NonStop Insider

job types


Site navigation


Recent articles


Editions


Subscribe


For monthly updates and news.
Subscribe here
NonStop Insider

Insider Technologies announces updated release of MultiBatch

Insider Technologies

DanDan

Insider Technologies logo

Insider Technologies is pleased to announce the latest release of MultiBatch – a mainframe class batch scheduler product for the HPE NonStop server arena.

MultiBatch allows users to configure and automatically sequence any number of batch jobs to create an applications off-line processing flow.

This batch program execution can be monitored for start time, length of execution and successful and unsuccessful completion states. Based on these criteria, problems can be escalated to other management applications, or other batch jobs can be started.

The package comes equipped with a configuration GUI and/or Pathway interface, with a full set of reporting and management tools.

Following extensive customer research, MultiBatch has been enhanced to offer extended benefits in coldload, disaster recovery and a new feature in DS-EventTimer: Crontab.

CRONTAB

Crontab is a powerful, multi date/time based scheduling component of the MultiBatch
DS-EventTimer. A user can configure a job, segment, or unit to run at any time of the day, several times a day, for 24 x 7 days of the year / years, e.g. start job at 08:45, 09:30, 11:12, only on a Wednesday, Friday, if date is 30th. Any combination is possible. CRONTAB is equivalent to the Unix Cron scheduler.

System Coldload

Traditionally, the coldload of a NonStop system invokes TACL based ‘coldload’ files. Contents of these files include commands to start TMF, DISKS, TCP/IP, LAN, EXPAND, RDF, Spoolers, etc. with each command line being invoked sequentially.

Although each command line in the startup files can be invoked sequentially and in turn load up the NonStop system, the parallelism of MultiBatch can speed up the coldload operation. This is especially beneficial when, following a site-swap, the ‘new’ Production node needs to be started up as soon as possible.

The MultiBatch units can execute the required site-swap commands using different user ids, e.g. a unit for STMF.AA (Start TMF) is run as 255,255, whereas the units for starting the various Pathways may be run under different ‘Manager’ ids.

The parallel approach of MultiBatch will facilitate the rapid coldload of a NonStop system, thereby reducing downtime and improving SLAs for customers.

Disaster Recovery

MultiBatch allows jobs to be placed on hold, or in a ‘runoff’ situation, whereby any units that are configured with a ‘runoff’ value will be skipped. So, if a site-swap is required, operations can amend all units simultaneously on the standby system from ‘runoff’ to ‘runon’ and the schedule can be started.

MultiBatch can be used in a site-swap situation with the batch schedule executed on the new Production node. As with a Coldload MultiBatch schedule, the units can run under different user ids.

Batch Recovery

In addition to fully tokenised MultiBatch EMS events and subsequent updates to the status tables by Watcher, the status of a batch schedule including waiting, running, failed jobs is dynamically recorded in a TMF audited BMON Recovery file.

If a BMON process is shutdown whilst the schedule is still running or perhaps due to a system crash, then recovery techniques are available to recover the BMON back to the point of failure.

Once recovered, this enables a user to determine the state of the batch schedule, e.g. jobs JOBA to JOBG are complete but JOBH is marked as incomplete as it was still running when the BMON process failed. A user can then determine the effect this has on their application.

Additionally, as the BMON Recovery file is TMF audited, data replication products can be used to maintain a copy of this file on a standby system. As the BMON Recovery file is regularly updated during the running of the batch schedule, the copy will be simultaneously updated on the standby system.

Therefore, if a system crash occurs, then following a site-swap, the BMON on the new live system can be rebuilt and users can check the status of the batch schedule up to the point of system failure.

Summary

MultiBatch provides mainframe batch scheduling capability for the NonStop platform. This functionality can be applied to both GUARDIAN and OSS executables.

Users can exploit the parallel nature of this architecture by building a parallel batch schedule that can execute concurrently across one or more networked nodes.

Configuration, execution and monitoring can be performed from a single point in a secure audited environment.

All the existing program execution options, such as CPU selection, high pin nomination, params, assigns and defines are supported.

Sophisticated scheduling and calendaring facilities can be used to provide a flexible selection of jobs for any given day.

Job execution can be performed through TACL processes, but the product also utilizes a low-level procedure call interface to invoke new processes. This approach provides a more efficient solution and negates the need to build and maintain TACL macros.

The system will monitor itself by scanning for “not executed” or “over executed” steps. Alternatively, failures can be escalated to EMS so that they can be displayed on management consoles.

You can follow Insider Technologies on Twitter @insidertech, on LinkedIn, or get in touch via hello@insidertech.co.uk

Karl Gilbank | Sales & Marketing Director
+44 161 868 6658
www.insidertech.co.uk