MongoDB Smasher

This project is a tool to fill up MongoDB databases with huge sets of randomized data to test your architecture and your data model performances.

How to build

The only local dependency is boost, be sure it is installed along with a modern C++ compiler and cmake then:

git clone https://github.com/duckie/mongo_smasher.git
cd mongo_smasher
git submodule update --init --recursive
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE:STRING=Release ../
make -j${number_of_machine_CPUs}

The generated executable is found as mongo_smasher/build/src/mongo_smasher. Optional:

make install

Tip: Do not use more CPUs than your machine has with makeor the compilation of the mongo C++11 driver could fail.

Tip: If you use Windows, make yourself a favor and install cash.

How to use

mongo_smasher must be fed a json file containing your collections and schemas. Check the examples directory. Here is one of them:

{
  "values":{
    "id":{"type":"incremental_id"},
    "restaurant_seed":{"type":"string","min_size":10,"max_size":150},
    "score":{"type":"int","min":-1,"max":131},
    "grade":{"type":"pick", "values":["A","B","C","D","E","F"]},
    "grade_date":{"type":"date","min":"2012-01-01 00:00:00","max":"2016-02-29 00:00:00"},
    "cuisine":{"type":"pick","file":"ny/cuisines.txt"},
    "nyzip":{"type":"pick","file":"ny/zipcodes.txt"},
    "nyst":{"type":"pick","file":"ny/street-names.txt"},
    "nybuilding":{"type":"pick","file":"ny/buildings.txt"},
    "borough":{"type":"pick","values":["Missing","Queens","Bronx","Brooklyn","Manhattan" ,"Staten Island"]},
    "geo_x":{"type":"double","min":-28.0168595,"max":52.5388779},
    "geo_y":{"type":"double","min":-157.8887924,"max":153.1628795},
    "price":{"type":"double","min":1.0,"max":500.0},
    "firstname":{"type":"pick","file":"dictionaries/first-names.txt"},
    "lastname":{"type":"pick","file":"dictionaries/names.txt"}
  },
  "collections":{
    "restaurants":{
      "schema":{
        "address":{
          "building":"$nybuilding",
          "street":"$nyst",
          "coord":["$geo_x","$geo_y"],
          "zipcode":"$nyzip"
        },
        "borough":"$borough",
        "cuisine?0.7":"$cuisine",
        "grades[0:4]":{
          "date":"$grade_date",
          "grade":"$grade",
          "score":"$score"
        },
        "name":"At $firstname $lastname's $cuisine $estaurant_seed",
        "restaurant_id":"$id"
      },
      "actions":{
        "insert":{
          "bulk_size":1000
        }
      }
    }
  }
}

Use --help switch to get some details about command line options.

Contribute

The tool is in very active development. Any contribution is welcome.

Some planned features:

• Updates and finds
• Oplog replayer