On Tue, Nov 20, 2012 at 10:12 AM, Shaun Thomas <sthomas@optionshouse.com> wrote:
> On 11/20/2012 10:03 AM, Merlin Moncure wrote:
>
>> Shared buffer manipulation changing contention is suggesting you're
>> running into free list lock issues. How many active backends/cores?
>
>
> Oh, the reason I wanted to point it out was that we see a lot more than CPU
> contention with higher shared_buffer settings. The scheduler starts
> aggressively clearing out the file cache *and* freeing memory, driving load
> way higher than it should be. That doesn't result in a very high iowait, but
> all that page swapping sure isn't free.
have you ruled out numa issues?
(http://frosty-postgres.blogspot.com/2012/08/postgresql-numa-and-zone-reclaim-mode.html)
merlin