Re: Using pgAdmin and pgAgent with Greenplum

From: Dave Page <dpage(at)postgresql(dot)org>
To: "Roberts, Jon" <Jon(dot)Roberts(at)asurion(dot)com>
Cc: pgadmin-hackers <pgadmin-hackers(at)postgresql(dot)org>
Subject: Re: Using pgAdmin and pgAgent with Greenplum
Date: 2007-10-29 13:52:33
Message-ID: 4725E5A1.8020909@postgresql.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgadmin-hackers

Roberts, Jon wrote:
> I've looked at this code all day long and tried many hacks to make it work
> with GP but there isn't a way.
>
> The easiest way I can think of to handle this with the least amount of
> change to the architecture is to add another column to pg_job. Maybe a
> Boolean called jobcompleted. Then create a view called vw_pg_job which
> executes the function pga_next_schedule that returns the jobnextrun value.

I'm not sure why you'd need the extra column, but there are some issues
with using a view to replace the nextrun column:

- pgAgent queries each job every minute to see if it needs to run. That
could lead to a lot of cpu being used on that rather nasty pl/pgsql
function.

- When we spec'ed pgAgent, it was decided that if a schedule were
missed, that instance should run immediately on startup. The proposed
change would prevent that behaviour, though personally I'm not sure
that's necessarily a bad thing.

- The 'Run Now' feature in pgAdmin would need to be reworked - perhaps
by adding a flag to pg_job to indicate the job should run once
regardless of scheduling.

The second are more minor issues, but the first doesn't seem at all good
to me.

Regards, Dave.

Browse pgadmin-hackers by date

  From Date Subject
Next Message Roberts, Jon 2007-10-29 17:17:50 pgAgent bugs?
Previous Message Raphaël Enrici 2007-10-29 13:40:38 Re: keyboard shortcuts in context menus are not