1. Designing a Highly Available SQL Server Infrastructure
Ross Mistry, SQL MVP, Author and Consultant

2. Agenda Introductions The Need for High Availability
High Availability Design Goals
Understanding the SQL 2008 HA Alternatives
Designing the SQL Server Infrastructure
Real World Examples
Q&A
Twitter Book Drawing
Ross Mistry, SQL MVP 2009

3. Ross Mistry – Bio Summary
4/28/2019 7:15 PM
Ross Mistry – Bio Summary
Bio:
Principal Consultant & Partner w/ Convergent Computing (CCO)
Located in the San Francisco Bay Area / Silicon Valley.
SQL Server MVP
SQL Server Specialties - HA, DR, Security & Virtualization.
Frequent Speaker - Microsoft, PASS, Connections and User Groups
SQL Server Consulting for companies such as Yahoo, EBay, EA. Etc
Author:
SQL Server 2008/2005 Management and Administration
Windows Server 2008 Unleashed
Contributing Author:
Hyper-V Unleashed
Exchange Server 2010/2007 Unleashed
SharePoint Server 2007/2003 Unleashed
Upcoming Books :
Windows Server 2008 R2 Unleashed
Introducing SQL Server 2008 R2 (Microsoft Press)
Contact:
Blog Site:
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

4. The Need for High Availability & HA Design Goals
4/28/2019 7:15 PM
PART 1
The Need for High Availability & HA Design Goals
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

5. High Availability - Defined
To increase the availability of servers that host mission critical applications.
Achieved by grouping servers.
If one server fails, another server will assume its responsibilities.
Failover can be Automatic or Seamless
Client Redirect can be Automatic or Seamless
Ross Mistry, SQL MVP 2009

6. Protection and HA Design Goals
Availability during Planned Downtime
Patching & Service Pack Installation
Hardware and Software Upgrade
System Reconfiguration
Database Maintenance
Application Upgrade
Protection against Unplanned Downtime
Human Error #1 Cause of Failure
Site Disasters
Hardware Malfunction
Data Corruption
Software Crash
Ross Mistry, SQL MVP 2009 6

7. Understand and Define a SLA
Availability Percentage
24-hour day
8-hour day
90 %
876 hours (36.5 days)
291.2 hours (12.13 days)
95 %
438 hours (18.25 days)
145.6 hours (6.07 days)
99 %
87.6 hours (3.65 days)
29.12 hours (1.21 days)
99.9%
8.76 hours
2.91 hours
99.99%
52.56 minutes
17.47 minutes
99.999%
5.256 minutes
1.747 minutes %
seconds
seconds
Main Goal - Set Reasonable Expectations
Ensure you have the hardware, software and administration to achieve goals
Ross Mistry, SQL MVP 2009

8. SQL Server 2008 High Availability
Redundancy at server, database and data page levels reduces costs associated with downtime
Highly credible, cost effective high availability solutions
Reduced downtime for database maintenance and hardware upgrades
Reduced contention for resources
Increased availability of data in the event of a partial failure
No need for 3rd party tools
Ross Mistry, SQL MVP 2009

9. Understanding SQL Server 2008 High availability alternatives
4/28/2019 7:15 PM
PART 2
Understanding SQL Server 2008 High availability alternatives
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

10. Understanding the SQL Server 2008 High Availability Alternatives
Database Mirroring
Transactional Replication
Log Shipping
Failover Clustering
Virtualization, Live Migration & Guest Clustering
Ross Mistry, SQL MVP 2009

11. Failover Clustering Server hardware redundancy
4/28/2019 7:15 PM
Failover Clustering
Server hardware redundancy
Using a shared disk subsystem
Entire instance virtualized and
fails over as a unit
Can include non-SQL Server resources
Clustering can be combined with Database Mirroring, Log Shipping, or Replication
Single Instance or Multiple Instance Failover Clusters
Geographically Dispersed Failover Clustering provides protection even if the disk array fails
Ross Mistry, SQL MVP 2009
© 2006 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.

