Hi,

We are using Mongo 3.0.4 with 3 member replica set in production. Our primary and one of the secondary is in one AWS zone(A) and other secondary is in different zone (B). All the queries have proper index and works fine otherwise. The problem is once in a while, we see high load on one of the secondary in B zone and when we check the currentOp, all the queries are waitingForLock : true. We are using read preference as "secondary preferred" in our application.

After looking at different forums, is there a possibility that the replica worker thread is taking long time to sync and acquire the lock? Is there any issue with the particular version? Should we upgrade it to the latest version?