Home > Workload Solutions > SQL Server > White Papers > Deploying Dell EMC SRDF/Metro Smart DR with Microsoft SQL Server for Physical and Virtual Environments > Test Case 2: Moving entire SQL Server role to another node
In this scenario we will demonstrate a failover scenario in which SRDF/Metro allows an application to restart very quickly even after the SQL Server role moves to another surviving node. Because SRDF/Metro provides high availability for shared disk resources, the Windows failover cluster can quickly bring up all resources online on the surviving node. The application can be restarted right after that failover completes. During this test, SQL Server will move from the R1 side to another node with proximity to the R2 side, and the PowerMax system on the R2 side will start servicing I/Os. No changes to the SRDF group are necessary because SRDF/Metro provides active/active availability to storage devices on either side.
This step is similar to Step 1: SQL Server running normally on all shared resources, where the SQL Server based application runs normally on the R1 side of the SRDF/Metro setup. The captured application TPM information is shown in the figure below as a baseline.
Figure 35. SQL Server TPM on R1 side of SRDF/Metro
In this step, we demonstrate the failover test by manually moving the SQL Server role from one node to another. To perform manual failover follow these steps:
During an automatic failover, this operation will be identical. The Windows failover cluster will choose the surviving node and resume the SQL Server role on that node. The storage system that is local to the SQL Server after the failover will continue to service all the I/Os.
Figure 36. Moved SQL Server role from one node to another
After all resources including disk resources and the IP SQL Server agent come up on the surviving node, the SQL Server roles come back up and the application can resume its normal operation. The figure below shows the application life cycle before the failover (940K), at the time of failover (some brief downtime as SQL Server role moves), and after the failover (steady state 966K).
Figure 37. Application life cycle with failover