0 SQL Server AlwaysOn – SharePoint 2013 High Availability and Disaster Recovery Sal Bawany,...

Preview:

Citation preview

1

SQL Server AlwaysOn – SharePoint 2013 High Availability and Disaster Recovery

Sal Bawany, Solutions Architectwww.bawanyconsulting.com

2

IntroductionSal Bawany, Solutions ArchitectSpecialties: SharePoint, Office 365, Cloud Computing • Independent Consultant• Previously Senior Consultant at Microsoft

Contact:• salb@bawanyconsulting.com• LinkedIn: Sal Bawany• Twitter: @fbawany• Blog: blogs.technet.com/b/salbawany

(moving to salbawany.wordpress.com)

3

Welcome to SharePoint Saturday Houston

• Please turn off all electronic devices or set them to vibrate• If you must take a phone call, please do so in the hall so as not

to disturb others• Special thanks to our Title Sponsor, ProSymmetry

Thank you for being a part of the 5th Annual SharePoint Saturday

for the greater Houston area!

4

Thanks to all our Sponsors!

5

Information• Speaker presentation slides should be available

from the SPSHOU website within a week or so

• The Houston SharePoint User Group will be having it’s next meeting Wednesday April 15th. Please join us at www.h-spug.org

6

Agenda• Concepts• Overview• Benefits• Requirements• Setup and Configuration• Failover and Failback Scenarios

7

ConceptsHigh Availability and Disaster Recovery

8

System Outages: Planned vs. Unplanned

Planned outage• Used for maintenance• A time window is pre-announced and coordinated for these activities

Unplanned outage• The result of system-level, infrastructure, or process failures• Can be unplanned and uncontrollable

9

High AvailabilityThe ability to continue operations when a component fails in primary data centre

Disaster Recovery The process to restore operations after primary data centre becomes unavailable

Disaster Recovery and High Availability

10

Service Level Agreement (SLA)Recovery Point Objective (RPO)• The maximum period in which data might be tolerably

lost due to some major incident

Recovery Time Objective (RTO)• The time within which the business service must be

restored after a disruption

Example:• RPO of 1 hour• RTO of 6 hours

11

High Availability concepts• Redundancy across switches, routers, load balancer,

AD/DNS, SMTP, power, cooling• Redundant web, application, and DB servers• Redundant VM hosts• Backup and Restore

12

Disaster Recovery ConceptsHot standby• A second data centre that can provide availability within seconds

or minutes• Highly dependent on network bandwidth and latency

Warm standby• A second data centre that can provide availability within minutes

or hours• Duplicated VMs, refreshed frequently

Cold standby• A second data centre that can provide availability within hours or

days

13

Disaster Recovery dependenciesRemote site with reliable WAN connectionFailover farm should match production farm: • VM hosts and their clusters• Number of servers• Operating system, Patch/updates• SQL server version and updates• SharePoint version and updates• Configuration, and Central admin content database• Web applications and service applications and their databases• Any customizations deployed in the production farm

Infrastructure: HA/DR for switches, routers, load balancer, Active Directory domains DNS, Exchange Server, SQL Servers, and power and cooling.

14

SQL Server AlwaysOn Availability Groups

15

SQL Server HA Solutions• AlwaysOn Failover Cluster Instances• AlwaysOn Availability Groups• Database mirroring• Log shipping

16

What is AlwaysOn Availability Group?

• Enterprise-level high-availability and disaster recovery solution introduced in SQL Server 2012– Listener, DNS name– Databases– Replicas– Replica roles– Replica Availability Mode– Replica Failover Mode

SQL Node1Primary Replica

Automatic failover

SQL Node3Secondary Replica

Manual failover

SQL Node2Secondary ReplicaAutomatic failover

Listner

Windows Cluster

AG1

AG2

AG1

AG2

AG1

AG2Async

Sync

WAN

SharePoint Farm

Backup

17

High level Planning

AlwaysOn Availability Groups

Availability Group Listeners

Availability Databases

Availability Replicas

Replica Role

Replica Availability Mode

Replica Failover Mode

Readable Secondary

AG for Content Databases

SP-AG1 Content Databases

PROD SQL 1 Primary Sync commit Automatic No

PROD SQL 2 Secondary Sync commit Automatic No

DR SQL 1 Secondary Async commit Manual No

AG for Service Application Databases

SP-AG2 Service Apps Databases

PROD SQL 1 Primary Sync commit Automatic No

PROD SQL 2 Secondary Sync commit Automatic No

18

Benefits• No need to setup SQL Instance Cluster• No need for SAN storage• Failover multiple databases in seconds• Up to eight secondary replicas• Compression and Encryption• Automatic and Manual Failover• Easy setup using the configuration wizard• AlwaysOn health dashboard

