Database Reference
In-Depth Information
Learning about a replica set
The
replica set
is a collection of
mongod
processes. Each replica set has one and only one
primary
mongod
process that accepts all write and read operations from the clients. All
other
mongod
processes are known as secondaries; they copy all operations from the
primary
node and apply them so that all secondaries have the same data at once, along
with the replication log at varying speeds. This lag depends on the amount of data, network
speed, and so on.
The data synchronization procedure happens with the help of replication. To support replic-
ation, the
primary
node will record all write operations in its
oplog
.
The following diagram illustrates the relations between clients,
primary
, and
second-
ary
nodes: