|
HP OpenVMS Systemsask the wizard |
|
The Question is: How do I port my applications over to a VMScluster environment to take advantage of the auto failover capabilities? The applications are written in Digital Fortran 77. The Answer is : Assuming that your applications already permit the sharing of access to the application data, congratulations, you've likely already done almost everything necessary to take advantage of basic application failover within an OpenVMS Cluster. Application failover is a continuum. In a simple case, it can mean one node taking over when another fails, either through operator intervention or under program control. Automated techniques for failover can involve little more than a restartable batch job and a set of batch queues on cluster members. Further along in the continuum of application failover are application environments that operate entirely in parallel, recovering from failures but not particularly failing over the processing. How much additional work might be involved in adding application failover support depends greatly on what specific features of an OpenVMS Cluster configuration you might want to take advantage of, and what specific requirements are in place for your application(s). As a start, the OpenVMS Wizard would encourage you to read the OpenVMS Programming Concepts Manual -- as you read, you will find that there is not a particularly significant difference between programming for a single node and programming for operations within an OpenVMS Cluster. You will want to learn more about OpenVMS Cluster configurations, as well -- there are manuals for this purpose, as well. As you are asking this question, the OpenVMS Wizard would also strongly recommend learning about the synchronization issues discussed elsewhere here in Ask The Wizard -- search for the title "Debugging synchronization problems? (SMP)".
|