How do I configure storage for vHANA in scale-up and scale-out deployments? How do I implement application aware SAP HANA host auto-failover in virtual environments? What steps need to be taken in order to setup a simple and efficient VMware HA solution? What are the pro’s and con’s of these two options? RDMs vs vmfs? Do you want to see an elegant solution to build a HANA shared filesystem with fault tolerance?
The new white paper “VMware virtualized SAP HANA with EMC storage” explains this perfectly in detail. To make you some appetite to read this excellent white paper let me highlight some aspects.
HIGH AVAILABILITY FOR VIRTUAL HANA
There are two ways to implement high availability in a virtual HANA environment
- Application aware SAP HANA host auto-failover using EMC’s virtualized HANA storage connector (EMCvHConnector). This enables use of a HANA standby node in all scale-out deployments on vSphere.
- VMware HA, a mechanism that detects failures of the virtual machine at OS level and can also detect ESXi host failures. If required, it restarts the virtual machine and moves it to another physical server.
The following table from the white paper summarizes the pro’s and con’s of these two options:
HANA SHARED FILESYSTEM WITH FAULT TOLERANCE
HANA scale-out scenarios require a shared file system that is mounted on every HANA node within the cluster and stores the installation binaries, configuration files, and trace files. In physical environments, this file system is generally provided as an NFS share by a NAS array such as a VNX series unified storage system.
If a NAS array is not available, vSphere with native Linux functionality provides a viable alternative. A Linux virtual machine (non-HANA node) running an NFS server process provides the NFS share. This process exports a file system that is mounted on all the HANA cluster nodes. You can use vSphere Fault Tolerance to achieve reliability. This is also a very good approach for solutions using EMC ScaleIO software-defined storage.
Because the HANA nodes depend on the shared file system, it is recommended that you turn on the fault-tolerance function for the NFS server virtual machine. To do this, go to the vSphere user interface, highlight the NFS Server virtual machine, and select and enable Fault Tolerance. It’s just that easy…
vSphere fault tolerance provides continuous availability for the HANA share during an NFS server failure. vSphere creates a live-shadow virtual-machine instance that is always up to date with the primary virtual machine. If there is a hardware outage, vSphere automatically triggers failover, ensuring zero downtime for the HANA share and preventing data loss. After failover, vSphere creates a secondary virtual machine to deliver continuous protection for the HANA share.
Again, for more aspects and details, please take a look at the white paper “VMware virtualized SAP HANA with EMC storage” and enjoy the reading.