Job scheduling typically relies on using the UNIX clock daemon called “cron” or the Windows NT command line interface scheduler service. These utilities are relatively primitive when it comes to scheduling critical production jobs, because they are limited to a single machine; require cryptic instructions in a “crontab” file, and lack of any error-handling, logging, or notification capabilities.
AutoSys provides extensive scheduling functionality. Jobs can be scheduled to run based on time and date, file arrival, and/or job dependencies such as the SUCCESS, FAILURE, or exit code of other jobs. Jobs can also be scheduled to run (or not to run) according to custom calendars you create using the easy-to-use Graphical Calendar Facility. Scheduling power can be applied across multiple platforms anywhere on the network. In addition, AutoSys provides error handling, logging, notification, and a centralized RDBMS for jobs run by any user across the network.
One of the most important aspects of implementing a job control scheme is how to monitor the progress (or failure) of multiple jobs across heterogeneous platforms. Jobs may fail in the middle of the night, preventing the receipt of an important report in the morning, causing pagers to go off, signaling the need for human intervention. When multiple, interdependent jobs are scheduled to run, being able to quickly identify and rectify malfunctions is imperative.
AutoSys provides extensive Alarm and Monitor Services that comprehensively monitor every potential point of failure. Should a problem occur, alarms are immediately generated for jobs that encounter processing problems. And, if appropriate, AutoSys automatically takes the necessary actions to correct these problems. Managing jobs is simple using the AutoSys Operator Console, which is designed to provide system-wide control from a single graphical interface.
Another important aspect of managing job execution is the ability to generate and inspect intelligent reports on the progression of events. All too often, to avoid a “blind” expectation that everything occurred successfully and on time, additional programming must be done to create the necessary report mechanisms for a job control scheme.
AutoSys provides a suite of reporting utilities that clearly detail the various aspects of job execution. Real-time inspection of job status can be viewed using command line reports or the Operator Console, job dependencies can be listed, and the current state of a job is clearly displayed— regardless of the network machine on which it is being run.
High Availability and Error-Recovery
System down-time can impede important processing functions, and even stop the time-critical computing that is vital to your business. If and when a “crash” occurs, your hardware and software systems must stay operable and be recoverable.
AutoSys has engineered recovery and restart capabilities at many levels, from user-defined error recovery to the AutoSys engine itself. If a problem occurs, a job may be restarted, or paging and e-mail notification may be invoked. AutoSys can also run on two machines simultaneously, eliminating a single point-of failure.
When using “Dual Server Mode,” all information is written to both systems, which are actively mirrored by AutoSys. If an AutoSys component becomes unavailable, AutoSys will perform an unattended rollover to single machine operation, ensuring that jobs will continue processing without interruption. This “High Availability Option” ensures that your jobs will run on time in the event of a machine failure and that jobs will continue processing without interruption.