Continuous data protection, maximizing uptime of a VMware vSphere virtualized data center. Enterprise downtime is costly and sometimes irrecoverable. Global-Active Device (GAD) is an active-active data protection model that can effectively host transparent storage-site failover without reconfiguration. In this exercise, we are combing Global-Active Device with VMware’s vMotion, conducting vMotion and HA test cases on virtual machines running SAP HANA.
This paper is written for IT professionals such as storage administrators, ESX host administrators, and application administrators who are charged with managing large, dynamic environments. It assumes familiarity with SAN-based storage systems, VMware ESX and general IT storage practices. For more information about Global-Active Device, please visit refer to Use Hitachi Global-Active Device to create a VMware Metro Storage Cluster for Active-Active vSphere Workloads and Implement vSphere Metro Storage Cluster with Hitachi Virtual Storage Platform (VSP G /F) Storage Array Platforms (2145375).
Note — Testing of this configuration was in a lab environment. Many things affect production environments beyond prediction or duplication in a lab environment. Follow the recommended practice of conducting proof-of-concept testing for acceptable results in a non-production, isolated test environment that otherwise matches your production environment before your production implementation of this solution.
Diagram below shows the overview of the high-level logical design with the VMware Metro Storage Cluster solution implemented using global-active device.
Testing and Results
128 GB SAP HANA VM
Duration of Test Case
vMotion VM from Site 1 to 2 (vSphere HA)
vMotion VM from Site 1 to 2 with 103 GB of data in memory (vSphere HA)
vMotion VM from Site 1 to 2 while loading 103 GB of data to memory (vSphere HA)
Site A unexpected shutdown. Site B to discover and bring up VM and loaded SAP HANA database into memory (vSphere HA)
We tested four different scenarios. In each of the cases, we moved the 128 GB in-memory database virtual machine from Site 1 to Site 2. In the first three test cases, we noticed the data successfully cut over from one storage and continued using the secondary storage. There was no disruption nor inconsistency, so the virtual machine did not require a reboot. The fourth test case was a disruptive case, but the VM was able to quickly come up since the secondary had a copy of the virtual machine. With the combination of vMotion/vSphere HA and global-active device, a very robust and highly available solution is created.