Browse Prior Art Database

A solution to automatically reconfigure crontab files on a High Availability cluster of LINUX or AIX systems

IP.com Disclosure Number: IPCOM000030971D
Original Publication Date: 2004-Sep-03
Included in the Prior Art Database: 2004-Sep-03
Document File: 3 page(s) / 60K

Publishing Venue

IBM

Abstract

The article that I am introducing proposes a simple system in order to automatically manage the job's scheduling within a cluster of computers configured in high availability

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 48% of the total text.

Page 1 of 3

A solution to automatically reconfigure crontab files on a High Availability cluster of LINUX or AIX systems

The availability of mission critical applications can be maximized by using a system with redundant computers and with special software to manage them. The invention deals with this kind of system as High Availability cluster (HA cluster), to the computers of an HA cluster as nodes, to the management software as HA software and to the mission critical applications as HA applications.

If a HA application runs on a node (primary node) and this node stops for some hardware or software failure, the HA software causes that another node (backup node) takes over the HA application. The HA cluster has various issues regarding the alignment of the operating system configurations between the nodes, for example user definitions, software updates, pheripheral configurations, etc. Many of these issues have been solved by embedding within the HA software the capability to automatically maintain the alignment of different configuration aspects. Except the situations where it is utilized an expensive system dedicated to the job's scheduling on several computers, like as an example Tivoli Workload Scheduler (TWE), the HA clusters based on IBM AIX and LINUX operating systems still have the problem of the reconfiguration of the job schedulers, based on crontab (CRON TABle) files (' cron' is a utility that allows tasks to be automatically run in the background at regular intervals by the cron daemon often termed as cron jobs. Crontab is a file which contains the schedule of cron entries to be run and at specified times). Each job scheduled within a crontab file corresponds to a row of this file and each user of the operating system can have his own crontab file defining the jobs that the user schedules. If a crontab file of a node schedules jobs regarding an HA application, these jobs must be also scheduled on the node that eventually takes over this application. Normally, this task is manually performed by simply preventively copying the crontab file defining such scheduled jobs on all nodes each time some of these schedulations have been modified. The raw alignment of the crontabs files prevents the possibility of job scheduling not regarding the HA applications and different for each node. Therefore, the alignment of the crontab files between the cluster nodes must be carefully performed by manually inserting within the crontab file of each node only the rows that schedule jobs regarding the HA applications. Moreover, this kind of alignment of corntab files between the cluster nodes creates the problem that the scheduled jobs regarding an HA application can be correctly executed only on the node where this application actually runs. In fact, on the other nodes errors will be generated because resources like scripts and data files, necessary to the scheduled jobs, are missing. Such kind of errors can be avoided by embedding the job to schedu...