Shell Programming and Scripting

Hi,

Since I staterd working as Unix sysadmin (about 3 years ago) I always used to trigger a process evaluating the conditions needed to this process to be executed. Recently I've change the company where I work, and they usually create a trigger file to start a process or to stop a process while another one is doing something.

Now I'm developing a little script that should do things if a field in a Oracle database table is changed from "idle" to "started". Which do you thing is the better way to do it? Connecting each 5 minutes to the DB and checking? Or creating a trigger file from de database when the value is changed in the table and let the script check if file exist?

not so much a matter of best practices, as it's really more situational in nature...

What's the most reliable option under the circumstances? Is there a stored procedure that would be creating the spool file that drives this setting? Couldn't this SP just go ahead and initiate whatever it is that the "watcher" script is processing? Is there a need to keep it functioning even if there's a DB outage? Could the "watcher" script just do a quick look-up on the DB to capture the state as a ${variable:-default}? Do you mind having to manage the framework required to keep a file watcher in sync with the DB, and vice versa?

It's up to your scenario and most of this is purely academic, because the real best practice here is what works most effectively under your environment.

Hope this helps. Best of luck...!