Ralph: Who am I?
• An Enterprise Architect
• at an iGaming Industry
• focus on Data Platforms
• A Microsoft Certified Trainer
• deliver MTA, MCSA, MCSE locally
• covering Windows, SQL Server, C#
• I’m here for a full-blown hands-on demonstration on possible high
availability deployment setups using SQL Server 2016
Log Shipping
Transaction Log Shipping
• TLS is a backup & restore configuration
• is mature, reliable, widely adopted technology
• requires a full database recovery model
• can be restored with norestore or standby
• Does not offer automation for failover scenarios
• used to verify backups and maintain lagged db copies
• also helps to minimise migrations between clusters
• Can be automated via SSMS wizard (sqllogship.exe)
• Or performed manually/scripted
• Perform backups on primary instance
• Copy full & tlog backups to other servers
• Restore backups on all the desired servers
AlwaysOn Failover Cluster Instances
• An SQL Server Instance (FCI) is installed across Windows
Server Failover Cluster (WSFC) nodes
• Protects against server failure by providing a virtual
server for client applications to connect to
• Requires shared storage, thus a single point of failure
which needs to be mitigated
• A quorum is required to keep cluster online
• Provides automated failover at instance level
Failover Cluster Instance
Active
Windows Cluster
Passive
Availability Group
Windows Cluster
Primary
replica
Active Secondary
replica
AlwaysOn Availability Groups
• A “shared nothing” approach to high availability on
top of Failover Cluster through replication
• Replicas can support alternative availability modes
• asynchronous – high performance
• synchronous – high availability
• Provides automated failover at group level
• Secondary replicas can provide additional support
• read-only access
• backup operations
2016 Availability Groups Upgrades
• Changes in number of replicas:
• Optimised log transport
• Load balanced read-only routing
• Support Microsoft Distributed Transaction Coordinator (MSDTC)
• Basic Availability Groups replaces Database Mirroring (deprecated)
• Distributed availability groups
• replication between 2 availability groups residing on different
Windows Server Failover Clusters.
• only manual failovers allowed
Version Replicas Synchronous
Commit
Automatic
Failover
2012 4 2 2
2014 8 3 2
2016 8 3 3
ff877884
mt651673
Events line up
• The Global Azure Bootcamp
• What: a one day, in-person community driven event focused on developers and IT professionals.
• When: April 22nd, 2017.
• http://global.azurebootcamp.net/
Sponsorships
• A big thanks to Microsoft Malta who sponsored today’s event; location, refreshments, parking!
• At Tech-Spark, we are constantly looking for sponsors to help in organising events by providing
meeting places, refreshments, and topic speakers amongst others. Feel free to contact us should
you want to help!
Questions?
Ralph Attard
raland@raland.net
Tech Spark
https://www.facebook.com/groups/TechSparkMalta

Tech-Spark: SQL Server 2016 High Availability

  • 2.
    Ralph: Who amI? • An Enterprise Architect • at an iGaming Industry • focus on Data Platforms • A Microsoft Certified Trainer • deliver MTA, MCSA, MCSE locally • covering Windows, SQL Server, C# • I’m here for a full-blown hands-on demonstration on possible high availability deployment setups using SQL Server 2016
  • 3.
    Log Shipping Transaction LogShipping • TLS is a backup & restore configuration • is mature, reliable, widely adopted technology • requires a full database recovery model • can be restored with norestore or standby • Does not offer automation for failover scenarios • used to verify backups and maintain lagged db copies • also helps to minimise migrations between clusters • Can be automated via SSMS wizard (sqllogship.exe) • Or performed manually/scripted • Perform backups on primary instance • Copy full & tlog backups to other servers • Restore backups on all the desired servers
  • 4.
    AlwaysOn Failover ClusterInstances • An SQL Server Instance (FCI) is installed across Windows Server Failover Cluster (WSFC) nodes • Protects against server failure by providing a virtual server for client applications to connect to • Requires shared storage, thus a single point of failure which needs to be mitigated • A quorum is required to keep cluster online • Provides automated failover at instance level Failover Cluster Instance Active Windows Cluster Passive
  • 5.
    Availability Group Windows Cluster Primary replica ActiveSecondary replica AlwaysOn Availability Groups • A “shared nothing” approach to high availability on top of Failover Cluster through replication • Replicas can support alternative availability modes • asynchronous – high performance • synchronous – high availability • Provides automated failover at group level • Secondary replicas can provide additional support • read-only access • backup operations
  • 6.
    2016 Availability GroupsUpgrades • Changes in number of replicas: • Optimised log transport • Load balanced read-only routing • Support Microsoft Distributed Transaction Coordinator (MSDTC) • Basic Availability Groups replaces Database Mirroring (deprecated) • Distributed availability groups • replication between 2 availability groups residing on different Windows Server Failover Clusters. • only manual failovers allowed Version Replicas Synchronous Commit Automatic Failover 2012 4 2 2 2014 8 3 2 2016 8 3 3 ff877884 mt651673
  • 7.
    Events line up •The Global Azure Bootcamp • What: a one day, in-person community driven event focused on developers and IT professionals. • When: April 22nd, 2017. • http://global.azurebootcamp.net/
  • 8.
    Sponsorships • A bigthanks to Microsoft Malta who sponsored today’s event; location, refreshments, parking! • At Tech-Spark, we are constantly looking for sponsors to help in organising events by providing meeting places, refreshments, and topic speakers amongst others. Feel free to contact us should you want to help!
  • 9.