12. Single Instance Clustering
4/28/2019 7:15 PM
Single Instance Clustering
Hot Spare System with Maximum Availability.
1st Node acts as the primary node and supports all the client requests.
2nd Node is idle.
The 2nd Node is a dedicated hot spare and is accessed only when a failover occurs.
If the primary node fails, the hot spare node immediately takes over all operations and continues to service the client requests.
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

13. Single Instance Failover
4/28/2019 7:15 PM
Single Instance Failover
Passive Node
Active Node
Active Node
Passive Node
Heartbeat
Shared Storage
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

14. Multiple Instance Cluster
4/28/2019 7:15 PM
Multiple Instance Cluster
1st Node acts as the primary node and supports all the client requests for SQL Instance 1.
2nd Node acts as the primary node and supports all the client requests for SQL Instance 2.
Either System can be a Hot Spare System
If either node fails, the hot spare node immediately takes over all operations and continues to service the client requests.
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

15. Multiple Instance Failover
4/28/2019 7:15 PM
Multiple Instance Failover
Active Node 1 – SQL Instance 1
Active Node 2 – SQL Instance 2
Active Node 2 – SQL Instance 2
Active Node 2 – SQL Instance 1
Shared Storage
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

16. When To Implement a Failover Cluster
4/28/2019 7:15 PM
When To Implement a Failover Cluster
To reduce downtime during routine maintenance or unplanned failures.
To perform upgrades to nodes, resources, and applications without interrupting client access.
To perform rolling upgrades.
To increase server availability for mission critical applications & network services.
To take advantage of cluster aware applications such as SQL Server
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

17. Failover Clustering SQL Server Editions
Standard Edition
Enterprise Edition
2-node clusters only
As many nodes as the operating system supports
Ross Mistry, SQL MVP 2009

18. Windows Server 2008 & R2 Clustering
Support for 16 nodes within a cluster
Nodes can reside on different subnets
Cluster Validation tool
Verifies adequate hardware resources for clustering
Hyper-V and Live Migration
Ross Mistry, SQL MVP 2009

19. Database Mirroring Overview
Hot Standby protects against database or server failure
Provides a fault-tolerant database
Cost effective as no specialized hardware is required
Straightforward setup and administration
Automatic or manual failover
SQL Server 2008 does not require a database restart after manual failover
Automatic, transparent client redirect
No shared components; two separate copies of data
SQL Server 2008 compresses the log stream from Principal to Mirror
Ross Mistry, SQL MVP 2009

20. Synchronous, high availability configuration
4/28/2019 7:15 PM
Database Mirroring
Synchronous, high availability configuration
Data is mirrored synchronously as part of a transaction
Automatic failover if Principal server fails
Automatic Client redirection
Witness
Client
Safety Full; synchronous operation
Commit when logged on Mirror
No automatic failover; manual failover only
Database quorum formed by Principal and Mirror
If Principal loses quorum, it stops servicing the database
Ensures high protection; database is never in ‘exposed’ state
No Witness present – no automatic failover
Principal
Mirror
Ross Mistry, SQL MVP 2009
© 2006 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.

21. Synchronous, high protection configuration
4/28/2019 7:15 PM
Database Mirroring
Synchronous, high protection configuration
Data is mirrored synchronously as part of a transaction
Manual failover if Principal server fails
Automatic Client redirection
Client
Safety Full; synchronous operation
Commit when logged on Mirror
No automatic failover; manual failover only
Database quorum formed by Principal and Mirror
If Principal loses quorum, it stops servicing the database
Ensures high protection; database is never in ‘exposed’ state
No Witness present – no automatic failover
Principal
Mirror
Ross Mistry, SQL MVP 2009
© 2006 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.

22. Asynchronous, high performance configuration
Database Mirroring
Asynchronous, high performance configuration
Data is mirrored asynchronously outside of transactions
Manual failover if Principal server fails
Automatic Client redirection
Client
Principal
Mirror
Ross Mistry, SQL MVP 2009

23. Transactional Replication Overview
High performance – latency measured in seconds
Minimal load on the server
Cost effective as no specialist hardware is required
Straightforward setup and administration
Can be implemented at database or table level
Two types
Standard transactional replication
Easy to design, setup & manage
Subscriber (standby) can be used for reporting
Peer-to-peer transactional replication
Multi-master model; schema is identical on all sites
Supports distributed applications with data partitioning; enables load balancing
Does not handle conflicts; design to avoid/prevent conflicts
Ross Mistry, SQL MVP 2009

24. Transactional Replication Replication Options
4/28/2019 7:15 PM
Transactional Replication Replication Options
Transactional Replication
Reporting + Redundancy
Peer-to-Peer Replication
Query Scale Out + Redundancy
London
New York
Boston
Shanghai
New Jersey
Seattle
Tokyo
Ross Mistry, SQL MVP 2009
© 2006 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.

