MongoDB Devops Madrid February 2012
0 likes760 views
This document provides an introduction to MongoDB, including when to use and not use MongoDB, features of MongoDB like document structure, indexing, replication, sharding, and architecture. It outlines the basic MongoDB scheme including databases, collections, and documents. It describes concepts like replication, auto-sharding, and the logical and physical architecture of MongoDB. Finally, it provides steps for initial sharding configuration and backup recommendations.
MongoDB Devops Madrid February 2012
- 1. First steps with MongoDB
- 2. Index When you should use NoSQL When you should not use NoSQL A quick look at MongoDB Other features Main scheme Replication Autosharding Architecture First steps
- 3. When should I use NoSQL? Caching High data amount Horizontal scalability(without joins or complex transactions) The Slogan is ”not only SQL” and ”not only NOSQL”
- 4. When should not I use NoSQL High transaction level Business Intelligence Problems that need relational queries
- 5. A quick look at MongoDB Document database oriented (BSON) Flexible scheme Dinamic queries Replication Autosharding Atomic operations Full indexes
- 6. A quick look at MongoDB Not relational Distributed Opensource Horizontal scalability Drivers for most known programming languages Replication and automatic sharding GridFS
- 7. Other features Map Reduce Queries are executed in parallel in all of the shards Adittion and process of flexible data Geospatial index Find objects based on their geographical location
- 8. Basic scheme Database Server Shell Sharding router Mongod Mongo Database server Interactive shell Mongos Sharding router
- 9. Replication Master: mongod –master Slave: mongod –source url:port or pasive –slavedelay numberofseconds Replica Sets Full support in dealing with failures(recommended by each shard) You can have passive members that never will be masters. Mastermaster Only recommended for insert and delete queries by id.
- 10. Auto-sharding Defined by a shard key Response in dealing with failures is handled by each replica set. Queries map reduce could run in paralell between many shards. You can add servers as shards and automatically data are distributed along of all shards.
- 11. Fragmentation
- 12. Fragmentation
- 13. Architecture Shards Servers: mongod –shardsvr You can configure replication to dealing with failures. Shards Routers: mongos –configdb urlconfig:portconfig It is the shell for the rest of clients. It can be executed in an application server for reduce the load of the server. Config Routers:mongod –configsvr Recommended a group of 3
- 15. Important note _id Special key Present in all stored documents Unique to each collection It can be the type you want
- 16. First step with sharding use admin db.runCommand( { addshard : "premongo1:10000" } ); db.runCommand( { addshard : "premongo2:10000" } ); db.runCommand( { enablesharding : "activities" } ); db.runCommand( { shardcollection : "activities.userWall" , key : { userId: 1 } } ); db.printShardingStatus()
- 17. Backup EC2 → Snapshot EBS LVM→ Snapshot LV FS→mysqldump, dd (with sharding not recommended) Weak point of sharding. More details > http://bit.ly/wNdlIA
- 18. Questions?
- 19. Who am I? Spanish web: http://www.juanvicenteherrera.es English web: http://www.juanvicenteherrera.eu Twitter: http://twitter.com/jvicenteherrera LinkedIn: http://es.linkedin.com/in/jvherrera