19

SharePoint Databases not supported in Async

• Configuration database (Farm specific)• Central Administration content database (Farm specific)• Search Administration database (copy and re-create)• Analytics Reporting database• Crawl database• Link database• Usage and Health Data Collection database (Farm

specific)• User Profile Synchronization database• State Service database

20

Requirements• SQL Server 2012 Enterprise Edition• Windows Server Failover Cluster (WSFC)– Failover processing and management– Heartbeat detection– Keeps track of changes in the replica roles

• At least three nodes– two at the local site to provide high availability– third node at the remote site to provide disaster recovery

• Replica Availability Mode– Sync = local high availability– Async = remote, disaster recovery

21

Additional considerations• All cluster members must belong to the same domain or

trusted domain• Cluster Quorum – For odd number of votes, use “Node Majority”– For even number of votes, use either “Node Majority” or add

file share and use “Node and File Majority”

• Follow the best practices for setting up the Clustering Service

• New databases need to be added manually to the Availability Group

• Do not use the Failover Cluster Manager to manipulate Availability Group

22

Setup

23

Steps1. Setup Windows Failover Clustering and make each SQL instance from local and

remote disaster recovery site member of that cluster2. Enable AlwaysOn on each node3. Create the Availability Group(s)4. Add Replicas to each Availability Group5. Verify Endpoints, open inbound port 5022 (default) on each SQL node6. Verify backup preferences7. Create DNS A record for the “Listener”8. Create Availability Group Listener, provide port 1433(default), and an IP

address from each of the local and remote DR subnets. These IPs should map to an A record for the Listener.

Note: Make sure that the same farm service accounts to each of the Prod and DR SQL Instances login and server roles have been added, otherwise the AlwaysOn synchronization process will Fail.

24

Steps (Cont.)9. Create a shared folder for data synchronization for the

Availability Group10. Install SharePoint 11. After SharePoint has been installed and configured, change

Recovery Model for each of its databases to Full. This is a requirement before a database can be added to the Availability Group

12. Backup all of the databases. This is a requirement before a database can be added to the Availability Group

13. Add databases to their Availability Group(s)14. Test failover and failback on each Availability Group

25

Failover and Failback Scenarios

26

High Availability Scenarios• Manual Failover and Failback from primary to

secondary replica and vice versa• Automatic Failover due to primary replica outage

and manual Failback from the secondary replica

SQL Node1Primary Replica

Automatic failover

SQL Node3Secondary Replica

Manual failover

SQL Node2Secondary ReplicaAutomatic failover

Listner

Windows Cluster

AG1

AG2

AG1

AG2

AG1

AG2Async

Sync

WANPRIMARY Secondary Secondary

27

Disaster Recovery Scenarios• Manual Failover and Failback from primary to DR replica and vice

versa• Forcing Quorum and Force Failover due to primary datacenter

outage and manual Failback from the secondary replica. Data loss possible. Be aware of the Spilt Brain quorum.

SQL Node1Primary Replica

Automatic failover

SQL Node3Secondary Replica

Manual failover

SQL Node2Secondary ReplicaAutomatic failover

Listner

Windows Cluster

AG1

AG2

AG1

AG2

AG1

AG2Async

Sync

WAN

DR SP Farm

PRIMARY Secondary Secondary

28

Demos

29

Demos• Availability Group walkthrough• Creating Availability Group• Dashboard• Adding Database to the Availability Group• Failover and Failback process

30

Summary• Maintain a Service Level Agreement (SLA) that

meets your business goal• Time is Money!• Maintain redundancy at all levels, software and

hardware• AlwaysOn Availability Group is easy to setup and

maintain• Run scheduled dry runs to make sure Disaster

Recovery is working• Happy Servers, Happy Admins

31

Q & A

32

Please Leave Feedback During Q&A

Please fill out the session survey with your feedback. You can scan the QR Code to launch the survey or use the following link: www.WhatsYourAnswer.com?S20154715202

Recommended