From: | Yann Michel <yann-postgresql(at)spline(dot)de> |
---|---|
To: | "Jonah H(dot) Harris" <jharris(at)tvi(dot)edu> |
Cc: | Josh Berkus <josh(at)agliodbs(dot)com>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: User Quota Implementation |
Date: | 2005-06-15 07:18:05 |
Message-ID: | 20050615071805.GA26713@zoom.spline.inf.fu-berlin.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Hi folks!
On Tue, Jun 14, 2005 at 11:39:06AM -0600, Jonah H. Harris wrote:
> >>On second thought, we need to have a GUC for this, whether I want it or
> >>not. It needs to be optional to the log, yes? So it would be:
> >>log_tablespace_full = %
> >>with the default being "0" (don't log).
> >
> >On third thought, could we do this as part of the maximum size
> >declaration? Like:
> >
> >ALTER TABLESPACE tbsp1 SET MAXSIZE 128M MAXWARN 80
> >
> >That would be ideal, since the % you might want could vary per tablespace.
> >This would be emitted as a WARNING to the log every time you run a check
> >(e.g. after each commit).
Yes, that is the best idea, I think. What I don't like with the GUC
variable is, that it is _one_ warninglevel for all tablespaces
independent of their character. If I have two tables in different
tablespaces both with the global 90 percent threshold I may become
warned that tablespace A is 90 percent full, but if this tbsp. will only
grow one percent per day I still have 10 days left. Otherwise if I have
a tablespace B with 9 percent full but 5 percent growth per day that
will only be two days. So normaly id would have been fine to be warned 3
days ago for tablespace B but 5 "to early" for tablespace A.
Did you get the idea?
Regards,
Yann
From | Date | Subject | |
---|---|---|---|
Next Message | Dennis Bjorklund | 2005-06-15 08:42:46 | Re: LGPL |
Previous Message | Qingqing Zhou | 2005-06-15 05:41:52 | Re: Autovacuum in the backend |