MySQL Group Replication
0 likes195 views
This document discusses MySQL Group Replication, which provides high availability and redundancy for MySQL databases. It describes the key features and architecture of MySQL Group Replication, including its support for multi-master replication where any server can handle writes, automatic recovery of servers that join or leave the replication group, and its use of GTIDs to track transactions across the group. The document is intended to outline the capabilities and benefits of MySQL Group Replication.
![Copyright © 2016, Oracle and/or its affiliates. All rights reserved. |
Some Theory Behind It...
• ImplementaIon based in Replicated Database State Machines
– Group CommunicaIon PrimiIves resemble properIes of Databases.
• Deferred update replicaIon: propagate atomically, check conflicts,
eventually apply
– Distributed state machine requires agreed delivery – implies total order;
– DeterminisIc cerIficaIon requires total order delivery.
• Membership Service
– Which servers are parIcipaIng in the replicaIon group at a given moment in Ime?
(associated with a logical Imestamp [ - view idenIfier]).
10](https://image.slidesharecdn.com/highavailabilityusingmysqlgroupreplication-170615143455/85/MySQL-Group-Replication-10-320.jpg)
MySQL Group Replication
- 9. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | MySQL Group ReplicaIon 9 M M M M M ReplicaKon Group Clients
- 10. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Some Theory Behind It... • ImplementaIon based in Replicated Database State Machines – Group CommunicaIon PrimiIves resemble properIes of Databases. • Deferred update replicaIon: propagate atomically, check conflicts, eventually apply – Distributed state machine requires agreed delivery – implies total order; – DeterminisIc cerIficaIon requires total order delivery. • Membership Service – Which servers are parIcipaIng in the replicaIon group at a given moment in Ime? (associated with a logical Imestamp [ - view idenIfier]). 10
- 12. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | MulI-Master update everywhere! • Any two transacIons on different servers can write to the same tuple. • Conflicts will be detected and dealt with. – First commiker wins rule. 12 M M M M M UPDATE t1 SET a=4 WHERE a=2 UPDATE t1 SET a=3 WHERE a=1
- 13. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | MulI-Master update everywhere! • Any two transacIons on different servers can write to the same tuple. • Conflicts will be detected and dealt with. – First commiker wins rule. 13 M M M M M UPDATE t1 SET a=4 WHERE a=2 UPDATE t1 SET a=3 WHERE a=1 OK OK
- 14. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | MulI-Master update everywhere! • Any two transacIons on different servers can write to the same tuple. • Conflicts will be detected and dealt with. – First commiker wins rule. 14 M M M M M UPDATE t1 SET a=2 WHERE a=1 UPDATE t1 SET a=3 WHERE a=1
- 15. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | MulI-Master update everywhere! • Any two transacIons on different servers can write to the same tuple. • Conflicts will be detected and dealt with. – First commiker wins rule. 15 M M M M M UPDATE t1 SET a=2 WHERE a=1 UPDATE t1 SET a=3 WHERE a=1 OK
- 17. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | AutomaKc distributed server recovery! • Server that joins the group will automaIcally synchronize with the others. 17 M M M M M N I want to play with you
- 18. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | AutomaKc distributed server recovery! • Server that joins the group will automaIcally synchronize with the others. 18 M M M M M N ONLINE RECOVERING
- 20. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | AutomaKc distributed server recovery! • If a server leaves the group, the others will automaIcally be informed. 20 M M M M M M My machine needs maintenance or a system crash happens Each membership configuraKon is idenKfied by a logical Kmestamp, i.e., view_id view_id: 4
- 21. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | AutomaKc distributed server recovery! • If a server leaves the group, the others will automaIcally be informed. 21 M M M M M view_id: 5
- 22. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | AutomaKc distributed server recovery! • Server that (re)joins the group will automaIcally synchronize with the others. 22 M M M M M M RECOVERING -> ONLINE view_id: 6
- 24. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | MySQL/InnoDB look & feel! • Load the plugin and start replicaIng. • Monitor group replicaIon stats though Performance Schema tables. 24 mysql> SET GLOBAL group_replication_group_name= "9eb07c6d-5e24-11e5-854b-34028662c0cd"; mysql> START GROUP_REPLICATION; mysql> SELECT * FROM performance_schema.replication_group_membersG *************************** 1. row *************************** CHANNEL_NAME: group_replication_applier MEMBER_ID: 597dbb72-3e2c-11e4-9d9d-ecf4bb227f3b MEMBER_HOST: nightfury MEMBER_PORT: 13000 MEMBER_STATE: ONLINE *************************** 2. row *************************** ...
