Hi,
On 2019-04-11 15:39:15 -0400, Jeff Janes wrote:
> But I don't think I would recommend starting at 25% of RAM larger server.
> Is that really good advice? I would usually start out at 1GB even if the
> server has 128GB, and increase it only if there was evidence it needed to
> be increased. D
On Thu, Dec 13, 2018 at 11:51 PM Thomas Munro
wrote:
> On Fri, Dec 14, 2018 at 2:17 AM Ron wrote:
> > https://www.postgresql.org/docs/9.6/runtime-config-resource.html
> >
> > The docs say, "If you have a dedicated database server with 1GB or more
> of
> > RAM, a reasonable starting value for sha
I tested. The shared buffers works better, then an OS level filesystem cache.
The more shared_buffers (but less then database size), the better. With
huge_pages is more better. But you must reserve enough free memory for OS and
PostgeSQL itself.
> 13 дек. 2018 г., в 18:17, Ron написал(а):
>
Le ven. 14 déc. 2018 à 07:00, Ron a écrit :
> On 12/13/2018 08:25 PM, Rene Romero Benavides wrote:
> > This topic seems to be always open to discussion. In my opinion, it
> > depends on how big your work dataset is, there's no use in sizing
> > shared_buffers beyond that size. I think, the most r
On 12/13/2018 08:25 PM, Rene Romero Benavides wrote:
This topic seems to be always open to discussion. In my opinion, it
depends on how big your work dataset is, there's no use in sizing
shared_buffers beyond that size. I think, the most reasonable thing is
analyzing each case as proposed here:
On Fri, Dec 14, 2018 at 2:17 AM Ron wrote:
> https://www.postgresql.org/docs/9.6/runtime-config-resource.html
>
> The docs say, "If you have a dedicated database server with 1GB or more of
> RAM, a reasonable starting value for shared_buffers is 25%".
>
> But that's pretty archaic in 2018. What i
This topic seems to be always open to discussion. In my opinion, it depends
on how big your work dataset is, there's no use in sizing shared_buffers
beyond that size. I think, the most reasonable thing is analyzing each case
as proposed here:
https://www.keithf4.com/a-large-database-does-not-mean-l
