Mysql High Availability With Group Replication

#MySQL #oow16
MySQL High Availability

with Group Replication
Nuno Carvalho (nuno.carvalho@oracle.com)
Principal Software Engineer, MySQL Replication Service Team Lead
Copyright
Copyright2016,
2016,Oracle
Oracleand/or
and/oritsitsaffiliates.
affiliates.All
Allrights
rightsreserved.
reserved.|
Safe Harbor Statement
The following is intended to outline our general product direction. It is intended for
information purposes only, and may not be incorporated into any contract. It is not a
commitment to deliver any material, code, or functionality, and should not be relied upon
in making purchasing decisions. The development, release, and timing of any features or
functionality described for Oracles products remains at the sole discretion of Oracle.
Wednesday, September 21, 2016 Copyright 2016, Oracle and/or its affiliates. All rights reserved. | Oracle Open World 2016 - San Francisco, California, US 2
Program Agenda
Wednesday, September 21, 2016 Copyright 2016, Oracle and/or its affiliates. All rights reserved. | Oracle Open World 2016 - San Francisco, California, US
Program Agenda
1 Background
2 MySQL Group Replication
3 Architecture
4 Use cases
5 Performance
6 Conclusion
1
Background
Background: What is Replication Used For?
Read scale-out
More
reads? M S
M S More
slaves!
S
write clients
write clients read clients
read clients
Background: What is Replication Used For?
Redundancy: If master crashes, promote slave to master
B B B
Uh Oh! Whew!
A A
Crash A
C C C
B is the
new master
MySQL Group Replication
What is MySQL Group Replication?
Multi-master update everywhere replication plugin for MySQL with built-in automatic
distributed recovery, conflict detection and group membership.
What does the MySQL Group Replication plugin do for the user?
Removes the need for handling server fail-over.
Provides fault tolerance.
Enables update everywhere setups.
Automates group reconfiguration (handling of crashes, failures, re-connects).
Provides a highly available replicated database.
Clients
M M M M M
Replication Group
Timeline - Server GA and GR Releases
MySQL 5.6.10 GR 0.6.0 labs

multi-platform support
GR 0.9.0 labs (RC)
Paxos-based consensus
GR 0.2.0 labs GR 0.5.0 labs multi-threaded applier support
Hello world! auto-inc fields handling single primary mode
GR 0.4.0 labs recovery enhancements
version handling GR 0.8.0 labs (beta)
GR 0.7.0 labs performance enhancements
bug fixes replication stream compression
GR 0.3.0 labs SSL support
support for corosync 2.x MySQL 5.7.9
lifecycle interfaces IP whitelisting
P_S tables for GR read-only mode
Server side changes error log enhancements
split brain handling
2
2.1 Multi-Master
Multi-Master update everywhere!
Any two transactions on different servers can write to the same tuple.
Conflicts will be detected and dealt with.
First committer wins rule.
UPDATE t1 SET a=3 WHERE a=1 UPDATE t1 SET a=4 WHERE a=2
M M M M M
http://mysqlhighavailability.com/mysql-group-replication-transaction-life-cycle-explained/
OK
UPDATE t1 SET a=3 WHERE a=1
OK
UPDATE t1 SET a=4 WHERE a=2
M M M M M
M M M M M
OK
M M M M M
2
2.1 Multi-Master
2.2 Automatic distributed server recovery
Automatic distributed server recovery!
Server that joins the group will automatically synchronize with the others.
I want to play with you
M M M M M N
http://mysqlhighavailability.com/distributed-recovery-behind-the-scenes/
ONLINE
RECOVERING
M M M M M N
ONLINE
M M M M M N
If a server leaves the group, the others will automatically be informed.
Each membership configuration My machine needs maintenance

is identified by a view_id or a system crash happens
M M M M M M
view_id: 4
If a server leaves the group, the others will automatically be informed.
M M M M M
Server that (re)joins the group will automatically synchronize with the
others.
RECOVERING -> ONLINE
M M M M M M
view_id: 5
2
2.1 Multi-Master
2.3 MySQL/InnoDB look & feel
MySQL look & feel!
MySQL Plugin
Regular MySQL Plugin. Nothing new.
MySQL InnoDB
Use InnoDB as normally you would. Nothing new.
Transparent optimizations in InnoDB to better support Group Replication.
MySQL Performance Schema

