vSphere High Availability
The content of this document is based on test results of Unify Deutschland GmbH & Co. KG and demonstrates how to set up an HA environment with the IACBOX and VMware vSphere.
Product Information
VMware vSphere
The Software vSphere is the virtualization platform of VMware which can be used to virtualize and manage servers. VMware ESX serves as operating system to assign virtual machines to a target system. The administration of the virtual machines and the ESX servers is handled by the vCenter server, which can be accessed with the vSphere client. This management interface, together with shared network- and storage resources, enables a wide variety of possibilities. For example vMotion, which can be used to move and reallocate virtual machines to other servers on the fly.
Procedure
With VMware solutions there are several possibilities to monitor systems and react in case of system failures. Therefore the differences between HA (High Availability) and FT (Fault Tolerance) must be considered.
vSphere High Availability (HA)
The HA feature of vSphere can react to failures and automatically restart the affected VM in the cluster on another ESX server.
In case of failure of a server the virtual machines will be moved on to another server and restarted. The downtime is limited to the restart of the virtual machine. It should be noted that a virtual machine failure can result in data loss. To use vSphere High Availability, a centralized SAN or NFS storage is required, so it can be used on all ESX servers. Further, a management network connection between all ESX servers, as well as an identical network configuration for the virtual machines on all ESX servers in the cluster is required. More information about this can be found on the VMware website
vSphere Fault Tolerance (FT)
With vSphere Fault Tolerance virtual machines along with their memory content will be replicated to other servers continuously, so that a hardware failure can, in the best case, be compensated without any downtime.
For each running virtual machine, an exact copy is being replicated while both, the original and the copy share the same memory and data. In case of failure, the replicated system becomes active and enables a seamless transition without the loss of data or downtime.