Re: dbms_jobs versus cron

Rauf Sarwar wrote:

> Hanne Iren Midttun <hannem_at_tihlde.org> wrote in message news:<Pine.LNX.4.21.0301080901580.19558-100000_at_colargol.tihlde.org>...
> > Hi,
> > from another thread (Oracle stored procedures vs Running from a flat .sql
> > file), I saw that it was stated that dbms_jobs was better then cron.
> >
> > I have only runned dbms_jobs on 7.3.4 AIX (I think it was the right
> > version, it's been a while), and on that version cron was far better.
> > (ie the dbms_jobs did not run at all and so on)
> > I was really sorry to migrate to cron because moving a database from one
> > machine to another was really a lot of work - I allways forgot a script, a
> > directory, some rights on the filesystem and so on.
> >
> > Now I am running 8.1.7.4, AIX and thinking of migrating to 9i. does
> > somebody have any experience with dbms_jobs on this versions?
> >
> > regards Hanne
>
> Any non-interactive process that can be scheduled inside the database
> will always be portable, safe and secure. On the other hand, same
> process from outside the database will run the risk of password
> exposure.
>
> Also... what if the database user's password is changed that is being
> used by the cron job? Cron job is hosed but dbms_job is owned by the
> user and will continue to run no matter what the password is.
>
> Just a few observations to tip the balance in favor of dbms_job.
>
> Regards
> /Rauf Sarwar

To which I will add another in favor of dbms_job.

All actions coded in a single language meaning any PL/SQL programmer can develop, test, and maintain the code. This is not true of a shell solution that requires a SysAdmin, usually far fewer in number, to write and maintain.

What is in the database can easily be written in development, moved to test, and then to production because the machine environments are irrelevant. How many times do we find dev, test, and prod boxes with identical configurations in the real world? Not often enough.

Daniel Morgan Received on Sat Jan 11 2003 - 13:15:16 CST