Relying of fine-grained time for sheduling tasks is usually a bad idea... If for some
reason the previous task will take longer, you will load your server with the next
one, which will take even longer, and over-load it with the following one...
You can rely on cron, and serialize your tasks, using the standard Linux
/etc/cron.hourly /etc/cron.daily scheme : put your scripts in these directories, and
tasks will be simply scheduled, without interfering...
Nicolas huillard
-----Message d'origine-----
De: dialFree administration [SMTP:[EMAIL PROTECTED]]
Date: jeudi 25 mai 2000 16:23
À: [EMAIL PROTECTED]
Objet: [ADMIN] Scheduling Database Tasks
Hi there.
I am a PostgreSQL newbie recently rescued from the depths of despair (MS SQL
Server). In this state however I find my self pining for some of the
functionality I have become accustomed too. We have a number of tasks that
we like to perform using our database like sending emails to folk or billing
them or generating invoices.
MS SQL Server has a Jobs interface as part of it's database management
'suite' that allows various types of scripts (SQL, Active X) to be performed
with fine granularity (1 Second and up). Does PsotsgreSQL have any similar
functionality?
Crontab does not provide the granularity we require (3 seconds or so for
some of our tasks). We are also investigating Expect. If this functionality
is not available can it be created efficiently with any of the PL/tcl etc.
scripting/functional languages?
Thanks for your time.
-Ian
Ian Marshall
[EMAIL PROTECTED]
www.talarion.com
(617) 527-1994
<< Fichier: ATT00023.html>>
