MongoDB: Optimising for Performance, Scale & Analytics
•
1 like•557 views
MongoDB is easy to download and run locally but requires some thought and further understanding when deploying to production. At scale, schema design, indexes and query patterns really matter. So does data structure on disk, sharding, replication and data centre awareness. This talk will examine these factors in the context of analytics, and more generally, to help you optimise MongoDB for any scale. Presented at MongoDB Days London 2013 by David Mytton.
Report
Share
MongoDB: Optimising for Performance, Scale & Analytics
- 1. Optimising for performance, scale, analytics
- 2. ~3333 write ops/s 0.07 - 0.05 ms response
- 5. MongoDB at Server Density
- 6. MongoDB at Server Density •27 nodes
- 7. MongoDB at Server Density •27 nodes • June 2009 - +3yrs
- 8. MongoDB at Server Density •27 nodes • June 2009 - +3yrs •MySQL -> MongoDB
- 9. MongoDB at Server Density •27 nodes • June 2009 - +3yrs •MySQL -> MongoDB •17TB data per month
- 10. MongoDB at Server Density Queues Primary data store Time series
- 11. Why?
- 13. Why? • Replication • Official drivers
- 14. Why? • Replication • Official drivers • Easy deployment
- 15. Why? • Replication • Official drivers • Easy deployment • Fast out of the box (sort of)
- 16. Fast out of the box? Photo: dannychoo.com
- 17. ~3333 write ops/s 0.07 - 0.05 ms response
- 18. Fast out of the box? • Softlayer cloud (1 core, 8GB) Photo: dannychoo.com
- 19. Fast out of the box? • Softlayer cloud (1 core, 8GB) • Local instance storage Photo: dannychoo.com
- 20. Fast out of the box? • Softlayer cloud (1 core, 8GB) • Local instance storage • Ubuntu 10/12.04 LTS Photo: dannychoo.com
- 22. Fast out of the box? • Journaling Photo: dannychoo.com
- 23. Fast out of the box? • Journaling • Replication Photo: dannychoo.com
- 24. Fast out of the box? Picture is unrelated! Mmm, ice cream.
- 25. Fast out of the box? • Fast network Picture is unrelated! Mmm, ice cream.
- 26. Fast out of the box? • Fast network • Working set in RAM Picture is unrelated! Mmm, ice cream.
- 27. mongos> db.metrics_20120508_15_1m.stats() { "sharded" : true, "flags" : 1, "ns" : "metrics.metrics_20120508_15_1m", "count" : 2752934, "numExtents" : 46, "size" : 746837640, "storageSize" : 823717888, "totalIndexSize" : 517581680, "indexSizes" : { "_id_" : 130358144, "a_1_i_1" : 155711920, "a_1_i_1_m_1_t_1" : 231511616 }, "avgObjSize" : 271.2878841265355, "nindexes" : 3, "nchunks" : 61,
- 28. "size" : 746837640, "totalIndexSize" : 517581680 Indexes = 493MB Data = 712MB
- 29. "size" : 746837640, "totalIndexSize" : 517581680 Indexes = 493MB Data = 712MB
- 30. "size" : 746837640, "totalIndexSize" : 517581680 Indexes = 493MB Data = 712MB Total = 1205MB
- 31. Where should it go? Should it be in What? memory? Indexes Always Data If you can http://www.flickr.com/photos/comedynose/4388430444/
- 32. How you’ll know 1) Slow queries Thu Oct 14 17:01:11 [conn7410] update sd.apiLog query: { c: "android/setDeviceToken", a: 1466, u: "blah", ua: "Server Density Android" } 51926ms www.flickr.com/photos/tonivc/2283676770/
- 33. How you’ll know 2) Timeouts cursor timed out (20000 ms)
- 34. How you’ll know 3) Disk i/o spikes www.flickr.com/photos/daddo83/3406962115/
- 35. Fast out of the box? • Fast network • Working set in RAM Picture is unrelated! Mmm, ice cream.
- 36. Fast out of the box? • Fast network • Working set in RAM • Fast disks (optional) Picture is unrelated! Mmm, ice cream.
- 39. Growing documents is bad
- 40. Updates in place db.my_collection.update( { _id : ... }, { $inc : { y : 2 } } )
- 41. BSON
- 42. BSON
- 43. Growing documents is bad > db.coll.stats() { "ns" : "...", ..., "paddingFactor" : 1, ..., "ok" : 1 }
- 44. Scaling writes • Global DB lock
- 45. Scaling writes • Global DB lock • Concurrency
- 49. Fast out of the box? • Fast network • Working set in RAM • Fast disks (optional) • Sharding (optional) Picture is unrelated! Mmm, ice cream.