25. Peer-To-Peer Transactional Replication
4/28/2019 7:15 PM
Peer-To-Peer Transactional Replication
“West”
“East”
Distribution
Agent
Dist DB
Logreader
Distribution
Agent
Dist DB
Logreader
“South”
Distribution
Agent
Dist DB
Logreader
Ross Mistry, SQL MVP 2009

26. Log Shipping Provides database redundancy
Cost effective as no specialist hardware is required
Straightforward setup and administration
Running read operations such as reports on secondary server is permitted
Users are disconnected when log restore occurs
Can maintain multiple secondary servers
Optional Monitor server
Records history and status of backup/
restore jobs
May be setup to raise alerts when jobs fail
Ross Mistry, SQL MVP 2009

27. SQL Server HA w\ Virtualization
4/28/2019 7:15 PM
SQL Server HA w\ Virtualization
Hyper-V, Cluster Shared Volumes (CSV) & Live Migration:
Cluster Shared Volumes (CSV)
Distributed file access solution for Hyper-V
Concurrent access to disk from any node
VMs do not know their host
VMs no longer bound to storage
VMs can share a CSV disk to reduce LUNs
Ross Mistry, SQL MVP 2009
© 2006 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.

28. SQL Server HA w\ Virtualization
4/28/2019 7:15 PM
SQL Server HA w\ Virtualization
Live Migration
Move a running VM from one host to another host with no downtime
Client is not aware of the migration
Maintain open TCP connections of the guest OS
Clients stay connected
Planned failover
Failover Clustering still recovers VM from a disk in an unplanned failover
Ross Mistry, SQL MVP 2009
© 2006 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.

29. Designing the SQL Server High Availability INFRATRUCTURE
4/28/2019 7:15 PM
PART 3
Designing the SQL Server High Availability INFRATRUCTURE
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

30. Choosing The Right Solution
4/28/2019 7:15 PM
Choosing The Right Solution
Cost
Complexity of setup and administration
Ease of failover and failback
Transparency of failover
Choice of operating systems and SQL Server edition
Transaction latency
Geographical Dispersion
Ross Mistry, SQL MVP 2009
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

31. HA Design Decisions to be Made
Determine Business Functionality Requirements:
Will Database Engine will be required
Will Integration Services will be required
Will Analysis Services will be required
Will Reporting Services will be required
Protection – DC, Server, Instance, Database or Table
Standby – Hot, Warm, Cold
Failover – Automatic or Manual
Client Redirect – Automatic or Manual
Ross Mistry, SQL MVP 2009

32. Summary: HA Technologies
Technology
Standby
Protection
Failover
Client Redirect
Failover Clustering
HOT
Server
Automatic
Database Mirroring
High Availability Configuration
Database
Latest SQL Server Client Required
Log Shipping
WARM
Manual
High Protection Configuration
High Performance Configuration
Replication
Ross Mistry, SQL MVP 2009

33. Putting it All Together
4/28/2019 7:15 PM
Putting it All Together
Database Mirroring
Primary disaster site for databases
Log Shipping
Additional disaster sites for databases
Logical Recovery
Replication
Database reporting and read scale out with redundancy
Clustering
Local server redundancy
Database
Scale Out
For Queries
Replication
Database Mirroring
Clustering
Hot
Standby
Production
Database
Log Shipping
Warm
Standby
Log Shipping
with Restore Delay
Logical
Recovery
Standby
Backup
Ross Mistry, SQL MVP 2009
© 2006 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.

34. Real World Example PART 4 4/28/2019 7:15 PM
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

35. Database Mirroring DR Design

36. Scaled Out SSRS Deployment

37. 8 Node Multiple Instance Failover Cluster

38. SQL Server HA w\ Virtualization
4/28/2019 7:15 PM
SQL Server HA w\ Virtualization
Ross Mistry, SQL MVP 2009
© 2006 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.

39. Consolidated Scalable Infrastructure with HA and DR
Ross Mistry, SQL MVP 2009

40. 4/28/2019 7:15 PM
Resources and Contact Info
Twitter
SQL Server 2008 Management & Administration
Covers Administration, Monitoring, Management and Security
4 Chapters Dedicated on HA
Windows Server 2008 Unleashed
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

41. Complete the Evaluation Form & Win!
You could win a Dell Mini Netbook – every day – just for handing in your completed form! Each session form is another chance to win!
Pick up your Evaluation Form:
Within each presentation room
At the PASS Booth near registration area
Drop off your completed Form:
Near the exit of each presentation room
Sponsored by Dell

42. for attending this session and the 2009 PASS Summit in Seattle
Thank you
for attending this session and the PASS Summit in Seattle