Scott Carey wrote:
> the OS can either quickly allocate to the process or
> the page cache from the free buffers, or more slowly take from the page
> cache, or even more slowly page out a process page.
Aha, now it all makes sense.
> I like to use the '5 second rule'. dirty_background_ratio should be sized
> so that it takes about 5 seconds to flush to disk in optimal conditions.
> dirty_ratio should be 2x to 5x this depending on your application's needs --
> for a system with well tuned postgres checkpoints, smaller tends to be
> better to limit stalls while waiting for the checkpoint fsync to finish.
Thanks a lot, this is invaluable information.
In response to
pgsql-performance by date
|Next:||From: Rui Carvalho||Date: 2009-07-01 09:40:02|
|Subject: - Slow Query|
|Previous:||From: Scott Carey||Date: 2009-07-01 00:11:19|
|Subject: Re: random slow query|