On Tue, Oct 6, 2009 at 9:43 AM, Alvaro Herrera
<alvherre@commandprompt.com> wrote:
> Jaime Casanova wrote:
>
>> now i think that what Euler shows me [1] is a fair compromise (this is
>> to allow this only when in standalone mode with system catalogs
>> allowed) otherwise we will have diferent behaviour for specific
>> (random) catalogs...
>>
>> [1] http://listas.postgresql.org.br/pipermail/pgbr-geral/2009-March/014374.html
>
> Hmm, I don't necessarily agree that having the system effectively shut
> down for the duration of the pg_largeobject move is a good idea.
>
well, my thinking was that if you know how to start in standalone and
know to allow system catalogs changes is more probable you did your
homework and read about the dangers it implies in other catalogs...
but yeah! the size of the pg_largeobject could be large enough to make
this something to worry about... let me ask the opinion of the bottle
of coke that is supporting me...
> I don't agree that pg_largeobject is a random catalog either -- it is,
> in fact, the only catalog in which an interesting size is to be
> expected.
>
i have just read Tom's comments and yes that question was around my
mind to: a system catalog that doesn't behaves like other system
catalogs and in which we want different sets of permissions (see
kaigai san's patch about largeobject controls in which he actually add
syntax for row level permission in that catalog, something we don't
have in any other place yet) is really a system catalog?
--
Atentamente,
Jaime Casanova
Soporte y capacitación de PostgreSQL
Asesoría y desarrollo de sistemas
Guayaquil - Ecuador
Cel. +59387171157