MongoDB Administrator's Guide

How it works...

We begin by creating the directories for the arbiter process. As mentioned in the beginning of this recipe, an arbiter is nothing but a mongod process that will not store any data. However, it does need to store some metadata about itself and hence a minimal amount of state has to be maintained.

For this purpose, in step 2, we provide the --dbpath parameter with a location to store its data along with an arbitrary port 30000.

In step 3, we connect to the primary node of our replica set, and in step 4, we use the rs.addArb() wrapper to add the new arbiter.

Next, in step 4, we check the status of the replica set; lo and behold, ...