- 25. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | MySQL/InnoDB look & feel! • Load the plugin and start replicaIng. • Monitor group replicaIon stats though Performance Schema tables. 25 mysql> SELECT * FROM performance_schema.replication_group_member_statsG *************************** 1. row *************************** CHANNEL_NAME: group_replication_applier VIEW_ID: 1428497631:3 MEMBER_ID: e38fdea8-dded-11e4-b211-e8b1fc3848de COUNT_TRANSACTIONS_IN_QUEUE: 0 COUNT_TRANSACTIONS_CHECKED: 12 COUNT_CONFLICTS_DETECTED: 5 COUNT_TRANSACTIONS_VALIDATING: 6 TRANSACTIONS_COMMITTED_ALL_MEMBERS: 8a84f397-aaa4-18df-89ab-c70aa9823561:1-7 LAST_CONFLICT_FREE_TRANSACTION: 8a84f397-aaa4-18df-89ab-c70aa9823561:7
- 26. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | MySQL/InnoDB look & feel! • Load the plugin and start replicaIng. • Monitor group replicaIon stats though Performance Schema tables. 26 mysql> SELECT * FROM performance_schema.replication_connection_statusG *************************** 1. row *************************** CHANNEL_NAME: group_replication_applier GROUP_NAME: 8a94f357-aab4-11df-86ab-c80aa9429563 SOURCE_UUID: 8a94f357-aab4-11df-86ab-c80aa9429563 THREAD_ID: NULL SERVICE_STATE: ON ...
- 28. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Full GTID support! • All group members share the same UUID, the group name. 28 M M M M M INSERT y; Will have GTID: group_name:2 INSERT x; Will have GTID: group_name:1
- 29. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Full GTID support! • Users can specify the idenIfier for the transacIon. 29 M M M M M INSERT y; Will have GTID: group_name:1 SET GTID_NEXT= “UUID:50” INSERT x;
- 30. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Full GTID support! • You can even replicate from a outside server to a group, global idenIfiers will be preserved. 30 M M M M M Conflicts will be detected!
- 33. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | New Distributed Agreement and CommunicaIons Engine • MulIple OS support. – Linux, but also Windows, OSX, Solaris, FreeBSD… • No third-party souware required. • No network mulIcast support required. – MySQL Group ReplicaIon can now operate on cloud based installaIons on which mulIcast is disallowed. • No message size limit. • No separate process. – MySQL Group ReplicaIon is now self-contained on the same souware stack. 33
- 35. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Requirements (by design) • Support for InnoDB only. • Primary key is required on every table. • Requires global transacIon idenIfiers turned on. • OpImisIc execuIon: transacIons may abort on COMMIT due to conflicts with concurrent transacIons on other members. 35
- 38. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Architecture IntroducKon 3 3.1
- 43. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Use Cases • ElasKc ReplicaKon – Environments that require a very fluid replicaIon infrastructure, where the number of servers has to grow or shrink dynamically and with as likle pain as possible. • Highly Available Shards – Sharding is a popular approach to achieve write scale-out. Users can use MySQL Group ReplicaIon to implement highly available shards. Each shard can map into a ReplicaIon Group. • AlternaKve to Master-Slave replicaKon – It may be that a single master server makes it a single point of contenIon. WriIng to an enIre group may prove more scalable under certain circumstances. 43
- 45. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Dependable and Scalable MySQL Setups M App M M Orchestrate & Manage M App M M Simple Shard Mapping, State and Extra metadata. Control, Coordinate, Provision ... Monitoring (MEM) MySQL Router Group ReplicaIon – Shard 1 Group ReplicaIon – Shard N C, PHP, Perl, ... 45 MySQL Router
- 47. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Summary • Cloud Friendly – Great techonology for deployments where elasIcity is a requirement, such as cloud based infrastructures. • Integrated – With server core through a well defined API. – With GTIDs, row based replicaIon, performance schema tables. • Autonomic and OperaKons Friendly – It is self-healing: no admin overhead for handling server fail-overs. – Provides fault-tolerance, enables mulI-master update everywhere and a dependable MySQL service. 47