Those of you who have done a project with MongoDB will notice that it functions and behaves quite differently than traditional RDBMS systems. From super fast queries to all of a sudden taking forever to return 10 documents is something beginners always face with MongoDB. I am no expert but these are the steps I took and Mongo worked much nicer than it had earlier.
- Configure “Mongod” to run as a service Many beginners make this mistake and its a very common one. Make sure you run it as a service which allows MongoDB to do better performance management and handle incoming queries a lot better.
- Indexing This should not even need to be mentioned, but with Mongo don’t do a blind indexing. Think of the fields you group the documents the most in your queries and set the indexes with that in mind. This will do a lot to speed up your MongoDB
- Start using _id Again this is something people do a lot, i.e. they don’t use the inbuilt _id field. You should using that over your own ids. Since it’s an ObjectID, it indexes better and is truly unique reducing programmer headache of creating unique id fields
- Create a re-indexer service Like any other database MongoDB needs to be re-indexed occasionally. One of the easiest ways is to create a daemon or service in your favorite language and make it do some maintenance like re-indexing and data cleanups.
- Implement Paging in your queries This is good to do in most projects. When showing large data sets, try to page your data so that you only show enough to start with, and then fetch more as you go. Mongo has an advantage over other databases in this regard in terms of speed. Please keep in mind the field you page on is a unique index
So these are a few observations I had while designing my project in MongoDB. I will be adding more improvement techniques as I go forward. If you think some of my above points are erroneous do let me know. Also share your tricks with me!