For most reasons, the main database applications in the public cloud are often the most complex use cases. They need to run 24×7 under the scene of failure. Consequently, they require complete unmatched, which include an example of the supply standby server and the data constantly copies. Settings that work well in the private cloud may not be possible in the public cloud. And providing high availability can cost a lot of money for licensing advanced software.
Of course, Amazon Web Services are the main high availability of SQL Server mission and how to protect the disaster recovery. But it is possible to select (and all commonly) settings that fail to result in the failure of the project.
AWS offers two basic features to run SQL Server applications; a relative database service and elastic computing cloud. RDS is a systematic service that is basically suitable for basic applications. While the RDD has offered six different database engines, its support for SQL Server requires more expensive Enterprise Edition to overcome the more expensive limits, such as tracking the failure of the application software.
For the main SQL Server applications mission, most of the capabilities available with EC2 prefer that the HA and DR are of importance. But there are also some limitations in EC2, especially the lack of shared storage used in traditional HA layout. And as with RDS, the availability of the available online edition of Enterprise Edition may be necessary to get the required level of protection.
AWS also offers the selection of SQL Server running on Windows or Linux. Windows Server Failure Clustering is a powerful and proven capability that is essential for Windows. But due to the need of joint storage of WSFC, data copy is required for the protection of HA / DC so that storage of separate commercial or customized software to the server during the server server To use
For Linux, which does not lack a feature of WSFC, additional HA / DRRs are also needed. Using multiple open source software, there is a need to meet many capabilities, which must include data copy, server clustering, and heart control surveillance with minimum, failure / failure terms. . But because the entire HA may be unusual to prevent working properly under the approval of the failure, only a large number of organizations need to consider taking this task.
Clustering for the failure of the cloud – build the purpose
The increasing popularity of private, public and hybrid clouds has increased the use of insufficient clusting solutions specifically designed for the cloud environment. These HA solutions are fully implemented in the software, as their nomination means, in order to ensure high availability at the server’s level of application with the server’s storage server and storage. Failure.
Most of these solutions provide a complete HA / DD, which includes real-time block-level data movements, continuous application monitoring and configuration failure configuration / failure policies. included. Some more innovative solutions also support for high-failure clustering such as in the lower-cost standard edition of SQL Server, maximizing double-multi- regional performance and manual switching of basic and secondary server assignments for both Windows and Linux WAN correction to optimize. For planning maintenance facilities, including the ability to perform regular back-up without interruption in the application.
Although the HA / D solutions prepared by these goals are generally storage-agenist, they enable them to work with shared storage area networks, jointly anything neutral failure clustering is usually the only possible failure of failure. Prefer the ability to finish points. Most neutral failure clusters are also applications – aggressive, for active organizations alone, Universal HA / DR solution. It protects the entire SQL Server application, including database, logs, agent jobs etc. in the same capacity.
Examples of EC2, for example, in general, a common two-node shows that the neutral failure of the cluster works with Windows or Linux. Cluster is configured as virtual private cloud with two SQL Server nodes in different available zones. The use of block-level movements compatible with two available zones has high availability and high performance. A part of the file, which needs to get a course, is done by the domain controller in a separate availability zone. If any zones go offline, retaining every server example quorum in a different zone eliminates the possibility of losing multiple votes.