1 architecture & design

MySQL Webinar Series 1/4
Architecture & Design
Mark Swarbrick
Principle Presales Consultant

Program
Agenda
§ Causes, Effects and Impacts of Downtime
§ Methodology to Select Right HA Solution
§ Guided Tour of MySQL HA Solutions
§ Operational Best Practices

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
decision. The development, release, and timing of any features or
functionality described for Oracle’s products remains at the sole
discretion of Oracle.

40% DATA GROWTH
PER YEAR
5.9BN MOBILE SUBS IN 2011
1 BILLION iOS & ANDROID APPS
DOWNLOADED PER WEEK
$1TR BY 2014
$700BN IN 2011
85% HANDSETS SHIPPED
WITH A BROWSER
2.1BN USERS
8X DATA GROWTH IN 5 YRS
70+ NEW DOMAINS EVERY 60 SECONDS
350m TWEETS PER
DAY
955M USERS
20M APPS PER DAY
60 HOURS
UPLOADED EVERY
MINUTE

Industry Leaders Rely on MySQL
Web & Enterprise OEM & ISVs
Cloud

The Lowdown on Downtime….
Failures Disasters Maintenance Operator / User
Revenue Customer Sat Employee Regulatory
CAUSE
EFFECT
IMPACT
Copyright © 2012, Oracle and/or its affiliates. All rights reserved. September 29th, 2012

HA Considerations
• SLA requirements to support business objectives
• Operational capabilities
• Service agility & time to market
• Budgetary constraints

Not Everything Needs 99.999% Uptime
• Recovery Time Objective
– Maximum length of downtime before there is break in “business continuity”
• Recovery Point Objective
– Point in time to which data must be recovered when service is re-established
• Tier 1, Mission-critical services
• Tier 2, Business-critical services
• Tier 3, Task-critical services
• Tier 4, Non-critical services

Mapping Uptime to Availability
9" 9" 9" 9" 9" %"
35 days 4 days 8 hours 50 mins 5 mins
Cost & Complexity
SPs &
Line of
Business
Web &
Cloud
Services
eCommerce
Telecoms
Military
.
Shared-Nothing,
Geo-Replicated
Clusters
Clustered &
Virtualized Systems
Replicated Systems

MySQL HA
Solutions
9" 9" 9" 9" 9" %"
35 days 4 days 8 hours 50 mins
Operational Complexity
SPs &
Line of
Business
Web &
Cloud
Services
eCommerce
Telecoms
Military
.
Replication
5 mins

Scale Up
• Proprietary SMP hardware
• Closed or Open Source software
• Platform lock-in
• “Fork Lift” to increase capacity
Scalability with MySQL
Scale Out
• Commodity Intel / AMD
• Open source software
• Platform freedom
• Add commodity servers to increase
capacity

Why Replicate?
• Duplicates database from a “master” to a “slave”
– Redundant copies of the data provide foundation for High Availability
– Scale out by distributing queries across the replication cluster
Master
Slaves
Web / App Servers
Writes & Reads Reads

Slave
Database
Replication Workflow
• Session thread: processes queries from
the application – writes data to master
database & associated events to binary
log
• Dump thread: reads events from binary log
and sends them to a slave
• I/O thread: receives replication events and
stores them in slave’s relay log
• SQL thread: reads replication events from
slave’s relay log and applies them to slave
database
Session
Binary
Log
Master
Database
Dump I/O
Relay
Log
SQL

Current Generation of MySQL Replication
Pros
§ Simple to configure
§ Heterogeneous
§ Low TCO – no shared storage
§ Configure over LAN and WAN
§ Fast failover
Cons
§ Asynchronous: risk of of data loss
§ Semi-Synchronous is configurable
§ No automated failover or
switchover
§ Performance overhead to master

MySQL 5.6: Evolving Replication
PERFORMANCE
• Multi-Threaded Slaves
• Binary Log Group Commit
• Optimized Row-Based Replication
FAILOVER & RECOVERY
• Global Transaction Identifiers
• Replication Failover & Admin Utilities
• Crash Safe Slaves & Binlogs
DATA INTEGRITY
• Replication Event Checksums
DEV/OPS AGILITY
• Replication Utilities
• Time-Delayed Replication
• Remote Binlog Backup
• Informational Log Events
• Server UUIDs

Global Transaction Identifiers
• Simple to track & compare replication across the cluster
– Unique identifier for each transaction written to the Binlog
• Automatically identify the most up-to-date slave for failover
• Deploy n-tier replication hierarchies
Master
GTID=123456
GTID=123456
GTID=123456 GTID=123456

