Greetings! Here at MapMyFitness.com, we're using MongoDB as a
document store backing for a humongous amount of route data, fronted
by a lightweight rest api served up by apache. We upgraded to version
1.2.6 of the driver, but we're still encountering unbalanced reads
across the replica set. setSlaveOkay(true) is being set at the query
level. Anybody know what might cause a distribution like this?
server1: 156 queries/s
server2: 0 queries/s #primary
server3: 12 queries/s