"Ron Mayer" <[EMAIL PROTECTED]> writes:
> One thing I've been wondering, is if "sort_mem" could be
> per connection/backend-process instead of per sorting operation
> so that sort_mem could be set more aggressivelly without running out
> of memory so easily with large queries.
Not very readily.
>
(B>
(B>Actually, the results are completely expected once you know what's
(B>exactly is going on. I found it weird that my sorts were also slowing
(B>down with more sort memory until Tom or Bruce or someone pointed out to
(B>me that my stats said my sorts were swapping.
(B>
(B>
(B
(Bth
On Wed, 12 Feb 2003, [ISO-8859-1] Hans-J$B|(Brgen Sch$Bv(Bnig wrote:
(B
(B> Be careful with sort_mem - this might lead to VERY unexpected results. I
(B> did some testing on my good old Athlon 500 with a brand new IBM 120 Gigs
(B> HDD. Reducing the sort_mem gave me significantly faster resul
Christopher Kings-Lynne wrote:
>
>I reckon that sort_mem is the hardest thing to optimise1
>
Agreed... in part because it depends a lot on the query.
Also, if I understand correctly sort_mem not only affects sorts
but also hash table stuff as well, right? If that's true for
the new hash aggrega
Gavin Sherry wrote:
Hi Chris,
On Wed, 12 Feb 2003, Christopher Kings-Lynne wrote:
Machine:
256MB RAM, FreeBSD 4.7, EIDE HDD, > 1 Ghz
Seems like a small amount of memory to be memory based tests with.
What about testing sort_mem as well. It would system to me that there
would be no neg
> > Machine:
> > 256MB RAM, FreeBSD 4.7, EIDE HDD, > 1 Ghz
>
> Seems like a small amount of memory to be memory based tests with.
Perhaps, but I'm benchmarking for that machine, not for any other. The
results have to include the 256MB spec.
Also, the peak was 25MB of SHM, which still leave 231MB
Hi Chris,
On Wed, 12 Feb 2003, Christopher Kings-Lynne wrote:
> Machine:
> 256MB RAM, FreeBSD 4.7, EIDE HDD, > 1 Ghz
Seems like a small amount of memory to be memory based tests with.
What about testing sort_mem as well. It would system to me that there
would be no negative to having infinite s
> > I am now going to leave it on 5000 and play with wal_buffers.
> > Is there anything else people are interested in me trying?
>
> Keenly interested. Who wouldn't want to know how to optimize it?
> That's the hardest guideline to find.
Oops - what that sentence was supposed to say is "Is there
> -Original Message-
> From: Christopher Kings-Lynne [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, February 11, 2003 8:54 PM
> To: Hackers; Advocacy
> Subject: [HACKERS] PostgreSQL Tuning Results
>
>
> Hi Everyone,
>
> I have just completed a basic
Hi Everyone,
I have just completed a basic set of benchmarking on our new database
server. I wanted to figure out a good value for shared_buffers before we go
live.
We are a busy ecommerce-style website and so we probably get 10 or 20 to 1
read transactions vs. write transactions. We also don't
10 matches
Mail list logo