Utilizing a Replica Set - MongoDB High Availability - page 47

Architecture of replica sets
After learning the replication process, we can go ahead and learn about the replica set ar-
chitecture. In order to create a readily available MongoDB server, you can simply use the
replica set network.
So, what is replica set?
A replica set is a group of mongod instances that host the same dataset ( ht-
tp://docs.mongodb.org/manual/core/replication-introduction/ ) .
Each replica set consists of different members and each one is responsible for a particular
task. The following list shows you all the possible roles for each member:
• Primary : A primary node is responsible for accepting all read/write operations
from clients. Each replica set can have one and only one primary node. Under
some conditions, a primary node can become secondary or a secondary
node can become primary by an election process.
• Secondary : The secondary nodes host the same dataset as primary nodes.
Each replica set can have one or more secondary nodes. It's possible that a
secondary node changes its state and becomes a primary node by the election
process.
• Arbiter : An arbiter node is used to facilitate the election procedure. An ar-
biter node won't change its state. The arbiter node doesn't host any dataset
and only votes in the election process. On the other hand, the arbiter node
doesn't need any dedicated hardware.
The following screenshot illustrates the architecture of primary , secondary , and ar-
biter nodes: