Module 10: Maintaining High-Availability
Overview Introduction to Availability Increasing Availability Using Failover Clustering Standby Servers and Log Shipping
Introduction to Availability Defining Availability Determining Availability Requirements Availability and Scalability Maximizing Availability with Microsoft.NET Enterprise Servers
Defining Availability SQL Server
Determining Availability Requirements Hours of Operation Business hours vs. all of the time Connectivity Requirements Online vs. offline Tight/Loose Coupling Synchronous vs. asynchronous
Availability and Scalability Scaling Up Increasing RAM Multiprocessing Scaling Out Replication Read-only standby servers Server federations
Maximizing Availability with Microsoft.NET Enterprise Servers DataData SQL Server RAID Replication Failover Clustering Standby Servers PresentationPresentation Network Load Balancing Web Farms BusinessBusiness Component Load Balancing Queued Components
Increasing Availability Using Failover Clustering Windows Clustering SQL Server Failover Clustering Active/Passive Clustering Active/Active Clustering Failover Clustering Considerations
Windows Clustering CLUSTER Network Secondary Node Primary Node RAID Array
SQL Server Failover Clustering Installing a Failover Cluster Run SQL Server Setup Specify network name and IP addresses Administering a Failover Cluster Install additional nodes Remove nodes Recover from failure
Active/Passive Clustering Windows 2000 Primary Node Secondary Node \\VirtualServer1 Client Windows 2000 SQL Server SQL Server RAID Array
Active/Active Clustering Windows 2000 Primary Node for //VirtualServer1 Secondary Node for //VirtualServer2 Primary Node for //VirtualServer2 Secondary Node for //VirtualServer1 //VirtualServer1 Windows 2000 Client //VirtualServer2 SQL Server SQL Server SQL Server SQL Server
Failover Clustering Considerations Make the MS DTC Cluster Aware Run Comclust.exe Consider Memory Configuration When Using Address Windowing Extensions Give each node the same amount of RAM Sum of max server memory should be less than minimum physical RAM on a node Use a Cluster Administration Share for Replication
Standby Servers and Log Shipping Maintaining High Availability with a Standby Server and Log Shipping Configuring Log Shipping with the Database Maintenance Plan Wizard Changing Log Shipping Roles
Maintaining High Availability with a Standby Server and Log Shipping Log Shipping Automates: Backing up the primary server transaction log Copying the transaction log to the standby server Restoring the transaction log backup to standby server
Configuring Log Shipping with the Database Maintenance Plan Wizard Using the Wizard You Can: Create the destination database Specify which destination servers can become source servers Define how often transaction log backups are generated Specify a monitor server to store status information
Changing Log Shipping Roles Create a DTS Package to Copy Logins Perform the Role Change sp_change_primary_role sp_change_secondary_role sp_change_monitor_role sp_resolve_logins Allow Clients to Reconnect
Maximize Availability at Each Tier of the Application Ensure All Servers in an Active/Active Cluster Can Handle the Combined Workload Ensure All Servers in an Active/Active Cluster Can Handle the Combined Workload Keep Your Standby Servers on a Different Subnet Mask and Different Power Supply to the Primary Server Keep Your Standby Servers on a Different Subnet Mask and Different Power Supply to the Primary Server For Maximum Availability, Use a Combination of Failover Clustering and Log Shipping For Maximum Availability, Use a Combination of Failover Clustering and Log Shipping Test Whether Your Availability Solution Works Recommended Practices
Lab A: Automating the Maintenance of a Standby Server
Review Introduction to Availability Increasing Availability Using Failover Clustering Standby Servers and Log Shipping