Replication HA Utilities
Promoted
Master
• Enabling self-healing replication topologies
• Automated failover & recovery
– mysqlfailover Utility
• Switchover & administration
– mysqlrpladmin Utility
Monitoring HA Utilities
• Delivers HA within the core MySQL
distribution
Failed
Master
Slaves

Clustering Overview
• Tightly coupled clusters of servers/storage providing service to an
application
– Use heartbeating and management software to monitor hardware, OS,
database and network
– Node failures detected and failover automatically initiated to redundant
systems
– Implements Virtual IP (VIP) so failover is transparent to the application
• Couple with Virtualization for higher agility
– Migrate between servers with load balancing

Oracle Certified Solutions
Oracle
VM
Template
Solaris
Cluster
Windows
Cluster DRBD
Auto-‐Failover
&
Migra1on
✔
✔
✔
✔
Zero
Data
Loss
✔
✔
✔
✔
Ac1ve
/
Passive
✔
✔
✔
✔
Pla:orm
Support
Linux
Solaris
Windows
Linux
Shared
Storage
✔
✔
✔
✖
Single
Vendor
Support
✔
✔
✖
✔

3rd Party HA Solutions
• Oracle supports MySQL on open source and commercial
HA technologies
– Support for HA mechanisms from respective communities or
vendors
• Symantec Veritas Cluster
• Red Hat Cluster Suite

MySQL Fabric
An extensible and easy-to-use
framework for
managing a farm of MySQL
server supporting high-availability
Copyright Oracle Corporation 24 2014 | <date>
and sharding

MySQL Fabric 1.4
High Availability + Sharding-Based Scale-out
Application
Application
§ High Availability:
– Server monitoring with auto-promotion
and transparent application failover
§ Optionally scale-out through sharding
– Application provides shard key
– Range or Hash
– Tools for resharding
– Global updates & tables
§ Fabric-aware connectors rather than
proxy: Python, Java & PHP
– Lower latency, bottleneck-free
MySQL Fabric
Connector
Read-slaves
mappings
SQL
HA group
Read-slaves
HA group
Connector

Asynchronous vs. Synchronous Replication
§ Asynchronous (MySQL Default)
– In parallel: Master acks to app and sends transaction to slave
§ Fast
§ Risk of lost changes if master dies
§ Synchronous (Only available with MySQL Cluster)
– Serially: Master waits for change to be applied on all slaves before ack
to app
§ Higher latency
§ If Active/Active, best suited to small transactions
§ Loss-less
§ Semi-Synchronous (MySQL 5.5+ - Enhanced in MySQL 5.7)
– Serially: Master waits for change to be received by slave then In
parallel ack to app and apply changes on slave
§ Intermediate latency
§ Lossless

MySQL Fabric Framework (HA)
All Data
Primary Secondary
Extra Read
Replicas
MySQL Fabric
Controller
SQL Queries
State &
Routing Info
HA Group
Coordination
and Control

MySQL Replication & MySQL Fabric HA
& how this effects failover
§ MySQL Replication is the initial implementation used in HA Groups
– PRIMARY = Replication Master & receives all writes
§ Failover
– MySQL Fabric detects failure of PRIMARY/Master
– Selects a SECONDARY/Slave and promotes it
– Updates State Store
– Updated state fetched by Fabric-aware connectors

High-Availability Group Concept
§ Abstract Concept
– Set of servers
– Server attributes
§ Connector Attributes
– Connection information
– Mode: read-only, read-write, ...
– Weight: distribute load
§ Management Attributes
– State: state/role of the server State: Primary
Mode: Read-Write
Host: server-1.example.com

The Path to Scalability
Scaling-Up can take you a long way
Scaling on dense,
multi-core, multi-thread
servers
• 10s - 100GBs RAM
• SSDs
Scale across cores
within a single instance
You can get a long
way with MySQL 5.6!

