On Wed, Feb 11, 2009 at 1:10 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Alvaro Herrera <alvherre@commandprompt.com> writes:
>> Peter Eisentraut wrote:
>>> More generally, it was pointed out to me that users apparently do
>>> updates of pg_autovacuum to change settings on a bunch of tables at
>>> once. We might get some complaints if we remove that facility.
>
>> Hmm, argh. Maybe we do need the rule on a fake pg_autovacuum that
>> Itagaki-san was proposing.
>
> AFAIR we pointed out from day one that pg_autovacuum was a temporary
> API that we were not promising to keep around. Anybody who was coding
> against it with the expectation that they'd not have to change that code
> later was willfully ignoring the warning label.
>
>> There's a problem however; for pg_autovacuum you used to need to insert
>> some -1 values on columns on which you wanted to keep as defaults. On
>> the new code you need to skip the value altogether, and a -1 is rejected
>> with an error. Not sure how would we translate that.
>
> Maybe use a real table with an ON INSERT trigger that could contain some
> actual logic? But it'd probably still have to be custom-tailored to
> whatever application code was inserting things into pg_autovacuum,
> so it's not clear there's much point to writing that instead of fixing
> the application.
In any case it's not difficult to write a script that loops over all
of your tables with ALTER TABLE. It's probably not as fast as a
single UPDATE statement, but I suspect you'd need to have an enormous
number of tables for that to matter much.
...Robert
