate here.

b. Master-slave replication and sharding can not be combined.

  • FALSE -- Have several Primaries, each with a subset of the data. Then have Replica(s) hanging off each of those Primaries.

c. By using sharding, each site acts as the single source for a subset of data.

  • TRUE. Well, what do you mean by "site"? The Primary + Replicas?

d. Peer-to-peer replication and sharding can not be combined.

  • FALSE -- same as (b). You could have half the Peplicas hanging off each Primary. Note that when one Primary dies, all of its Replicas quickly become stale. (Clustering solutions are similar.)

e. Peer-to-peer replication is a single point of failure.

  • TRUE -- Because of the "split brain" scenario. (See MHA, Galera, InnoDB Cluster)

Not the answer you're looking for? Browse other questions tagged or ask your own question.