| Previous | TContents | YIndex |
H SCHEDULE also enables you to differentiate between application G programming and operation analysis. That means application programmers H can concentrate on program logic and operation analysts can develop job ? schedules best able to cope with the pressures of a production environment.
F SCHEDULE can automate the running of all types of repetitive @ production jobs on the appropriate day of the week or month, at n specific intervals, or according to a defined calendar. See Section 2.1.1 D for more on job rescheduling. A wide variety of repeat patterns and - interdependencies between jobs can be set up.
A Applications destined for automated `lights-out' and `hands-off' F operations can be assembled from simple batch jobs into sophisticated E job streams with conditional branching base on application or system detected error conditions.
H The SCHEDULE System is partitioned into two basic functions. A F server process that performs all primary operations. Several G different client foreground programs that send basic function A and action requests to the server process. Three types of client & programs are included with the system:
F Several types of data are maintained by the SCHEDULE System. E To simplify the management of a large number of jobs the database is A organized with a top level directory and optionally one sublevel n directory. Any number or type of directories can exist. See Section 2.3.3 for more on directories.
H These directories can be assigned to specific users or groups of users. H Inside each directory jobs and calendars can be defined. Once a job has F become active a controlling entry is maintained in the schedule queue F system. The scheduling queues represent the various states that a job F can be in (for example waiting for disk space). As the job progresses C it is moved from one queue to the next. Whenever such a transition D takes place an entry is made into the history file. A replay of the C history data can be made to determine when and in what order these transitions took place.
G An active job is one that has an entry in the scheduling queue system. @ Only jobs that are active will ever be submitted into batch for D execution. For a job to become active any one of the following must happen.
H The SCHEDULE System creates batch jobs. These jobs are entered A into the standard VMS batch system. As such they can be E directed into generic cluster wide queues or node specific execution H queues. The single biggest advantage to using this approach is that all D jobs, whether created by the SCHEDULE System or created by G other means appear in a common queue structure and can be managed in a = cohesive and centralized fashion by all standard Digital and non-Digital utilities.
G All such jobs are properly recorded in the VMS accounting log G as batch jobs. Jobs that are directed into cluster wide generic queues F are readily revectored from one node to another as nodes go down. The @ execution attributes (such things as quota and priority) can be / controlled for both the queue and for the user.
B Scheduled jobs consist of operational control information and DCL E commands, which are executed in any designated standard VMS . batch queue. The control information includes:
| Previous | VNext | TContents | YIndex |