vSphere HA Cluster

VMWare's vSphere is a virtualization environment that allows users to create clusters consisting of multiple hosts.

A host is a physical computer which delivers the computing power to run VMs. In order to coordinate the resource allocation, vSphere installs the ESXi hypervisor on each host. The VMs themselves are saved on storage devices which are combined into one or multiple virtual DataStores. In addition to normal clusters vSphere also allows the configuration of HA clusters. VMs in a HA cluster are automatically restarted on a different host if a failure is detected.

Requirements

  • Valid WinCC OA license.
  • SEH Dongle Server (only necessary when High Availability is used)

Installation

Since vSphere HA automatically restarts VMs on different hosts, WinCC OA should be configured as a service. Additionally, if a server is also used as a workstation with UIs, the operating system of the VM can be configured to not require user input during startup. However, this imposes a security risk for your server and is not advised.

Furthermore a SEH Dongle Server has to be set up for licensing purposes when High Availability is used.

WinCC OA

  • Start WinCC OA as a service. Information can be found here and here.

VMware vSphere

For general questions regarding VMware vSphere please refer to the vSphere documentation: https://docs.vmware.com/en/VMware-vSphere/index.html

For additional information on networking and availability please refer to:

http://pubs.vmware.com/vsphere-60/topic/com.vmware.ICbase/PDF/vsphere-esxi-vcenter-server-60-networking-guide.pdf

http://pubs.vmware.com/vsphere-60/topic/com.vmware.ICbase/PDF/vsphere-esxi-vcenter-server-60-availability-guide.pdf

VMware Tools

WinCC OA supports VMware Tools 12.4.5.

For general questions regarding VMware tools, please refer to the documentation: https://docs.vmware.com/de/VMware-Tools/index.html

Notes and Restrictions

Transaction logging

In order to minimize the probability of database corruption due to a non proper shutdown of the system, transaction logging is recommended.

Since data is written block by block into the RAIMA database, any unsaved block can be lost at a system crash and the stored data can be corrupted. Additionally the database has to be repaired for a few minutes after every crash and therefore the startup is slower. Transaction logging, however, prevents a permanent damage of the database since an internal (data internal) transaction is only finished after all data is written to the disk. When transaction logging is activated, no database repair is necessary after a crash.

To enable transaction logging, add the following entry to your config file:

[data]

transactionLogging = 1

Note:

The use of transaction logging is recommended in any case. Though it is most useful in Standard Projects, since in a Redundant Project, the database does not need to be repaired, unless both server fail at the same time. As a consequence of the securer functionality, transaction logging causes the Data Manager to operate slower during some engineering steps (e.g. Ascii import).

Note:

In order to disable transactionLogging, set the config entry to 0. TransactionLogging cannot be disabled by deleting the config entry.

Redundancy

vSphere HA provides high availability through automatic restarts of failing VMs or VMs on failing hosts, but does not provide any redundancy. In order to ensure data integrity in addition to high availability, a redundant WinCC OA project is recommended.

To avoid a single point of failure, the two VMs with the redundant WinCC OA projects should be stored on separate DataStores. Additionally the two VMs should not run on the same host. This has to be set in the VM/Host-Rules:

  • Browse to the cluster in the vSphere Web Client navigator
  • Click the Manage tab
  • Click Settings and scroll down to Configuration
  • Click on VM/Host-Rules and click on Add
  • In the dialog box, enter a name for the rule
  • From the Type drop-down menu, select Separate Virtual Machines
  • Click Add
  • Select the two VMs to which the rule will apply and click OK
  • Click OK

With a redundant WinCC OA project and these VM/Host-Rules, at least three hosts are needed to take advantage of the automatic failover feature provided by vSphere HA. With two hosts vSphere HA would not be able to restart a VM in case of a host failure.

Restrictions

  • The VMWare Fault Tolerance feature is not supported.