Roger Breu @rogerbreuTechnology Solution Professional DataPlatformMicrosoft Corporation

SQL Server 2012AlwaysOn Availability Groups

What is an Availability Group?Availability Group Setup Step by StepActive SecondariesTuning and MonitoringQ&A

Agenda

SQL Server 2012 AlwaysOn Availability Groups

What is an Availability Group?

HA/DR Features beforeFailover Clusterin

gVirtual Networkname

(transparent)

Multidatabase failover

Shared storage needed

No active secondaries available

DR only with Storage Replication or additional Logshipping/ Database

Mirroring

Database Mirroring

Fast Failover

No Shared Storage needed

Application has to be “mirroring-aware” (SNAC or connectstring modification)

No virtual network name to connect

Only 1 active secondary with database snapshots on

the mirror

Log Shipping

Multiple Secondaries

No Shared Storage needed

Needs good monitoring

No automatic failover

Secondaries offline while applying transactionlog

backups

... List is not complete, Replication only for Table Level HA...

AlwaysOn Availability Groups is a new feature that enhances and combines failover cluster, database mirroring and log shipping capabilities

AlwaysOn Availability Groups

Multi-database failoverMultiple secondaries

Total of 4 secondaries2 synchronous secondaries

1 automatic failover pairSynchronous and asynchronous data movementBuilt in compression and encryptionFlexible failover policyAutomatic Page Repair

Application failover using virtual nameConfiguration WizardDashboardSystem Center IntegrationRich diagnostic infrastructureFile-stream replicationReplication publisher failover

Flexible Integrated Efficient

Active SecondaryReadable SecondaryBackup from SecondaryAutomation using power-shell

Availability Group Scenarios

Synchronous Data Movement

Asynchcronous Data Movement

AG AG

AG

AG

AG

Direct Attached Storagelocal, regional and geo secondaries

ExamplePrimary in CalgaryFailover Partner in VancouverSync DR in Toronto

Async Secondary in Montreal (Reporting)Async Secondary in London (Geo DR)

Availability Group Scenarios

Synchronous Data Movement

Asynchcronous Data Movement

Shared Storagelocal, regional and geo secondaries

mixed with Direct Attached Storage

ExamplePrimary Failover Cluster in CalgarySync Secondary Failover Cluster in Vancouver

AG

AG

Async DR Standalone Server in Toronto

AG

«Container»1 or more DatabasesListener (aka virtual networkname)

1 or more IP Addresses (DHCP or fix, aka virtual IP)

Replica DefinitionPrimary/SecondariesAutomatic Failover PartnerSync/Async Secondaries

The Availability GroupDB

DBDB

AG

AG_VNN

AG_IP

Availability Group Architecture

Availability Group uses Windows Server Failover Cluster (WSFC) for

Inter-node health detection Failover coordinationPrimary health detection Distributed data store for Availability Group settings and state Distributed change notifications

Windows Server Failover Cluster

DatabaseActive Log Synchronization

DatabaseActive Log Synchronization

SQL Server synchronizes databases

Based on Transaction LogVia TCP EndpointSynchronous or asynchronousCompressed LogstreamEncrypted Logstrem

DBDB

DBDB

DBDB

AG

AG_VNN

AG_IP

SQL Server 2012 AlwaysOn Availability Groups

Availability Group Setup Step by Step

Windows Server 2008/2008 R2Failover Cluster Feature installedWindows Server Failover Cluster created

Needs Cluster Networkname and IP AddressNo Shared Storage neededQuorum can be configured with Fileshare Witness

OS Requirements

SQL Server 2012Installed as Standalone (or SQL Server Failover Cluster*) Instance on every Windows Server Failover Cluster Node

Can be named or default instanceInstance names don’t have to be the sameAll Nodes have to be part of the same Windows Server Failover ClusterAll Instances must use the same collationDomain Service Account

AlwaysOn Availability Groups feature enabled via SQL Server Service Configuration Manager

SQL Server Requirements

Each availability group name must be unique on the WSFC clusterFor a given availability group, the availability replicas must be hosted by server instances running on different nodes of the same WSFC cluster# of availability groups on a given server instance is unlimited

Recommended maximum is ten availability groups per instance

# of database in a given availability group is unlimited

Recommended maximum is 100 databases per availability group

Availability Group Requirements

Must be a user databaseSystem databases cannot belong to an availability group

Must be a read-write databaseRead-only databases cannot be added to an availability group

