| From: | Cédric Villemain <cedric(dot)villemain(at)dalibo(dot)com> |
|---|---|
| To: | Bill Moran <wmoran(at)collaborativefusion(dot)com> |
| Cc: | "Campbell, Lance" <lance(at)uiuc(dot)edu>, pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: work_mem and shared_buffers |
| Date: | 2007-11-12 10:17:57 |
| Message-ID: | 47382855.1090902@dalibo.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
Bill Moran a écrit :
> On Fri, 9 Nov 2007 12:08:57 -0600
> "Campbell, Lance" <lance(at)uiuc(dot)edu> wrote:
>
>
>> How do you know when you should up the value of work_mem? Just play
>> with the number. Is there a query I could do that would tell me if
>> PostgreSql is performing SQL that could use more memory for sorting?
>>
>
> 8.2 and older, it can be difficult to know, and I don't have a specific
> recommendation.
>
>
I haven't use it in that context before, but perhaps inotify can be used
to catch postgresql usage of temp files. ( http://inotify.aiken.cz/ ,
http://inotify.aiken.cz/?section=incron&page=about&lang=en )
> 8.3 includes a parameter to log the usage of temporary files by Postgres.
> When a sort can't fit in the available memory, it uses a temp file, thus
> you could use this new feature to track when sorts don't fit in
> work_mem.
>
>
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Shane Ambler | 2007-11-12 10:23:13 | Re: PostgreSQL vs MySQL, and FreeBSD |
| Previous Message | adrobj | 2007-11-12 06:59:05 | Clustered/covering indexes (or lack thereof :-) |