On Mon, Oct 31, 2005 at 02:50:31PM -0600, Scott Marlowe wrote:
> > Your point was about cache efficiency as an argument for not increasing
> > shared_buffers. Politely, I don't accept that argument. Clearly, there
> > are some other considerations (for which I agree completely) but those
> > don't prevent you increasing shared_buffers, they just place limits on
> > your overall memory budget which could effect shared_buffers of course.
>
> As I understand it, when the last backend referencing a collection of
> data stops referencing it, that the buffers holding that data are
> released, and if, a second later, another backend wants the data, then
> it has to go to the Kernel for it again.
>
> Is this still the case in 8.1?
Depends what you mean. What one backend uses stays in the shared
buffers when it's done. It's only removed to make room for other blocks
that have been requested. Whether it's still there after a second kind
of depends on how much other data you read in the meantime and whether
the caching algorithm decided the data was old enough that you wern't
likely to need it soon.
It's kind of like the kernel cache, once you've been running for a
while it's always full of blocks of data. There's no point forgetting
perfectly good data. The only time you don't need to throw away blocks
is if your database is smaller than your memory,
You mentioned something about those OSDL tests, where can we download
the results? I just get told khack.osdl.org is unreachable...
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
> tool for doing 5% of the work and then sitting around waiting for someone
> else to do the other 95% so you can sue them.