- 50. Scaling writes • Global DB lock • Concurrency • Sharding
- 51. Scaling writes
- 53. Scaling writes • Collection location
- 54. Scaling writes • Collection location • Pre-split / moveChunk
- 55. Scaling writes • Collection location • Pre-split / moveChunk • Hashing (v2.4)
- 57. Failover •Replica sets •Master/slave
- 58. Failover •Replica sets •Master/slave •Min 3 nodes
- 59. Failover •Replica sets •Master/slave •Min 3 nodes •Automatic failover
- 60. rs.status() { ! "_id" : 1, ! "name" : "rs3b:27018", ! "health" : 1, ! "state" : 2, ! "stateStr" : "SECONDARY", ! "uptime" : 1886098, ! "optime" : { ! ! "t" : 1291252178000, ! ! "i" : 13 ! }, ! "optimeDate" : ISODate("2010-12-02T01:09:38Z"), "lastHeartbeat" : ISODate("2010-12-02T01:09:38Z") }, www.ex-astris-scientia.org/inconsistencies/ent_vs_tng.htm (yes it’s a replicator from Star Trek)
- 61. rs.status() 1) myState Value Meaning 0 Starting up (phase 1) 1 Primary 2 Secondary 3 Recovering 4 Fatal error 5 Starting up (phase 2) 6 Unknown state 7 Arbiter 8 Down
- 62. rs.status() 2) Optime "optimeDate" : ISODate("2010-12-02T01:09:38Z") www.flickr.com/photos/robbie73/4244846566/
- 63. rs.status() 3) Heartbeat "lastHeartbeat" : ISODate("2010-12-02T01:09:38Z") www.flickr.com/photos/drawblindfaith/3400981091/
- 64. Scaling reads
- 67. Scaling reads •Replica slaves •Consistency •w flag / tags
- 68. WriteConcern Changed Nov 27 2012
- 69. WriteConcern • Safe by default >>> from pymongo import MongoClient >>> connection = MongoClient()
- 70. WriteConcern • Safe by default >>> from pymongo import MongoClient >>> connection = MongoClient(w=int/str) Value Meaning 0 Unsafe 1 Primary 2 Primary + x1 secondary 3 Primary + x2 secondaries
- 71. Tags { _id : "someSet", members : [ {_id : 0, host : "A", tags : {"dc": "ny"}}, {_id : 1, host : "B", tags : {"dc": "ny"}}, {_id : 2, host : "C", tags : {"dc": "sf"}}, {_id : 3, host : "D", tags : {"dc": "sf"}}, {_id : 4, host : "E", tags : {"dc": "cloud"}} ] settings : { getLastErrorModes : { veryImportant : {"dc" : 3}, sortOfImportant : {"dc" : 2} } } } > db.foo.insert({x:1}) > db.runCommand({getLastError : 1, w : "veryImportant"})
- 72. Tags { _id : "someSet", (A or B) + (C or D) + E members : [ {_id : 0, host : "A", tags : {"dc": "ny"}}, {_id : 1, host : "B", tags : {"dc": "ny"}}, {_id : 2, host : "C", tags : {"dc": "sf"}}, {_id : 3, host : "D", tags : {"dc": "sf"}}, {_id : 4, host : "E", tags : {"dc": "cloud"}} ] settings : { getLastErrorModes : { veryImportant : {"dc" : 3}, sortOfImportant : {"dc" : 2} } } } > db.foo.insert({x:1}) > db.runCommand({getLastError : 1, w : "veryImportant"})
- 73. Tags { _id : "someSet", (A + C) or (D + E) ... members : [ {_id : 0, host : "A", tags : {"dc": "ny"}}, {_id : 1, host : "B", tags : {"dc": "ny"}}, {_id : 2, host : "C", tags : {"dc": "sf"}}, {_id : 3, host : "D", tags : {"dc": "sf"}}, {_id : 4, host : "E", tags : {"dc": "cloud"}} ] settings : { getLastErrorModes : { veryImportant : {"dc" : 3}, sortOfImportant : {"dc" : 2} } } } > db.foo.insert({x:1}) > db.runCommand({getLastError : 1, w : "sortOfImportant"})
- 74. WriteConcern • Safe by default •J
- 75. WriteConcern • Safe by default •J • fsync
- 76. Bottlenecks •EC2
- 77. Bottlenecks •EC2 •Local storage
- 78. Bottlenecks •EC2 •Local storage •EBS: RAID10 4-8 volumes
- 79. Bottlenecks •EC2 •Local storage •EBS: RAID10 4-8 volumes •i/o: rand but not sequential
- 82. Bottlenecks •CPU •Index building
- 85. Tips: rand() •_id •Field names •Covered indexes
- 86. Tips: rand() •_id •Field names •Covered indexes •Collections / databases
- 87. mongostat
- 90. Current operations db.currentOp(); { ! ! ! "opid" : "shard1:299939199", ! ! ! "active" : true, ! ! ! "lockType" : "write", ! ! ! "waitingForLock" : false, ! ! ! "secs_running" : 15419, ! ! ! "op" : "remove", ! ! ! "ns" : "sd.metrics", ! ! ! "query" : { ! ! ! ! "accId" : 1391, ! ! ! ! "tA" : { ! ! ! ! ! "$lte" : ISODate("2010-11-24T19:53:00Z") ! ! ! ! } ! ! ! }, ! ! ! "client" : "10.121.12.228:44426", ! ! ! "desc" : "conn" ! ! }, www.flickr.com/photos/jeffhester/2784666811/
- 91. Monitoring tools Run yourself Ganglia
- 93. Fast out of the box? • Fast network • Working set in RAM Picture is unrelated! Mmm, ice cream.
- 94. Fast out of the box? • Fast network • Working set in RAM • bit.ly/benchrun Picture is unrelated! Mmm, ice cream.