Must be a multi-user databaseNot use AUTO_CLOSEUse the full recovery modelPossess a full database backupReside on the SQL Server instance where you are creating the availability group and be accessible to the server instanceNot belong to another availability groupNot be configured for database mirroring

Database Requirements

Setup Windows Server Failover Cluster NodesConfigure WSFCSetup SQL Server Instances

Same collationRecommended to configure same DB/Tlog File pathes

Enable AlwaysOn Availability Groups via SQL Server Service Configuration Manager

Instance restart neededStart deploying Availability Groups

Checklist

Application reconnects using a virtual name (listener) after a failover to a secondary

Primary Secondary

-server Techdays1

Application retry during failover Connect to new primary

once failover is complete and the listener is online

Primary SecondarySecondary

CALGARY VANCOUVER TORONTO

Demo Scenario – Application Failover

DB1

DB2 DB

1

DB2

DB1

DB2

TechdaysAG1

Techdays1 (VNN)

AG_IP

WSFC DENALIWINCLU1

Via SQL Server Management StudioExpand the Management node, the AlwaysOn High Availability node, and the Availability Groups node. Right-click the availability group to be failed over, and select the Failover command.

Via T-SQLALTER AVAILABILITY GROUP TechdaysAG1 FORCE_FAILOVER_ALLOW_DATA_LOSS;

Via PowerShellSwitch-SqlAvailabilityGroup -Path SQLSERVER:\SQL\secondarymachine\secondaryinstance\AvailabilityGroups\TechdaysAG1 -AllowDataLoss

Via Failover Cluster Manager

Initiate a failover

Introducing Contained Databases or CDB’s

Unit of application programmability in DenaliA DB which establishes a boundary between application and server

CDBs sever the user–login relationship

Windows users no longer need matching loginsUsers with passwords replace SQL logins

What about Server Objects?

CDB can be part of an availability group and can failover across replicasAuthentication information moves with the CDBLimitation

Other objects, like jobs are not contained and have to be managed

SQL Server 2012 AlwaysOn Availability Groups

Active Secondaries

AlwaysOn Active Secondary

IT efficiency and cost-effectiveness is critical for businesses

Idle hardware is not an option anymore

AlwaysOn Active Secondary enables efficient utilization of high availability hardware resources thereby improving overall IT efficiencyActive Secondary can be utilized for

Balancing read-only workloadsOffloading Backup Operations

Read-Only client connection behavior determined by Availability Replica Option+ ApplicationIntent Property

ApplicationIntent is a connection property Replica option determines whether a replica is enabled for read access when in a secondary role

Read-Only Routing enables redirection of client connection to new secondary on role change

Enable seamless redirection of application connection across replicas without manual interventionConfiguration with Read-Only Routing Lists

Read-Only Client Connectivity

Readable secondary allows offloading read queries to secondaryAutomatic redirect with Routing Lists or manual connect to Secondary

Primary Secondary

-server Techdays1Application retry during failover Connect to new

primary/new secondary once failover is complete and the listener is online

Primary SecondarySecondary

CALGARY VANCOUVER TORONTO

Demo Scenario – Active Secondary

DB1

DB2 DB

1

DB2

DB1

DB2

TechdaysAG1

Techdays1 (VNN)

AG_IP

Reports

-server Techdays1; ApplicationIntent=ReadOnly

WSFC DENALIWINCLU1

DB1 Data

Log Cache

Log Cache

Close to real-time reads

Log is first hardened and then applied Redo thread is asynchronous and runs in the background

Latency (typically seconds) can be larger for log intensive operations like bulk import or index create/rebuildSync Replica minimizes latency due to network issues

DB1

DB1 Log

DB1

Primary SecondaryLog

Capture Log Recieve

DB1 Data

Network

Redo Thread

Redo Pages

DB1 Log

Page Updated

Log HardenLog Flush

Commit

Acknowledge Commit

Log Pool

Backups can be done on any replica of a database (Full WITH COPY_ONLY or Tlog)

Backups on primary replica still works

Log backups done on all replicas form a single log chain

Database Recovery Advisor makes restores simple

Enabling Backup On Secondary

R/W workload

Primary

Backups

Secondary

Backups

Secondary

Backups

SQL Server 2012 AlwaysOn Availability Groups

Tuning and Monitoring

All the availability replicas should run on comparable systems that can handle identical workloadsFor best performance, use a dedicated network adapter for AlwaysOn Availability Groups (LogStream)Every computer on which a server instance hosts an availability replica must possess sufficient disk space for all the databases in the availability group