Monitor Group Replication using regular Performance Schema tables. Nothing new.
http://mysqlhighavailability.com/mysql-group-replication-monitoring/
MySQL look & feel!
Outcome
Group Replication is no alien component.
Existing MySQL users feel right at home.
New MySQL users only have to learn MySQL tech, nothing else.
http://mysqlhighavailability.com/mysql-group-replication-monitoring/
Full GTID support!
All group members share the same UUID, the group name.
INSERT x;
Will have GTID: group_name:1
INSERT y;
M M M M M
Full GTID support!
Users can specify the identifier for the transaction.
SET GTID_NEXT= UUID:50

INSERT x;
INSERT y;
M M M M M
Full GTID support!
You can even replicate from a outside server to a group, global identifiers
will be preserved.
Conflicts will be detected!
M M M M M
Full GTID support!
You can also replicate from a group to a outside server, global identifiers
will be preserved.
M M M M M
2
2.1 Multi-Master
2.4 Full GTID support
2.5 Auto-increment configuration/handling
Auto-increment configuration/handling
Group is configured not to generate the same auto-increment value on all
members.
INSERT y;
INSERT x; y: 4
x: 1
INSERT z;
z: 11
M M M M M
By default, the offset is provided by server_id and increment is 7 [1].
INSERT y;
INSERT x; y: 4
x: 1
INSERT z;
z: 11
server_id: 1
server_id: 4
M M M M M
[1]: http://mysqlhighavailability.com/mysql-group-replication-auto-increment-configuration-handling/
Users can change the increment size to their needs using
GROUP_REPLICATION_AUTO_INCREMENT_INCREMENT option.
INSERT y;
INSERT x; y: 4
x: 1
INSERT z;
z: 11
server_id: 1
server_id: 4
M M M M M
2
2.1 Multi-Master
2.6 Plugin version access control
Plugin Version Access Control
When joining, versions are crucial when determining if a member is
compatible with a group.
Member with different patch

version is allowed to join
5.7.13 5.7.13 5.7.14 5.7.15 5.7.12
Member with lower major version

than the major version in
the group is not allowed to join
8.0.0 8.0.0 8.0.0 8.0.0 5.7.15
Member with higher major

version is allowed to join but
is not allowed to do writes
5.7.14 5.7.14 5.7.15 5.7.15 8.0.0
2
2.1 Multi-Master
2.6 Plugin version access control
2.7 Built-in communication engine
Built-in Communication Engine
Feature rich new replication plugin based on proven distributed systems
algorithms (Paxos).
Compression, multi-platform, dynamic membership, distributed agreement, quorum
based message passing, SSL, IP whitelisting.
No third-party software required.
No network multicast support required.
MySQL Group Replication can operate on cloud based installations on which multicast
is not allowed.
http://mysqlhighavailability.com/mysql-group-replication-for-mysql-5-7-9/
2
2.8 Read-only mode
Read-only mode
When a member joins the group, during distributed recovery, read-only
mode is set automatically.
On a unlikely event of a member failure, read-only mode is set
automatically to prevent inconsistency with the group and member state
changes to ERROR.
M M M M M
http://mysqlhighavailability.com/mysql-group-replication-important-behavior-and-configuration-changes/
2
2.8 Read-only mode
2.9 Full stack secure connections
Full stack secure connections
Following the industry standards, Group Replication supports secure
connections along the complete stack
Client connections
Distributed recovery connections
Connections between members
IP Whitelisting
Restrict which hosts are allowed to connect to the group
By default it is set to the value AUTOMATIC, which allows connections from private
subnetworks active on the host
http://mysqlhighavailability.com/mysql-group-replication-securing-the-perimeter/
2
2.8 Read-only mode
2.10 Parallel applier support
Parallel applier support
Group Replication now also takes full advantage of parallel binary log
applier infrastructure
Reduces applier lag and improves replication performance considerably
Configured in the same way as in asynchronous replication
--slave_parallel_workers=NUMBER
--slave_parallel_type=logical_clock
--slave_preserve_commit_order=ON
2
2.8 Read-only mode
2.11 Single Primary Mode
Single Primary Mode
Configuration mode that makes a single member act as a writeable master
(PRIMARY) and the rest of the members act as hot-standbys (SECONDARY).
The group itself coordinates automatically to figure out which is the member that will
act as the PRIMARY, through a leader election mechanism.
Default mode
Closer to classic asynchronous replication setups, simpler to reason about from the
beginning.
Avoids some of the limitations of multi-master mode.
Single Primary Mode
Automatic leader election mechanism
Secundaries are automatically set to read-only
Primary
S S S S P S
Single Primary Mode
Automatic leader election mechanism
Primary
S S S P S
Single Primary Mode
The current primary member UUID can be known by executing the
following SQL statement
mysql> SELECT * FROM performance_schema.global_status WHERE

