Databases : understand how to organize and to use data

First step toward performance: Capped collections

• Capped Collections are like a "rolling" collection
• Items in capped collections are limited to a certain number of entries
• What can be the usage of those collections?

Capped collections: 1st implementation

• Implement a capped collection:

  db.createCollection("lastUpdatedQuiz", {capped: true, size: 100000, max: 10})

• Every time the maximum number of documents is reached, the documents are replaced in the collection.
• Exercise: try to create a smaller capped collection (3 items max) and check that the 1st inserted item disappears when the maximum is reached
• Exercise 2: what is the display order when you use the find() method? How to display it in the reverse direction?

Ensuring High Availability databases: Introduction

• What are the possible strategies to ensure the best possible availability for databases?

Replicasets: intro

Replicasets: implementation

in your mongodb executable folder:

mkdir ./mongoA ./mongoB ./mongoC

then :

mongod --replSet quiz --dbpath ./mongoA --port 27001 --rest

mongod --replSet quiz --dbpath ./mongoB --port 27002 --rest

mongod --replSet quiz --dbpath ./mongoC --port 27003 --rest

Replicasets: implementation (2)

Then

mongo localhost:27001

rs.initiate({
    _id: 'quiz',
    members: [
        {_id: 1, host: 'localhost:27001'},
        {_id: 2, host: 'localhost:27002'},
        {_id: 3, host: 'localhost:27003'}
    ]
})

Replicasets: Experiments

Try out the following on each server

db.isMaster()

db.tests.insertOne({insertTest:"sample text"})

Replicasets: Experiments (2)

• Shutdown the primary server, what happens?
• How replicaset is working? Need odd numbers of servers to "elect" the primary.
• you can define supplemental nodes for the election, that will not contain data {_id: 3, host: 'localhost:27003', arbiterOnly : true}