SQL Server High Availability (iii) shared disk

A. Shared disk

Shared disks may be used in cluster technology. This type of disk can be accessed concurrently by multiple nodes, but only the primary node has access to the shared disk at any one time.

Second, use the shared disk scene

1. Quorum disk

When building MSFC, if it is an even number of nodes, then you can add a quorum disk, so that the vote can form a "majority."

2. SQL Server cluster data disks

The essence of SQL Server cluster is to put the data on a disk shared by all nodes, and when the primary node fails, the next node can successfully start the SQL Server instance (service) by acquiring the rights to the shared disk. From the client's point of view, the data does not seem to change (because the data is on a shared disk) and the service is not interrupted (MSFC redirects the client connection to the new node).

For AlwaysOn availability groups, you do not need to use shared disks. Each node of an AlwaysOn availability group has its own separate instance of SQL Server (service), and the instance accesses a local disk (or a network disk as a locally exclusive disk).

This article from "We chased the MSSQL" blog, declined reprint!