MongoDB aggregation framework on secondary node - c++

Question

How to execute aggregation framework task on secondary node using c++ driver?
Here`s an example that always executes on primary node:
DBClientConnection c;
bo res;
vector<bo> pipeline;
pipeline.push_back( BSON( "$match" << BSON( "firstName" << "Stephen" ) ) );
c.connect( "localhost:12345" );
c.runCommand( "test", BSON( "aggregate" << "people" << "pipeline" << pipeline ), res );
cout << res.toString() << endl;
I need to execute it on secondary.

Answer 1

While I haven't worked with the C++ driver for MongoDB, running aggregations on secondary is easily possible by simply setting the read preference to secondary. For e.g. on the shell:
mongo -u admin -p <pwd> --authenticationDatabase admin --host
RS-repl0-0/server-1.servers.example.com:27017,server-2.servers.example.com:27017
RS-repl0-0:PRIMARY> use test
switched to db test
RS-repl0-0:PRIMARY> db.setSlaveOk() // Ok to run commands on a slave
RS-repl0-0:PRIMARY> db.getMongo().setReadPref('secondary') // Set read pref
RS-repl0-0:PRIMARY> db.getMongo().getReadPrefMode()
secondary
RS-repl0-0:PRIMARY> db.zips.aggregate( [
... { $group: { _id: "$state", totalPop: { $sum: "$pop" } } },
... { $match: { totalPop: { $gte: 10*1000*1000 } } }
... ] )
{ "_id" : "CA", "totalPop" : 29754890 }
{ "_id" : "FL", "totalPop" : 12686644 }
...
One can verify from the MongoDB logs that this indeed ran on the secondary:
...
2016-12-05T06:20:14.783+0000 I COMMAND [conn200] command test.zips command: aggregate { aggregate: "zips", pipeline: [ { $group: { _id: "$state", totalPop: { $sum: "$pop" } } }, {
$match: { totalPop: { $gte: 10000000.0 } } } ], cursor: {} } keyUpdates:0 writeConflicts:0 numYields:229 reslen:338 locks:{ Global: { acquireCount: { r: 466 } }, Database: { acquire
Count: { r: 233 } }, Collection: { acquireCount: { r: 233 } } } protocol:op_command 49ms
...
Note that this is applicable on secondaries of a sharded MongoDB cluster as well.