Re: Trigger Fail in Job execution

FC wrote:

> "Howard J. Rogers" <dba_at_hjrdba.com> wrote in message
> news:aghcl6$dhv$1_at_lust.ihug.co.nz...
> > Not really.
> >
> > The only tip I have, in all seriousness, is that I would sooner shag a
> dead
> > sheep than use 8.1.5. Like the sheep, it is riddled with bugs. You need to
> > upgrade to something supported. Like 8.1.7.
> >
> > Regards
> > HJR
> >
>
> I might agree,
> but I encountered similar problems when using something like that on 8.1.7
> (Win2K and PE on Win98).
> If I remember well, the manual (Admin Guide) states that unlike other
> background processes, scheduled jobs are automatically restarted when they
> fail, up to 16 times, after that they are flagged as "broken".
> Now, if your job fails and for some reason when it is restarted it quits
> because some required condition is not met, it is like hello, no trace of
> anything, not even the "broken" job in the view user_jobs. Well, almost no
> trace,
> probably if you turn on oracle tracing and/or logging you might see
> something somewhere.
>
> By the way, is the trigger body declared an "autonomous transaction" ?
> Is there an error trapping clause in the trigger body ?
>
> Bye,
> Flavio

If a developer writes a proc, creates a job, whatever, and doesn't have a table that logs start_time, end_time, and any generated ORA messages (as the first and last things the proc does) it says something about the importance placed on the application. Logging doesn't just catch what runs and how long it runs, it catches what doesn't run, what breaks, and why it breaks.

Daniel Morgan Received on Wed Jul 10 2002 - 16:14:49 CDT