VARIABLE_NAME='group_replication_primary_member';
VARIABLE_NAME VARIABLE_VALUE
group_replication_primary_member dcd3b36b-79c5-11e6-97b8-00212844d44e
2
2.8 Read-only mode
2.11 Single Primary Mode
2.12 Requirements
Requirements (by design) Forbidden
Requires InnoDB storage engine Serializable (on multi-master)
Primary key is required on every table Cascading Foreign Keys (on multi-master)
Requires global transaction identifiers Transaction savepoints
turned on
Binary log events checksum
Requires binary log turned on
Requires binary log row format
Optimistic execution: transactions may Warnings
abort on COMMIT due to conflicts
with concurrent transactions on other Concurrent DDL (on multi-master)
members
Up to 9 servers in the group
3
Architecture
3.1 Introduction
MySQL Group Replication is
Built on top of proven technology!
Shares many pieces of MySQL Replication.
Multi-Master approach to replication.
Built on reusable components!
Layered implementation approach.
Interface driven development.
Decoupled from the server core.
The plugin registers as listener to server events.
Reuses the capture procedure from regular replication.
Provides further decoupling from the communication infrasctructure.
3
Architecture
3.1 Introduction
3.2 Major Building Blocks
Major Building Blocks (1)
MySQL
Server
API
Replication
Plugin
Com. API
M M M M M
Group Comm.
Group Com. Engine
System (Corosync)
Server calls into the plugin through a generic
interface
MySQL
Server (Most of server) internals are hidden from the plugin.
Some of the semi-sync interfaces were reused. Others
were created.
API
Replication Plugin interacts with the server through a generic

Plugin
interface
Com. API
Replication plugin determines the fate of the commit
operation through a well defined server interface.
Group Comm.
Group Com. Engine The plugin makes use of the relay log infrastructure to
System (Corosync)
inject changes in the receiving server.
The plugin is responsible for
MySQL Maintaining distributed execution context.
Server
Detecting conflicts.
API Handling distributed recovery:
Detect membership changes;
Replication Donate state if needed;
Plugin Collect state if needed.
Receiving and handling transactions from other
Com. API
members.
Deciding the fate of on-going transactions.
Group Comm.
Group Com. Engine
System (Corosync)
The communication API (and bindings) is
MySQL responsible for:
Server
Abstracting the underlaying communication engine
from the plugin itself.
API
Replication Mapping the interface to a specific communication
Plugin engine.
Com. API
Group Comm.
Group Com. Engine
System (Corosync)
The communication engine:
MySQL Variant of Paxos developed at MySQL.
Server
Building block to provide distributed agreement
API between servers.
Replication
Plugin
Com. API
Group Comm.
Group Com. Engine
System (Corosync)
3
Architecture
3.1 Introduction
3.2 Major Building Blocks
3.3 The Complete Stack
The Complete Stack Performance Schema Tables: Monitoring
MySQL
MySQL
Server
APIs: Lifecycle / Capture / Applier
API
Conflicts
Replication Capture Applier Recovery
Plugin
InnoDB Handler
Plugin
Replication Protocol
API
Group Com. API
Group Com. Binding

