From: | daveg <daveg(at)sonic(dot)net> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | "Matthew T(dot) O'Connor" <matthew(at)zeut(dot)net>, pgsql-patches(at)postgresql(dot)org |
Subject: | Re: patch for pg_autovacuum 8.0.x prevent segv for dropped tables |
Date: | 2005-10-20 19:38:42 |
Message-ID: | 20051020193842.GD2329@sonic.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-patches |
On Thu, Oct 20, 2005 at 12:30:27PM -0400, Tom Lane wrote:
> "Matthew T. O'Connor" <matthew(at)zeut(dot)net> writes:
> > Tom Lane wrote:
> >> Surely this is completely broken? AFAICT you are testing the result
> >> from a VACUUM or ANALYZE command, which is not going to return any
> >> tuples.
>
> > Upon further inspection, I think you are right. I would think that
> > instead of checking the query result with PQntuples, it should probably
> > be checked with |PQresultStatus.
>
> ISTM this is the wrong place to test at all. I put a PQntuples check
> into update_table_thresholds() instead, which seems a far more direct
> defense against trouble. (Consider eg the case where someone drops the
> table just after your VACUUM completes successfully. Also there are
> drop/rename scenarios to think about: success of the VACUUM proves that
> there is a table named FOO, not that there is still a table with the OID
> you have on record.)
Yes, I agree, update_table_thresholds() is the right place for the check.
Please ignore the earlier patch.
Thanks
-dg
--
David Gould daveg(at)sonic(dot)net
If simplicity worked, the world would be overrun with insects.
From | Date | Subject | |
---|---|---|---|
Next Message | David Fetter | 2005-10-20 23:39:08 | Error in trigger example |
Previous Message | Andrew Dunstan | 2005-10-20 16:33:48 | Re: patch for pg_autovacuum 8.0.x prevent segv for dropped |