hubert depesz lubaczewski writes:
> On Fri, Aug 28, 2009 at 12:47:15PM -0400, Tom Lane wrote:
>> Ouch. You need to update to 8.4 --- the SI messaging stuff will
>> definitely be hurting you with that many backends. Or consider
>> using connection pooling or something to cut the number of backend
On Fri, Aug 28, 2009 at 12:47:15PM -0400, Tom Lane wrote:
> hubert depesz lubaczewski writes:
> > On Fri, Aug 28, 2009 at 12:28:06PM -0400, Tom Lane wrote:
> >> Hm, do you have forty or so idle backends hanging around while this
> >> happens? The only thing I can think of that might be causing th
On Fri, Aug 28, 2009 at 12:47:15PM -0400, Tom Lane wrote:
> hubert depesz lubaczewski writes:
> > On Fri, Aug 28, 2009 at 12:28:06PM -0400, Tom Lane wrote:
> >> Hm, do you have forty or so idle backends hanging around while this
> >> happens? The only thing I can think of that might be causing th
hubert depesz lubaczewski writes:
> On Fri, Aug 28, 2009 at 12:28:06PM -0400, Tom Lane wrote:
>> Hm, do you have forty or so idle backends hanging around while this
>> happens? The only thing I can think of that might be causing this is
>> shared cache invalidation messages being broadcast to all
On Fri, Aug 28, 2009 at 12:28:06PM -0400, Tom Lane wrote:
> Hm, do you have forty or so idle backends hanging around while this
> happens? The only thing I can think of that might be causing this is
> shared cache invalidation messages being broadcast to all the other
> sessions.
I have about 100
hubert depesz lubaczewski writes:
> One of operations that happens on the database is:
> begin;
> call function();
> commit;
> where function is plpgsql function which does:
> - drop several (n) views/tables
> - rename ~2n views and tables (and related objects like indexes and
> constraints) -
