Jesse Robinson wrote: >Depending on the one-jobname restriction was indeed a 'convenient' method of >serialization--but a method fraught with peril.
Yaaa, tell me. ;-) >One-jobname was always a poor man's solution to a complex problem. With even a >modestly sophisticated job scheduling system, serialization was achieved far >more capably. Very true. Use Automation Software to schedule jobs based on Return Code (from SYSLOG or from jobs themselves) and success/failure of previous jobs ran on same or other LPARs. Use the Automation Software handling of 'incoming' and 'outgoing' condition codes. With this setup there is not a problem with same or different jobnames and the order in which they are running. One example: every day we run SMF Dump, Automation Software collects the results of those SMF jobs and pass it on as conditions to release other jobs which post process SMF data. The jobnames can be the same or not, it does not matter. Before we got Automation Software, it was a real PITA to ensure jobs with the same names ran in the right sequence. The Operators sometimes released the wrong job, surely all h*ll broke loose when the angry client had to restore their data. Groete / Greetings Elardus Engelbrecht ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