Group Comm.
Group Com. Engine
System (Corosync) Group Com. Engine
Network
4
Use cases
Use Cases
Elastic Replication
Environments that require a very fluid replication infrastructure, where the number
of servers has to grow or shrink dynamically and with as little pain as possible.
M M M M M
Use Cases
Highly Available Shards
Sharding is a popular approach to achieve write scale-out. Users can use MySQL
Group Replication to implement highly available shards. Each shard can map into a
Replication Group.
M M M M M M M M M M M M M M M
Use Cases
Alternative to Master-Slave replication
Single Primary mode provides further automation on such setups
Automatic PRIMARY/SECONDARY roles assignment
Automatic new PRIMARY election on PRIMARY failures
Automatic setup of read/write modes on PRIMARY and SECONDARIES
Global consistent view of which server is the PRIMARY
S S S S P S
Primary
5
Performance
Performance
Group Replication Throughput Peak Throughput
(as perceived by the client application) The number of transactions that
writers can propagate to the group
Single-master Sustained Single-master Peak Multi-master Sustained (per second).
Multi-master Peak Single-server (MySQL 5.7.14) Sustained Throughput
30000 The number of transactions that
25000
can be propagated to the group
without increasing the replication
THROUGHPUT (TPS)
20000 lag on any member (per second).

15000
Servers
10000 9 Dual Xeon E5-2660-v3
5000 Enterprise SSD Storage
10Gbps Ethernet Network
2 3 5 7 9 Client
NUMBER OF SERVERS IN THE GROUP. 1 Dual Xeon E5-2699-v3
Replication Performance blogs at: 10Gbps Ethernet Network
http://mysqlhighavailability.com/category/performance/ Sysbench 0.5 RW workload
Performance
On a sustained throughput:
Multi-master performance degrades gracefully while going from a group with 2
servers to a group with 9 servers.
Single primary performance degrades marginally when growing the group size.
On a peak throughput:
Multi-master exhibits 1.8X speedup when compared to the single server.
Read load is balanced across the servers in the group.
Write load is lower since execution is balanced across the group, whereas in single primary mode the
primary becomes a bottleneck.
With a single primary there is no lag on the other members.
6
Conclusion
Summary
Cloud Friendly
Great techonology for deployments where elasticity is a requirement, such as cloud
based infrastructures.
Integrated
With server core through a well defined API.
With GTIDs, row based replication, performance schema tables.
Autonomic and Operations Friendly
It is self-healing: no admin overhead for handling server fail-overs.
Provides fault-tolerance, enables multi-master update everywhere and a dependable
MySQL service.
Lab releases provide a sneak peek at what is coming - a new replication plugin which is
the core for the new MySQL InnoDB Cluster.
MySQL InnoDB Cluster: The End Goal
Application Application Application Application
MySQL Connector MySQL Connector MySQL Connector MySQL Connector
MySQL Router MySQL Router MySQL Router MySQL Router
M HA M HA M HA
MySQL Shell
M M M M M M
ReplicaSet 1
ReplicaSet 2
ReplicaSet 3
S1 S2 S3 S4 S S1 S2 S3 S4 S S1 S2 S3 S4
Where to go from here?
Packages
http://labs.mysql.com
Blogs from the Engineers (news, technical information, and much more)
http://mysqlhighavailability.com

Mysql High Availability With Group Replication

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Mysql High Availability With Group Replication

Загружено:

Авторское право:

Доступные форматы

#MySQL #oow16

MySQL High Availability

MySQL 5.6.10 GR 0.6.0 labs

I want to play with you

Each membership configuration My machine needs maintenance

RECOVERING -> ONLINE

MySQL Performance Schema

SET GTID_NEXT= UUID:50

Conflicts will be detected!

Member with different patch

5.7.13 5.7.13 5.7.14 5.7.15 5.7.12

Member with lower major version

8.0.0 8.0.0 8.0.0 8.0.0 5.7.15

Member with higher major

5.7.14 5.7.14 5.7.15 5.7.15 8.0.0

mysql> SELECT * FROM performance_schema.global_status WHERE

Replication Plugin interacts with the server through a generic

Group Com. Binding

20000 lag on any member (per second).

MySQL Router MySQL Router MySQL Router MySQL Router

Вам также может понравиться