MySQL Fabric Features
§ Connector API Extensions
– Support Transactions
– Support full SQL
§ Decision logic in connector
– Reducing network load
§ Shard Multiple Tables
– Using same key
§ Global Updates
– Global tables
– Schema updates
§ Sharding Functions
– Range
– (Consistent) Hash
§ Shard Operations
– Shard move
– Shard split

MySQL Fabric Node
MySQL
Protocols
? ?
MySQL Fabric
Framework
Executor State Store
(Persister)
Sh
HA
AMQP MySQL XML-RPC
Extensions
Backing
Store
Extensible Architecture

MySQL Fabric: Goals & Features
§ Connector API Extensions
– Support Transactions
– Support full SQL
§ Fabric-Aware Connectors at GA:
– PHP + Doctrine, Python, Java +
Hibernate
§ Decision logic in connector
– Reducing network load
§ Load Balancing
– Read-Write Split
– Distribute transactions
§ Global Updates
– Global tables
– Schema updates
§ Shard Multiple Tables
– Using same key
§ Sharding Functions
– Range
– (Consistent) Hash
§ Shard Operations
– Shard move
– Shard split

MySQL Fabric – Current Limitations
§ Routing is dependent on Fabric-aware
connectors
– Currently Java (+ Hibernate), PHP (+
Doctrine) & Python
§ MySQL Fabric node is a single (non-redundant
process)
– HA Maintained as connectors
continue to route using local caches
§ Establishes asynchronous replication
– Manual steps to switch to
semisynchronous
§ Sharding not completely
transparent to application (must
provide shard key – column
from application schema)
§ No cross-shard joins or other
queries
§ Management is through CLI or
XML/RPC API
– No GUI

Who’s Using MySQL Cluster?

MySQL Cluster Architecture:
High Performance
Data Layer
MySQL Cluster Data Nodes
Clients
Applica1on
Layer
Management
Management

MySQL Cluster Architecture:
Extreme Resilience
Data Layer
Clients
Applica1on
Layer
Management
Management

Multi-Data Availability
Geographic Replication
• DR and Data Locality
• Replicate complete clusters
across data centers
– Fully active/active
– No passive resources
• Split individual clusters across
data centers
– Synchronous replication &
auto-failover between sites
Geographic
Replication

Handling Scheduled Maintenance
On-Line Operations
Scale
Backup
Upgrade
Evolve

SLAs
Platforms
Application requirements (storage engines)
Performance expectations vs tolerance to data loss
Skill sets

Mapping Applications to HA Technology
Applica'ons Database/
Replica'on
Clustered///
Virtualized
Shared:Nothing,/
Geo:Replicated/
Cluster
E"Commerce)/)Trading ))))))))))))))))(1))
Session)Management ))))))))))))))))(1))
User)Authen<ca<on)/)Accoun<ng )))))))))))))))))(1))
Feeds,)Blogs,)Wikis
OLTP )))))))))))))))))(1))
Data)Warehouse/BI
Content)Management
CRM
Collabora<on
Packaged)SoJware
Network)Infrastructure)
Core)Telco)Apps)(HLR/HSS/SDP…))
1: Replication used in combination with cluster or virtualization – based HA

MySQL HA Solutions
Note that MySQL Fabric is currently at Release Candidate Status

Small: Web Reference Architecture
Single server supporting:
Users & sessions
eCommerce
Content management
MySQL replication
Analytics and Backups
If traffic volumes increase, scale session
management first on a separate server
Limited Scalability
Recommended to start with the Medium Architecture
MySQL
Enterprise Monitor
MySQL
Enterprise Backup
MySQL Master
MySQL Replication
Analytics
Slave 2
Backup
Slave 1

Medium: Web Reference Architecture

Best Practices - Overview
Medium Web Reference Architecture
Server ratio: 10 application servers to each MySQL Server
More for PHP applications, fewer for Java
Add more slaves as the application tier scales
Caching layer deployed in session & content management components
Memcached or Redis are common
Reads fulfilled from cache, relieving load on the source database servers
NoSQL Memcached APIs can enable compression of caching and data layers

Best Practices – Content Management
Each slave can handle around 5,000 concurrent users
Each master handles 20 slaves
More slaves are often used in larger environments
MySQL Replication for high availability & scale-out
Meta data of content assets managed by MySQL
Distributed File System / CDN / Cloud (i.e. S3) for physical
storage of assets on
High quality SAN (redundancy for HA)
Across commodity nodes
XOR

Best Practices – Sessions & eCommerce
Session Management & eCommerce
Session data maintained for up to 1 hour in a dedicated partition, rolling
partitions used to delete aged data
NoSQL Memcached API for InnoDB can be used to scale session management
eCommerce HA
Semi-Synchronous replication or clustering (DRBD, shared storage, etc.)
If web traffic grows, move Session Management to MySQL Cluster
Persist session data for real-time personalization of user experience
99.999% availability and in-memory data management can reduce need for DRBD & caching
§ Data is captured in Analytics Database, increasingly replicated to Big Data
system for analysis