Keep in mind that as primary databases grow, their corresponding secondary databases grow the same amount

For the New Availability Group wizard to automatically start initial data synchronization, every secondary database must use the same file path as its corresponding primary database

If you add a file to a primary database, secondary replica(s) will be able to add the new file on the corresponding secondary database(s)

Fine Tuning/Tipps

Make sure, WSFC network name has „add workstation to domain“ rights otherwise it will only add up to 10 VNNsFailover Cluster Failure Threshold should be adjusted for Test-runs

In the Failover Cluster Management snap-in navigation pane, expand one of the managed clusters that has a highly available application or service configured.Expand the Services and Applications category.Select and then right-click one of the groups, and then click Properties.Click the Failover tab, and then view the Maximum failures in the specified period setting

TraceFlag 9532 needed for all SQL Server 2012 CTP3 instances to allow more than one replica

Configuration via SQL Server Configuration Manager on every nodeAdd T9532 on the startup tabWon’t be necessary with RTM Bits

Fine Tuning/Tipps

Please consult following KB article for configuration and download Hotfix

http://support.microsoft.com/kb/2494036/en-us http://msdn.microsoft.com/en-us/library/hh270279(v=SQL.110).aspxCluster.exe DENALIWINCLU1 node TORONTO /prop NodeWeight=0

Remove the DR side vote so that a network interruption doesn’t affect the servers running on the main site for HA

Node weight Quorum Settings

WSFC Service

SQL Server Failover Cluster

Instance

AG Res DLL

IsAlive /LooksAliveWSFC asks Res DLL if SQL AG alive

exec sp_server_diagnostics

Diagnostics generated for Health State Components• System• Resource• Query Processing• IO Subsystem • Events

Diagnostics

User sets new Cluster propertiesHealthCheckTimeout and FailureConditionLevel

IsAlive/ LooksAlive result based on diagnostics and FailureConditionLevel

FailureConditionLevel (0 to 5)5 – Failover or restart on any qualified failure4 – Failover or restart on moderate SQL Server

errors3 – Failover or restart on critical SQL Server errors2 – Failover or restart on SQL Server unresponsive1 – Failover or restart on SQL Server down0 – No Automatic Failover or restart

Flexible Failover Policy

Availability Group Dashboard in Management StudioMonitoring will be fully integrated into the SQL Server 2012 Management Pack for SCOM

Including Database Mirroring Monitoring and Replication Monitoring for SQL Server 2008 R2

Lots of new DMVs:http://msdn.microsoft.com/en-us/library/ff878305(v=SQL.110).aspx

Availability Group Monitoring

Virtualization provides best consolidation isolationVirtualization without AlwaysOn:

Simplest management story for limited HA/DR:

When to use AlwaysOn for the guest:Need better HA/DR protection than standalone VM

Virtualization with AlwaysOn

Planned Unplanned

Host Live Migration VM failover (OS restart, Quick Migration)

Guest Downtime during patch

No protection from virtualization

SQL Server 2012 AlwaysOn Availability Groups

Conclusion

SQL Server AlwaysOn Availability Groups is a comprehensive high availability and disaster recovery solution

IntegratedFlexibleCost-Efficient

«Classical» Failover Clustering, Database Mirroring and Log Shipping are not dead

A combination of all solutions can bring you to meet your SLAs

Conclusion

SQL Server 2012 AlwaysOn Availability Groups

Additional Resources

SQL Server 2012 Developer Training Kithttp://www.microsoft.com/download/en/details.aspx?displaylang=en&id=27721

SQL Server 2012 CTP3 Product Guidehttp://www.microsoft.com/download/en/details.aspx?id=27069

swissSQL Bloghttp://blogs.technet.com/swissSQL

AlwaysOn Resource Centerhttp://msdn.microsoft.com/en-us/sqlserver/gg490638(en-us,MSDN.10)

AlwaysOn Availability Group Prerequisiteshttp://msdn.microsoft.com/en-us/library/ff878487(v=SQL.110).aspx

AlwaysOn Availability Groups together with SQL Server Failover Cluster Instance

http://msdn.microsoft.com/en-us/library/ff929171(v=SQL.110).aspx

Links

Please help us make TechDays even better by Evaluating this Session. Thank you!

Give us your feedback!

© 2011 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.

The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after

the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.