Browse Prior Art Database

Smart job actions sequency dynamic analyzer

IP.com Disclosure Number: IPCOM000241273D
Publication Date: 2015-Apr-10
Document File: 5 page(s) / 87K

Publishing Venue

The IP.com Prior Art Database

Abstract

Normally a scheduler system is able to submit batch jobs according to a user pre-defined plan. As a consequence, of the batch job activity, a sequence of actions is performed. Under particular circumstances it can occur that the sequence of actions is not the "normal" and expected one and this can cause errors in the future. A system that realizes a sequence outside the normal flow, might help to alert and take timely actions.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 37% of the total text.

Page 01 of 5

Smart job actions sequency dynamic analyzer

    Usually, a scheduler system is able to submit batch jobs according to a user pre-defined plan. As a consequence of the batch job activity, a sequence of actions is performed and a set of messages is issued in the system log to notify the user about the execution of these actions.

But, even if the sequence of job execution is the same (same as in the schedule plan), it can occur that some events (as a consequence of the job execution) may occur out of sequence for several reasons (like system workload, data quantity, resources unavailability, modified data in batch, etc). The user can be notified as usually by messages (like job action start/end), but there is no explicit notification about event flow possible inconsistency. User can be not aware that a recovery action might be needed in front of the actual flow sequence.

    Hereafter is disclosed a System which, based on normal sequence of jobs caused action patterns, (for example dataset operations like Read, Write and Delete sequence: RW RW RW RW D) identifies anomalies when modifications over the "normal" are dynamically detected ( ex dataset RW RW RW W RW D vs RW RW RW RW WR D). A warning notification can be sent to the user, highlighting where an alteration occurred in the scheduled plan compared with the "normal" pattern, and highlighting the involved elements and eventually suggesting some actions which might be needed to recover (the easiest one might be to contact the correct batch

job owner/expert).

    The proposed method might be part of a scheduling system, which owns the scheduling plan, submits the jobs, and monitors for the activity normal completion.

    The system is able to verify, besides the normal completion of planned jobs, also the flow of operations originated by these jobs, identifying if they are outside a "normal" sequence.

    As part of this system there are the following major components:
- Monitor
- Pattern builder
- Pattern analyzer
- Alert builder
- Recovery identifier

The following Figure 1 depicts the described system components and their relations.

1


Page 02 of 5

Figure 1

Monitor

    This component is responsible for monitoring the system log and the involved products message logs to retrieve the actions in progress, originated by the job execution, based on the messages. There is a certain number of pre-defined messages needed to be monitored (Monitoring Msg List file). These messages can span from standard operating system messages to specific application messages. The list of the messages to be monitored is filled in by default with some of these messages but can be customized by the user to exclude or include some messages (like user defined application messages).

     When a message present in the Messages monitoring list is encountered in the logs, its presence, date, time and job/task name are recorded into a record written in a issued messages data base (Recognized Messages data base).

As part of the input customizatio...