Configuring Microsoft HPC Pack for High Availability with SOA Applications

Article
11/20/2015

Applies To: Microsoft HPC Pack 2012, Microsoft HPC Pack 2012 R2

This guide provides procedures and guidance for deploying HPC Pack in a configuration designed for high availability for applications based on service-oriented architecture (SOA). For this configuration, you configure the head node (specifically, a set of head node services) in a failover cluster so that the head node can run on one of the servers. Then you can configure one or more Windows Communication Foundation (WCF) broker nodes (specifically, the HPC Broker Service) in one or more separate failover clusters.

Each failover cluster contains multiple servers that work together, so that if there is a failure of one server in a failover cluster, another server automatically begins providing service (in a process known as failover). Even if an outage occurs on a server, existing jobs can continue to run and new jobs can be submitted.

Setting up failover clusters enhances the availability of head node and WCF broker node services for unscheduled and scheduled outages.

Unscheduled outages If there is a failure of a resource such as a hardware or software failure on a server that is acting as the head node or as a WCF broker node, the result is a brief interruption during failover. After failover, you can work on the server that has the problem while the needed functionality continues to run on another server in that cluster.
Scheduled outages You might not want to lose functionality during routine node maintenance (for example, applying a service pack or hotfix, upgrading hardware, or updating an application). If your head node or WCF broker nodes are configured in a failover cluster, one server in a given failover cluster can support needed functionality while you perform maintenance on another server in that failover cluster.

Important

Before you begin your deployment, review the Release Notes for your version of HPC Pack.