If you’re running a 3 member ReplicaSet that uses an Arbiter as its third member, then you will be presented with this warning when you connect to the shell.
When a node goes down in a PSA architecture (Primary, Secondary, Arbiter) the wiredTiger storage engine starts generating what’s known as cache pressure. If your readConcern is set to majority that means it must read from the primary and one secondary before success. If your secondary is down, a majority can no longer be established and cache pressure starts to build.
Disabling this feature allows your cluster to run smoothly and won’t cause problems if your secondary fails. This doesn’t mean you can’t use majority on your queries, that is still allowed even if you disable it as the default.
What readConcern majority provides is data consistency on your reads. If some queries require data consistency then apply the readConcern on those queries alone.
If you’re running a configuration file, you will need to set the readConcern inside. Common locations are /etc/mongod.conf or the data directory.
Add the parameter under replication for enableMajorityReadConcern:
This can also be set from the command line when you launch mongod. This will need to be added along with the other parameters you use to launch your ReplicaSet.
mongod --enableMajorityReadConcern false