Database Mirroring
"Database Mirroring (Database Mirroring)" is a kind of technology can improve database availability. When the principal database is changed, those changes will be automatically applied to the mirror database. When the main server fails, the client application can be automatically redirected to the mirror server, but the application without any changes.
Database Mirroring can be configured to change the real-time applications on the mirror server, called "synchronous"; or near real-time application changes, known as "asynchronous", so you can specify the level of protection in the system and performance overhead.
With SQL Server 2008 Enterprise Edition, the mirror can be extended to protect the data page. If you find that the main server or mirror server on the data page corruption, from the partner server to retrieve the corresponding data page, and data manipulation will steadily continue. SQL Server 2008 database mirroring also improved, it will compress the main server and data flow between the mirror server, and the implementation of fault recovery in the manual, do not restart the database.
Internet-based sports goods company's case, when there is a server, database, or page-level failure, Database Mirroring will provide very fast fault recovery. Although the need to buy a new server, but without having to rewrite applications, standby server is standard equipment, no SQL Server license.
Log transfer
"Log Shipping" is a high-availability technology that provides a warm standby server. Backup is performed on the main server, while reduction in one or more secondary servers to implement. Then, the "log shipping" will be applied to the secondary server log backups on a regular basis. Although database mirroring can only have a mirror server, but "log shipping" can have many secondary server, which can increase the level of protection.
"Log Shipping" in accordance with the timetable for implementation, so the main data on the server changes and these changes to the secondary transmission time delay between the server. This delay will result in data loss, but in many scenarios, these delays can be used as a method of reducing user error, because the log data may delay the application of the secondary server (so as to preserve the original data on the secondary server).
To reduce the workload of the primary server, secondary server can be used as read-only reporting server, but in the application log backup can not run the report.
In the case of executive search firm can be used to log back up and restore the backup server to perform the delay between the application of "log shipping." This can be caused by user error to reverse the situation, because if the primary server fails, the log can be directly applied to reduce data loss.
Failover cluster
"Failover Clustering" is a Windows feature that provides protection to the entire server, not just the protection of databases. More than one server (called "nodes") shared disk array, so if a node fails, can another node in the cluster to start its service. The solution can prevent any data loss, and provides automatic client redirection service, but the main server to secondary server database mirroring recovery operation is not as fast, and require specialized hardware.
Extends the earlier version of Windows Server 2008 cluster Windows feature, which reduces the hardware and infrastructure requirements, and support in a cluster of 16 nodes. In addition, all nodes are not required to reside on the same subnet and supports OR dependencies (for example, when two IP address resources in any one of the available, to ensure that the network name resource is also available), thus strengthening the cluster function . Overall, in Windows Server 2008 can be more easily set up and manage the cluster, and the introduction of the cluster validation tool will help to ensure that the cluster solutions are sufficient hardware resources. By using Windows Server 2008 enhancements in the cluster, SQL Server 2008 SQL Server instance as a whole provides a robust high availability solution. SQL Server 2008 also improved the earlier version of the cluster function, it does not require every SQL Server instance must have a drive letter, which enhances the clustering solution can support the number of instances.
In the insurance case, the "failover cluster" can be used to provide near continuous availability. "Failover Clustering" does not affect system performance, because it is a server-level solution can provide other resources on the server availability. If the concern shared disk array can not provide sufficient redundancy, you can use the "geographically dispersed clusters."
Geographically dispersed failover cluster
"Geographically dispersed failover cluster" in accordance with the certification of Microsoft Geographically Dispersed Cluster Services configuration provides server-level redundancy, this configuration requires that each site has one or more storage arrays. If the site, the server node or disk failure, system and disk full redundancy failover cluster will be able to handle the follow-up on another site. The shared disk array configuration can avoid the risk of failure, which would prevent the standard cluster configuration function.
On the other copy
"The other copy" to copy multiple databases to each other. Can change any database implementation, but also the way in near real time replication of these changes apply to the other nodes in the topology. Application design requirements, if the principal server is not available, the application must be able to connect to other nodes in the topology.
SQL Server 2008 enhanced replication of such function, it can increase the replication solution node, while maintaining the copy process is online. In earlier versions of SQL Server, to add new nodes to be executed in the offline copy process. In addition, SQL Server 2008 can use the graphical Topology Viewer, right, right to set such replication, monitoring and management. SQL Server 2008 also introduced a conflict detection mechanism to prevent multiple replication nodes update the same row of occasional conflicts arise.
Training in the management of the company's case, on the other copy can access the orders from the local database to another site if the database is not available, each site can be used for recovery.