Large: Web Reference Architecture
Session Management eCommerce
Content Management
MySQL Master
Slave N
Slave 6 Slave 7 Slave 8 Slave 9 Slave 10
Slave 1
MySQL Servers
Node Group 2
Analytics
MySQL Master
MySQL Servers
MySQL
Enterprise Monitor
Slave 2 Slave 3 Slave 4 Slave 5
Slave 2
F2
F4
Node 4 Node 3
F2
F4
Node Group 1
F1
F3
Node 4 Node 3
F1
F3
Node Group 2
F2
F4
Node 4 Node 3
F2
F4
Node Group 1
F1
F3
Node 4 Node 3
F1
F3
Slave 1 Slave 3
Hadoop
Cluster
MySQL
Enterprise Backup
Caching / Application Servers
Distributed
Storage
XOR

Extra Large: Social Network
Shard Catalog
MySQL Servers
Authentication
MySQL Master
Slave 2
Analytics
Node Group 2
F2
F4
Node 4 Node 3
F2
F4
Node Group 1
F1
F3
Node 4 Node 3
F1
F3
Slave 1 Slave 3
Hadoop
Cluster
MySQL
Enterprise Backup
MySQL Servers
Node Group 2
F2
F4
Node 4 Node 3
F2
F4
Node Group 1
F1
F3
Node 4 Node 3
F1
F3
App /Caching
Central Databases Servers
Shards
Customers
1% – 33%
Customers
34% – 66%
Customers
67% – 100%
Slave 1 Slave 2 Slave N
MySQL
Enterprise Monitor

Operational Best Practices
Training
Consul1ng
24
x
7
x
365
Global
Support
Database
Management
&
Monitoring

MySQL Enterprise Edition
Highest Levels of Security, Performance and Availability
MySQL Enterprise
Oracle Premier
Lifetime Support
Oracle Product
Certifications/Integrations
MySQL Enterprise
High Availability
Security
MySQL Enterprise
Scalability
MySQL Enterprise
Monitor/Query Analyzer
MySQL Enterprise
Backup
MySQL Workbench
MySQL Enterprise
Audit

MySQL Enterprise Monitor
• Web-based, global view of MySQL/
Cluster applications (on-premise
and Cloud deployments)
• Automated, rules-based monitoring
and alerts (SMTP, SNMP enabled)
• Query capture, monitoring, analysis
and tuning, correlated with Monitor
graphs
• Real-time Replication Monitor with
auto-discovery of master-slave
topologies
• Integrated with Oracle Support
A Virtual MySQL DevOps + Admin!

MySQL Enterprise Backup
§ Online Backup for InnoDB (scriptable interface)
§ Full, Incremental, Partial Backups (with compression)
§ Point in Time, Full, Partial Recovery options
§ Parallel, multi-threaded backup/recovery operations
§ Metadata on status, progress, history
§ Unlimited Database Size
§ Cross-Platform
- Windows, Linux, Unix
§ Certified with Oracle Secure Backup
MEB Backup
Files
MySQL
Database Files
mysqlbackup
Ensures quick, online backup and recovery of your on premise and Cloud
based MySQL applications.

Oracle Premier Support for MySQL
Rely on The Experts - Get Unique Benefits
• Straight from the Source
• Largest Team of MySQL Experts
• Backed by MySQL Developers
• Forward Compatible Hot Fixes
• MySQL Maintenance Releases
• MySQL Support in 29 Languages
• 24/7/365
• Unlimited Incidents
• Knowledge Base
• MySQL Consultative Support
Only From
Oracle
"The MySQL support service has been
essential in helping us with troubleshooting
and providing recommendations for the
production cluster, Thanks."
-- Carlos Morales – Playfulplay.com

Next Steps
HA Solutions
Guide
http://bit.ly/NB8up1
Learn More
Evaluate MySQL 5.6 Book a Workshop
Download Today
http://dev.mysql.com/
downloads/mysql/
High Availability
Jumpstart
http://www.mysql.com/
consulting/
Copyright 2011 Oracle Corporation

mark.swarbrick@oracle.com

1 architecture & design

Recommended

More Related Content

What's hot (20)

Viewers also liked (18)

Similar to 1 architecture & design (20)

More from Mark Swarbrick (20)

Recently uploaded (20)

1 architecture & design