Re: [PERFORM] buffercache/bgwriter

2011-03-28 Thread Uwe Bartels
OK. Thanks. Uwe On 28 March 2011 08:02, Greg Smith wrote: > On 03/24/2011 05:19 AM, Uwe Bartels wrote: > >> It would also be good to see how many buffers were written by backend >> processes grouped by Buffer Access Strategy - to better distinguish evil >> backend writes from wanted backend wr

Re: [PERFORM] buffercache/bgwriter

2011-03-27 Thread Greg Smith
On 03/24/2011 05:19 AM, Uwe Bartels wrote: It would also be good to see how many buffers were written by backend processes grouped by Buffer Access Strategy - to better distinguish evil backend writes from wanted backend writes. Since all these writes are being cached by the operating system,

Re: [PERFORM] buffercache/bgwriter

2011-03-24 Thread Uwe Bartels
Hi Cédric, OK, sounds promising. But all of these improvements are for the postgres developers. For me as an administrator I can't do a thing right now. OK. Thanks for you suggestions. I think for batchjobs other that just COPY they could speed up the process quite well if now the backend process

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Cédric Villemain
2011/3/23 Uwe Bartels : > On 23 March 2011 16:36, Jeff Janes wrote: >> >> On Wed, Mar 23, 2011 at 6:19 AM, Jochen Erwied >> wrote: >> > Wednesday, March 23, 2011, 1:51:31 PM you wrote: >> > >> > [rearranged for quoting] >> > >> >> background writer stats >> >>  checkpoints_timed | checkpoints_req

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Jeff Janes
On Wed, Mar 23, 2011 at 8:26 AM, Uwe Bartels wrote: > Hi Brad, > > yes. that's the question > in the source code in freelist.c there is something that I don't understand. > > This is the first try to get a free page. The second try scans used buffers. > What makes me wonder is the why postgres

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Uwe Bartels
On 23 March 2011 16:36, Jeff Janes wrote: > On Wed, Mar 23, 2011 at 6:19 AM, Jochen Erwied > wrote: > > Wednesday, March 23, 2011, 1:51:31 PM you wrote: > > > > [rearranged for quoting] > > > >> background writer stats > >> checkpoints_timed | checkpoints_req | buffers_checkpoint | > buffers_cl

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Jeff Janes
On Wed, Mar 23, 2011 at 6:19 AM, Jochen Erwied wrote: > Wednesday, March 23, 2011, 1:51:31 PM you wrote: > > [rearranged for quoting] > >> background writer stats >>  checkpoints_timed | checkpoints_req | buffers_checkpoint | buffers_clean | >> maxwritten_clean | buffers_backend | buffers_alloc >>

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Uwe Bartels
l-performance-ow...@postgresql.org [mailto:pgsql-performance- > > ow...@postgresql.org] On Behalf Of t...@fuzzy.cz > > Sent: Wednesday, March 23, 2011 10:42 AM > > To: Uwe Bartels > > Cc: pgsql-performance@postgresql.org > > Subject: Re: [PERFORM] buffercache/bgwriter &

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Nicholson, Brad (Toronto, ON, CA)
> -Original Message- > From: pgsql-performance-ow...@postgresql.org [mailto:pgsql-performance- > ow...@postgresql.org] On Behalf Of t...@fuzzy.cz > Sent: Wednesday, March 23, 2011 10:42 AM > To: Uwe Bartels > Cc: pgsql-performance@postgresql.org > Subject: Re:

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Uwe Bartels
Hi Thomas, thanks, but there were no new informations in there for me. this article http://www.westnet.com/~gsmith/content/postgresql/chkp-bgw-83.htm I know and others on his website. Best... Uwe On 23 March 2011 15:41, wrote: > > Hi, > > > > I have very bad bgwriter statistics on a server wh

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread tv
> Hi, > > I have very bad bgwriter statistics on a server which runs since many > weeks > and it is still the same after a recent restart. > There are roughly 50% of buffers written by the backend processes and the > rest by checkpoints. > The statistics below are from a server with 140GB RAM, 32GB

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Uwe Bartels
Hi Jochen, yes, I had that impression too. But it is running. ...And has almost no effect. I changed all parameter to the most aggressive, but Before I restarted the server I had a percentage of writes by the bgwriter of less that 1 percent. postgres=# select name,setting from pg_settings whe

Re: [PERFORM] buffercache/bgwriter

2011-03-23 Thread Jochen Erwied
Wednesday, March 23, 2011, 1:51:31 PM you wrote: [rearranged for quoting] > background writer stats > checkpoints_timed | checkpoints_req | buffers_checkpoint | buffers_clean | > maxwritten_clean | buffers_backend | buffers_alloc > ---+-++-

[PERFORM] buffercache/bgwriter

2011-03-23 Thread Uwe Bartels
Hi, I have very bad bgwriter statistics on a server which runs since many weeks and it is still the same after a recent restart. There are roughly 50% of buffers written by the backend processes and the rest by checkpoints. The statistics below are from a server with 140GB RAM, 32GB shared_buffers