I have a test.foo database, which contains a single document, in a replica set. I then connect to a secondary of that replica set in standalone mode: REPLICASET:SECONDARY> db.runCommand({distinct: "foo" ,key: "x" }) { "values" : [ NumberLong(1) ], "stats" : { "n" : 1, "nscanned" : 1, "nscannedObjects" : 1, "timems" : 0, "cursor" : "BasicCursor" }, "ok" : 1 } REPLICASET:SECONDARY> db.runCommand({aggregate: "foo" ,pipeline:[{$match:{x:1}}]}) { "result" : [ { "_id" : ObjectId( "5176e98ce84df1db0e30f327" ), "x" : NumberLong(1) } ], "ok" : 1 } Map/reduce and count commands both fail unless I set slaveOk: REPLICASET:SECONDARY> db.runCommand({count: "foo" }) { "errmsg" : "not master" , "note" : "from execCommand" , "ok" : 0 } REPLICASET:SECONDARY> db.runCommand({mapReduce: "foo" ,map:function(){},reduce:function(){},out:{inline: true }}) { "errmsg" : "not master" , "note" : "from execCommand" , "ok" : 0 } REPLICASET:SECONDARY> rs.slaveOk( true ) REPLICASET:SECONDARY> db.runCommand({count: "foo" }) { "n" : 1, "ok" : 1 } REPLICASET:SECONDARY> db.runCommand({mapReduce: "foo" ,map:function(){},reduce:function(){},out:{inline: true }}) { "results" : [ ], "timeMillis" : 17, "counts" : { "input" : 1, "emit" : 0, "reduce" : 0, "output" : 0 }, "ok" : 1 }